Schedule of Events | Symposia

The Connectivity Crisis

Poster Session F - Tuesday, April 1, 2025, 8:00 – 10:00 am EDT, Back Bay Ballroom/Republic Ballroom
Also presenting in Data Blitz Session 4 - Saturday, March 29, 2025, 10:30 am – 12:00 pm EDT, Constitution B.

Ole Jensen¹ (ole.jensen@psych.ox.ac.uk), Hyojin Park², Oscar Ferrante²; ¹Department of Experimental Psychology, Department of Psychiatry, University of Oxford, ²Centre for Human Brain Health, University of Birmingham

The case is often made that the brain should be investigated as a network rather than studying individual brain regions in isolation. It has been proposed that interregional functional connectivity identified from electrophysiological data is reflected by phase-synchronization in the gamma band and that the degree of synchronisation is task-dependent thus reflecting communication. Other ‘non-oscillatory’ methodologies have also been tested e.g. information theoretical measures, Granger causality and dynamical causal modelling. Nevertheless, the field of cognitive neuroscience has after more than three decades not converged on a commonly accepted measure of task-related functional brain connectivity estimated from electrophysiological data. We suggest that the lack of convergence is anchored in the observation that few findings on task-modulated connectivity are reproduced across laboratories. We challenge the community to prove us wrong by pointing to a MEG dataset collected by the Cogitate consortium (https://www.arc-cogitate.com/). This dataset allows for testing and verifying measures of functional connectivity associated with visual perception. In contrast to measures of interregional connectivity, the application of multivariate approaches relying on spatially distributed activity patterns has proven very powerful for identifying representational and task-specific neuronal activity. These multivariate patterns are highly distributed and reflect network interactions across the brain. We argue that theories based on circuit diagrams relying on estimating functional connectivity between a handful of regions should be abandoned in favour of testable models embracing neuronal computations distributed across the brain.

Topic Area: METHODS: Electrophysiology