Although individual neurons can be highly selective to particular stimuli and certain upcoming
actions, they can provide a complex representation of stimuli and actions at the level of
population. The ability to dynamically allocate neural resources is crucial for cognitive
flexibility. However, it is unclear whether cognitive flexibility emerges from changes in activity
at the level of individual neurons, population, or both. By applying a combination of decoding
and encoding methods to simultaneously recorded neural data, we show that while maintaining
their stimulus selectivity, neurons in prefrontal cortex alter their correlated activity during
various cognitive states, resulting in an enhanced representation of visual space. During a task
with various cognitive states, individual prefrontal neurons maintained their limited spatial
sensitivity between visual encoding and saccadic target selection whereas the population
selectively improved its encoding of spatial locations far from the neurons’ preferred locations.
This ‘encoding expansion’ relied on high-dimensional neural representations and was
accompanied by selective reductions in noise correlation for non-preferred locations. Our results
demonstrate that through recruitment of less-informative neurons and reductions of noise
correlation in their activity, the representation of space by neuronal ensembles can be
dynamically enhanced, and suggest that cognitive flexibility is mainly achieved by changes in
neural representation at the level of population of prefrontal neurons rather than individual
neurons.
Browse: Home
/ Publications
/ Enhanced representation of space by prefrontal neuronal ensembles and its dependence on cognitive states