Ranking

Discord

Join

Support revolution crtpto

Least funded crypto

Information vote

Dashboard

Unite

Soft side human

Jorge Arango talks about his writing process. He actually co-authored a book on Information Architecture in 2015. This vid more about his latest. uses lots of paper notebooks Mentions 3 (obvious) phases in his non-fiction writing research structuring writing Says all three req diff tools. (he uses Obsidian f research)

Stage 1 research, vague idea about what topic, (s paper notebooks, index/sequence them) His book duly noted is on this phase.

Stage 2 structuring. mistake to go directly to writing. (vgl [[A System for Writing by Bob Doto]] ) Mentions author Robert Caro who had book structure on his wall story book like, vgl what I do when writing reports w wall of post its for structure. Limitation is size of wall. But physcial movement helpt. He does this type of wall in [[Tinderbox]] !!!! canvas, with the sticky notes having metadata. His chapter and heading structure emerges from it! Prevents getting lost in the details. Shows him what is missing too (looping back to research phase sometimes). It's not writing but getting a sense of the narrative flow of the entire thing. Seems like outlining of talks, but a book is bigger and thus the outlining is more filled out with links to material

What question would a reader have entering chapter 2 after reading chapter 1, and go from there. Has opening story/anecdote for each chapter. Mwah. Uses the story board view like how the periodic table suggested the existence of atoms that had not been discovered yet. Dmitri Mendleev.

Stage 3 writing After structuring the writing tools come into play. He used Scrivener, avoiding having a monolithic text. Now Ulysses, in markdown. Uses word count targets in these tools for sections/chapters

set up supportive environments for each pahse structure first but keep flexible, don't make it a linear process, but structure ulnocks the writing swith modalities - walking, free form in [[Tinderbox]] or another tool, to avoid getting stuck.

The problem with this page is that the author is playing fast and loose with the words "parser" and "parsing". It's also not totally honest (or is yet again confused) about e.g. what the IETF RFCs say about parsing high precision numbers.

Tip from colleague C for images on AI that break the usual (anthropomorphic) frame. CC-licensed for re-use usually.

This line connects nature’s season winter, to the idea of sleep or rest. Romantic poets often use nature to express human feelings.

I thought about the connections between Lie groups and the monogram notation for a while and put together some notes here. As it turns out, the \(^A\mathbf v^{B_A}_F\) velocities add like vectors, if expressed in the same frame! The quantity \(^A\mathbf v^{B}_F\) has the advantage of being more physically meaningful.

Reminds me of the infatigable environment of Rain World, Death Stranding, and Frostpunk. Death looms.

Not just reading, though, we do things instrumentally, as we are biological machines. Reading is tough, so we scrutinise it severly. Watching TV isn't tough, so the usefulness dilutes and becomes not so present.

DOI: 10.1016/j.isci.2025.114132

Resource: RRID:IMSR_JAX:000664

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:000664

What is this?

DOI: 10.1016/j.immuni.2025.10.024

Resource: RRID:IMSR_JAX:000664

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:000664

What is this?

DOI: 10.1016/j.immuni.2025.10.024

Resource: (IMSR Cat# JAX_005023,RRID:IMSR_JAX:005023)

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:005023

What is this?

DOI: 10.1016/j.cub.2025.10.062

Resource: (IMSR Cat# JAX_006474,RRID:IMSR_JAX:006474)

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:006474

What is this?

DOI: 10.1016/j.cub.2025.10.062

Resource: (IMSR Cat# JAX_017320,RRID:IMSR_JAX:017320)

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:017320

What is this?

DOI: 10.1016/j.cub.2025.10.062

Resource: RRID:IMSR_JAX:000664

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:000664

What is this?

DOI: 10.1016/j.cub.2025.10.062

Resource: (IMSR Cat# JAX_013044,RRID:IMSR_JAX:013044)

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:013044

What is this?

DOI: 10.1016/j.celrep.2025.116588

Resource: RRID:IMSR_JAX:000664

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:000664

What is this?

DOI: 10.1016/j.celrep.2025.116576

Resource: (IMSR Cat# JAX_003232,RRID:IMSR_JAX:003232)

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:003232

What is this?

DOI: 10.1016/j.celrep.2025.116576

Resource: (IMSR Cat# JAX_021160,RRID:IMSR_JAX:021160)

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:021160

What is this?

DOI: 10.1016/j.celrep.2025.116576

Resource: (IMSR Cat# JAX_018956,RRID:IMSR_JAX:018956)

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:018956

What is this?

DOI: 10.1016/j.celrep.2025.116576

Resource: (IMSR Cat# JAX_026965,RRID:IMSR_JAX:026965)

Curator: @areedewitt04

SciCrunch record: RRID:IMSR_JAX:026965

What is this?

DOI: 10.6084/m9.figshare.29801927

Resource: Wisconsin-Madison University Biotechnology Center Gene Expression Center Core Facility (RRID:SCR_017757)

Curator: @scibot

SciCrunch record: RRID:SCR_017757

What is this?

DOI: 10.6084/m9.figshare.29801927

Resource: Wisconsin-Madison University Biotechnology Center Bioinformatics Resource Center Core Facility (RRID:SCR_017799)

Curator: @scibot

SciCrunch record: RRID:SCR_017799

What is this?

DOI: 10.1007/s00424-014-1662-4

Resource: Analysis of Functional NeuroImages (RRID:SCR_005927)

Curator: @scibot

SciCrunch record: RRID:SCR_005927

What is this?

DOI: 10.1007/s00424-014-1662-4

Resource: MRIcroGL (RRID:SCR_024413)

Curator: @scibot

SciCrunch record: RRID:SCR_024413

What is this?

DOI: 10.1007/s00424-014-1662-4

Resource: R Project for Statistical Computing (RRID:SCR_001905)

Curator: @scibot

SciCrunch record: RRID:SCR_001905

What is this?

DOI: 10.1007/s00424-014-1662-4

Resource: GraphPad Prism (RRID:SCR_002798)

Curator: @scibot

SciCrunch record: RRID:SCR_002798

What is this?

DOI: 10.1007/s00424-014-1662-4

Resource: SPSS (RRID:SCR_002865)

Curator: @scibot

SciCrunch record: RRID:SCR_002865

What is this?

DOI: 10.1007/s00424-014-1662-4

Resource: FSL (RRID:SCR_002823)

Curator: @scibot

SciCrunch record: RRID:SCR_002823

What is this?

Lymph isn’t actively pumped by the heart but due to pressure gradient caused by skeletal muscle contraction and breathing. This pressure gradient causes the vessels to open

follow me

the core concepts that Global Linguistic Flow presents in terms of composition and Hip-hop as the Cipha and "The dusty-foot philosopher" both have major connections in terms of hip-hop, The Cipha is more important being able to tie in the end with the beginning is a key compositional technique that allows writers and rappers to make unique and interesting songs and text, for example an artist like MFDOOM (a hip hop artist) who is known to make connections from the beginning and end, he also has multiple alter egos that he references and has call back conversations with in his songs.

this echo's Ashanti Young's "code-meshing" from his 2010 article, “Should Writers Use They Own English?” the main connection it has is preserving authenticity and personality in your lyrics or texts.

I think this sentence is really strong and it connect the topic of hip-hop to the world and the people in it, but also shows how hip-hop is an amalgamation of different cultures and meanings.

using the cipha really connects hip-hop to composition by having that call and response formatting. The book also uses the cipha by referring to each paragraphs and topics as tracks, which even further connects how hip-hop tracks aren't just sound and words they have meaning and information in them and reading into those meaning can help the listener better understand what the artist might be trying to communicate..

Because perception of an environment can shape mental well-being as much as the environment itself, reframing negative beliefs, practicing gratitude, making small comforting changes, and processing emotions can help improve mental health even when the external situation cannot be changed.

A person's social environment, including relationships, social support, socioeconomic factors, and exposure to conflict or abuse, strongly shapes mental health, with healthy connections promoting resilience and coping, while harmful relationships increases the risk of stress, anxiety, depression, and substance abuse.

Because people spend over half of their waking hours at work, the work environment–including factors like high demands, stress, bullying, low support, lack of control, and feeling undervalued–plays a major role in increasing the risk of mental health issues such as depression and anxiety.

The home environment, such as physical conditions of the home, community factors, and exposures such as crime, climate, pollution, toxins, environmental racism, and poverty significantly shapes mental health by influencing stress levels, safety, and overall psychological well-being.

Anything that can be perceived through the senses has the ability to impact how one feels.

Neuroscience shows that environmental factors can physically affect brain functions.

I see how more players in the field can lead to competition over limited resources, it will be interesting to me to see how being extremely supportive and open with their methods with the intention of helping others do what they've done can remain as a focus with this limitation in mind.

This is so important in my opinion!

I am interested in how scaling up is different from increasing the output of a press this significantly, which I don't fully understand. Does this mean drastically increasing efficiency?

This doesn't seem to me to be a tiny event at all, but I see what they mean. The appearance of these presses opening up possibilities that were not even considered before, which is exciting.

Although deserts usually offer low restorative benefits due to evolutionary survival instincts, environmental majors experience greater recovery in desert environments because their education, familiarity, and ecological understanding overrides their innate avoidance. Again, showing that restoration depends not just on the environment bur also the viewers background and learned experiences.

Environmental majors experience stronger physical and psychological recovery from natural environments than non-majors because of their education, familiarity, ecological awareness, and emotional connection to nature enhances their ability to relax and recover. Suggests that nature based mental health interventions should account for students' backgrounds and environmental values

Juniors experience stronger psychological and physical restoration from natural environments than freshman, possibly due to the fact that they are older, more emotionally mature, have better coping skills and social support, and therefore better able to benefit from nature. Suggests that age and developmental stage should guide how schools design nature-based mental health programs.

Psychological restoration varies by landscape type and by viewer characteristics: • lawn scenes are the most mentally restorative and deserts are the least – from most to least restorative: Grass/Lawn, Waterscapes, Forest, Plants, Field, Mountains, Artificial Nature, Deserts • gender does not significantly affect psychological restoration • juniors report stronger psychological recovery than freshmen • environmental majors consistently experience higher psychological restoration than non-majors

Visual preferences depend on the type of landscape and viewer's background: • waterscapes are the most preferred, and deserts are the least • gender does not influence preference • juniors prefer lawn and artificial environments more than freshmen • landscape majors show higher preference for artificial environments than non-majors

Heart rate recovery depends on who the viewer is: • females exhibit in lawn and artificial landscapes • juniors show dramatically stronger HR restoration across almost all landscape types compared to freshmen • environmental majors recover more deeply, especially in forest or artificial nature scenes

Gender and grade level had no effect on systolic BP recovery. The only major-related difference was that landscape/environmental majors showed a significantly larger SBP reduction than non-environment majors when viewing desert landscapes, but not for any other landscape type.

Hypotheses presented for study.

Mass is any physical thing that is dependent on the quantity of the object. Mass is its weight.

a program or initiative that supports the development of neurotechnology by providing resources, funding, and training, often with a focus on ethical considerations, equity, and societal impact

There appear to be some missing data in the supplements? I can only access figs S1-S5.

"Summary and future directions" margin notes: Clarifies the scope: applies to LLM-based tools like ChatGPT, not all AI.

Acknowledges rapid evolution of AI: guidelines may need updates over time.

Introduces academic-specific AI tools: suggests they may reduce issues like hallucinations.

Warns against losing human expertise: emphasizes importance of reasoning and deep knowledge.

Connects to earlier caution about over-reliance: reinforces theme of responsible use.

Calls for ongoing discussion : encourages future research and community engagement.

Good material for your own conclusion: especially if you discuss long-term ethical implications.

"Author checklist for ethical use of AI generative tools" Margin Notes: Ethics require reflection, not fixed rules

Human responsibility remains central

AI cannot be the source of original ideas

Protect human research + writing abilities

Must verify facts and references

Always disclose AI use clearly

If any question = “no,” revise process

Checklist acts as a self-audit tool

Goal: maintain integrity, accuracy, and transparency

Together, these questions provide a practical self-check system that helps authors use AI without compromising academic integrity. They highlight key responsibilities: ensuring true intellectual ownership, protecting human learning and critical thinking, verifying the accuracy and fairness of all AI-generated material, and openly reporting how AI was used. If any answer is “no,” it signals that the author may be misusing AI or relying on it too heavily, and must adjust their approach before continuing.

The authors emphasize that ethical AI use is complex and cannot be reduced to a strict list of rules. Instead, scholars must apply thoughtful, case-by-case judgment.

This explains why outlining, summarizing, and brainstorming must be done carefully and checked thoroughly.

Even simple translation requires human review, reinforcing that AI must be used with oversight.

his defines the lowest-risk types of AI usage; things involving editing, clarity, and structure rather than generating new content.

The authors warn that overuse of AI can weaken critical thinking and writing skills, especially for novice researchers.

This explains why AI can produce false information and why authors must verify all AI-generated text.

"Issues with AI assisted writing" margin note Usefulness vs. risk: AI can help, but only if humans critically evaluate and revise the content.

Medical context: These issues are especially serious in healthcare writing, where misinformation could cause harm.

Citation danger: AI is weakest at references, with extremely high error/fabrication rates.

Ethical responsibility: Humans remain fully accountable for verifying accuracy and ensuring originality.

AI cannot be blamed for errors. The author must verify content, citations, and integrity.

AI struggles significantly with citations. This data demonstrates that AI-generated references are unreliable and often unusable.

AI can sound confident even when it is completely wrong. This is especially dangerous in medical writing, where accuracy is essential.

This highlights a risk; once plagiarized AI text enters the literature, others may unknowingly cite it, multiplying the academic integrity problem.

AI may unintentionally create text that is too similar to published work because it pulls from existing content. This reinforces the need for human oversight to avoid accidental plagiarism

This sentence outlines the core dangers of relying on AI tools in academic writing. It frames the entire section: AI can copy text, make things up, and cite nonexistent sources. All of which jeopardize academic integrity.

"What do the journals and publishers say?" Margin notes: Hybrid research approach: Authors combine literature search + ChatGPT to gather initial perspectives. Shows real-world use of AI in research.

Publishers’ priorities: Integrity, ethical practice, and preventing misconduct shape AI guidelines.

Central principle = transparency: Full disclosure of AI use is essential; hiding it is unethical.

AI ≠ author: AI can create content but cannot take accountability → cannot qualify as an author.

Human creativity still primary: AI can support technical tasks but cannot replace intellectual contribution.

Evolving policies: AI guidelines are new and changing; scholars must stay informed.

The authors emphasize that AI is a tool, not a replacement for original intellectual labor. Academic writing still requires human creativity, decision-making, and ethical judgment.

This sentence reinforces one of the strongest uniform rules across journals: AI cannot be listed as an author. It can assist, but it cannot take responsibility or be accountable for the work.

This tells scholars exactly where and how to disclose AI use. The argument is that methods sections provide clarity and accessibility, unlike acknowledgements which readers may skip.

This is a major thematic point. Transparency is repeatedly highlighted across organizations. Essentially, hiding AI use = ethical problem. This becomes a guiding principle for academic writers.

This makes clear that the foundation for all publisher guidelines (even those related to AI) is protecting academic integrity. This frames AI policy as an extension of long-standing ethical practices, not something completely new.

This shows the authors using both traditional research methods and AI-generated responses, highlighting the hybrid research environment scholars now navigate. It also shows transparency because they disclose how they gathered their info.

"Introduction" Margin Notes: Benefits for paper: Shows how AI can improve efficiency and support the research process when used ethically.

Drawbacks: Raises concerns about AI producing novel content, which challenges traditional academic integrity.

Relevancy: Directly addresses how AI impacts academic writing, which is central to my topic.

One-sentence summary: The introduction outlines the rise of generative AI in academic writing, highlighting both its potential benefits and the ethical challenges it presents, while emphasizing the need for responsible engagement.

Relevant for my advantages section. Demonstrates the authors’ stance on ethical benefits and responsible use of AI

Highlights a major drawback of AI in writing: potential ethical concerns.

Explains the technical foundation of ChatGPT and LLMs; useful for defining key AI concepts in my paper.

Shows the structure and scope of the paper, helping identify where to find discussions of both benefits and challenges.

Points out significant drawbacks of AI, which is important to balance the discussion of AI in writing in my paper.

Highlights the main benefit of AI in academic writing efficiency. Relevant to my paper because it supports the argument that AI can save time and streamline research tasks.

Above all else, it is crucial to understand that humans are essential to all the innovations AI will be a part of. Especially in terms of ethics.

Battelle's involvement with AI ethics

Human machine collaboration!

AI is useful in tracking mass amounts of data quickly

Reiterating point above but this is highly interesting and aligns with what were thinking of talking about

Could be an interesting segment related to AI and breaking what seems impossible

Connection to Battelle

Huge in terms of finding new cures and medicines

While AI will have helpful uses its important to always remember that human involvement remains the most important constant.

Interesting

Potential question

Most general AI isn't capable of free thought and must rely on humanity in order to function

ChatGPT

.

This was most likely a prevalent thing in ancient times, as simply having the access to the tools to learn was the difference in social class

If English gets to a point where only people with advanced educational resources can fluently understand it, it is at risk of dissolving

This quote serves to introduce us to the topic that the author discusses, accents. She states that her mother isn't ashamed of her accent despite it being hard to understand sometimes.

This means that the mother's accent is a mix of both her mother's accent and the accent of her home in Puerto Rico.

Her accent is one of the only things left from her home, encasing everything from how she speaks, to how she acts.

WOW! I've been out of the classroom for quite awhile and never considered this scenario regarding AI. This hit a nerve in me as I'm sure it will with many. I get it!

How do we respond and mitigate the isolation, the loss of human dialogue, mentorship and connection?

Personally, I love this; Image annotations.

This can be used as an effective analysis tool. Visual stimulus is powerful and for many individuals constitutes a deeper connection to fundamental learning. Balanced learning across the three basic modes; written, auditory and visual is implied in academics and the workplace but the reality is that one mode is dominant. Recognizing that mode and enhancing reaps far more advantage than forcing an inferior mode. @Byrnesz

with this p value does the unit increase in feelings really indicate the increase in care-based intervention support?

figure 6 does not appear labeled, maybe "figure 6" is cut off by the image being cut off? either way i dont see figure captions

going to need to fix presentation of the graphs, things are cut off, the legend is blocking a bit of the graph too

from here on i wont comment on output, just get rid of it unless it contains something that you dont report in a figure caption/paragraph

this output isn't awful but since you're reporting the findings already i would say there's no need for this

these do not match your figure?

output

unnecessary

get rid of this output too

see if you can get rid of this output

We are grateful to Eva Lüdi Kong for bringing these additional German translations to our attention.

• Hauer, Erich. Das Ts'ien-tze-wên in vier chinesischen Schriftformen (The Thousand Character Essay in four Chinese scripts). Berlin: W. de Gruyter, 1925.
• Lüdi Kong, Eva. Der 1000 Zeichen Klassiker (The 1000 character classic). Ditzingen: Philipp Reclam jun. Verlag GmbH, 2018.
• Maag, Barbara. Tausend Zeichen Klassiker (Thousand character classic) (undated): https://www.barbara-maag.de/Qianziwen.htm

Deep Concept: Modern America is mostly corrupt from it's own creation of wealth. Wealth is power, power corrupts and absolute power corrupts absolutely! Money and wealth have completely changed the underlying foundation of America. Modern America is the corrupted result of wealth. Morality and ethics in modern American have been reshaped to "fit" European Aristocracy, ironically the same European aristocracy America fled in the Revolutionary War.

Billions and billions of tax payer money is spent on projects that could never pass rigorous examination and best public ROI use. Political authoritative conditions rule public tax money for the benefit of a few at the expense of the many. The public "cult-like" sheep have no clue how they are being abused.

The authoritative abusers (politicians) follow the "mostly" corrupt American (fuck-you) form of government and individual power tactics that have been conveniently embedded in corrupt modern morality and ethics, used by corrupted lawyers and judges to codify the fundamental moral code that underpins the original American Constitution.

Claims white Americans are hypocrites, urging them to listen to their own words.

America is built on the backs, blood and tears, of slaves.

David Walker was a black man; this is important to note so we understand how his audience viewed him.

Summarizes the limitations

Summarizes the key advantages of AI tools

Emphasizes both oversight and ethical responsibility; connects directly to discussions of drawbacks and necessary safeguards in AI use.

Highlights the balanced perspective advocated by the article (AI as a support tool, not a replacement) which is critical when discussing benefits vs. limitations.

Reinforces the theme of transparency and human responsibility. useful for citing standards and guidelines in research.

Introduces the ethical focus of this section, highly relevant for my paper when discussing accountability and responsible AI use.

Identifies a key drawback of AI in writing, useful for discussing the potential limitations and ethical considerations in my paper.

Highlights concrete benefits of AI tools in research efficiency and literature analysis, which is directly relevant to discussing the advantages of AI in writing.

Even with helpful tools, researchers must ensure honesty, transparency, and academic integrity when using AI.

The study includes a visual tool (the decision tree) to help researchers choose the right AI tool depending on their specific needs.

Grammarly strengthens the writing by improving grammar, flow, and clarity—important for the final stages of a scientific paper.

ChatGPT helps writers create initial drafts and develop ideas, making the early stages of writing easier.

"Materials and Methods" margin notes: Benefit: AI tools help with summarizing articles, organizing information, drafting text, and improving grammar.

Relevancy: Shows how the researchers tested AI writing tools (supports credibility of the article.)

Theme: Evaluation criteria (accuracy, accessibility) connect to both benefits & drawbacks in your paper.

Note: Mentions major AI tools used in academic writing → good background for my topic.

