Bifurcation in space: Functional modularity and distributed cognition in the neocortex

Abstract

Recent reports of widespread neural representations challenge the notion that brain areas are specialized in distinct aspects of cognition. Here we tackle this challenge, using connectome-based neocortex models endowed with macro-scopic gradients of neurobiological properties. We show that a bifurcation (an abrupt change of behavior) occurring locally in the cortical space gives rise to specialization of a subset of areas, that defines a multi-regional functional module, for subjective decision making or working memory coding. We found that mnemonic activity exhibits an inverted-V shaped pattern of timescales across the cortical hierarchy, which represents an experimentally testable prediction. A plethora of bifurcations in space coexist, corresponding to a variety of functional modules not well predicted by network analysis of structural modules; their associated timescale profiles could be observed in animals performing tasks that require different mental processes. These findings suggest bifurcation in space as a fundamental principle for understanding brains modular organization. TeaserThe theory of bifurcation in space mechanistically accounts for the emergence of functional specialization of cortical areas dedicated to distinct cognitive functions that is compatible with distributed neural representations in a multiregional cortex.