Synergistic population encoding and precise coordinated variability across interlaminar ensembles in the early visual system

Abstract

Sensory stimuli are represented by the joint activity of large populations of neurons across the mammalian cortex. Information in such responses is limited by trial-to-trial variability. Because that variability is not independent between neurons, it has the potential to improve or degrade the amount of sensory information in the population response. How visual information scales with population size remains an open empirical question. Here, we use Neuropixels to simultaneously record tens to hundreds of single neurons in primary visual cortex (V1) and lateral geniculate nucleus (LGN) of mice and estimate population information. We found a mix of synergistic and redundant coding: synergy predominated in small populations (2-12 cells) before giving way to redundancy. The shared variability of this coding regime included global shared spike count variability at longer timescales, layer specific shared spike count variability at finer timescales, and shared variability in spike timing (jitter) that linked ensembles that span layers. Such ensembles defined by their shared variability carry more information. Our results suggest fine time scale stimulus encoding may be distributed across physically overlapping but distinct ensembles in V1.