Supports the benefit of AI as a writing enhancement tool.

Relevant because these criteria relate directly to AI’s benefits and drawbacks in writing.

.

Explains research approach. shows how the article gathered its evidence.

Margin notes on the Introduction section: Benefit: AI improves clarity, organization, and literature search efficiency.

Drawback: Risk of inaccurate information + writing sounding too standardized.

Relevancy: Direct connection to my topic (outlines BOTH benefits and drawbacks of AI in writing.)

Clear statement on drawbacks of AI in writing

This quote is saying that AI improves scientific writing but also introduces risks that require strong human oversight.

1. Congress should appropriate $10 million to establish an interagency coordination office within OSTP that is co-chaired by the DOC.
2. Congress should then appropriate $500 million to DOC and DOE to fund a biomanufacturing moonshot that includes creating the pilot network of three nodes to form the BioNETWORK in regions of the U.S. within six months of receiving funding. Funding managed by the interagency coordination office in collaboration with a not-for-profit organization.
3. Continued investment of an additional $500 million should be appropriated by Congress to create economic incentives to sustain and transition the BioNETWORK from public funding to full commercial operation.

Three themes: 1. Provide alternative supply chain pathways 2. Explore distributed biomanufacturing innovation to enhance supply chain resilience 3. Address standards and data infrastructure to support biotech and biomanufacturing commercialization and trade

The BioNETWORK is a physical and informational layer including manufacturing innovation, generating market forces, and leverages data capture and feedback loops. This allows for a much needed scale up of production of novel biomaterials. The BioNETWORK is a physical and digital backbone allowing the translation of information to be efficient, protected, and nationwide.

The CHIPS and Science Act authorizes a suite of new regional technology development initiatives, reflecting substantial congressional interest in expanding the geographical distribution of innovative activity in the U.S. Enhancing regional innovation will also be central to implementing the legislation’s semiconductor provisions, according to the Department of Commerce.

BioNETWORK delivers decentralized and resilient network offering rapid access to manufacturing, expertise, and data repositories wherever they reside

Biotech as a pillar of domestic manufacturing

Reasons why centralized biomanufacturing is less secure and does not deliver on the full potential of biotech.

tie into bruce rittman episode?

Test

Good enough... that's what I'm talking about... and 'customize later' encapsulates the escape route needed to enhance as brand stakeholders will often require, and ultimately where traditional DS efforts break down.

This could be a good thing, and allowable in certain contexts. Marketing sites for example. Often brand wants to push a DS beyond its stated definitions, and as such, this 'abuse' of 'cn' could be considered acceptable practice in these bleeding-edge cases of marketing/brand expression

This article is a little outdated but as per my read through, the only aspect I have found that is truly 'out of date' is the way forms are implemented as that has been recently changed. The method mentioned in the article is still valid but being sunsetted as of the time I'm making this comment.

This comment further underscores the separation of 'DS as Doc', and puts the responsibility of code expression of said DS on the front-end team (implicitly the front-end team of a particular product)

I find it interesting (especially given recent discussions on this topic), that the author has reduced the concept of a design system from a technical toolset or implementation to mere documentation (e.g. style guide)

This is a key point of what slows down our DS development. While we do take care to ensure these usability concerns, and we expect reusability through adoption to spread out the cost, being able to start from a place where these concerns have already been solved for could reduce early development friction.

Is there a way to format this more clearly?

Might be better to format as "5 (excellent)"

remove warnings

Gaps in research are more conceptual, I would change to something like "dietary quality of university students."

eLife Assessment

This important study combines brain stimulation with fMRI and behavioural modelling to probe the role of the left superior frontal sulcus in perceptual and value-based decision making. The evidence that the left SFS plays a key role in perceptual decision making is convincing; the results also suggest that the value-based decision process was largely unaffected by the stimulation, despite a change in response times.

Reviewer #1 (Public review):

Summary:

In this study, participants completed two different tasks. A perceptual choice task in which they compared the sizes of pairs of items and a value-different task in which they identified the higher value option among pairs of items with the two tasks involving the same stimuli. Based on previous fMRI research, the authors sought to determine whether the superior frontal sulcus (SFS) is involved in both perceptual and value-based decisions or just one or the other. Initial fMRI analyses were devised to isolate brain regions that were activated for both types of choices and also regions that were unique to each. Transcranial magnetic stimulation was applied to the SFS in between fMRI sessions and it was found to lead to a significant decrease in accuracy and RT on the perceptual choice task but only a decrease in RT on the value-different task. Hierarchical drift diffusion modelling of the data indicated that the TMS had led to a lowering of decision boundaries in the perceptual task and a lower of non-decision times on the value-based task. Additional analyses show that SFS covaries with model derived estimates of cumulative evidence, that this relationship is weakened by TMS.

Strengths:

The paper has many strengths, including the rigorous multi-pronged approach of causal manipulation, fMRI and computational modelling, which offers a fresh perspective on the neural drivers of decision making. Some additional strengths include the careful paradigm design, which ensured that the two types of tasks were matched for their perceptual content while orthogonalizing trial-to-trial variations in choice difficulty. The paper also lays out a number of specific hypotheses at the outset regarding the behavioural outcomes that are tied to decision model parameters and well justified.

Weaknesses:

In my previous comments (1.3.1 and 1.3.2) I noted that key results could be potentially explained by cTBS leading to faster perceptual decision making in both the perceptual and value-based tasks. The authors responded that if this were the case then we would expect either a reduction in NDT in both tasks or a reduction in decision boundaries in both tasks (whereas they observed a lowering of boundaries in the perceptual task and a shortening of NDT in the value task). I disagree with this statement. First, it is important to note that the perceptual decision that must be completed before the value-based choice process can even be initiated (i.e. the identification of the two stimuli) is no less trivial than that involved in the perceptual choice task (comparison of stimulus size). Given that the perceptual choice must be completed before the value comparison can begin, it would be expected that the model would capture any variations in RT due to the perceptual choice in the NDT parameter and not as the authors suggest in the bound or drift rate parameters since they are designed to account for the strength and final quantity of value evidence specifically. If, in fact, cTBS causes a general lowering of decision boundaries for perceptual decisions (and hence speeding of RTs) then it would be predicted that this would manifest as a short NDT in the value task model, which is what the authors see.

Reviewer #2 (Public review):

Summary:

The authors set out to test whether a TMS-induced reduction in excitability of the left Superior Frontal Sulcus influenced evidence integration in perceptual and value-based decisions. They directly compared behaviour-including fits to a computational decision process model---and fMRI pre and post TMS in one of each type of decision-making task. Their goal was to test domain-specific theories of the prefrontal cortex by examining whether the proposed role of the SFS in evidence integration was selective for perceptual but not value-based evidence.

Strengths:

The paper presents multiple credible sources of evidence for the role of the left SFS in perceptual decision making, finding similar mechanisms to prior literature and a nuanced discussion of where they diverge from prior findings. The value-based and perceptual decision-making tasks were carefully matched in terms of stimulus display and motor response, making their comparison credible.

Weaknesses:

-I was confused about the model specification in terms of the relationship between evidence level and drift rate. While the methods (and e.g. supplementary figure 3) specify a linear relationship between evidence level and drift rate, suggesting, unless I misunderstood, that only a single drift rate parameter (kappa) is fit. However, the drift rate parameter estimates in the supplementary tables (and response to reviewers) do not scale linearly with evidence level.

-The fit quality for the value-based decision task is not as good as that for the PDM, and this would be worth commenting on in the paper.

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public Review):

Summary:

In this study, participants completed two different tasks. A perceptual choice task in which they compared the sizes of pairs of items and a value-different task in which they identified the higher value option among pairs of items with the two tasks involving the same stimuli. Based on previous fMRI research, the authors sought to determine whether the superior frontal sulcus (SFS) is involved in both perceptual and value-based decisions or just one or the other. Initial fMRI analyses were devised to isolate brain regions that were activated for both types of choices and also regions that were unique to each. Transcranial magnetic stimulation was applied to the SFS in between fMRI sessions and it was found to lead to a significant decrease in accuracy and RT on the perceptual choice task but only a decrease in RT on the value-different task. Hierarchical drift-diffusion modelling of the data indicated that the TMS had led to a lowering of decision boundaries in the perceptual task and a lower of non-decision times on the value-based task. Additional analyses show that SFS covaries with model-derived estimates of cumulative evidence and that this relationship is weakened by TMS.

Strengths:

The paper has many strengths including the rigorous multi-pronged approach of causal manipulation, fMRI and computational modelling which offers a fresh perspective on the neural drivers of decision making. Some additional strengths include the careful paradigm design which ensured that the two types of tasks were matched for their perceptual content while orthogonalizing trial-to-trial variations in choice difficulty. The paper also lays out a number of specific hypotheses at the outset regarding the behavioural outcomes that are tied to decision model parameters and are well justified.

Weaknesses:

(1.1) Unless I have missed it, the SFS does not actually appear in the list of brain areas significantly activated by the perceptual and value tasks in Supplementary Tables 1 and 2. Its presence or absence from the list of significant activations is not mentioned by the authors when outlining these results in the main text. What are we to make of the fact that it is not showing significant activation in these initial analyses?

You are right that the left SFS does not appear in our initial task-level contrasts. Those first analyses were deliberately agnostic to evidence accumulation (i.e., average BOLD by task, irrespective of trial-by-trial evidence). Consistent with prior work, SFS emerges only when we model the parametric variation in accumulated perceptual evidence.

Accordingly, we ran a second-level GLM that included trial-wise accumulated evidence (aE) as a parametric modulator. In that analysis, the left SFS shows significant aE-related activity specifically during perceptual decisions, but not during value-based decisions (SVC in a 10-mm sphere around x = −24, y = 24, z = 36).

To avoid confusion, we now:

(i) explicitly separate and label the two analysis levels in the Results; (ii) state up front that SFS is not expected to appear in the task-average contrast; and (iii) add a short pointer that SFS appears once aE is included as a parametric modulator. We also edited Methods to spell out precisely how aE is constructed and entered into GLM2. This should make the logic of the two-stage analysis clearer and aligns the manuscript with the literature where SFS typically emerges only in parametric evidence models.

(1.2) The value difference task also requires identification of the stimuli, and therefore perceptual decision-making. In light of this, the initial fMRI analyses do not seem terribly informative for the present purposes as areas that are activated for both types of tasks could conceivably be specifically supporting perceptual decision-making only. I would have thought brain areas that are playing a particular role in evidence accumulation would be best identified based on whether their BOLD response scaled with evidence strength in each condition which would make it more likely that areas particular to each type of choice can be identified. The rationale for the authors' approach could be better justified.

We agree that both tasks require early sensory identification of the items, but the decision-relevant evidence differs by design (size difference vs. value difference), and our modelling is targeted at the evidence integration stage rather than initial identification.

To address your concern empirically, we: (i) added session-wise plots of mean RTs showing a general speed-up across the experiment (now in the Supplement); (ii) fit a hierarchical DDM to jointly explain accuracy and RT. The DDM dissociates decision time (evidence integration) from non-decision time (encoding/response execution).

After cTBS, perceptual decisions show a selective reduction of the decision boundary (lower accuracy, faster RTs; no drift-rate change), whereas value-based decisions show no change to boundary/drift but a decrease in non-decision time, consistent with faster sensorimotor processing or task familiarity. Thus, the TMS effect in SFS is specific to the criterion for perceptual evidence accumulation, while the RT speed-up in the value task reflects decision-irrelevant processes. We now state this explicitly in the Results and add the RT-by-run figure for transparency.

(1.2.1) The value difference task also requires identification of the stimuli, and therefore perceptual decision-making. In light of this, the initial fMRI analyses do not seem terribly informative for the present purposes as areas that are activated for both types of tasks could conceivably be specifically supporting perceptual decision-making only.

Thank you for prompting this clarification.

The key point is what changes with cTBS. If SFS supported generic identification, we would expect parallel cTBS effects on drift rate (or boundary) in both tasks. Instead, we find: (a) boundary decreases selectively in perceptual decisions (consistent with SFS setting the amount of perceptual evidence required), and (b) non-decision time decreases selectively in the value task (consistent with speed-ups in encoding/response stages). Moreover, trial-by-trial SFS BOLD predicts perceptual accuracy (controlling for evidence), and neural-DDM model comparison shows SFS activity modulates boundary, not drift, during perceptual choices.

Together, these converging behavioral, computational, and neural results argue that SFS specifically supports the criterion for perceptual evidence accumulation rather than generic visual identification.

(1.2.2) I would have thought brain areas that are playing a particular role in evidence accumulation would be best identified based on whether their BOLD response scaled with evidence strength in each condition which would make it more likely that areas particular to each type of choice can be identified. The rationale for the authors' approach could be better justified.

We now more explicitly justify the two-level fMRI approach. The task-average contrast addresses which networks are generally more engaged by each domain (e.g., posterior parietal for PDM; vmPFC/PCC for VDM), given identical stimuli and motor outputs. This complements, but does not substitute for, the parametric evidence analysis, which is where one expects accumulation-related regions such as SFS to emerge. We added text clarifying that the first analysis establishes domain-specific recruitment at the task level, whereas the second isolates evidence-dependent signals (aE) and reveals that left SFS tracks accumulated evidence only for perceptual choices. We also added explicit references to the literature using similar two-step logic and noted that SFS typically appears only in parametric evidence models.

(1.3) TMS led to reductions in RT in the value-difference as well as the perceptual choice task. DDM modelling indicated that in the case of the value task, the effect was attributable to reduced non-decision time which the authors attribute to task learning. The reasoning here is a little unclear.

(1.3.1) Comment: If task learning is the cause, then why are similar non-decision time effects not observed in the perceptual choice task?

Great point. The DDM addresses exactly this: RT comprises decision time (DT) plus non-decision time (nDT). With cTBS, PDM shows reduced DT (via a lower boundary) but stable nDT; VDM shows reduced nDT with no change to boundary/drift. Hence, the superficially similar RT speed-ups in both tasks are explained by different latent processes: decision-relevant in PDM (lower criterion → faster decisions, lower accuracy) and decision-irrelevant in VDM (faster encoding/response). We added explicit language and a supplemental figure showing RT across runs, and we clarified in the text that only the PDM speed-up reflects a change to evidence integration.

(1.3.2) Given that the value-task actually requires perceptual decision-making, is it not possible that SFS disruption impacted the speed with which the items could be identified, hence delaying the onset of the value-comparison choice?

We agree there is a brief perceptual encoding phase at the start of both tasks. If cTBS impaired visual identification per se, we would expect longer nDT in both tasks or a decrease in drift rate. Instead, nDT decreases in the value task and is unchanged in the perceptual task; drift is unchanged in both. Thus, cTBS over SFS does not slow identification; rather, it lowers the criterion for perceptual accumulation (PDM) and, separately, we observe faster non-decision components in VDM (likely familiarity or motor preparation). We added a clarifying sentence noting that item identification was easy and highly overlearned (static, large food pictures), and we cite that nDT is the appropriate locus for identification effects in the DDM framework; our data do not show the pattern expected of impaired identification.

(1.4) The sample size is relatively small. The authors state that 20 subjects is 'in the acceptable range' but it is not clear what is meant by this.

We have clarified what we mean and provided citations. The sample (n = 20) matches or exceeds many prior causal TMS/fMRI studies targeting perceptual decision circuitry (e.g., Philiastides et al., 2011; Rahnev et al., 2016; Jackson et al., 2021; van der Plas et al., 2021; Murd et al., 2021). Importantly, we (i) use within-subject, pre/post cTBS differences-in-differences with matched tasks; (ii) estimate hierarchical models that borrow strength across participants; and (iii) converge across behavior, latent parameters, regional BOLD, and connectivity. We now replace the vague phrase with a concrete statement and references, and we report precision (HDIs/SEs) for all main effects.

Reviewer #2 (Public Review):

Summary:

The authors set out to test whether a TMS-induced reduction in excitability of the left Superior Frontal Sulcus influenced evidence integration in perceptual and value-based decisions. They directly compared behaviour - including fits to a computational decision process model - and fMRI pre and post-TMS in one of each type of decision-making task. Their goal was to test domain-specific theories of the prefrontal cortex by examining whether the proposed role of the SFS in evidence integration was selective for perceptual but not value-based evidence.

Strengths:

The paper presents multiple credible sources of evidence for the role of the left SFS in perceptual decision-making, finding similar mechanisms to prior literature and a nuanced discussion of where they diverge from prior findings. The value-based and perceptual decision-making tasks were carefully matched in terms of stimulus display and motor response, making their comparison credible.

Weaknesses:

(2.1) More information on the task and details of the behavioural modelling would be helpful for interpreting the results.

Thank you for this request for clarity. In the revision we explicitly state, up front, how the two tasks differ and how the modelling maps onto those differences.

(1) Task separability and “evidence.” We now define task-relevant evidence as size difference (SD) for perceptual decisions (PDM) and value difference (VD) for value-based decisions (VDM). Stimuli and motor mappings are identical across tasks; only the evidence to be integrated changes.

(2) Behavioural separability that mirrors task design. As reported, mixed-effects regressions show PDM accuracy increases with SD (β=0.560, p<0.001) but not VD (β=0.023, p=0.178), and PDM RTs shorten with SD (β=−0.057, p<0.001) but not VD (β=0.002, p=0.281). Conversely, VDM accuracy increases with VD (β=0.249, p<0.001) but not SD (β=0.005, p=0.826), and VDM RTs shorten with VD (β=−0.016, p=0.011) but not SD (β=−0.003, p=0.419).

(3 How the HDDM reflects this. The hierarchical DDM fits the joint accuracy–RT distributions with task-specific evidence (SD or VD) as the predictor of drift. The model separates decision time from non-decision time (nDT), which is essential for interpreting the different RT patterns across tasks without assuming differences in the accumulation process when accuracy is unchanged.

These clarifications are integrated in the Methods (Experimental paradigm; HDDM) and in Results (“Behaviour: validity of task-relevant pre-requisites” and “Modelling: faster RTs during value-based decisions is related to non-decision-related sensorimotor processes”).

(2.2) The evidence for a choice and 'accuracy' of that choice in both tasks was determined by a rating task that was done in advance of the main testing blocks (twice for each stimulus). For the perceptual decisions, this involved asking participants to quantify a size metric for the stimuli, but the veracity of these ratings was not reported, nor was the consistency of the value-based ones. It is my understanding that the size ratings were used to define the amount of perceptual evidence in a trial, rather than the true size differences, and without seeing more data the reliability of this approach is unclear. More concerning was the effect of 'evidence level' on behaviour in the value-based task (Figure 3a). While the 'proportion correct' increases monotonically with the evidence level for the perceptual decisions, for the value-based task it increases from the lowest evidence level and then appears to plateau at just above 80%. This difference in behaviour between the two tasks brings into question the validity of the DDM which is used to fit the data, which assumes that the drift rate increases linearly in proportion to the level of evidence.

We thank the reviewer for raising these concerns, and we address each of them point by point:

2.2.1. Comment: It is my understanding that the size ratings were used to define the amount of perceptual evidence in a trial, rather than the true size differences, and without seeing more data the reliability of this approach is unclear.

That is correct—we used participants’ area/size ratings to construct perceptual evidence (SD).

To validate this choice, we compared those ratings against an objective image-based size measure (proportion of non-black pixels within the bounding box). As shown in Author response image 3, perceptual size ratings are highly correlated with objective size across participants (Pearson r values predominantly ≈0.8 or higher; all p<0.001). Importantly, value ratings do not correlate with objective size (Author response image 2), confirming that the two rating scales capture distinct constructs. These checks support using participants’ size ratings as the participant-specific ground truth for defining SD in the PDM trials.

Author response image 1.

Objective size and value ratings are unrelated. Scatterplots show, for each participant, the correlation between objective image size (x-axis; proportion of non-black pixels within the item box) and value-based ratings (y-axis; 0–100 scale). Each dot is one food item (ratings averaged over the two value-rating repetitions). Across participants, value ratings do not track objective size, confirming that value and size are distinct constructs.

Author response image 2.

Perceptual size ratings closely track objective size. Scatterplots show, for each participant, the correlation between objective image size (x-axis) and perceptual area/size ratings (y-axis; 0–100 scale). Each dot is one food item (ratings averaged over the two perceptual ratings). Perceptual ratings are strongly correlated with objective size for nearly all participants (see main text), validating the use of these ratings to construct size-difference evidence (SD).

(2.2.2) More concerning was the effect of 'evidence level' on behaviour in the value-based task (Figure 3a). While the 'proportion correct' increases monotonically with the evidence level for the perceptual decisions, for the value-based task it increases from the lowest evidence level and then appears to plateau at just above 80%. This difference in behaviour between the two tasks brings into question the validity of the DDM which is used to fit the data, which assumes that the drift rate increases linearly in proportion to the level of evidence.

We agree that accuracy appears to asymptote in VDM, but the DDM fits indicate that the drift rate still increases monotonically with evidence in both tasks. In Supplementary figure 11, drift (δ) rises across the four evidence levels for PDM and for VDM (panels showing all data and pre/post-TMS). The apparent plateau in proportion correct during VDM reflects higher choice variability at stronger preference differences, not a failure of the drift–evidence mapping. Crucially, the model captures both the accuracy patterns and the RT distributions (see posterior predictive checks in Supplementary figures 11-16), indicating that a monotonic evidence–drift relation is sufficient to account for the data in each task.

Author response image 3.

