黑料社

Neural correlates of optimized invariant coding in sensory systems

Amin Akhshi, 黑料社
Tuesday September 9, 12-1pm
Zoom Link:听
In Person: 550 Sherbrooke, Room 189

Abstract:听It has been widely observed that neurons ubiquitously exhibit substantial heterogeneities in their firing patterns, including variability in burst firing (i.e., clusters of action potentials followed by quiescence). However, the factors driving these firing heterogeneities and their functional role in population coding of incoming sensory information remain a central challenge in neuroscience.

We combined computational modeling with in vivo recordings from pyramidal cells in the electrosensory system of weakly electric fish to investigate the sources and functional impact of this heterogeneity. Specifically, we examined how the brain maintains stable representations of sensory objects despite contextual changes. Our conductance-based model accurately reproduced diverse spiking activities in vivo and demonstrated that extrinsic synaptic input, rather than intrinsic cellular properties, primarily drives spiking heterogeneity at the population level. We then investigated the functional consequences of this diversity for invariant coding, which allows for stable object recognition across different contexts. Our results revealed that burst firing, caused by the synergistic interaction between extrinsic input and intrinsic properties, is a key mechanism that promotes these invariant representations of natural stimuli at the population level, while the underlying heterogeneities contribute to discrimination between details of the input context.

Given the shared anatomical and physiological features across sensory modalities, these findings likely generalize broadly. We thus propose a novel function for burst firing in enhancing the invariant neural representation of sensory input at the population level.