Coordination Dynamics in Cognitive Neuroscience

Abstract

Many researchers and clinicians in cognitive neuroscience hold to a modular view of cognitive function in which the cerebral cortex operates by the activation of areas with circumscribed elementary cognitive functions. Yet an ongoing paradigm shift to a dynamic network perspective is underway. This new viewpoint treats cortical function as arising from the coordination dynamics within and between cortical regions. Cortical coordination dynamics arises due to the unidirectional influences imposed on a cortical area by inputs from other areas that project to it, combined with the projection reciprocity that characterizes cortical connectivity and gives rise to reentrant processing. As a result, cortical dynamics exhibits both segregative and integrative tendencies and gives rise to both cooperative and competitive relations within and between cortical areas that are hypothesized to underlie the emergence of cognition in brains.

Keywords: HKB model; cerebral cortex; computational context; event-related potential; interareal interaction; local field potential; neuronal communication; relative coordination.

Figure 1

Coordination dynamics of a coupled dynamical system. The coordination dynamics is represented by the relation between the relative phase of the coupled components Φ, and its first time derivative $\dot{\varphi }$ (vertical axis). It is captured by the extended HKB model, which generated these graphs. Thick solid and broken lines correspond to attractive and repelling fixed points of the dynamics. At the left, the symmetric condition is shown with parameter δω = 0.00. At the right is the broken symmetry condition with parameter δω = 0.50. Slices through the surface in the broken-symmetry case show different dynamics, depending on the control parameter, k. For high values of k, representing a low frequency (long period) of oscillation, two stable fixed points near Φ = 0° and Φ = 180°, and two unstable fixed points, appear. For intermediate values of k, one stable fixed point disappears in a saddle-node bifurcation. For low values of k, the remaining stable fixed point disappears the same way. Metastable, intermittent dynamics is observed for low values of k: although there are no longer any fixed points, there is still attraction to the remnants of the previously stable states (Modified figure reproduced with permission from Kelso, 1994).