HDDM parameters by evidence level. Group-level posterior means (± posterior SD) for drift (δ), boundary (α), and non-decision time (τ) across the four evidence levels, shown (a) collapsed across TMS sessions, (b) for PDM (blue) pre- vs post-TMS (light vs dark), and (c) for VDM (orange) pre- vs post-TMS. Crucially, drift increases monotonically with evidence in both tasks, while TMS selectively lowers α in PDM and reduces τ in VDM (see Supplementary Tables for numerical estimates).

(2.3) The paper provides very little information on the model fits (no parameter estimates, goodness of fit values or simulated behavioural predictions). The paper finds that TMS reduced the decision bound for perceptual decisions but only affected non-decision time for value-based decisions. It would aid the interpretation of this finding if the relative reliability of the fits for the two tasks was presented.

We appreciate the suggestion and have made the quantitative fit information explicit:

(1) Parameter estimates. Group-level means/SDs for drift (δ), boundary (α), and nDT (τ) are reported for PDM and VDM overall, by evidence level, pre- vs post-TMS, and per subject (see Supplementary Tables 8-11).

(2) Goodness of fit and predictive adequacy. DIC values accompany each fit in the tables. Posterior predictive checks demonstrate close correspondence between simulated and observed accuracy and RT distributions overall, by evidence level, and across subjects (Supplementary Figures 11-16).

Together, these materials document that the HDDM provides reliable fits in both tasks and accurately recovers the qualitative and quantitative patterns that underlie our inferences (reduced α for PDM only; selective τ reduction in VDM).

(2.4) Behaviourally, the perceptual task produced decreased response times and accuracy post-TMS, consistent with a reduced bound and consistent with some prior literature. Based on the results of the computational modelling, the authors conclude that RT differences in the value-based task are due to task-related learning, while those in the perceptual task are 'decision relevant'. It is not fully clear why there would be such significantly greater task-related learning in the value-based task relative to the perceptual one. And if such learning is occurring, could it potentially also tend to increase the consistency of choices, thereby counteracting any possible TMS-induced reduction of consistency?

Thank you for pointing out the need for a clearer framing. We have removed the speculative label “task-related learning” and now describe the pattern strictly in terms of the HDDM decomposition and neural results already reported:

(1) VDM: Post-TMS RTs are faster while accuracy is unchanged. The HDDM attributes this to a selective reduction in non-decision time (τ), with no change in decision-relevant parameters (α, δ) for VDM (see Supplementary Figure 11 and Supplementary Tables). Consistent with this, left SFS BOLD is not reduced for VDM, and trialwise SFS activity does not predict VDM accuracy—both observations argue against a change in VDM decision formation within left SFS.

(2) PDM: Post-TMS accuracy decreases and RTs shorten, which the HDDM captures as a lower decision boundary (α) with no change in drift (δ). Here, left SFS BOLD scales with accumulated evidence and decreases post-TMS, and trialwise SFS activity predicts PDM accuracy, all consistent with a decision-relevant effect in PDM.

Regarding the possibility that faster VDM RTs should increase choice consistency: empirically, consistency did not change in VDM, and the HDDM finds no decision-parameter shifts there. Thus, there is no hidden counteracting increase in VDM accuracy that could mask a TMS effect—the absence of a VDM accuracy change is itself informative and aligns with the modelling and fMRI.

Reviewer #3 (Public Review):

Summary:

Garcia et al., investigated whether the human left superior frontal sulcus (SFS) is involved in integrating evidence for decisions across either perceptual and/or value-based decision-making. Specifically, they had 20 participants perform two decision-making tasks (with matched stimuli and motor responses) in an fMRI scanner both before and after they received continuous theta burst transcranial magnetic stimulation (TMS) of the left SFS. The stimulation thought to decrease neural activity in the targeted region, led to reduced accuracy on the perceptual decision task only. The pattern of results across both model-free and model-based (Drift diffusion model) behavioural and fMRI analyses suggests that the left SLS plays a critical role in perceptual decisions only, with no equivalent effects found for value-based decisions. The DDM-based analyses revealed that the role of the left SLS in perceptual evidence accumulation is likely to be one of decision boundary setting. Hence the authors conclude that the left SFS plays a domain-specific causal role in the accumulation of evidence for perceptual decisions. These results are likely to add importance to the literature regarding the neural correlates of decision-making.

Strengths:

The use of TMS strengthens the evidence for the left SFS playing a causal role in the evidence accumulation process. By combining TMS with fMRI and advanced computational modelling of behaviour, the authors go beyond previous correlational studies in the field and provide converging behavioural, computational, and neural evidence of the specific role that the left SFS may play.

Sophisticated and rigorous analysis approaches are used throughout.

Weaknesses:

(3.1) Though the stimuli and motor responses were equalised between the perception and value-based decision tasks, reaction times (according to Figure 1) and potential difficulty (Figure 2) were not matched. Hence, differences in task difficulty might represent an alternative explanation for the effects being specific to the perception task rather than domain-specificity per se.

We agree that RTs cannot be matched a priori, and we did not intend them to be. Instead, we equated the inputs to the decision process and verified that each task relied exclusively on its task-relevant evidence. As reported in Results—Behaviour: validity of task-relevant pre-requisites (Fig. 1b–c), accuracy and RTs vary monotonically with the appropriate evidence regressor (SD for PDM; VD for VDM), with no effect of the task-irrelevant regressor. This separability check addresses differences in baseline RTs by showing that, for both tasks, behaviour tracks evidence as designed.

To rule out a generic difficulty account of the TMS effect, we relied on the within-subject differences-in-differences (DID) framework described in Methods (Differences-in-differences). The key Task × TMS interaction compares the pre→post change in PDM with the pre→post change in VDM while controlling for trialwise evidence and RT covariates. Any time-on-task or unspecific difficulty drift shared by both tasks is subtracted out by this contrast. Using this specification, TMS selectively reduced accuracy for PDM but not VDM (Fig. 3a; Supplementary Fig. 2a,c; Supplementary Tables 5–7).

Finally, the hierarchical DDM (already in the paper) dissociates latent mechanisms. The post-TMS boundary reduction appears only in PDM, whereas VDM shows a change in non-decision time without a decision-relevant parameter change (Fig. 3c; Supplementary Figs. 4–5). If unmatched difficulty were the sole driver, we would expect parallel effects across tasks, which we do not observe.

(3.2) No within- or between-participants sham/control TMS condition was employed. This would have strengthened the inference that the apparent TMS effects on behavioural and neural measures can truly be attributed to the left SFS stimulation and not to non-specific peripheral stimulation and/or time-on-task effects.

We agree that a sham/control condition would further strengthen causal attribution and note this as a limitation. In mitigation, our design incorporates several safeguards already reported in the manuscript:

· Within-subject pre/post with alternating task blocks and DID modelling (Methods) to difference out non-specific time-on-task effects.

· Task specificity across levels of analysis: behaviour (PDM accuracy reduction only), computational (boundary reduction only in PDM; no drift change), BOLD (reduced left-SFS accumulated-evidence signal for PDM but not VDM; Fig. 4a–c), and functional coupling (SFS–occipital PPI increase during PDM only; Fig. 5).

· Matched stimuli and motor outputs across tasks, so any peripheral sensations or general arousal effects should have influenced both tasks similarly; they did not.

Together, these converging task-selective effects reduce the likelihood that the results reflect non-specific stimulation or time-on-task. We will add an explicit statement in the Limitations noting the absence of sham/control and outlining it as a priority for future work.

(3.3) No a priori power analysis is presented.

We appreciate this point. Our sample size (n = 20) matched prior causal TMS and combined TMS–fMRI studies using similar paradigms and analyses (e.g., Philiastides et al., 2011; Rahnev et al., 2016; Jackson et al., 2021; van der Plas et al., 2021; Murd et al., 2021), and was chosen a priori on that basis and the practical constraints of cTBS + fMRI. The within-subject DID approach and hierarchical modelling further improve efficiency by leveraging all trials.

To address the reviewer’s request for transparency, we will (i) state this rationale in Methods—Participants, and (ii) ensure that all primary effects are reported with 95% CIs or posterior probabilities (already provided for the HDDM as pmcmcp_{\mathrm{mcmc}}pmcmc). We also note that the design was sensitive enough to detect RT changes in both tasks and a selective accuracy change in PDM, arguing against a blanket lack of power as an explanation for null VDM accuracy effects. We will nevertheless flag the absence of a formal prospective power analysis in the Limitations.

Recommendations for the Authors:

Reviewer #1 (Recommendations For The Authors):

Some important elements of the methods are missing. How was the site for targeting the SFS with TMS identified? The methods described how M1 was located but not SFS.

Thank you for catching this omission. In the revised Methods we explicitly describe how the left SFS target was localized. Briefly, we used each participant’s T1-weighted anatomical scan and frameless neuronavigation to place a 10-mm sphere at the a priori MNI coordinates (x = −24, y = 24, z = 36) derived from prior work (Heekeren et al., 2004; Philiastides et al., 2011). This sphere was transformed to native space for each participant. The coil was positioned tangentially with the handle pointing posterior-lateral, and coil placement was continuously monitored with neuronavigation throughout stimulation. (All of these procedures mirror what we already report for M1 and are now stated for SFS as well.)

Where to revise the manuscript:

Methods → Stimulation protocol. After the first sentence naming cTBS, insert:<br /> “The left SFS target was localized on each participant’s T1-weighted anatomical image using frameless neuronavigation. A 10-mm radius sphere was centered at the a priori MNI coordinates x = −24, y = 24, z = 36 (Heekeren et al., 2004; Philiastides et al., 2011), then transformed to native space. The MR-compatible figure-of-eight coil was positioned tangentially over the target with the handle oriented posterior-laterally, and its position was tracked and maintained with neuronavigation during stimulation.”

It is not clear how participants were instructed that they should perform the value-difference task. Were they told that they should choose based on their original item value ratings or was it left up to them?

We agree the instruction should be explicit. Participants were told_: “In value-based blocks, choose the item you would prefer to eat at the end of the experiment.”_ They were informed that one VDM trial would be randomly selected for actual consumption, ensuring incentive-compatibility. We did not ask them to recall or follow their earlier ratings; those ratings were used only to construct evidence (value difference) and to define choice consistency offline.

Where to revise the manuscript:

Methods → Experimental paradigm.

Add a sentence to the VDM instruction paragraph:

“In value-based (LIKE) blocks, participants were instructed to choose the item they would prefer to consume at the end of the experiment; one VDM trial was randomly selected and implemented, making choices incentive-compatible. Prior ratings were used solely to construct value-difference evidence and to score choice consistency; participants were not asked to recall or match their earlier ratings.”

Line 86 Introduction, some previous studies were conducted on animals. Why it is problematic that the studies were conducted in animals is not stated. I assume the authors mean that we do not know if their findings will translate to the human brain? I think in fairness to those working with animals it might be worth an extra sentence to briefly expand on this point.

We appreciate this and will clarify that animal work is invaluable for circuit-level causality, but species differences and putative non-homologous areas (e.g., human SFS vs. rodent FOF) limit direct translation. Our point is not that animal studies are problematic, but that establishing causal roles in humans remains necessary.

Revision:

Introduction (paragraph discussing prior animal work). Replace the current sentence beginning “However, prior studies were largely correlational”

“Animal studies provide critical causal insights, yet direct translation to humans can be limited by species-specific anatomy and potential non-homologies (e.g., human SFS vs. frontal orienting fields in rodents). Therefore, establishing causal contributions in the human brain remains essential.”

Line 100-101: "or whether its involvement is peripheral and merely functionally supporting a larger system" - it is not clear what you mean by 'supporting a larger system'

We meant that observed SFS activity might reflect upstream/downstream support processes (e.g., attentional control or working-memory maintenance) rather than the computation of evidence accumulation itself. We have rephrased to avoid ambiguity.

Revision:

Introduction. Replace the phrase with:

“or whether its observed activity reflects upstream or downstream support processes (e.g., attention or working-memory maintenance) rather than the accumulation computation per se.”

The authors do have to make certain assumptions about the BOLD patterns that would be expected of an evidence accumulation region. These assumptions are reasonable and have been adopted in several previous neuroimaging studies. Nevertheless, it should be acknowledged that alternative possibilities exist and this is an inevitable limitation of using fMRI to study decision making. For example, if it turns out that participants collapse their boundaries as time elapses, then the assumption that trials with weaker evidence should have larger BOLD responses may not hold - the effect of more prolonged activity could be cancelled out by the lower boundaries. Again, I think this is just a limitation that could be acknowledged in the Discussion, my opinion is that this is the best effort yet to identify choice-relevant regions with fMRI and the authors deserve much credit for their rigorous approach.

Agreed. We already ground our BOLD regressors in the DDM literature, but acknowledge that alternative mechanisms (e.g., time-dependent boundaries) can alter expected BOLD–evidence relations. We now add a short limitation paragraph stating this explicitly.

Revision:

Discussion (limitations paragraph). Add:

“Our fMRI inferences rest on model-based assumptions linking accumulated evidence to BOLD amplitude. Alternative mechanisms—such as time-dependent (collapsing) boundaries—could attenuate the prediction that weaker-evidence trials yield longer accumulation and larger BOLD signals. While our behavioural and neural results converge under the DDM framework, we acknowledge this as a general limitation of model-based fMRI.”

Reviewer #2 (Recommendations For The Authors):

Minor points

I suggest the proportion of missed trials should be reported.

Thank you for the suggestion. In our preprocessing we excluded trials with no response within the task’s response window and any trials failing a priori validity checks. Because non-response trials contain neither a choice nor an RT, they are not entered into the DDM fits or the fMRI GLMs and, by design, carry no weight in the reported results. To keep the focus on the data that informed all analyses, we now (i) state the trial-inclusion criteria explicitly and (ii) report the number of analysed (valid) trials per task and run. This conveys the effective sample size contributing to each condition without altering the analysis set.

Revision:

Methods → (at the end of “Experimental paradigm”): “Analyses were conducted on valid trials only, defined as trials with a registered response within the task’s response window and passing pre-specified validity checks; trials without a response were excluded and not analysed.”

Results → “Behaviour: validity of task-relevant pre-requisites” (add one sentence at the end of the first paragraph): “All behavioural and fMRI analyses were performed on valid trials only (see Methods for inclusion criteria).”

Figure 4 c is very confusing. Is the legend or caption backwards?

Thanks for flagging. We corrected the Figure 4c caption to match the colouring and contrasts used in the panel (perceptual = blue/green overlays; value-based = orange/red; ‘post–pre’ contrasts explicitly labeled). No data or analyses were changed, just the wording to remove ambiguity.

Revision:

Figure 4 caption (panel c sentence). Replace with:

“(c) Post–pre contrasts for the trialwise accumulated-evidence regressor show reduced left-SFS BOLD during perceptual decisions (green overlay), with a significantly stronger reduction for perceptual vs value-based decisions (blue overlay). No reduction is observed for value-based decisions.”

Even if not statistically significant it may be of interest to add the results for Value-based decision making on SFS in Supplementary Table 3.

Done. We now include the SFS small-volume results for VDM (trialwise accumulated-evidence regressor) alongside the PDM values in the same table, with exact peak, cluster size, and statistics.

Revision:

Supplementary Table 3 (title):

“Regions encoding trialwise accumulated evidence (parametric modulation) during perceptual and value-based decisions, including SFS SVC results for both tasks.”

Model comparisons: please explain how model complexity is accounted for.

We clarify that model evidence was compared using the Deviance Information Criterion (DIC), which penalizes model fit by an effective number of parameters (pD). Lower DIC indicates better out-of-sample predictive performance after accounting for model complexity.

Revision:

Methods → Hierarchical Bayesian neural-DDM (last paragraph). Add:

“Model comparison used the Deviance Information Criterion (DIC = D̄ + pD), where pD is the effective number of parameters; thus DIC penalizes model complexity. Lower DIC denotes better predictive accuracy after accounting for complexity.”

Reviewer #3 (Recommendations For The Authors):

The following issues would benefit from clarification in the manuscript:

- It is stated that "Our sample size is well within acceptable range, similar to that of previous TMS studies." The sample size being similar to previous studies does not mean it is within an acceptable range. Whether the sample size is acceptable or not depends on the expected effect size. It is perfectly possible that the previous studies cited were all underpowered. What implications might the lack of an a priori power analysis have for the interpretation of the results?

We agree and have revised our wording. We did not conduct an a priori power analysis. Instead, we relied on a within-participant design that typically yields higher sensitivity in TMS–fMRI settings and on convergence across behavioural, computational, and neural measures. We now acknowledge that the absence of formal power calculations limits claims about small effects (particularly for null findings in VDM), and we frame those null results cautiously.

Revision:

Discussion (limitations). Add:

“The within-participant design enhances statistical sensitivity, yet the absence of an a priori power analysis constrains our ability to rule out small effects, particularly for null results in VDM.”

- I was confused when trying to match the results described in the 'Behaviour: validity of task-relevant pre-requisites' section on page 6 to what is presented in Figure 1. Specifically, Figure 1C is cited 4 times but I believe two of these should be citing Figure 1B?

Thank you—this was a citation mix-up. The two places that referenced “Fig. 1C” but described accuracy should in fact point to Fig. 1B. We corrected both citations.

Revision:

Results → Behaviour: validity… Change the two incorrect “Fig. 1C” references (when describing accuracy) to “Fig. 1B”.

- Also, where is the 'SD' coefficient of -0.254 (p-value = 0.123) coming from in line 211? I can't match this to the figure.

This was a typographical error in an earlier draft. The correct coefficients are those shown in the figure and reported elsewhere in the text (evidence-specific effects: for PDM RTs, SD β = −0.057, p < 0.001; for VDM RTs, VD β = −0.016, p = 0.011; non-relevant evidence terms are n.s.). We removed the erroneous value.

Revision:

Results → Behaviour: validity… (sentence with −0.254). Delete the incorrect value and retain the evidence-specific coefficients consistent with Fig. 1B–C.

- It is reported that reaction times were significantly faster for the perceptual relative to the value-based decision task. Was overall accuracy also significantly different between the two tasks? It appears from Figure 3 that it might be, But I couldn't find this reported in the text.

To avoid conflating task with evidence composition, we did not emphasize between-task accuracy averages. Our primary tests examine evidence-specific effects and TMS-induced changes within task. For completeness, we now report descriptive mean accuracies by task and point readers to the figure panels that display accuracy as a function of evidence (which is the meaningful comparison in our matched-evidence design). We refrain from additional hypothesis testing here to keep the analyses aligned with our preregistered focus.

Revision:

Results → Behaviour: validity… Add:

“For completeness, group-mean accuracies by task are provided descriptively in Fig. 3a; inferential tests in the manuscript focus on evidence-specific effects and TMS-induced changes within task.”

relationship is used when exposure and outcome are continious, difference when atleast one is a catergory

I think it’s amazing how connected the world was back then! China introduced tea to Japan in 1191, and Chinese ships were sailing all the way to East Africa—something we know from both Chinese and Arab writers. It really shows how far their influence and trade networks reached.

I find it so cool how these schools taught the seven liberal arts—the trivium (grammar, rhetoric, logic) and the quadrivium (arithmetic, geometry, music, astronomy)—along with scripture and liturgy. Plus, big cathedral centers like Chartres and Paris opened their schools to students who weren’t monks, which helped education spread even more during this time.

I think it’s really striking how different Saladin’s approach was in 1187. Unlike the Crusaders who brutally sacked Jerusalem in 1099, Saladin didn’t destroy the city and even allowed the Christians living there to leave safely and return home. It really shows how differently the two sides handled victory.

It’s so interesting how around the mid-11th century, Constantine had to flee North Africa—either because of political chaos or even accusations of sorcery—and ended up in Salerno, Italy. Since Salerno was already a famous medical center, he started translating important Arabic and Persian medical texts into Latin, which helped spread that knowledge across Europe.

I don't know what this part is doing, I assume you want to returns a matrix of p-values, but I don't think you need p-values in a matrix table. Cause I can tell from your plot, If p < .05, it show the correlation, but If p > .05, it won't show anything.

If your goal is to evaluate the relationships between all variables, it may be more informative to display the full correlation matrix without masking non-significant values. In that case, you can remove the significance filter so the plot shows every correlation.

resize the number that shows on the plot

if you want, you can add another layer "geom_jitter()", adds a small amount of random variation to the location of each point, so we can clearly visualizes data density.

(But maybe ask Dr.Shane before you add this)

same here, explain the meaning of this diagram and what the differences are before and after the intervention.

Please explain the meaning of this diagram and what the differences are before and after the intervention.

add one or two sentence to explain this alpha results, something like: "PREZEROSUM scale has borderline internal consistency, and the POSTZEROSUM scale has weak internal consistency, meaning the items may not be measuring the construct reliably."

use #| warning: false to force the chunk not showing the warning message.

Same here, Transform should be in the results section, so double ##Trandform

Import should be in the results section, so double ##Import

Discussion section should be at the end of the report

This is the thesis. The nursery is the villain from this perspective. This Thesis also aligns with my own essay in terms of emotional detachment from the parents.

This is the thesis. The nursery is the villain from this perspective. This Thesis also aligns with my own essay in terms of emotional detachment from the parents.

It seems the author will be using a Formalist approach to this essay.

eLife Assessment

This important study presents a cross-species and cross-disciplinary analysis of cortical folding. The authors use a combination of physical gel models, computational simulations, and morphometric analysis, extending prior work in human brain development to macaques and ferrets. The findings support the hypothesis that mechanical forces driven by differential growth can account for major aspects of gyrification. The evidence presented is overall strong and convincingly supports the central claims; the findings will be of broad interest in developmental neuroscience.

