This is a striking demonstration of dynamic reversal of information flow between IT and PFC under perceptual uncertainty. The temporal precedence of PFC representations for boundary stimuli suggests not merely top-down modulation, but a conditional inversion of inferential primacy within the same cortical circuit.

One way to interpret this reversal is in terms of question inference rather than answer inference. When forward prediction-error minimization becomes ill-posed, higher-order regions may need to infer and impose candidate questions or hypotheses that structure subsequent sensory representations before classical forward inference can proceed. This contrasts an “answer → question” regime with the more familiar “question → answer” regime and may be essential for resolving ambiguity and enabling representational re-framing.

It would be interesting to examine whether similar reversals occur in domains requiring abstraction or cognitive flexibility, and whether failures of such reversibility contribute to phenotypes marked by concreteness or rigidity. Herbert Harris, MD, PhD