The predictive power of intrinsic timescale during the perceptual decision-making process across the mouse brain

Across the cortical hierarchy, single neurons are characterized by differences in the extent to which they can sustain their firing rate over time (i.e., their "intrinsic timescale"). Previous studies have demonstrated that neurons in a given brain region mostly exhibit either short or long intrinsic timescales. In this study, we sought to identify populations of neurons that accumulate information over different timescales in the mouse brain and to characterize their functions in the context of a visual discrimination task. Thus, we separately examined the neural population dynamics of neurons with long or short intrinsic timescales across different brain regions. More specifically, we looked at the decoding performance of these neural populations aligned to different task variables (stimulus onset, movement). Taken together, our population-level findings support the hypothesis that long intrinsic timescale neurons encode abstract variables related to decision formation. Furthermore, we investigated whether there was a relationship between how well a single neuron represents the animal's choice or stimuli and their intrinsic timescale. We did not observe any significant relationship between the decoding of these task variables and a single neuron's intrinsic timescale. In summary, our findings support the idea that the long intrinsic timescale population of neurons, which appear at different levels of the cortical hierarchy, are primarily more involved in representing the decision variable.