Reviewer #1 (Public review):

The manuscript by Yin and colleagues addresses a long-standing question in the field of cortical morphogenesis, regarding factors that determine differential cortical folding across species and individuals with cortical malformations. The authors present work based on a computational model of cortical folding evaluated alongside a physical model that makes use of gel swelling to investigate the role of a two-layer model for cortical morphogenesis. The study assesses these models against empirically derived cortical surfaces based on MRI data from ferret, macaque monkey, and human brains.

The manuscript is clearly written and presented, and the experimental work (physical gel modeling as well as numerical simulations) and analyses (subsequent morphometric evaluations) are conducted at the highest methodological standards. It constitutes an exemplary use of interdisciplinary approaches for addressing the question of cortical morphogenesis by bringing together well-tuned computational modeling with physical gel models. In addition, the comparative approaches used in this paper establish a foundation for broad-ranging future lines of work that investigate the impact of perturbations or abnormalities during cortical development.

The cross-species approach taken in this study is a major strength of the work. However, correspondence across the two methodologies did not appear to be equally consistent in predicting brain folding across all three species. The results presented in Figures 4 (and Figures S3 & S4) show broad correspondence in shape index and major sulci landmarks across all three species. Nevertheless, the results presented for the human brain lack the same degree of clear correspondence for the gel model results as observed in the macaque and ferret. While this study clearly establishes a strong foundation for comparative cortical anatomy across species and the impact of perturbations on individual morphogenesis, further work that fine-tunes physical modeling of complex morphologies, such as that of the human cortex, may help to further understand the factors that determine cortical functionalization and pathologies.

Reviewer #2 (Public review):

This manuscript explores the mechanisms underlying cerebral cortical folding using a combination of physical modelling, computational simulations, and geometric morphometrics. The authors extend their prior work on human brain development (Tallinen et al., 2014; 2016) to a comparative framework involving three mammalian species: ferrets (Carnivora), macaques (Old World monkeys), and humans (Hominoidea). By integrating swelling gel experiments with mathematical differential growth models, they simulate sulcification instability and recapitulate key features of brain folding across species. The authors make commendable use of publicly available datasets to construct 3D models of fetal and neonatal brain surfaces: fetal macaque (ref. [26]), newborn ferret (ref. [11]), and fetal human (ref. [22]).

Using a combination of physical models and numerical simulations, the authors compare the resulting folding morphologies to real brain surfaces using morphometric analysis. Their results show qualitative and quantitative concordance with observed cortical folding patterns, supporting the view that differential tangential growth of the cortex relative to the subcortical substrate is sufficient to account for much of the diversity in cortical folding. This is a very important point in our field, and can be used in the teaching of medical students.

Brain folding remains a topic of ongoing debate. While some regard it as a critical specialization linked to higher cognitive function, others consider it an epiphenomenon of expansion and constrained geometry. This divergence was evident in discussions during the Strüngmann Forum on cortical development (Silver et al., 2019). Though folding abnormalities are reliable indicators of disrupted neurodevelopmental processes (e.g., neurogenesis, migration), their relationship to functional architecture remains unclear. Recent evidence suggests that the absolute number of neurons varies significantly with position-sulcus versus gyrus-with potential implications for local processing capacity (e.g., https://doi.org/10.1002/cne.25626). The field is thus in need of comparative, mechanistic studies like the present one.

This paper offers an elegant and timely contribution by combining gel-based morphogenesis, numerical modelling, and morphometric analysis to examine cortical folding across species. The experimental design - constructing two-layer PDMS models from 3D MRI data and immersing them in organic solvents to induce differential swelling - is well-established in prior literature. The authors further complement this with a continuum mechanics model simulating folding as a result of differential growth, as well as a comparative analysis of surface morphologies derived from in vivo, in vitro, and in silico brains.

Conclusion:

This is a well-executed and creative study that integrates diverse methodologies to address a longstanding question in developmental neurobiology. While a few aspects-such as regional folding peculiarities, sensitivity to initial conditions, and available human data-could be further elaborated, they do not detract from the overall quality and novelty of the work. I enthusiastically support this paper and believe that it will be of broad interest to the neuroscience, biomechanics, and developmental biology communities.

[Editor's note: The reviewers were satisfied with the authors' response. The eLife Assessment was slightly updated to reflect the author's response.]

Author response:

The following is the authors’ response to the original reviews

Public Reviews:

Reviewer #1 (Public review):

The manuscript by Yin and colleagues addresses a long-standing question in the field of cortical morphogenesis, regarding factors that determine differential cortical folding across species and individuals with cortical malformations. The authors present work based on a computational model of cortical folding evaluated alongside a physical model that makes use of gel swelling to investigate the role of a two-layer model for cortical morphogenesis. The study assesses these models against empirically derived cortical surfaces based on MRI data from ferret, macaque monkey, and human brains.

The manuscript is clearly written and presented, and the experimental work (physical gel modeling as well as numerical simulations) and analyses (subsequent morphometric evaluations) are conducted at the highest methodological standards. It constitutes an exemplary use of interdisciplinary approaches for addressing the question of cortical morphogenesis by bringing together well-tuned computational modeling with physical gel models. In addition, the comparative approaches used in this paper establish a foundation for broad-ranging future lines of work that investigate the impact of perturbations or abnormalities during cortical development.

The cross-species approach taken in this study is a major strength of the work. However, correspondence across the two methodologies did not appear to be equally consistent in predicting brain folding across all three species. The results presented in Figures 4 (and Figures S3 and S4) show broad correspondence in shape index and major sulci landmarks across all three species. Nevertheless, the results presented for the human brain lack the same degree of clear correspondence for the gel model results as observed in the macaque and ferret. While this study clearly establishes a strong foundation for comparative cortical anatomy across species and the impact of perturbations on individual morphogenesis, further work that fine-tunes physical modeling of complex morphologies, such as that of the human cortex, may help to further understand the factors that determine cortical functionalization and pathologies.

We thank the reviewer for positive opinions and helpful comments. Yes, the physical gel model of the human brain has a lower similarity index with the real brain. There are several reasons.

First, the highly convoluted human cortex has a few major folds (primary sulci) and a very large number of minor folds associated with secondary or tertiary sulci (on scales of order comparable to the cortical thickness), relative to the ferret and macaque cerebral cortex. In our gel model, the exact shapes, positions, and orientations of these minor folds are stochastic, which makes it hard to have a very high similarity index of the gel models when compared with the brain of a single individual.

Second, in real human brains, these minor folds evolve dynamically with age and show differences among individuals. In experiments with the gel brain, multiscale folds form and eventually disappear as the swelling progresses through the thickness. Our physical model results are snapshots during this dynamical process, which makes it hard to have a concrete one-to-one correspondence between the instantaneous shapes of the swelling gel and the growing human brain.

Third, the growth of the brain cortex is inhomogeneous in space and varying with time, whereas, in the gel model, swelling is relatively homogeneous.

We agree that further systematic work, based on our proposed methods, with more fine-tuned gel geometries and properties, might provide a deeper understanding of the relations between brain geometry, and growth-induced folds and their functionalization and pathologies. Further analysis of cortical pathologies using computational and physical gel models can be found in our companion paper (Choi et al., 2025), also published in eLife:

G. P. T. Choi, C. Liu, S. Yin, G. Séjourné, R. S. Smith, C. A. Walsh, L. Mahadevan, Biophysical basis for brain folding and misfolding patterns in ferrets and humans. eLife, 14, RP107141, 2025. doi:10.7554/eLife.107141

Reviewer# 2 (Public review):

This manuscript explores the mechanisms underlying cerebral cortical folding using a combination of physical modelling, computational simulations, and geometric morphometrics. The authors extend their prior work on human brain development (Tallinen et al., 2014; 2016) to a comparative framework involving three mammalian species: ferrets (Carnivora), macaques (Old World monkeys), and humans (Hominoidea). By integrating swelling gel experiments with mathematical differential growth models, they simulate sulcification instability and recapitulate key features of brain folding across species. The authors make commendable use of publicly available datasets to construct 3D models of fetal and neonatal brain surfaces: fetal macaque (ref. [26]), newborn ferret (ref. [11]), and fetal human (ref. [22]).

Using a combination of physical models and numerical simulations, the authors compare the resulting folding morphologies to real brain surfaces using morphometric analysis. Their results show qualitative and quantitative concordance with observed cortical folding patterns, supporting the view that differential tangential growth of the cortex relative to the subcortical substrate is sufficient to account for much of the diversity in cortical folding. This is a very important point in our field, and can be used in the teaching of medical students.

Brain folding remains a topic of ongoing debate. While some regard it as a critical specialization linked to higher cognitive function, others consider it an epiphenomenon of expansion and constrained geometry. This divergence was evident in discussions during the Strungmann Forum on cortical development (Silver¨ et al., 2019). Though folding abnormalities are reliable indicators of disrupted neurodevelopmental processes (e.g., neurogenesis, migration), their relationship to functional architecture remains unclear. Recent evidence suggests that the absolute number of neurons varies significantly with position-sulcus versus gyrus-with potential implications for local processing capacity (e.g., https://doi.org/10.1002/cne.25626). The field is thus in need of comparative, mechanistic studies like the present one.

This paper offers an elegant and timely contribution by combining gel-based morphogenesis, numerical modelling, and morphometric analysis to examine cortical folding across species. The experimental design - constructing two-layer PDMS models from 3D MRI data and immersing them in organic solvents to induce differential swelling - is well-established in prior literature. The authors further complement this with a continuum mechanics model simulating folding as a result of differential growth, as well as a comparative analysis of surface morphologies derived from in vivo, in vitro, and in silico brains.

We thank the reviewer for the very positive comments.

I offer a few suggestions here for clarification and further exploration:

Major Comments

(1) Choice of Developmental Stages and Initial Conditions

The authors should provide a clearer justification for the specific developmental stages chosen (e.g., G85 for macaque, GW23 for human). How sensitive are the resulting folding patterns to the initial surface geometry of the gel models? Given that folding is a nonlinear process, early geometric perturbations may propagate into divergent morphologies. Exploring this sensitivity-either through simulations or reference to prior work-would enhance the robustness of the findings.

The initial geometry is one of the important factors that decides the final folding pattern. The smooth brain in the early developmental stage shows a broad consistency across individuals, and we expect the main folds to form similarly across species and individuals.

Generally, we choose the initial geometry when the brain cortex is still relatively smooth. For the human, this corresponds approximately to GW23, as the major folds such as the Rolandic fissure (central sulcus), arise during this developmental stage. For the macaque brain, we chose developmental stage G85, primarily because of the availability of the dataset corresponding to this time, which also corresponds to the least folded.

We expect that large-scale folding patterns are strongly sensitive to the initial geometry but fine-scale features are not. Since our goal is to explain the large-scale features, we expect sensitivity to the initial shape.

Below are some references of other researchers that are consistent with this idea. Figure 4 from Wang et al. shows some images of simulations obtained by perturbing the geometry of a sphere to an ellipsoid. We see that the growth-induced folds mostly maintain their width (wavelength), but change their orientations.

Reference:

Wang, X., Lefévre, J., Bohi, A., Harrach, M.A., Dinomais, M. and Rousseau, F., 2021. The influence of biophysical parameters in a biomechanical model of cortical folding patterns. Scientific Reports, 11(1), p.7686.

Related results from the same group show that slight perturbations of brain geometry, cause these folds also tend to change their orientations but not width/wavelength (Bohi et al., 2019).

Reference:

Bohi, A., Wang, X., Harrach, M., Dinomais, M., Rousseau, F. and Lefévre, J., 2019, July. Global perturbation of initial geometry in a biomechanical model of cortical morphogenesis. In 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) (pp. 442-445). IEEE.

Finally, a systematic discussion of the role of perturbations on the initial geometries and physical properties can be seen in our work on understanding a different system, gut morphogenesis (Gill et al., 2024).

We have added the discussion about geometric sensitivity in the section Methods-Numerical Simulations:

“Small perturbations on initial geometry would affect minor folds, but the main features of major folds, such as orientations, width, and depth, are expected to be conserved across individuals [49, 50]. For simplicity, we do not perturb the fetal brain geometry obtained from datasets.”

(2) Parameter Space and Breakdown Points

The numerical model assumes homogeneous growth profiles and simplifies several aspects of cortical mechanics. Parameters such as cortical thickness, modulus ratios, and growth ratios are described in Table II. It would be informative to discuss the range of parameter values for which the model remains valid, and under what conditions the physical and computational models diverge. This would help delineate the boundaries of the current modelling framework and indicate directions for refinement.

Exploring the valid parameter space is a key problem. We have tested a series of growth parameters and will state them explicitly in our revision. In the current version, we chose the ones that yield a relatively high similarity index to the animal brains. More generally, folding patterns are largely regulated by geometry as well as physical parameters, such as cortical thickness, modulus ratios, growth ratios, and inhomogeneity. In our previous work on a different system, gut morphogenesis, where similar folding patterns are seen, we have explored these features (Gill et al., 2024).

Reference:

Gill, H.K., Yin, S., Nerurkar, N.L., Lawlor, J.C., Lee, C., Huycke, T.R., Mahadevan, L. and Tabin, C.J., 2024. Hox gene activity directs physical forces to differentially shape chick small and large intestinal epithelia. Developmental Cell, 59(21), pp.2834-2849.

(3) Neglected Regional Features: The Occipital Pole of the Macaque

One conspicuous omission is the lack of attention to the occipital pole of the macaque, which is known to remain smooth even at later gestational stages and has an unusually high neuronal density (2.5× higher than adjacent cortex). This feature is not reproduced in the gel or numerical models, nor is it discussed. Acknowledging this discrepancy-and speculating on possible developmental or mechanical explanationswould add depth to the comparative analysis. The authors may wish to include this as a limitation or a target for future work.

Yes, we have added that the omission of the Occipital Pole of the macaque is one of our paper’s limitations. Our main aim in this paper is to explore the formation of large-scale folds, so the smooth region is not discussed. But future work could include this to make the model more complete.

The main text has been modified in Methods, Numerical simulations:

“To focus on fold formation, we did not discuss the relatively smooth region, such as the Occipital Pole of the macaque.”

and also in the caption of Figure 4: “... The occipital pole region of macaque brains remains smooth in real and simulated brains.”

(4) Spatio-Temporal Growth Rates and Available Human Data

The authors note that accurate, species-specific spatio-temporal growth data are lacking, limiting the ability to model inhomogeneous cortical expansion. While this may be true for ferret and macaque, there are high-quality datasets available for human fetal development, now extended through ultrasound imaging (e.g., https://doi.org/10.1038/s41586-023-06630-3). Incorporating or at least referencing such data could improve the fidelity of the human model and expand the applicability of the approach to clinical or pathological scenarios.

We thank the reviewer for pointing out the very useful datasets that exist for the exploration of inhomogeneous growth driven folding patterns. We have referred to this paper to provide suggestions for further work in exploring the role of growth inhomogeneities.

We have referred to this high-quality dataset in our main text, Discussion:

“...the effect of inhomogeneous growth needs to be further investigated by incorporating regional growth of the gray and white matter not only in human brains [29, 31] based on public datasets [45], but also in other species.”

A few works have tried to incorporate inhomogeneous growth in simulating human brain folding by separating the central sulcus area into several lobes (e.g., lobe parcellation method, Wang, PhD Thesis, 2021). Since our goal in this paper is to explain the large-scale features of folding in a minimal setting, we have kept our model simple and show that it is still capable of capturing the main features of folding in a range of mammalian brains.

Reference:

Xiaoyu Wang. Modélisation et caractérisation du plissement cortical. Signal and Image Processing. Ecole nationale superieure Mines-Télécom Atlantique, 2021. English. 〈NNT : 2021IMTA0248〉.

(5) Future Applications: The Inverse Problem and Fossil Brains

The authors suggest that their morphometric framework could be extended to solve the inverse growth problem-reconstructing fetal geometries from adult brains. This speculative but intriguing direction has implications for evolutionary neuroscience, particularly the interpretation of fossil endocasts. Although beyond the scope of this paper, I encourage the authors to elaborate briefly on how such a framework might be practically implemented and validated.

For the inverse problem, we could use the following strategies:

a. Perform systematic simulations using different geometries and physical parameters to obtain the variation in morphologies as a function of parameters.

b. Using either supervised training or unsupervised training (physics-informed neural networks, PINNs) to learn these characteristic morphologies and classify their dependence on the parameters using neural networks. These can then be trained to determine the possible range of geometrical and physical parameters that yield buckled patterns seen in the systematic simulations.

c. Reconstruct the 3D surface from fossil endocasts. Using the well-trained neural network, it should be possible to predict the initial shape of the smooth brain cortex, growth profile, and stiffness ratio of the gray and white matter.

As an example in this direction, supervised neural networks have been used recently to solve the forward problem to predict the buckling pattern of a growing two-layer system (Chavoshnejad et al., 2023). The inverse problem can then be solved using machine-learning methods when the training datasets are the folded shape, which are then used to predict the initial geometry and physical properties.

Reference:

Chavoshnejad, P., Chen, L., Yu, X., Hou, J., Filla, N., Zhu, D., Liu, T., Li, G., Razavi, M.J. and Wang, X., 2023. An integrated finite element method and machine learning algorithm for brain morphology prediction. Cerebral Cortex, 33(15), pp.9354-9366.

Conclusion

This is a well-executed and creative study that integrates diverse methodologies to address a longstanding question in developmental neurobiology. While a few aspects-such as regional folding peculiarities, sensitivity to initial conditions, and available human data-could be further elaborated, they do not detract from the overall quality and novelty of the work. I enthusiastically support this paper and believe that it will be of broad interest to the neuroscience, biomechanics, and developmental biology communities.

Note: The paper mentions a companion paper [reference 11] that explores the cellular and anatomical changes in the ferret cortex. I did not have access to this manuscript, but judging from the title, this paper might further strengthen the conclusions.

The companion paper (Choi et al., 2025) has also been submitted to eLife and can be found here:

G. P. T. Choi, C. Liu, S. Yin, G. Séjourné, R. S. Smith, C. A. Walsh, L. Mahadevan, Biophysical basis for brain folding and misfolding patterns in ferrets and humans. eLife, 14, RP107141, 2025. doi:10.7554/eLife.107141

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

This study was conducted and presented to the highest methodological standards. It is clearly written, and the results are thoroughly presented in the main manuscript and supplementary materials. Nevertheless, I would present the following minor points and comments for consideration by the authors prior to finalizing their work:

We thank the reviewer for positive opinions and helpful comments.

(1) Where did the MRI-based cortical surface data come from? Specifically, it would be helpful to include more information regarding whether the surfaces were reconstructed based on individual- or group-level data. It appears the surfaces were group-level, and, if so, accounting for individual-level cortical folding may be a fruitful direction for future work.

The surface data come from public database, which are stated in the Methods Section. “We used a publicly available database for all our 3d reconstructions: fetal macaque brain surfaces are obtained from Liu et al. (2020); newborn ferret brain surfaces are obtained from Choi et al. (2025); and fetal human brain surfaces are obtained from Tallinen et al. (2016).”

These surfaces are reconstructed based on group-level data. Specifically, the macaque atlas images are constructed for brains at gestational ages of 85 days (G85, N \=18_, 9 females), 110 days (G110, _N \=10_, 7 females) and 135 days (G135, _N \=16_,_ 7 females). And yes, future work may focus on individual-level cortical folding, and we expect that more specific results could be found.

(2) One methodological approach for assessing consistency of cortical folding within species might be an evaluation of cross-hemispheric symmetry. I would find this particularly interesting with respect to the gel models, as it could complement the quantification of variation with respect to the computationally derived and real surfaces.

Yes, the cross-hemispheric symmetry comparison can be done by our morphometric analysis method. We have added the results of ferret brain’s left-right symmetry for gel models, simulations, and real surfaces in the supplementary material. A typical conformal mapping figure and the similarity index table are shown here.

(3) Was there a specific reason to reorder the histogram plots in Figure 4c to macaque, ferret, human rather than to maintain the order presented in Figure 4a/b of ferret, macaque, human? I appreciate that this is a minor concern, and all subplots are indeed properly titled, but consistent order may improve clarity.

We have reordered the histogram plots to make all the figure orders consistent.

Reviewer #2 (Recommendations for the authors):

(1) Please consider revising the caption of Figure 1 (or equivalent figures) to explicitly state whether features such as the macaque occipital flatness were reproduced or not.

We thank the reviewer for pointing out the macaque occipital flatness.

Author response table 1.

Left-right similarity index evaluated by comparing the shape index of ferret brains, calculated with vector P-NORM p\=2,

Author response image 1.

Left-right similarity index of ferret brains

Occipital Pole of the macaque remains relatively smooth in both real brains and computational models. But our main aim in this paper is to explore the large-scale folds formation, so the smooth region is not discussed in depth. But future work could include this to make the model more complete.

(2) Some figures could benefit from clearer labelling to distinguish between in vivo, in vitro, and in silico results.

We have supplemented some texts in panels to make the labelling clearer.

(3) The manuscript would benefit from a short paragraph in the Discussion reflecting on how future incorporation of regional heterogeneities might improve model fidelity.

We have added a sentence in the Discussion Section about improving the model fidelity by considering regional heterogeneities.

“Future more accurate models incorporating spatio-temporal inhomogeneous growth profiles and mechanical properties, such as varying stiffness, would make the folding pattern closer to the real cortical folding. This relies on more in vivo measurements of the brain’s physical properties and cortical expansion.”

(4) Suggestions for improved or additional experiments, data, or analyses.

(5) Clarify and justify the selection of developmental stages: The authors should explain why particular gestational stages (e.g., G85 for macaque, GW23 for human) were chosen as starting points for the physical and computational models. A discussion of how sensitive the folding patterns are to the initial geometry would help assess the robustness of the model. If feasible, a brief sensitivity analysis-varying initial age or surface geometry-would strengthen the conclusions.

