A fundamental question in neuroscience is how the brain computes and codes using the spiking output of neural populations. This has driven rapid advances in large-scale recording technology at the cellular level, but revealed a fundamental bottleneck: once recorded, how do we decode the high-dimensional data-sets of activity? One solution may lie in the presumed modularity at all scales in the brain, providing a dramatic reduction in dimensional complexity. Using the Aplysia locomotion circuit as a model, we show how harnessing modularity of large-scale recordings can rapidly decompose both the functional and physical organisation of a neural system. We introduce powerful, general methods derived from network theory that decompose population activity into functional modules, and introduce a new parameter-free mixture-of-models approach to the problem of classifying types of neurons and modules based on their patterns of activity. Using these methods, we show a comprehensive mapping of the neural basis for the Aplysia’s escape locomotion programme (With Angela Bruno and Bill Frost at the Chicago Medical School).
Mark Humphries is a MRC Senior non-Clinical Fellow in the Faculty of Life Sciences, University of Manchester. Previously, he was a research fellow in the Group for Neural Theory at Ecole Normale Superieure (Paris) from 2009-2012, and did his PhD and postdoctoral training at the University of Sheffield with Kevin Gurney and Peter Redgrave. His research interests include the function and dysfunction of the basal ganglia and their components, analysis of large-scale recordings, and population coding; he also finds the simple pleasures of network theory to be a welcome distraction from the unfathomable complexities of the brain.