这是页面里最大胆的主张，也是最需要细节支撑的主张。传统超声波成像分辨率远低于MRI，且对骨骼和含气组织（如肺）的穿透能力很弱——这是物理限制，不是单纯的工程问题。「在许多方面优于MRI」意味着他们声称克服了某些根本性限制，但页面完全没有说明是如何做到的。60秒的扫描时间如果属实，在可及性上确实远优于MRI（通常需要30-90分钟）。这个主张的可信度在看到同行评审数据之前只能存疑。

Gollub, R. L. (2022). Brain changes after COVID revealed by imaging. Nature. https://doi.org/10.1038/d41586-022-00503-x

Each pixel is an output from 125 million cells. It's similar to a country that votes for a president or prime minister. If you try to figure out how that leader was voted into power without analyzing all the voters, that is quite an impossible task!

Voxel (unit)

volumetric and pixel

pixel for given slice thickness

Douaud, G., Lee, S., Alfaro-Almagro, F., Arthofer, C., Wang, C., McCarthy, P., Lange, F., Andersson, J. L. R., Griffanti, L., Duff, E., Jbabdi, S., Taschler, B., Winkler, A. M., Nichols, T. E., Collins, R., Matthews, P. M., Allen, N., Miller, K. L., & Smith, S. M. (2021). Brain imaging before and after COVID-19 in UK Biobank (p. 2021.06.11.21258690). https://doi.org/10.1101/2021.06.11.21258690

Roy-Gash, F., Marine, D. M., Jean-Michel, D., Herve, V., Raphael, B., & Nicolas, E. (2020). COVID-19-associated acute cerebral venous thrombosis: Clinical, CT, MRI and EEG features. Critical Care, 24(1), 419. https://doi.org/10.1186/s13054-020-03131-x

Very nice review. Please note the reference [94] (Pantazatos et al.) is misplaced because they did not argue that spatial biases cannot entirely explain the putative links between CGE and functional segregation. Instead, they argued there was insufficient evidence in the original Richiardi et al. study linking elevated CGE with resting state functional networks, and that spatial biases may in fact entirely account for their findings. To describe the debate/exchange more accurately, I would suggest replacing the below sentence

“While spatial biases may contribute to these findings, as nodes belonging to the same module tend to be anatomically colocalized [7,8], they cannot explain these effects entirely [94,95].”

with the below paragraph:

“Spatial biases may contribute to these findings, as nodes belonging to the same module tend to be anatomically colocalized [7,8]. Pantazatos et al. argued that these findings are entirely explained by spatial biases [94]. They showed that elevated CGE, as defined in the original Richiardi et al. study, falls monotonically as longer distance edges are removed. Moreover, they showed that 1,000 sets of randomly spaced modules all have significantly high CGE when using the same null distribution defined in the original Richiardi et al. analyses. Therefore, elevated CGE is not specifically related to functional segregation as defined by resting state functional networks, which is in direct contradiction to the main conclusion of the original Richiardi et al. study. Since randomly placed modules do not align (spatially) with any distributed pattern of functional segregation, the finding of elevated CGE may instead be attributed entirely to anatomical colocalization of the nodes within each module. In their rebuttal to [94], Richiardi et al. argue spatial biases cannot explain their findings entirely [95]. However, the authors do not offer an explanation for significantly high CGE observed for randomly spaced sets of modules, other than to note that nodes tend to be closer on average compared to when modules are defined by resting state fMRI. Future work is required to dissociate the effects of spatially proximity on relationships between CGE and spatially distributed functional networks.”