The initial geometry is one of the important factors that decides the final folding pattern. The smooth brain in the early developmental stage shows a broad consistency across individuals, and we expect the main folds to form similarly across species and individuals.

Generally, we choose the initial geometry when the brain cortex is still relatively smooth. For the human, this corresponds approximately to GW23, as the major folds such as the Rolandic fissure (central sulcus), arise during this developmental stage. For the macaque brain, we chose developmental stage G85, primarily because of the availability of the dataset corresponding to this time, which also corresponds to the least folded.

We expect that large-scale folding patterns are strongly sensitive to the initial geometry but fine-scale features are not. Since our goal is to explain the large-scale features, we expect sensitivity to the initial shape.

We have added the discussion about geometric sensitivity in the section Methods-Numerical Simulations: “Small perturbations on initial geometry would affect minor folds, but the main features of major folds, such as orientations, width, and depth, are expected to be conserved across individuals [49, 50]. For simplicity, we do not perturb the fetal brain geometry obtained from datasets.”

(6) Explore parameter boundaries more explicitly: The paper would benefit from a clearer account of the ranges of mechanical and geometric parameters (e.g., growth ratios, cortical thickness) for which the model holds. Are there specific conditions under which the physical and numerical models diverge? Identifying breakdown points would help readers understand the model’s limitations and applicability.

Exploring the valid parameter space is a key problem. We have tested a series of growth parameters and will state them explicitly in our revision. In the current version, we chose the ones that yield a relatively high similarity index to the animal brains. More generally, folding patterns are largely regulated by geometry as well as physical parameters, such as cortical thickness, modulus ratios, and growth ratios and inhomogeneity. In our previous work on a different system, gut morphogenesis, where similar folding patterns are seen, we have explored these features (Gill et al., 2024).

(7) Address species-specific cortical peculiarities: A striking omission is the flat occipital pole of the macaque, which is not reproduced in the physical or computational models. Given its known anatomical and cellular distinctiveness, this discrepancy warrants discussion. Even if not explored experimentally, the authors could speculate on what developmental or mechanical conditions would be needed to reproduce such regional smoothness.

Please refer to our answer to the public reviewer 2, question (3). From our results, the formation of smooth Occipital Pole might indicate that the spatio-temporal growth rate of gray and white matter are consistent in this region, such that there’s no much differential growth.

(8) Consider integration of available human growth data: While the authors note the lack of spatiotemporal growth data across species, such datasets exist for human fetal brain development, including those from MRI and ultrasound studies (e.g., Nature 2023). Incorporating these into the human model-or at least discussing their implications-would enhance biological relevance.

Yes, some datasets for fetal human brains have provided very comprehensive measurements on brain shapes at many developmental stages. This can surely be implemented in our current model by calculating the spatio-temporal growth rate from regional cortical shapes at different stages.

(9) Recommendations for improving the writing and presentation:

a) The manuscript is generally well-written, but certain sections would benefit from more explicit linksbetween the biological phenomena and the modeling framework. For instance, the Introduction and Discussion could more clearly articulate how mechanical principles interface with genetic or cellular processes, especially in the context of evolution and developmental variation.

We have briefly discussed the gene-regulated cellular process and the induced changes of mechanical properties and growth rules in SI, table S1. In the main text, to be clearer, we have added a sentence:

“Many malformations are related to gene-regulated abnormal cellular processes and mechanical properties, which are discussed in SI”

b) The Discussion could better acknowledge limitations and future directions, including regional dif-ferences in folding, inter-individual variability, and the model’s assumptions of homogeneous material properties and growth.

In the discussion section, we have pointed out four main limitations and open directions based on our current model, including the discussion on spatiotemporal growth and property. To be more complete, we have supplemented other limitations on the regional differences in folding and the interindividual variability. In the main text, we added the following sentence:

“In addition to the homogeneity assumption, we have not investigated the inter-individual variability and regional differences in folding. More accurate and specific work is expected to focus on these directions.”

c) The authors briefly mention the potential for addressing the inverse growth problem. Expanding this idea in a short paragraph - perhaps with hypothetical applications to fossil brain reconstructions-would broaden the paper’s appeal to evolutionary neuroscientists.

We have stated general steps in the response to public reviewer 2, question (5).

(10) Minor corrections to the text and figures:

a) Figures:

Label figures more clearly to distinguish between in vivo, in vitro, and in silico brain representations.– Ensure that the occipital pole of the macaque is visible or annotated, especially if it lacks the expected smoothness.

Add scale bars where missing for clarity in morphometric comparisons.

We thank the reviewer for suggestions to improve the readability of our manuscript.

The in vivo (real), in vitro (gel), and in silico (simulated) results are both distinguished by their labels and different color scheme: gray-white for real brain, pink-white for gel model, and blue-white for simulations, respectively.

The occipital pole of the macaque brain remains relatively smooth in our computational model but notin our physical gel model. We have clarified this in the main text: “To focus on fold formation, we did not discuss the relatively smooth region, such as the Occipital Pole of the macaque.”

All the brain models are rescaled to the same size, where the distance between the anterior-most pointof the frontal lobe and the posterior-most point of the occipital lobe is two units.

b) Text:

Consider revising figure captions to explicitly mention whether specific regional features (e.g., flatoccipital pole) were observed or absent in models.

In Table II (and relevant text), ensure parameter definitions are consistent and explained clearly for across-disciplinary audience.

Add citations to recent human fetal growth imaging work (e.g., ultrasound-based studies) to support claims about available data.

We have added some descriptions of the characters of the folding pattern in the caption of Figure 4,including major folds and smooth regions.

“Three or four major folds of each brain model are highlighted and served as landmarks. The occipital pole region of macaque brains remains smooth in real and simulated brains.”

We have clarified the definition of growth ratio gMsub>g</sub>/g_w and stiffness ratio µ_g/µ_w between gray matter and white matter, and the normalized cortical thickness h/L in Table 2.

We have referred to a high-quality dataset of fetal brain imaging work, the ultrasound-imaging method(Namburete et al. 2023), in our main text, Discussion:

“...the effect of inhomogeneous growth needs to be further investigated by incorporating regional growth of the gray and white matter not only in human brains [29, 31] based on public datasets [45], but also in other species.”

eLife Assessment

This important study provides insights into the neurodevelopmental trajectories of structural and functional connectivity gradients in the human brain and their potential associations with behaviour and psychopathology. The evidence supporting the findings is solid. This study will be of interest to neuroscientists interested in understanding functional connectivity across development.

Reviewer #2 (Public review):

Summary:

This study aims to show how structural and functional brain organization develops during childhood and adolescence using two large neuroimaging datasets. It addresses whether core principles of brain organization are stable across development, how they change over time, and how these changes relate to cognition and psychopathology. The study finds that brain organization is established early and remains stable but undergoes gradual refinement, particularly in higher-order networks. Structural-functional coupling is linked to better working memory but shows no clear relationship with psychopathology.

Comments on revisions:

Follow-up: I would like to thank the authors for their thoughtful and comprehensive revisions. The additional analyses addressing developmental differences in structure-function coupling between CALM and NKI are valuable and clearly strengthen the manuscript. I particularly appreciate the inclusion of the neurotypical subgroup within CALM to disentangle neurotypicality from potential site-related effects, as well as the expanded discussion of these findings in the context of individual variability and equifinality.

Regarding my earlier comment on the use of COMBAT, I realize that "exclusion" may have been a poor choice of wording. What I meant was that harmonization procedures like COMBAT can, in some cases, weaken extremes or reduce variability by shrinking values toward the mean, rather than literally excluding participants from the analysis. Nevertheless, I appreciate the authors' careful consideration of this point and their additional analysis examining sample coverage following motion-based exclusions.

Overall, I am satisfied with the revisions, and I believe the manuscript has been substantially improved.

Author response:

The following is the authors’ response to the original reviews.

Reviewer #1 (Public review):

Lack of Sensitivity Analyses for some Key Methodological Decisions: Certain methodological choices in this manuscript diverge from approaches used in previous works. In these cases, I recommend the following: (i) The authors could provide a clear and detailed justification for these deviations from established methods, and (ii) supplementary sensitivity analyses could be included to ensure the robustness of the findings, demonstrating that the results are not driven primarily by these methodological changes. Below, I outline the main areas where such evaluations are needed:

This detailed guidance is incredibly valuable, and we are grateful. Work of this kind is in its relative infancy, and there are so many design choices depending on the data available, questions being addressed, and so on. Help us navigate that has been extremely useful. In our revised manuscript we are very happy to add additional justification for design choices made, and wherever possible test the impact of those choices. It is certainly the case that different approaches have been used across the handful of papers published in this space, and, unlike in other areas of systems neuroscience, we have yet to reach the point where any of these approaches are established. We agree with the reviewer that wherever possible these design choices should be tested. 

Use of Communicability Matrices for Structural Connectivity Gradients: The authors chose to construct structural connectivity gradients using communicability matrices, arguing that diffusion map embedding "requires a smooth, fully connected matrix." However, by definition, the creation of the affinity matrix already involves smoothing and ensures full connectedness. I recommend that the authors include an analysis of what happens when the communicability matrix step is omitted. This sensitivity test is crucial, as it would help determine whether the main findings hold under a simpler construction of the affinity matrix. If the results significantly change, it could indicate that the observations are sensitive to this design choice, thereby raising concerns about the robustness of the conclusions. Additionally, if the concern is related to the large range of weights in the raw structural connectivity (SC) matrix, a more conventional approach is to apply a log-transformation to the SC weights (e.g., log(1+𝑆𝐶_𝑖𝑗)), which may yield a more reliable affinity matrix without the need for communicability measures.

The reason we used communicability is indeed partly because we wanted to guarantee a smooth fully connected matrix, but also because our end goal for this project was to explore structure-function coupling in these low-dimensional manifolds.  Structural communicability – like standard metrics of functional connectivity – includes both direct and indirect pathways, whereas streamline counts only capture direct communication. In essence we wanted to capture not only how information might be routed from one location to another, but also the more likely situation in which information propagates through the system. 

In the revised manuscript we have given a clearer justification for why we wanted to use communicability as our structural measure (Page 4, Line 179):

“To capture both direct and indirect paths of connectivity and communication, we generated weighted communicability matrices using SIFT2-weighted fibre bundle capacity (FBC). These communicability matrices reflect a graph theory measure of information transfer previously shown to maximally predict functional connectivity (Esfahlani et al., 2022; Seguin et al., 2022). This also foreshadowed our structure-function coupling analyses, whereby network communication models have been shown to increase coupling strength relative to streamline counts (Seguin et al., 2020)”.

We have also referred the reader to a new section of the Results that includes the structural gradients based on the streamline counts (Page 7, line 316):

“Finally, as a sensitivity analysis, to determine the effect of communicability on the gradients, we derived affinity matrices for both datasets using a simpler measure: the log of raw streamline counts. The first 3 components derived from streamline counts compared to communicability were highly consistent across both NKI  (r_s = 0.791, r_s = 0.866, r_s = 0.761) and the referred subset of CALM (r_s = 0.951, r_s = 0.809, r_s = 0.861), suggesting that in practice the organisational gradients are highly similar regardless of the SC metric used to construct the affinity matrices”. 

Methodological ambiguity/lack of clarity in the description of certain evaluation steps: Some aspects of the manuscript’s methodological description are ambiguous, making it challenging for future readers to fully reproduce the analyses based on the information provided. I believe the following sections would benefit from additional detail and clarification:

Computation of Manifold Eccentricity: The description of how eccentricity was computed (both in the results and methods sections) is unclear and may be problematic. The main ambiguity lies in how the group manifold origin was defined or computed. (1) In the results section, it appears that separate manifold origins were calculated for the NKI and CALM groups, suggesting a dataset-specific approach. (2) Conversely, the methods section implies that a single manifold origin was obtained by somehow combining the group origins across the three datasets, which seems contradictory. Moreover, including neurodivergent individuals in defining the central group manifold origin in conceptually problematic. Given that neurodivergent participants might exhibit atypical brain organization, as suggested by Figure 1, this inclusion could skew the definition of what should represent a typical or normative brain manifold. A more appropriate approach might involve constructing the group manifold origin using only the neurotypical participants from both the NKI and CALM datasets. Given the reported similarity between group-level manifolds of neurotypical individuals in CALM and NKI, it would be reasonable to expect that this combined origin should be close to the origin computed within neurotypical samples of either NKI or CALM. As a sanity check, I recommend reporting the distance of the combined neurotypical manifold origin to the centres of the neurotypical manifolds in each dataset. Moreover, if the manifold origin was constructed while utilizing all samples (including neurodivergent samples) I think this needs to be reconsidered. 

This is a great point, and we are very happy to clarify. Separate manifolds were calculated for the NKI and CALM participants, hence a dataset-specific approach. Indeed, in the long-run our goal was to explore individual differences in these manifolds, relative to the respective group-level origins, and their intersection across modalities, so manifold eccentricity was calculated at an individual level for subsequent analyses. At the group level, for each modality, we computed 3 manifold origins: one for NKI, one for the referred subset of CALM, and another for the neurotypical portion of CALM. Crucially, because the manifolds are always normal, in each case the manifold origin point is near-zero (extremely near-zero, to the 6^th or 7^th decimal place). In other words, we do indeed calculate the origin separately each time we calculate the gradients, but the origin is zero in every case. As a result, differences in the origin point cannot be the source of any differences we observe in manifold eccentricity between groups or individuals. We have updated the Methods section with the manifold origin points for each dataset and clarified our rationale (Page 16, Line 1296):

“Note that we used a dataset-specific approach when we computed manifold eccentricity for each of the three groups relative to their group-level origin: neurotypical CALM (SC origin = -7.698 x 10^-7, FC origin = 6.724 x 10^-7), neurodivergent CALM (SC origin = -6.422 x 10 , FC origin = 1.363 x 10 ), and NKI (SC origin = -7.434 x 10 , FC origin = 4.308 x 10^-6). Eccentricity is a relative measure and thus normalised relative to the origin. Because of this normalisation, each time gradients are constructed the manifold origin is necessarily near-zero, meaning that differences in manifold eccentricity of individual nodes, either between groups or individuals, are stem from the eccentricity of that node rather than a difference in origin point”. 

We clarified the computation of the respective manifold origins within the Results section, and referred the reader to the relevant Methods section (Page 9, line 446):

“For each modality (2 levels: SC and FC) and dataset (3 levels: neurotypical CALM, neurodivergent CALM, and NKI), we computed the group manifold origin as the mean of their respective first three gradients. Because of the normal nature of the manifolds this necessarily means that these origin points will be very near-zero, but we include the exact values in the ‘Manifold Eccentricity’ methodology sub-section”. 

Individual-Level Gradients vs. Group-Level Gradients: Unlike previous studies that examined alterations in principal gradients (e.g., Xia et al., 2022; Dong et al., 2021), this manuscript focuses on gradients derived directly from individual-level data. In contrast, earlier works have typically computed gradients based on grouped data, such as using a moving window of individuals based on age (Xia et al.) or evaluating two distinct age groups (Dong et al.). I believe it is essential to assess the sensitivity of the findings to this methodological choice. Such an evaluation could clarify whether the observed discrepancies with previous reports are due to true biological differences or simply a result of different analytical strategies.

This is a brilliant point. The central purpose of our project was to test how individual differences in these gradients, and their intersection across modalities, related to differences in phenotype (e.g. cognitive difficulties). These necessitated calculating gradients at the level of individuals and building a pipeline to do so, given that we could find no other examples. Nonetheless, despite this different goal and thus approach, we had expected to replicate a couple of other key findings, most prominently the ‘swapping’ of gradients shown by Dong et al. (2021). We were also surprised that we did not find this changing in order. The reviewer is right and there could be several design features that produce the difference, and in the revised manuscript we test several of them. We have added the following text to the manuscript as a sensitivity analysis for the Results sub-section titled “Stability of individual-level gradients across developmental time” (Page 7, Line 344 onwards):

“One possibility is that our observation of gradient stability – rather than a swapping of the order for the first two gradients (Dong et al., 2021) – is because we calculated them at an individual level. To test this, we created subgroups and contrasted the first two group-level structural and functional gradients derived from children (younger than 12 years old) versus those from adolescents (12 years old and above), using the same age groupings as prior work (Dong et al., 2021). If our use of individually calculated gradients produces the stability, then we should observe the swapping of gradients in this sensitivity analysis. Using baseline scans from NKI, the primary structural gradient in childhood (N = 99) as shown in Figure 1f, this was highly correlated (r_s = 0.995) with those derived from adolescents (N = 123). Likewise, the secondary structural gradient in childhood was highly consistent in adolescence (r_s = 0.988). In terms of functional connectivity, the principal gradient in childhood (N = 88) was highly consistent in adolescence (r_s = 0.990, N = 125). The secondary gradient in childhood was again highly similar in adolescence (r_s = 0.984). The same result occurred in the CALM dataset: In the baseline referred subset of CALM, the primary and secondary communicability gradients derived from children (N = 258) and adolescents (N = 53) were near-identical (r_s = 0.991 and r_s = 0.967, respectively). Alignment for the primary and secondary functional gradients derived from children (N = 130) and adolescents (N = 43) were also near-identical (r_s = 0.972 and r_s = 0.983, respectively). These consistencies across development suggest that gradients of communicability and functional connectivity established in childhood are the same as those in adolescence, irrespective of group-level or individual-level analysis. Put simply, our failure to replicate the swapping of gradient order in Dong et al. (2021) is not the result of calculating gradients at the level of individual participants.”

Procrustes Transformation: It is unclear why the authors opted to include a Procrustes transformation in this analysis, especially given that previous related studies (e.g., Dong et al.) did not apply this step. I believe it is crucial to evaluate whether this methodological choice influences the results, particularly in the context of developmental changes in organizational gradients. Specifically, the Procrustes transformation may maximize alignment to the group-level gradients, potentially masking individual-level differences. This could result in a reordering of the gradients (e.g., swapping the first and second gradients), which might obscure true developmental alterations. It would be informative to include an analysis showing the impact of performing vs. omitting the Procrustes transformation, as this could help clarify whether the observed effects are robust or an artifact of the alignment procedure. (Please also refer to my comment on adding a subplot to Figure 1). Additionally, clarifying how exactly the transformation was applied to align gradients across hemispheres, individuals, and/or datasets would help resolve ambiguity. 

The current study investigated individual differences in connectome organisation, rather than group-level trends (Dong et al., 2021). This necessitates aligning individual gradients to the corresponding group-level template using a Procrustes rotation. Without a rotation, there is no way of knowing if you are comparing  ‘like with like’: the manifold eccentricity of a given node may appear to change across individuals simply due to subtle differences in the arbitrary orientation of the underlying manifolds. We also note that prior work examining individual differences in principal alignment have used Procrustes (Xia et al., 2022), who demonstrated emergence of the principal gradient across development, albeit with much smaller effects than Dong and colleagues (2021). Nonetheless, we agree, the Procrustes rotation could be another source of the differences we observed with the previous paper (Dong et al. 2021). We explored the impact of the Procrustes rotation on individual gradients as our next sensitivity analysis. We recalculated everyone’s gradients without Procrustes rotation. We then tested the alignment of each participant with the group-level gradients using Spearman’s correlations, followed by a series of generalised linear models to predict principal gradient alignment using head motion, age, and sex. The expected swapping of the first and second functional gradient (Dong et al., 2021) would be represented by a decrease in the spatial similarity of each child’s principal functional gradient to the principal childhood group-level gradient, at the onset of adolescence (~age 12). However, there is no age effect on this unrotated alignment, suggesting that the lack of gradient swapping in our data does not appear to be the result of the Procrustes rotation. When you use unrotated individual gradients the alignment is remarkably consistent across childhood and adolescence. Alignment is, however, related to head motion, which is often related to age. To emphasise the importance of motion, particularly in relation to development, we conducted a mediation analysis between the relationship between age and principal alignment (without correcting for motion), with motion as a mediator, within the NKI dataset. Before accounting for motion, the relationship between age and principal alignment is significant, but this can be entirely accounted for by motion. In our revised manuscript we have included this additional analysis in the Results sub-section titled “Stability of individual-level gradients across developmental time”, following on from the above point about the effect of group-level versus individual-level analysis (Page 8, Line 400):

“A second possible discrepancy between our results and that of prior work examining developmental change in group-level functional gradients (Dong et al., 2021) was the use of Procrustes alignment. Such alignment of individual-level gradients to group-level templates is a necessary step to ensure valid comparisons between corresponding gradients across individuals, and has been implemented in sliding-window developmental work tracking functional gradient development (Xia et al., 2022). Nonetheless, we tested whether our observation of stable principal functional and communicability gradients may be an artefact of the Procrustes rotation. We did this by modelling how individual-level alignment without Procrustes rotation to the group-level templates varies with age, head motion, and sex, as a series of generalised linear models. We included head motion as the magnitude of the Procrustes rotation has been shown to be positively correlated with mean framewise displacement (Sasse et al., 2024), and prior group-level work (Dong et al., 2021) included an absolute motion threshold rather than continuous motion estimates. Using the baseline referred CALM sample, there was no significant relationship between alignment and age (β = -0.044, 95% CI = [-0.154, 0.066], p = 0.432) after accounting for head motion and sex. Interestingly, however head motion was significantly associated with alignment ( β = -0.318, 95% CI = [-0.428, -.207], p = 1.731 x 10^-8), such that greater head motion was linked to weaker alignment. Note that older children tended to have exhibit less motion for their structural scans (r_s = 0.335, p < 0.001). We observed similar trends in functional alignment, whereby tighter alignment was significantly predicted by lower head motion (β = -0.370, 95% CI = [-0.509, -0.231], p = 1.857 x 10^-7), but not by age (β= 0.049, 95% CI = [-0.090, 0.187], p = 0.490). Note that age and head motion for functional scans were not significantly related (r_s = -0.112, p = 0.137). When repeated for the baseline scans of NKI, alignment with the principal structural gradient was not significantly predicted by either scan age (β = 0.019, 95% CI = [-0.124, 0.163], p = 0.792) or head motion (β = -0.133, 95% CI = [-0.175, 0.009], p = 0.067) together in a single model, where age and motion were negatively correlated (r_s = -0.355, p < 0.001). Alignment with the principal functional gradient was significantly predicted by head motion (β = -0.183, 95% CI = [-0.329, -0.036], p = 0.014) but not by age (β= 0.066, 95% CI = [-0.081, 0.213], p = 0.377), where age and motion were also negatively correlated (r_s = -0.412, p < 0.001). Across modalities and datasets, alignment with the principal functional gradient in NKI was the only example in which there was a significant correlation between alignment and age (r_s = 0.164, p = 0.017) before accounting for head motion and sex. This suggests that apparent developmental effects on alignment are minimal, and where they do exist they are removed after accounting for head motion. Put together this suggests that the lack of order swapping for the first two gradients is not the result of the Procrustes rotation – even without the rotation there is no evidence for swapping”.

