New Frontiers in Population Recording
The advent of reliable simultaneous recording of the activity of many neurons has enabled the study of interactions between neurons at a large scale: the number of observed pairwise interactions is proportional to the square of the number of recorded neurons. The dominant phenomenon in these pairwise interactions is synchronization, reflecting a system where many observed variables have in common a smaller set of latent variables. This permits the possibility that the complex signals observed in the brain might be reducible to a simpler system. We used this insight to design a better signal processing scheme for neuroprosthetics; to identify the same neurons in many recording sessions from their pairwise interactions; to show that the tuning functions of neurons in motor and premotor cortex do not reflect simple coordinate frame models; and to identify error as a dominant signal during continuous movements.