Séminaire interne - Sébastien Wolf (IBENS)

Auditory selective attention is a cognitive ability enabling to adapt the processing of stimuli according to their relevance in a given context. The initial cortical processing of sounds is performed in the primary auditory cortex (A1) and it has been shown that attention contributes to its reorganization. How this flexibility is accomplished by neural circuits is still an open question. To study the neural coding of auditory selective attention, we used two-photon imaging in A1 L2/3 of mice performing a sound-discrimination task in different contexts. We discovered a unique neural population that shows attentional gain, encodes behavioral choice, drives the decision of the animal, and displays a ramping activity until the reward delivery. To characterize network representation of the context and the decision, we analyzed functional connectivity using Boltzmann Machine inference and neural decoding. Our results suggest that context dependent sensory-based decision-making involves two neural populations in A1 with strong recurrent connectivity and cross inhibition.