“To emphasise the importance of head motion in the appearance of developmental change in alignment, we examined whether accounting for head motion removes any apparent developmental change within NKI. Specifically, we tested whether head motion mediates the relationship between age and alignment (Figure 1X), controlling for sex, given that higher motion is associated with younger children (β= -0.429, 95% CI = [0.552, -0.305], p = 7.957 x 10^-11), and stronger alignment is associated with reduced motion (β = -0.211, 95% CI = [-0.344, -0.078], p = 2.017 x 10^-3). Motion mediated the relationship between age and alignment (β = 0.078, 95% CI = [0.006, 0.146], p = 1.200 x 10^-2), accounting for 38.5% variance in the age-alignment relationship, such that the link between age and alignment became non-significant after accounting for motion (β = 0.066, 95% CI = [-0.081, 0.214], p = 0.378). This firstly confirms our GLM analyses, where we control for motion and find no age associations. Moreover, this suggests that caution is required when associations between age and gradients are observed. In our analyses, because we calculate individual gradients, we can correct for individual differences in head motion in all our analyses. However, other than using an absolute motion threshold and motion-matched child and adolescent groups, individual differences in motion were not accounted for by prior work which demonstrated a flipping of the principal functional gradients with age (Dong et al., 2021)”. 

We further clarify the use of Procrustes rotation as a separate sub-section within the Methods (Page 25, Line 1273):

“Procrustes Rotation

For group-level analysis, for each hemisphere we constructed an affinity matrix using a normalized angle kernel and applied diffusion-map embedding. The left hemisphere was then aligned to the right using a Procrustes rotation. For individual-level analysis, eigenvectors for the left hemisphere were aligned with the corresponding group-level rotated eigenvectors. No alignment was applied across datasets. The only exception to this was for structural gradients derived from the referred CALM cohort. Specifically, we aligned the principal gradient of the left hemisphere to the secondary gradient of the right hemisphere: this was due to the first and second gradients explaining a very similar amount of variance, and hence their order was switched”. 

SC-FC Coupling Metric: The approach used to quantify nodal SC-FC coupling in this study appears to deviate from previously established methods in the field. The manuscript describes coupling as the "Spearman-rank correlation between Euclidean distances between each node and all others within structural and functional manifolds," but this description is unclear and lacks sufficient detail. Furthermore, this differs from what is typically referred to as SC-FC coupling in the literature. For instance, the cited study by Park et al. (2022) utilizes a multiple linear regression framework, where communicability, Euclidean distance, and shortest path length are independent variables predicting functional connectivity (FC), with the adjusted R-squared score serving as the coupling index for each node. On the other hand, the Baum et al. (2020) study, also cited, uses Spearman correlation, but between raw structural connectivity (SC) and FC values. If the authors opt to introduce a novel coupling metric, it is essential to demonstrate its similarity to these previous indices. I recommend providing an analysis (supplementary) showing the correlation between their chosen metric and those used in previous studies (e.g., the adjusted R-squared scores from Park et al. or the SC-FC correlation from Baum et al.). Furthermore, if the metrics are not similar and results are sensitive to this alternative metric, it raises concerns about the robustness of the findings. A sensitivity analysis would therefore be helpful (in case the novel coupling metric is not like previous ones) to determine whether the reported effects hold true across different coupling indices.

This is a great point, and we are happy to take the reviewer’s recommendation. There are multiple different ways of calculating structure-function coupling. For our set of questions, it was important that our metric incorporated information about the structural and functional manifolds, rather than being a separate approach that is unrelated to these low-dimensional embeddings. Put simply, we wanted our coupling measure to be about the manifolds and gradients outlined in the early sections of the results. We note that the multiple linear regression framework was developed by Vázquez-Rodríguez and colleagues (2019), whilst the structure-function coupling computed in manifold space by Park and colleagues (2022) was operationalised as a linear correlation between z-transformed functional connectomes and structural differentiation eigenvectors. To clarify how this coupling was calculated, and to justify why we developed a new coupling method based on manifolds rather than borrow an existing approach from the literature, we have revised the manuscript to make this far clearer for readers (Page 13, line 604):

“To examine the relationship between each node’s relative position in structural and functional manifold space, we turned our attention to structure-function coupling. Whilst prior work typically computed coupling using raw streamline counts and functional connectivity matrices, either as a correlation (Baum et al., 2020) or through a multiple linear regression framework (Vázquez-Rodríguez et al., 2019), we opted to directly incorporate low-dimensional embeddings within our coupling framework. Specifically, as opposed to correlating row-wise raw functional connectivity with structural connectivity eigenvectors (Park et al., 2022), our metric directly incorporates the relative position of each node in low-dimensional structural and functional manifold spaces. Each node was situated in a low-dimensional 3D space, the axes of which were each participant’s gradients, specific to each modality. For each participant and each node, we computed the Euclidean distance with all other nodes within structural and functional manifolds separately, producing a vector of size 200 x 1 per modality. The nodal coupling coefficient was the Spearman correlation between each node’s Euclidean distance to all other nodes in structural manifold space, and that in functional manifold space. Put simply, a strong nodal coupling coefficient suggests that that node occupies a similar location in structural space, relative to all other nodes, as it does in functional space”. 

We also agree with the reviewer’s recommendation to compare this to some of the more standard ways of calculating coupling. We compare our metric with 3 others (Baum et al., 2020; Park et al., 2022; VázquezRodríguez et al., 2019), and find that all metrics capture the core developmental sensorimotor-to-association axis (Sydnor et al., 2021). Interestingly, manifold-based coupling measures captured this axis more strongly than non-manifold measures. We have updated the Results accordingly (Page 14, Line 638):

“To evaluate our novel coupling metric, we compared its cortical spatial distribution to three others (Baum et al., 2020; Park et al., 2022; Vázquez-Rodríguez et al., 2019), using the group-level thresholded structural and functional connectomes from the referred CALM cohort. As shown in Figure 4c, our novel metric was moderately positively correlated to that of a multi-linear regression framework (r_s = 0.494, p_spin = 0.004; Vázquez-Rodríguez et al., 2019) and nodal correlations of streamline counts and functional connectivity (r_s = 0.470, p_spin = 0.005; Baum et al., 2020). As expected, our novel metric was strongly positively correlated to the manifold-derived coupling measure (r_s = 0.661, p_spin < 0.001; Park et al., 2022), more so than the first (Z(198) = 3.669, p < 0.001) and second measure (Z(198) = 4.012, p < 0.001). Structure-function coupling is thought to be patterned along a sensorimotor-association axis (Sydnor et al., 2021): all four metrics displayed weak-tomoderate alignment (Figure 4c). Interestingly, the manifold-based measures appeared most strongly aligned with the sensorimotor-association axis: the novel metric was more strongly aligned than the multi-linear regression framework (Z(198) = -11.564, p < 0.001) and the raw connectomic nodal correlation approach (Z(198) = -10.724, p < 0.001), but the previously-implemented structural manifold approach was more strongly aligned than the novel metric  (Z(198) = -12.242, p < 0.001). This suggests that our novel metric exhibits the expected spatial distribution of structure-function coupling, and the manifold approach more accurately recapitulates the sensorimotor-association axis than approaches based on raw connectomic measures”.

We also added the following to the legend of Figure 4 on page 15:

“d. The inset Spearman correlation plot of the 4 coupling measures shows moderate-to-strong correlations (p_spin < 0.005 for all spatial correlations). The accompanying lollypop plot shows the alignment between the sensorimotor-to-association axis and each of the 4 coupling measures, with the novel measure coloured in light purple (p_spin < 0.007 for all spatial correlations)”. 

Prediction vs. Association Analysis: The term “prediction” is used throughout the manuscript to describe what appear to be in-sample association tests. This terminology may be misleading, as prediction generally implies an out-of-sample evaluation where models trained on a subset of data are tested on a separate, unseen dataset. If the goal of the analyses is to assess associations rather than make true predictions, I recommend refraining from the term “prediction” and instead clarifying the nature of the analysis. Alternatively, if prediction is indeed the intended aim (which would be more compelling), I suggest conducting the evaluations using a k-fold cross-validation framework. This would involve training the Generalized Additive Mixed Models (GAMMs) on a portion of the data and training their predictive accuracy on a held-out sample (i.e. different individuals). Additionally, the current design appears to focus on predicting SC-FC coupling using cognitive or pathological dimensions. This is contrary to the more conventional approach of predicting behavioural or pathological outcomes from brain markers like coupling. Could the authors clarify why this reverse direction of analysis was chosen? Understanding this choice is crucial, as it impacts the interpretation and potential implications of the findings. 

We have replaced “prediction” with “association” across the manuscript. However, for analyses corresponding to Figure 5, which we believe to be the most compelling, we conducted a stratified 5-fold cross-validation procedure, outlined below, repeated 100 times to account for random variation in the train-test splits. To assess whether prediction accuracy in the test splits was significantly greater than chance, we compared our results to those derived from a null dataset in which cognitive factor 2 scores had been permuted across participants. To account for the time-series element and block design of our data, in that some participants had 2 or more observations, we permuted entire participant blocks of cognitive factor 2 scores, keeping all other variables, including covariates, the same. Included in our manuscript are methodological details and results pertaining to this procedure. Specifically, the following has been added to the Results (Page 16, Line 758):

“To examine the predictive value of the second cognitive factor for global and network-level structure-function coupling, operationalised as a Spearman rank correlation coefficient, we implemented a stratified 5-fold crossvalidation framework, and predictive accuracy compared with that of a null data frame with cognitive factor 2 scores permuted across participant blocks (see ‘GAMM cross-validation’ in the Methods). This procedure was repeated 100 times to account for randomness in the train-test splits, using the same model specification as above. Therefore, for each of the 5 network partitions in which an interaction between the second cognitive factor and age was a significant predictor of structure-function coupling (global, visual, somato-motor, dorsal attention, and default-mode), we conducted a Welch’s independent-sample t-test to compare 500 empirical prediction accuracies with 500 null prediction accuracies. Across all 5 network partitions, predictive accuracy of coupling was significantly higher than that of models trained on permuted cognitive factor 2 scores (all p < 0.001). We observed the largest difference between empirical (M = 0.029, SD = 0.076) and null (M = -0.052, SD = 0.087) prediction accuracy in the somato-motor network [t (980.791) = 15.748, p < 0.001, Cohen’s d = 0.996], and the smallest difference between empirical (M = 0.080, SD = 0.082) and null (M = 0.047, SD = 0.081) prediction accuracy in the dorsal attention network [t (997.720) = 6.378, p < 0.001, Cohen’s d = 0.403]. To compare relative prediction accuracies, we ordered networks by descending mean accuracy and conducted a series of Welch’s independent sample t-tests, followed by FDR correction (Figure 5X). Prediction accuracy was highest in the default-mode network (M = 0.265, SD = 0.085), two-fold that of global coupling (t(992.824) = 25.777, p_FDR = 5.457 x 10^-112, Cohen’s d = 1.630, M = 0.131, SD = 0.079). Global prediction accuracy was significantly higher than the visual network (t (992.644) = 9.273, p_FDR = 1.462 x 10^-19, Cohen’s d = 0.586, M = 0.083, SD = 0.085), but visual prediction accuracy was not significantly higher than within the dorsal attention network (t (997.064) = 0.554, p_FDR = 0.580, Cohen’s d = 0.035, M = 0.080, SD = 0.082). Finally, prediction accuracy within the dorsal attention network was significantly stronger than that of the somato-motor network [t (991.566) = 10.158, p_FDR = 7.879 x 10^-23, Cohen’s d = 0.642 M = 0.029, SD = 0.076]. Together, this suggests that out-of-sample developmental predictive accuracy for structure-function coupling, using the second cognitive factor, is strongest in the higher-order default-mode network, and lowest in the lower-order somatosensory network”. 

We have added a separate section for GAMM cross-validation in the Methods (Page 27, Line 1361):

GAMM cross-validation

“We implemented a 5-fold cross validation procedure, stratified by dataset (2 levels: CALM or NKI). All observations from any given participant were assigned to either the testing or training fold, to prevent data leakage, and the cross-validation procedure was repeated 100 times, to account for randomness in data splits. The outcome was predicted global or network-level structure-function coupling across all test splits, operationalised as the Spearman rank correlation coefficient. To assess whether prediction accuracy exceeded chance, we compared empirical prediction accuracy with that of GAMMs trained and tested on null data in which cognitive factor 2 scores were permuted across subjects. The number of observations formed 3 exchangeability blocks (N = 320 with one observation, N = 105 with two observations, and N = 33 with three observations), whereby scores from a participant with two observations were replaced by scores from another participant with two observations, with participant-level scores kept together, and so on for all numbers of observations. We compared empirical and null prediction accuracies using independent sample t-tests as, although the same participants were examined, the shuffling meant that the relative ordering of participants within both distributions was not preserved. For parallelisation and better stability when estimating models fit on permuted data, we used the bam function from the mgcv R package (Wood, 2017)”. 

We also added a justification for why we predicted coupling using behaviour or psychopathology, rather than vice versa (Page 27, Line 1349):

“When using our GAMMs to test for the relationship between cognition and psychopathology and our coupling metrics, we opted to predict structure-function coupling using cognitive or psychopathological dimensions, rather than vice versa, to minimise multiple comparisons. In the current framework, we corrected for 8 multiple comparisons within each domain. This would have increased to 16 multiple comparison corrections for predicting two cognitive dimensions using network-level coupling, and 24 multiple comparison corrections for predicting three psychopathology dimensions. Incorporating multiple networks as predictors within the same regression framework introduces collinearity, whilst the behavioural dimensions were orthogonal: for example, coupling is strongly correlated between the somato-motor and ventral attention networks (r_s = 0.721), between the default-mode and frontoparietal networks (r_s = 0.670), and between the dorsal attention and fronto-parietal networks (r_s = 0.650)”. 

Finally, we noticed a rounding error in the ages of the data frame containing the structure-function coupling values and the cognitive/psychopathology dimensions. We rectified this and replaced the GAMM results, which largely remained the same. 

In typical applications of diffusion map embedding, sparsification (e.g., retaining only the top 10  of the strongest connections) is often employed at the vertex-level resolution to ensure computational feasibility. However, since the present study performs the embedding at the level of 200 brain regions (a considerably coarser resolution), this step may not be necessary or justifiable. Specifically, for FC, it might be more appropriate to retain all positive connections rather than applying sparsification, which could inadvertently eliminate valuable information about lower-strength connections. Whereas for SC, as the values are strictly non-negative, retaining all connections should be feasible and would provide a more complete representation of the structural connectivity patterns. Given this, it would be helpful if the authors could clarify why they chose to include sparsification despite the coarser regional resolution, and whether they considered this alternative approach (using all available positive connections for FC and all non-zero values for SC). It would be interesting if the authors could provide their thoughts on whether the decision to run evaluations at the resolution of brain regions could itself impact the functional and structural manifolds, their alteration with age, and or their stability (in contrast to Dong et al. which tested alterations in highresolution gradients).

This is another great point. We could retain all connections, but we usually implement some form of sparsification to reduce noise, particularly in the case of functional connectivity. But we nonetheless agree with the reviewer’s point. We should check what impact this is having on the analysis. In brief, we found minimal effects of thresholding, suggesting that the strongest connections are driving the gradient (Page 7, Line 304):

“To assess the effect of sparsity on the derived gradients, we examined group-level structural (N = 222) and functional (N = 213) connectomes from the baseline session of NKI. The first three functional connectivity gradients derived using the full connectivity matrix (density = 92%) were highly consistent with those obtained from retaining the strongest 10% of connections in each row (r₁ = 0.999, r₂ = 0.998, r₃ < 0.999, all p < 0.001). Likewise, the first three communicability gradients derived from retaining all streamline counts (density = 83%) were almost identical to those obtained from 10% row-wise thresholding (r₁ = 0.994, r₂ = 0.963, r₃ = 0.955, all p < 0.001). This suggests that the reported gradients are driven by the strongest or most consistent connections within the connectomes, with minimal additional information provided by weaker connections. In terms of functional connectivity, such consistency reinforces past work demonstrating that the sensorimotor-toassociation axis, the major axis within the principal functional connectivity gradient, emerges across both the top- and bottom-ranked functional connections (Nenning et al., 2023)”.

Furthermore, we appreciate the nudge to share our thoughts on whether the difference between vertex versus nodal metrics could be important here, particularly regarding thresholds. To combine this point with R2’s recommendation to expand the Discussion, we have added the following paragraph (Page 19, Line 861): 

“We consider the role of thresholding, cortical resolution, and head motion as avenues to reconcile the present results with select reports in the literature (Dong et al., 2021; Xia et al., 2022). We would suggest that thresholding has a greater effect on vertex-level data, rather than parcel-level. For example, a recent study revealed that the emergence of principal vertex-level functional connectivity gradients in childhood and adolescence are indeed threshold-dependent (Dong et al., 2024). Specifically, the characteristic unimodal organisation for children and transmodal organisation for adolescents only emerged at the 90% threshold: a 95% threshold produced a unimodal organisation in both groups, whilst an 85% threshold produced a transmodal organisation in both groups. Put simply, the ‘swapping’ of gradient orders only occurs at certain thresholds. Furthermore, our results are not necessarily contradictory to this prior report (Dong et al., 2021): developmental changes in high-resolution gradients may be supported by a stable low-dimensional coarse manifold. Indeed, our decision to use parcellated connectomes was partly driven by recent work which demonstrated that vertex-level functional gradients may be derived using biologically-plausible but random data with sufficient spatial smoothing, whilst this effect is minimal at coarser resolutions (Watson & Andrews, 2023). We observed a gradual increase in the variance of individual connectomes accounted for by the principal functional connectivity gradient in the referred subset of CALM, in line with prior vertex-level work demonstrating a gradual emergence of the sensorimotor-association axis as the principal axis of connectivity (Xia et al., 2022), as opposed to a sudden shift. It is also possible that vertex-level data is more prone to motion artefacts in the context of developmental work. Transitioning from vertex-level to parcel-level data involves smoothing over short-range connectivity, thus greater variability in short-range connectivity can be observed in vertex-level data. However, motion artefacts are known to increase short-range connectivity and decrease long-range connectivity, mimicking developmental changes (Satterthwaite et al., 2013). Thus, whilst vertexlevel data offers greater spatial resolution in representation of short-range connectivity relative to parcel-level data, it is possible that this may come at the cost of making our estimates of the gradients more prone to motion”.

Evaluating the consistency of gradients across development: the results shown in Figure 1e are used as evidence suggesting that gradients are consistent across ages. However, I believe additional analyses are required to identify potential sources of the observed inconsistency compared to previous works. The claim that the principal gradient explains a similar degree of variance across ages does not necessarily imply that the spatial structure remains the same. The observed variance explanation is hence not enough to ascertain inconsistency with findings from Dong et al., as the spatial configuration of gradients may still change over time. I suggest the following additional analyses to strengthen this claim. Alignment to group-level gradients: Assess how much of the variance in individual FC matrices is explained by each of the group-level gradients (G1, G2, and G3, for both FC and SC). This analysis could be visualized similarly to Figure 1e, with age on the x-axis and variance explained on the y-axis. If the explained variance varies as a function of age, it may indicate that the gradients are not as consistent as currently suggested. 

This is another great suggestion. In the additional analyses above (new group-level analyses and unrotated gradient analyses) we rule-out a couple of the potential causes of the different developmental trends we observe in our data – namely the stability of the gradients over time. The suggested additional analysis is a great idea, and we have implemented it as follows (Page 8, Line 363):

“To evaluate the consistency of gradients across development, across baseline participants with functional connectomes from the referred CALM cohort (N = 177), we calculated the proportion of variance in individuallevel connectomes accounted for by group-level functional gradients. Specifically, we calculated the proportion of variance in an adjacency matrix A accounted for by the vector v_i as the fraction of the square of the scalar projection of v_i onto A, over the Frobenius norm of A. Using a generalised linear model, we then tested whether the proportion of variance explained varies systematically with age, controlling for sex and headmotion. The variance in individual-level functional connectomes accounted for by the group-level principal functional gradient gradually increased with development (β= 0.111, 95% CI = [0.022, 0.199], p = 1.452 x 10^-2, Cohen’s d = 0.367), as shown in Figure 1g, and decreased with higher head motion ( β = -10.041, 95% CI = [12.379, -7.702], p = 3.900 x 10^-17), with no effect of sex (β= 0.071, 95% CI = [-0.380, 0.523], p = 0.757). We observed no developmental effects on the variance explained by the second (r_s = 0.112, p = 0.139) or third (r_s = 0.053, p = 0.482) group-level functional gradient. When repeated with the baseline functional connectivity for NKI (N = 213), we observed no developmental effects (β = 0.097, 95% CI = [-0.035, 0.228], p = 0.150) on the variance explained by the principal functional gradient after accounting for motion (β= -3.376, 95% CI = [8.281, 1.528], p = 0.177) and sex (β = -0.368, 95% CI = [-1.078, 0.342], p = 0.309). However, we observed significant developmental correlations between age and variance (r_s = 0.137, p = 0.046) explained before accounting for head motion and sex. We observed no developmental effects on the variance explained by the second functional gradient (r_s = -0.066, p = 0.338), but a weak negative developmental effect on the variance explained by the third functional gradient (r_s = -0.189, p = 0.006). Note, however, the magnitude of the variance accounted for by the third functional gradient was very small (all < 1%). When applied to communicability matrices in CALM, the proportion of variance accounted for by the group-level communicability gradient was negligible (all < 1%), precluding analysis of developmental change”. 

“To further probe the consistency of gradients across development, we examined developmental changes in the standard deviation of gradient values, corresponding to heterogeneity, following prior work examining morphological (He et al., 2025) and functional connectivity gradients (Xia et al., 2022). Using a series of generalised linear models within the baseline referred subset of CALM, correcting for head motion and sex, we found that gradient variation for the principal functional gradient increased across development (= 0.219, 95% CI = [0.091, 0.347], p = 0.001, Cohen’s d = 0.504), indicating greater heterogeneity (Figure 1h), whilst gradient variation for the principal communicability gradient decreased across development (β = -0.154, 95% CI = [-0.267, -0.040], p = 0.008, Cohen’s d = -0.301), indicating greater homogeneity (Figure 1h). Note, a paired t-test on the 173 common participants demonstrated a significant effect of modality on gradient variability (t(172) = -56.639, p = 3.663 x 10^-113), such that the mean variability of communicability gradients (M = 0.033, SD = 0.001) was less than half that of functional connectivity (M = 0.076, SD = 0.010). Together, this suggests that principal functional connectivity and communicability gradients are established early in childhood and display age-related refinement, but not replacement”. 

The Issue of Abstraction and Benefits of the Gradient-Based View: The manuscript interprets the eccentricity findings as reflecting changes along the segregation-integration spectrum. Given this, it is unclear why a more straightforward analysis using established graph-theory metrics of segregationintegration was not pursued instead. Mapping gradients and computing eccentricity adds layers of abstraction and complexity. If similar interpretations can be derived directly from simpler graph metrics, what additional insights does the gradient-based framework offer? While the manuscript argues that this approach provides “a more unifying account of cortical reorganization”, it is not evident why this abstraction is necessary or advantageous over traditional graph metrics. Clarifying these benefits would strengthen the rationale for using this method. 

This is a great point, and something we spent quite a bit of time considering when designing the analysis. The central goal of our project was to identify gradients of brain organisation across different datasets and modalities and then test how the organisational principles of those modalities align. In other words, how do structural and functional ‘spaces’ intersect, and does this vary across the cortex? That for us was the primary motivation for operationalising organisation as nodal location within a low-dimensional manifold space (Bethlehem et al., 2020; Gale et al., 2022; Park et al., 2021), using a simple composite measure to achieve compression, rather than as a series of graph metrics. The reason we subsequently calculated those graph metrics and tested for their association was simply to help us interpret what eccentricity within that lowdimensional space means. Manifold eccentricity was moderately positively correlated to graph-theory metrics of integration, leaving a substantial portion of variance unaccounted for, but that association we think is nonetheless helpful for readers trying to interpret eccentricity. However, since ME tells us about the relative position of a node in that low-dimensional space, it is also likely capturing elements of multiple graph theory measures. Following the Reviewer’s question, this is something we decided to test. Specifically, using 4 measures of segregation, including two new metrics requested by the Reviewer in a minor point (weighted clustering coefficient and normalized degree centrality), we conducted a dominance analysis (Budescu, 1993) with normalized manifold eccentricity of the group-level referred CALM structural connectome. We also detail the use of gradient measures in developmental contexts, and how they can be complementary to traditional graph theory metrics. 

We have added the following to the Results section (Page 10, Lines 472 onwards): 

“To further contextualise manifold eccentricity in terms of integration and segregation beyond simple correlations, we conducted a multivariate dominance analysis (Budescu, 1993) of four graph theory metrics of segregation as predictors of nodal normalized manifold eccentricity within the group-level referred CALM structural and functional connectomes (Figure 2c). A dominance analysis assesses the relative importance of each predictor in a multilinear regression framework by fitting 2ⁿ – 1 models (where n is the number of predictors) and calculating the relative increase in adjusted R2 caused by adding each predictor to the model across both main effects and interactions. A multilinear regression model including weighted clustering coefficient, within-module degree Z-score, participation coefficient and normalized degree centrality accounted for 59% of variance in nodal manifold eccentricity in the group-level CALM structural connectome. Withinmodule degree Z score was the most important predictor (40.31% dominance), almost twice that of the participation coefficient (24.03% dominance) and normalized degree centrality (24.05% dominance) which made roughly equal contributions. The least important predictor was the weighted clustering coefficient (11.62% dominance). When the same approach was applied for the group-level referred CALM functional connectome, the 4 predictors accounted for 52% variability. However, in contrast to the structural connectome, functional manifold eccentricity seemed to incorporate the same graph theory metrics in different proportions. Normalized degree centrality was the most important predictor (47.41% dominance), followed by withinmodule degree Z-score (24.27%), and then the participation coefficient (15.57%) and weighted clustering coefficient (12.76%) which made approximately equal contributions. Thus, whilst structural manifold eccentricity was dominated most by within-module degree Z-score and least by the weighted clustering coefficient, functional manifold eccentricity was dominated most by normalized degree centrality and least by the weighted clustering coefficient. This suggests that manifold mapping techniques incorporate different aspects of integration dependent on modality. Together, manifold eccentricity acts as a composite measure of segregation, being differentially sensitive to different aspects of segregation, without necessitating a priori specification of graph theory metrics. Further discussion of the value of gradient-based metrics in developmental contexts and as a supplement to traditional graph theory analyses is provided in the ‘Manifold Eccentricity’ methodology sub-section”. 

We added further justification to the manifold eccentricity Methods subsection (Page 26, line 1283):

“Gradient-based measures hold value in developmental contexts, above and beyond traditional graph theory metrics: within a sample of over 600 cognitively-healthy adults aged between 18 and 88 years old, sensitivity of gradient-based within-network functional dispersion to age were stronger and more consistent across networks compared to segregation (Bethlehem et al., 2020). In the context of microstructural profile covariance, modules resolved by Louvain community detection occupied distinct positions across the principal two gradients, suggesting that gradients offer a way to meaningfully order discrete graph theory analyses (Paquola et al., 2019)”. 

We added the following to the Introduction section outlining the application of gradients as cortex-wide coordinate systems (Page 3, Line 121):

“Using the gradient-based approach as a compression tool, thus forgoing the need to specify singular graph theory metrics a priori, we operationalised individual variability in low-dimensional manifolds as eccentricity (Gale et al., 2022; Park et al., 2021). Crucially, such gradients appear to be useful predictors of phenotypic variation, exceeding edge-level connectomics. For example, in the case of functional connectivity gradients, their predictive ability for externalizing symptoms and general cognition in neurotypical adults surpassed that of edge-level connectome-based predictive modelling (Hong et al., 2020), suggesting that capturing lowdimensional manifolds may be particularly powerful biomarkers of psychopathology and cognition”. 

We also added the following to the Discussion section (Page 18, Line 839):

“By capitalising on manifold eccentricity as a composite measure of segregation across development, we build upon an emerging literature pioneering gradients as a method to establish underlying principles of structural (Paquola et al., 2020; Park et al., 2021) and functional (Dong et al., 2021; Margulies et al., 2016; Xia et al., 2022) brain development without a priori specification of specific graph theory metrics of interest”. 

It is unclear whether the statistical tests finding significant dataset effects are capturing effects of neurotypical vs. Neurodivergent, or simply different scanners/sites. Could the neurotypical portion of CALM also be added to distinguish between these two sources of variability affecting dataset effects (i.e. ideally separating this to the effect of site vs. neurotypicality would better distinguish the effect of neurodivergence).

At a group-level, differences in the gradients between the two cohorts are very minor. Indeed, in the manuscript we describe these gradients as being seemingly ‘universal’. But we agree that we should test whether we can directly attribute any simple main effects of ‘dataset’ are resulting from the different site or the phenotype of the participants. The neurotypical portion of CALM (collected at the same site on the same scanner) helped us show that any minor differences in the gradient alignments is likely due to the site/scanner differences rather than the phenotype of the participants. We took the same approach for testing the simple main effects of dataset on manifold eccentricity. To better parse neurotypicality and site effects at an individual-level, we conducted a series of sensitivity analyses. First, in response to the reviewer’s earlier comment, we conducted a series of nodal generalized linear models for communicability and FC gradients derived from neurotypical and neurodivergent portions of CALM, alongside NKI, and tested for an effect of neurotypicality above and beyond scanner. As at the group level, having those additional scans on a ‘comparison’ sample for CALM is very helpful in teasing apart these effects. We find that neurotypicality affects communicability gradient expression to a greater degree than functional connectivity. We visualised these results and added them to Figure 1. Second, we used the same approach but for manifold eccentricity. Again, we demonstrate greater sensitivity of neurotypicality to communicability at a global-level, but we cannot pin these effects down to specific networks because the effects do not survive the necessary multiple comparison correction. We have added these analyses to the manuscript (Page 13, Line 583): 

“Much as with the gradients themselves, we suspected that much of the simple main effect of dataset could reflect the scanner / site, rather than the difference in phenotype. Again, we drew upon the CALM comparison children to help us disentangle these two explanations. As a sensitivity analysis to parse effects of neurotypicality and dataset on manifold eccentricity, we conducted a series of generalized linear models predicting mean global and network-level manifold eccentricity, for each modality. We did this across all the baseline data (i.e. including the neurotypical comparison sample for CALM) using neurotypicality (2 levels: neurodivergent or neurotypical), site (2 levels: CALM or NKI), sex, head motion, and age at scan (Figure 3X). We restricted our analysis to baseline scans to create more equally-balanced groups. In terms of structural manifold eccentricity (N = 313 neurotypical, N = 311 neurodivergent), we observed higher manifold eccentricity in the neurodivergent participants at a global level (β = 0.090, p = 0.019, Cohen’s d = 0.188) but the individual network level effects did not survive the multiple comparison correction necessary for looking across all seven networks, with the default-mode network being the strongest (β = 0.135, p = 0.027, p_FDR = 0.109, Cohen’s d = 0.177). There was no significant effect of neurodiversity on functional manifold eccentricity (N = 292 neurotypical and N = 177 neurodivergent). This suggests that neurodiversity is significantly associated with structural manifold eccentricity, over and above differences in site, but we cannot distinguish these effects reliably in the functional manifold data”. 

Third, we removed the Scheirer-Ray-Hare test from the results for two reasons. First, its initial implementation did not account for repeated measures, and therefore non-independence between observations, as the same participants may have contributed both structural and functional data. Second, if we wanted to repeat this analysis in CALM using the referred and control portions, a significant difference in group size existed, which may affect the measures of variability. Specifically, for baseline CALM, 311 referred and 91 control participants contributed SC data, whilst 177 referred and 79 control participants contributed FC data. We believe that the ‘cleanest’ parsing of dataset and site for effects of eccentricity is achieved using the GLMs in Figure 3. 

We observed no significant effect of neurodivergence on the magnitude of structure-function coupling across development, and have added the following text (Page 14, Line 632):

“To parse effects of neurotypicality and dataset on structure-function coupling, we conducted a series of generalized linear models predicting mean global and network-level coupling using neurotypicality, site, sex, head motion, and age at scan, at baseline (N = 77 CALM neurotypical, N = 173 CALM neurodivergent, and N = 170 NKI). However, we found no significant effects of neurotypicality on structure-function coupling across development”. 

Since we demonstrated no significant effects of neurotypicality on structure-function coupling magnitude across development, but found differential dataset-specific effects of age on coupling development, we added the following sentence at the end of the coupling trajectory results sub-section (Page 14, line 664):

“Together, these effects demonstrate that whilst the magnitude of structure-function coupling appears not to be sensitive to neurodevelopmental phenotype, its development with age is, particularly in higher-order association networks, with developmental change being reduced in the neurodivergent sample”.  

Figure 1.c: A non-parametric permutation test (e.g. Mann-Whitney U test) could quantitatively identify regions with significant group differences in nodal gradient values, providing additional support for the qualitative findings.

This is a great idea. To examine the effect of referral status on nodal gradient values, whilst controlling for covariates (head motion and sex), we conducted a series of generalised linear models. We opted for this instead of a Mann-Whitney U test, as the former tests for differences in distributions, whilst the direction of the t-statistic for referral status from the GLM would allow us to specify the magnitude and direction of differences in nodal gradient values between the two groups. Again, we conducted this in CALM (referred vs control), at an individual-level, as downstream analyses suggested a main effect of dataset (which is reflected in the highly-similar group-level referred and control CALM gradients). We have updated the Results section with the following text (Page 6, Line 283):

“To examine the effect of referral status on participant-level nodal gradient values in CALM, we conducted a series of generalized linear models controlling for head motion, sex and age at scan (Figure 1d). We restricted our analyses to baseline scans to reduce the difference in sample size for the referred (311 communicability and 177 functional gradients, respectively) and control participants (91 communicability and 79 functional gradients, respectively), and to the principal gradients. For communicability, 42 regions showed a significant effect (p < 0.05) of neurodivergence before FDR correction, with 9 post FDR correction. 8 of these 9 regions had negative t-statistics, suggesting a reduced nodal gradient value and representation in the neurodivergent children, encompassing both lower-order somatosensory cortices alongside higher-order fronto-parietal and default-mode networks. The largest reductions were observed within the prefrontal cortices of the defaultmode network (t = -3.992, p = 6.600 x 10^-5, p_FDR = 0.013, Cohen’s d = -0.476), the left orbitofrontal cortex of the limbic network (t = -3.710, p = 2.070 x 10^-4, p_FDR = 0.020, Cohen’s d = -0.442) and right somato-motor cortex (t = -3.612, p = 3.040 x 10^-4, p_FDR = 0.020, Cohen’s d = -0.431). The right visual cortex was the only exception, with stronger gradient representation within the neurotypical cohort (t = 3.071, p = 0.002, p_FDR = 0.048, Cohen’s d = 0.366). For functional connectivity, comparatively fewer regions exhibited a significant effect (p < 0.05) of neurotypicality, with 34 regions prior to FDR correction and 1 post. Significantly stronger gradient representation was observed in neurotypical children within the right precentral ventral division of the defaultmode network (t = 3.930, p = 8.500 x 10^-5, p_FDR = 0.017, Cohen’s d = 0.532). Together, this suggests that the strongest and most robust effects of neurodivergence are observed within gradients of communicability, rather than functional connectivity, where alterations in both affect higher-order associative regions”. 

In the harmonization methodology, it is mentioned that “if harmonisation was successful, we’d expect any significant effects of scanner type before harmonisation to be non-significant after harmonisation”. However, given that there were no significant effects before harmonization, the results reported do not help in evaluating the quality of harmonization.

We agree with the Reviewer, and have removed the post-harmonisation GLMs, and instead stating that there were no significant effects of scanner type before harmonization. 

Figure 3: It would be helpful to include a plot showing the GAMM predictions versus real observations of eccentricity (x-axis: predictions, y-axis: actual values). 

To plot the GAMM-predicted smooth effects of age, which we used for visualisation purposes only, we used the get_predictions function from the itsadug R package. This creates model predictions using the median value of nuisance covariates. Thus, whilst we specified the entire age range, the function automatically chooses the median of head motion, alongside controlling for sex (default level: male) and, for each dataset-specific trajectory. Since the gamm4 package separates the fitted model into a gam and linear mixed effects model (which accounts for participant ID as a random effect), and the get_predictions function only uses gam, random effects are not modelled in the predicted smooths. Therefore, any discrepancy between the observed and predicted manifold eccentricity values is likely due to sensitivity to default choices of covariates other than age, or random effects. To prevent Figure 3 being too over-crowded, we opted to not include the predictions: these were strongly correlated with real structural manifold data, but less for functional manifold data especially where significant developmental change was absent.

The 30mm threshold for filtering short streamlines in tractography is uncommon. What is the rationale for using such a large threshold, given the potential exclusion of many short-range association fibres?

A minimum length of 30mm was the default for the MRtrix3 reconstruction workflow, and something we have previously used. In a previous project, we systematically varied the minimum fibre length and found that this had minimal impact on network organisation (e.g. Mousley et al. 2025). However, we accept that short-range association fibres may have been excluded and have included this in the Discussion as a methodological limitation, alongside our predictions for how the gradients and structure-function coupling may’ve been altered had we included such fibres (Page 20, Line 955):

“A potential methodological limitation in the construction of structural connectomes was the 30mm tract length threshold which, despite being the QSIprep reconstruction default (Cieslak et al., 2021), may have potentially excluded short-range association fibres. This is pertinent as tracts of different lengths exhibit unique distributions across the cortex and functional roles (Bajada et al., 2019) : short-range connections occur throughout the cortex but peak within primary areas, including the primary visual, somato-motor, auditory, and para-hippocampal cortices, and are thought to dominate lower-order sensorimotor functional resting-state networks, whilst long-range connections are most abundant in tertiary association areas and are recruited alongside tracts of varying lengths within higher-order functional resting-state networks. Therefore, inclusion of short-range association fibres may have resulted in a relative increase in representation of lower-order primary areas and functional networks. On the other hand, we also note the potential misinterpretation of short-range fibres: they may be unreliably distinguished from null models in which tractography is restricted by cortical gyri only (Bajada et al., 2019). Further, prior (neonatal) work has demonstrated that the order of connectivity of regions and topological fingerprints are consistent across varying streamline thresholds (Mousley et al., 2025), suggesting minimal impact”. 

Given the spatial smoothing of fMRI data (6mm FWHM), it would be beneficial to apply connectome spatial smoothing to structural connectivity measures for consistent spatial smoothness.

This is an interesting suggestion but given we are looking at structural communicability within a parcellated network, we are not sure that it would make any difference. The data structural data are already very smooth. Nonetheless we have added the following text to the Discussion (Page 20, Line 968): 

“Given the spatial smoothing applied to the functional connectivity data, and examining its correspondence to streamline-count connectomes through structure-function coupling, applying the equivalent smoothing to structural connectomes may improve the reliability of inference, and subsequent sensitivity to cognition and psychopathology. Connectome spatial smoothing involves applying a smoothing kernel to the two streamline endpoints, whereby variations in smoothing kernels are selected to optimise the trade-off between subjectlevel reliability and identifiability, thus increasing the signal-to-noise ratio and the reliability of statistical inferences of brain-behaviour relationships (Mansour et al., 2022). However, we note that such smoothing is more effective for high-resolution connectomes, rather than parcel-level, and so have only made a modest improvement (Mansour et al., 2022)”.

Why was harmonization performed only within the CALM dataset and not across both CALM and NKI datasets? What was the rationale for this decision?

We thought about this very carefully. Harmonization aims to remove scanner or site effects, whilst retaining the crucial characteristics of interest. Our capacity to retain those characteristics is entirely dependent on them being *fully* captured by covariates, which are then incorporated into the harmonization process. Even with the best set of measures, the idea that we can fully capture ‘neurodivergence’ and thus preserve it in the harmonisation process is dubious. Indeed, across CALM and NKI there are limited number of common measures (i.e. not the best set of common measures), and thus we are limited in our ability to fully capture the neurodivergence with covariates. So, we worried that if we put these two very different datasets into the harmonisation process we would essentially eliminate the interesting differences between the datasets. We have added this text to the harmonization section of the Methods (Page 24, Line 1225):

“Harmonization aims to retain key characteristics of interest whilst removing scanner or site effects. However, the site effects in the current study are confounded with neurodivergence, and it is unlikely that neurodivergence may be captured fully using common covariates across CALM and NKI. Therefore, to preserve variation in neurodivergence, whilst reducing scanner effects, we harmonized within the CALM dataset only”. 

The exclusion of subcortical areas from connectivity analyses is not justified. 

This is a good point. We used the Schaefer atlas because we had previously used this to derive both functional and structural connectomes, but we agree that it would have been good to include subcortical areas (Page 20, Line 977). 

“A potential limitation of our study was the exclusion of subcortical regions. However, prior work has shed light on the role of subcortical connectivity in structural and functional gradients, respectively, of neurotypical populations of children and adolescents (Park et al., 2021; Xia et al., 2022). For example, in the context of the primary-to-transmodal and sensorimotor-to-visual functional connectivity gradients, the mean gradient scores within subcortical networks were demonstrated to be relatively stable across childhood and adolescence (Xia et al., 2022). In the context of structural connectivity gradients derived from streamline counts, which we demonstrated were highly consistent with those derived from communicability, subcortical structural manifolds weighted by their cortical connectivity were anchored by the caudate and thalamus at one pole, and by the hippocampus and nucleus accumbens at the opposite pole, with significant age-related manifold expansion within the caudate and thalamus (Park et al., 2021)”. 

In the KNN imputation method, were uniform weights used, or was an inverse distance weighting applied?

Uniform weights were used, and we have updated the manuscript appropriately.

The manuscript should clarify from the outset that the reported sample size (N) includes multiple longitudinal observations from the same individuals and does not reflect the number of unique participants.

We have rectified the Abstract (Page 2, Line 64) and Introduction (Page 3, Line 138):

“We charted the organisational variability of structural (610 participants, N = 390 with one observation, N = 163 with two observations, and N = 57 with three) and functional (512 participants, N = 340 with one observation, N = 128 with two observations, and N = 44 with three)”.

The term “structural gradients” is ambiguous in the introduction. Clarify that these gradients were computed from structural and functional connectivity matrices, not from other structural features (e.g. cortical thickness).

We have clarified this in the Introduction (Page 3, Line 134):

“Applying diffusion-map embedding as an unsupervised machine-learning technique onto matrices of communicability (from streamline SIFT2-weighted fibre bundle capacity) and functional connectivity, we derived gradients of structural and functional brain organisation in children and adolescents…”

Page 5: The sentence, “we calculated the normalized angle of each structural and functional connectome to derive symmetric affinity matrices” is unclear and needs clarification.

We have clarified this within the second paragraph of the Results section (Page 4, Line 185):

“To capture inter-nodal similarity in connectivity, using a normalised angle kernel, we derived individual symmetric affinity matrices from the left and right hemispheres of each communicability and functional connectivity matrix. Varying kernels capture different but highly-related aspects of inter-nodal similarity, such as correlation coefficients, Gaussian kernels, and cosine similarity. Diffusion-map embedding is then applied on the affinity matrices to derive gradients of cortical organisation”. 

Figure 1.a: “Affine A” likely refers to the affinity matrix. The term “affine” may be confusing; consider using a clearer label. It would also help to add descriptive labels for rows and columns (e.g. region x region).

Thank you for this suggestion! We have replaced each of the labels with “pairwise similarity”. We also labelled the rows and columns as regions.

Figure 1.d: Are the cross-group differences statistically significant? If so, please indicate this in the figure.

We have added the results of a series of linear mixed effects models to the legend of Figure 1 (Page 6, line 252):

“indicates a significant effect of dataset (p < 0.05) on variance explained within a linear mixed effects model controlling for head motion, sex, and age at scan”.

The sentence “whose connectomes were successfully thresholded” in the methods is unclear. What does “successfully thresholded” mean? Additionally, this seems to be the first mention of the Schaefer 100 and Brainnetome atlas; clarify where these parcellations are used. 

We have amended the Methodology section (Page 23, Line 1138):

“For each participant, we retained the strongest 10% of connections per row, thus creating fully connected networks required for building affinity matrices. We excluded any connectomes in which such thresholding was not possible due to insufficient non-zero row values. To further ensure accuracy in connectome reconstruction, we excluded any participants whose connectomes failed thresholding in two alternative parcellations: the 100node Schaefer 7-network (Schaefer et al., 2018) and Brainnetome 246-node (Fan et al., 2016) parcellations, respectively”. 

We have also specified the use of the Schaefer 200-node parcellation in the first sentence on the second Results paragraph.

The use of “streamline counts” is misleading, as the method uses SIFT2-weighted fibre bundle capacity rather than raw streamline counts. It would be better to refer to this measure as “SIFT2-weighted fibre bundle capacity” or “FBC”.

We replaced all instances of “streamline counts” with “SIFT2-weighted fibre bundle capacity” as appropriate.

Figure 2.c: Consider adding plots showing changes in eccentricity against (1) degree centrality, and (2) weighted local clustering coefficient. Additionally, a plot showing the relationship between age and mean eccentricity (averaged across nodes) at the individual level would be informative.

We added the correlation between eccentricity and both degree centrality and the weighted local clustering coefficient and included them in our dominance analysis in Figure 2. In terms of the relationship between age and mean (global) eccentricity, these are plotted in Figure 3. 

Figure 2.b: Considering the results of the following sections, it would be interesting to include additional KDE/violin plots to show group differences in the distribution of eccentricity within 7 different functional networks.

As part of our analysis to parse neurotypicality and dataset effects, we tested for group differences in the distribution of structural and functional manifold eccentricity within each of the 7 functional networks in the referred and control portions of CALM and have included instances of significant differences with a coloured arrow to represent the direction of the difference within Figure 3. 

Figure 3: Several panels lack axis labels for x and y axes. Adding these would improve clarity.

To minimise the amount of text in Figure 3, we opted to include labels only for the global-level structural and functional results. However, to aid interpretation, we added a small schematic at the bottom of Figure 3 to represent all axis labels. 

The statement that “differences between datasets only emerged when taking development into account” seems inaccurate. Differences in eccentricity are evident across datasets even before accounting for development (see Fig 2.b and the significance in the Scheirer-Ray-Hare test).

We agree – differences in eccentricity across development and datasets are evident in structural and functional manifold eccentricity, as well as within structure-function coupling. However, effects of neurotypicality were particularly strong for the maturation of structure-function coupling, rather than magnitude. Therefore, we have rephrased this sentence in the Discussion (page 18, line 832):

“Furthermore, group-level structural and functional gradients were highly consistent across datasets, whilst differences between datasets were emphasised when taking development into account, through differing rates of structural and functional manifold expansion, respectively, alongside maturation of structure-function coupling”.

The handling of longitudinal data by adding a random effect for individuals is not clear in the main text. Mentioning this earlier could be helpful. 

We have included this detail in the second sentence of the “developmental trajectories of structural manifold contraction and functional manifold expansion” results sub-section (page 11, line 503):

“We included a random effect for each participant to account for longitudinal data”. 

Figure 4.b: Why were ranks shown instead of actual coefficient of variation values? Consider including a cortical map visualization of the coefficients in the supplementary material.

We visualised the ranks, instead of the actual coefficient of variation (CV) values, due to considerable variability and skew in the magnitude of the CV, ranging from 28.54 (in the right visual network) to 12865.68 (in the parietal portion of the left default-mode network), with a mean of 306.15. If we had visualised the raw CV values, these larger values would’ve been over-represented. We’ve also noticed and rectified an error in the labelling of the colour bar for Figure 4b: the minimum should be most variable (i.e. a rank of 1). To aid contextualisation of the ranks, we have added the following to the Results (page 14, line 626):

“The distribution of cortical coefficients of variation (CV) varied considerably, with the largest CV (in the parietal division of the left default-mode network) being over 400 times that of the smallest (in the right visual network). The distribution of absolute CVs was positively skewed, with a Fisher skewness coefficient g₁ of 7.172, meaning relatively few regions had particularly high inter-individual variability, and highly peaked, with a kurtosis of 54.883, where a normal distribution has a skewness coefficient of 0 and a kurtosis of 3”. 

Reviewer #2 (Public review):

Some differences in developmental trajectories between CALM and NKI (e.g. Figure 4d) are not explained. Are these differences expected, or do they suggest underlying factors that require further investigation?

This is a great point, and we appreciate the push to give a fuller explanation. It is very hard to know whether these effects are expected or not. We certainly don’t know of any other papers that have taken this approach. In response to the reviewer’s point, we decided to run some more analyses to better understand the differences. Having observed stronger age effects on structure-function coupling within the neurotypical NKI dataset, compared to the absent effects in the neurodivergent portion of CALM, we wanted to follow up and test that it really is that coupling is more sensitive to the neurodivergent versus neurotypical difference between CALM and NKI (rather than say, scanner or site effects). In short, we find stronger developmental effects of coupling within the neurotypical portion of CALM, rather than neurodivergent, and have added this to the Results (page 15, line 701):

“To further examine whether a closer correspondence of structure-function coupling with age is associated with neurotypicality, we conducted a follow-up analysis using the additional age-matched neurotypical portion of CALM (N = 77). Given the widespread developmental effects on coupling within the neurotypical NKI sample, compared to the absent effects in the neurodivergent portion of CALM, we would expect strong relationships between age and structure-function coupling with the neurotypical portion of CALM. This is indeed what we found: structure-function coupling showed a linear negative relationship with age globally (F = 16.76, p_FDR < 0.001, adjusted R² = 26.44%), alongside fronto-parietal (F = 9.24, p_FDR = 0.004, adjusted R² = 19.24%), dorsalattention (F = 13.162, p_FDR = 0.001, adjusted R²= 18.14%), ventral attention (F = 11.47, p_FDR  = 0.002, adjusted R²= 22.78), somato-motor (F = 17.37, p_FDR  < 0.001, adjusted R²= 21.92%) and visual (F = 11.79, p_FDR  = 0.002, adjusted R²= 20.81%) networks. Together, this supports our hypothesis that within neurotypical children and adolescents, structure-function coupling decreases with age, showing a stronger effect compared to their neurodivergent counterparts, in tandem with the emergence of higher-order cognition. Thus, whilst the magnitude of structure-function coupling across development appeared insensitive to neurotypicality, its maturation is sensitive. Tentatively, this suggests that neurotypicality is linked to stronger and more consistent maturational development of structure-function coupling, whereby the tethering of functional connectivity to structure across development is adaptive”. 

In conjunction with the Reviewer’s later request to deepen the Discussion, we have included an additional paragraph attempting to explain the differences in neurodevelopmental trajectories of structure-function coupling (Page 19, Line 924):

“Whilst the spatial patterning of structure-function coupling across the cortex has been extensively documented, as explained above, less is known about developmental trajectories of structure-function coupling, or how such trajectories may be altered in those with neurodevelopmental conditions. To our knowledge, only one prior study has examined differences in developmental trajectories of (non-manifold) structure-function coupling in typically-developing children and those with attention-deficit hyperactivity disorder (Soman et al., 2023), one of the most common conditions in the neurodivergent portion of CALM. Namely, using cross-sectional and longitudinal data from children aged between 9 and 14 years old, they demonstrated increased coupling across development in higher-order regions overlapping with the defaultmode, salience, and dorsal attention networks, in children with ADHD, with no significant developmental change in controls, thus encompassing an ectopic developmental trajectory (Di Martino et al., 2014; Soman et al., 2023). Whilst the current work does not focus on any condition, rather the broad mixed population of young people with neurodevelopmental symptoms (including those with and without diagnoses), there are meaningful individual and developmental differences in structure-coupling. Crucially, it is not the case that simply having stronger coupling is desirable. The current work reveals that there are important developmental trajectories in structure-function coupling, suggesting that it undergoes considerable refinement with age. Note that whilst the magnitude of structure-function coupling across development did not differ significantly as a function of neurodivergence, its relationship to age did. Our working hypothesis is that structural connections allow for the ordered integration of functional areas, and the gradual functional modularisation of the developing brain. For instance, those with higher cognitive ability show a stronger refinement of structurefunction coupling across development. Future work in this space needs to better understand not just how structural or functional organisation change with time, but rather how one supports the other”. 

The use of COMBAT may have excluded extreme participants from both datasets, which could explain the lack of correlations found with psychopathology.

COMBAT does not exclude participants from datasets but simply adjusts connectivity estimates. So, the use of COMBAT will not be impacting the links with psychopathology by removing participants. But this did get us thinking. Excluding participants based on high motion may have systematically removed those with high psychopathology scores, meaning incomplete coverage. In other words, we may be under-representing those at the more extreme end of the range, simply because their head-motion levels are higher and thus are more likely to be excluded. We found that despite certain high-motion participants being removed, we still had good coverage of those with high scores and were therefore sensitive within this range. We have added the following to the revised Methods section (Page 26, Line 1338):

“As we removed participants with high motion, this may have overlapped with those with higher psychopathology scores, and thus incomplete coverage. To examine coverage and sensitivity to broad-range psychopathology following quality control, we calculated the Fisher-Pearson skewness statistic g₁ for each of the 6 Conners t-statistic measures and the proportion of youth with a t-statistic equal to or greater than 65, indicating an elevated or very elevated score. Measures of inattention (g₁ = 0.11, 44.20% elevated), hyperactivity/impulsivity (g₁ = 0.48, 36.41% elevated), learning problems (g₁ = 0.45, 37.36% elevated), executive functioning (g₁ = 0.27, 38.16% elevated), aggression (g₁ = 1.65, 15.58% elevated), and peer relations (g₁ = 0.49, 38% elevated) were positively skewed and comprised of at least 15% of children with elevated or very elevated scores, suggesting sufficient coverage of those with extreme scores”. 

There is no discussion of whether the stable patterns of brain organization could result from preprocessing choices or summarizing data to the mean. This should be addressed to rule out methodological artifacts. 

This is a brilliant point. We are necessarily using a very lengthy pipeline, with many design choices to explore structural and functional gradients and their intersection. In conjunction with the Reviewer’s later suggestion to deepen the Discussion, we have added the following paragraph which details the sensitivity analyses we carried out to confirm the observed stable patterns of brain organization (Page 18, Line 863):

“That is, whilst we observed developmental refinement of gradients, in terms of manifold eccentricity, standard deviation, and variance explained, we did not observe replacement. Note, as opposed to calculating gradients based on group data, such as a sliding window approach, which may artificially smooth developmental trends and summarise them to the mean, we used participant-level data throughout. Given the growing application of gradient-based analyses in modelling structural (He et al., 2025; Li et al., 2024) and functional (Dong et al., 2021; Xia et al., 2022) brain development, we hope to provide a blueprint of factors which may affect developmental conclusions drawn from gradient-based frameworks”.

Although imputing missing data was necessary, it would be useful to compare results without imputed data to assess the impact of imputation on findings. 

It is very hard to know the impact of imputation without simply removing those participants with some imputed data. Using a simulation experiment, we expressed the imputation accuracy as the root mean squared error normalized by the range of observable data in each scale. This produced a percentage error margin. We demonstrate that imputation accuracy across all measures is at worst within approximately 11% of the observed data, and at best within approximately 4% of the observed data, and have included the following in the revised Methods section (Page 27, Line 1348):

“Missing data

To avoid a loss of statistical power, we imputed missing data. 27.50% of the sample had one or more missing psychopathology or cognitive measures (equal to 7% of all values), and the data was not missing at random: using a Welch’s t-test, we observed a significant effect of missingness on age [t (264.479) = 3.029, p = 0.003, Cohen’s d = 0.296], whereby children with missing data (M = 12.055 years, SD = 3.272) were younger than those with complete data (M = 12.902 years, SD = 2.685). Using a subset with complete data (N = 456), we randomly sampled 10% of the values in each column with replacement and assigned those as missing, thereby mimicking the proportion of missingness in the entire dataset. We conducted KNN imputation (uniform weights) on the subset with complete data and calculated the imputation accuracy as the root mean squared error normalized by the observed range of each measure. Thus, each measure was assigned a percentage which described the imputation margin of error. Across cognitive measures, imputation was within a 5.40% mean margin of error, with the lowest imputation error in the Trail motor speed task (4.43%) and highest in the Trails number-letter switching task (7.19%). Across psychopathology measures, imputation exhibited a mean 7.81% error margin, with the lowest imputation error in the Conners executive function scale (5.75%) and the highest in the Conners peer relations scale (11.04%). Together, this suggests that imputation was accurate”.

The results section is extensive, with many reports, while the discussion is relatively short and lacks indepth analysis of the findings. Moving some results into the discussion could help balance the sections and provide a deeper interpretation. 

We agree with the Reviewer and appreciate the nudge to expand the Discussion section. We have added 4 sections to the Discussion. The first explores the importance of the default-mode network as a region whose coupling is most consistently predicted by working memory across development and phenotypes, in terms of its underlying anatomy (Paquola et al., 2025) (Page 20, Line 977):

“An emerging theme from our work is the importance of the default-mode network as a region in which structure-function coupling is reliably predicted by working memory across neurodevelopmental phenotypes and datasets during childhood and adolescence. Recent neurotypical adult investigations combining highresolution post-mortem histology, in vivo neuroimaging, and graph-theory analyses have revealed how the underlying neuroanatomy of the default-mode network may support diverse functions (Paquola et al., 2025), and thus exhibit lower structure-function coupling compared to unimodal regions. The default-mode network has distinct neuroanatomy compared to the remaining 6 intrinsic resting-state functional networks (Yeo et al., 2011), containing a distinctive combination of 5 of the 6 von Economo and Koskinas cell types (von Economo & Koskinas, 1925), with an over-representation of heteromodal cortex, and uniquely balancing output across all cortical types. A primary cytoarchitectural axis emerges, beyond which are mosaic-like spatial topographies. The duality of the default-mode network, in terms of its ability to both integrate and be insulated from sensory information, is facilitated by two microarchitecturally distinct subunits anchored at either end of the cytoarchitectural axis (Paquola et al., 2025). Whilst beyond the scope of the current work, structure-function coupling and their predictive value for cognition may also differ across divisions within the default-mode network, particularly given variability in the smoothness and compressibility of cytoarchitectural landscapes across subregions (Paquola et al., 2025)”. 

The second provides a deeper interpretation and contextualisation of greater sensitivity of communicability, rather than functional connectivity, to neurodivergence (Page 19, Lines 907):

“We consider two possible factors to explain the greater sensitivity of neurodivergence to gradients of communicability, rather than functional connectivity. First, functional connectivity is likely more sensitive to head motion than structural-based communicability and suffers from reduced statistical power due to stricter head motion thresholds, alongside greater inter-individual variability. Second, whilst prior work contrasting functional connectivity gradients from neurotypical adults with those with confirmed ASD diagnoses demonstrated vertex-level reductions in the default-mode network in ASD and marginal increases in sensorymotor communities (Hong et al., 2019), indicating a sensitivity of functional connectivity to neurodivergence, important differences remain. Specifically, whilst the vertex-level group-level differences were modest, in line with our work, greater differences emerged when considering step-wise functional connectivity (SFC); in other words, when considering the dynamic transitions of or information flow through the functional hierarchy underlying the static functional connectomes, such that ASD was characterised by initial faster SFC within the unimodal cortices followed by a lack of convergence within the default-mode network (Hong et al., 2019). This emphasis on information flow and dynamic underlying states may point towards greater sensitivity of neurodivergence to structural communicability – a measure directly capturing information flow – than static functional connectivity”. 

The third paragraph situates our work within a broader landscape of reliable brain-behaviour relationships, focusing on the strengths of combining clinical and normative samples to refine our interpretation of the relationship between gradients and cognition, as well as the importance of equifinality in developmental predictive work (Page 20, line 994):

“In an effort to establish more reliable brain-behaviour relationships despite not having the statistical power afforded by large-scale, typically normative, consortia (Rosenberg & Finn, 2022), we demonstrated the development-dependent link between default-mode structure-function coupling and working memory generalised across clinical (CALM) and normative (NKI) samples, across varying MRI acquisition parameters, and harnessing within- and across-participant variation. Such multivariate associations are likely more reliable than their univariate counterparts (Marek et al., 2022), but can be further optimised using task-related fMRI (Rosenberg & Finn, 2022). The consistency, or lack of, of developmental effects across datasets emphasises the importance of validating brain-behaviour relationships in highly diverse samples. Particularly evident in the case of structure-function coupling development, through our use of contrasting samples, is equifinality (Cicchetti & Rogosch, 1996), a key concept in developmental neuroscience: namely, similar ‘endpoints’ of structure-function coupling may be achieved through different initialisations dependent on working memory. 

The fourth paragraph details methodological limitations in response to Reviewer 1’s suggestions to justify the exclusion of subcortical regions and consider the role of spatial smoothing in structural connectome construction as well as the threshold for filtering short streamlines”. 

While the methods are thorough, it is not always clear whether the optimal approaches were chosen for each step, considering the available data. 

In response to Reviewer 1’s concerns, we conducted several sensitivity analyses to evaluate the robustness of our results in terms of procedure. Specifically, we evaluated the impact of thresholding (full or sparse), level of analysis (individual or group gradients), construction of the structural connectome (communicability or fibre bundle capacity), Procrustes rotation (alignment to group-level gradients before Procrustes), tracking the variance explained in individual connectomes by group-level gradients, impact of head motion, and distinguishing between site and neurotypicality effects. All these analyses converged on the same conclusion: whilst we observe some developmental refinement in gradients, we do not observe replacement. We refer the reviewer to their third point, about whether stable patterns of brain organization were artefactual. 

The introduction is overly long and includes numerous examples that can distract readers unfamiliar with the topic from the main research questions. 

We have removed the following from the Introduction, reducing it to just under 900 words:

“At a molecular level, early developmental patterning of the cortex arises through interacting gradients of morphogens and transcription factors (see Cadwell et al., 2019). The resultant areal and progenitor specialisation produces a diverse pool of neurones, glia, and astrocytes (Hawrylycz et al., 2015). Across childhood, an initial burst in neuronal proliferation is met with later protracted synaptic pruning (Bethlehem et al., 2022), the dynamics of which are governed by an interplay between experience-dependent synaptic plasticity and genomic control (Gottlieb, 2007)”.

“The trends described above reflect group-level developmental trends, but how do we capture these broad anatomical and functional organisational principles at the level of an individual?”

We’ve also trimmed the second Introduction paragraph so that it includes fewer examples, such as removal of the wiring-cost optimisation that underlies structural brain development, as well as removing specific instances of network segregation and integration that occur throughout childhood.

Here we can argue that it was not just the edwardian conquest which sought to undermine identity - it was only really in the remaining north and west, Gwynedd in particular

So we're not going to be hearing about the Iroquois Confederacy, then.

An unproductive waste, because there were no public monuments. Convenient definition, for the invaders!