Discretized representations in V1 predictsuboptimal orientation discrimination
Corbo, J.; Erkat, O. B.; McClure, J. P.; Khdour, H. Y.; Polack, P.-O.
Show abstract
Neuronal population activity in sensory cortices is the substrate for perceptual decisions. Yet, we still do not understand how neuronal information content in sensory cortices relates to behavioral reports. To reconcile neurometric and psychometric performance, we recorded the activity of V1 neurons in mice performing a Go/NoGo orientation discrimination task. We found that, around the discrimination threshold, V1 does not represent the orientation of the stimuli as canonically expected. Instead, it forms categorical representations characterized by a relocation of activity at task-relevant domains of the orientation representational space. The relative neuronal activity at those discrete domains accurately predicted the probabilities of the animals decisions. Our results thus suggest that the categorical integration of discretized feature representations from sensory cortices explains perceptual decisions.
Matching journals
The top 2 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Layer 6 ensembles can selectively regulate the behavioral impact and layer-specific representation of sensory deviants 98%
- Slowly evolving dopaminergic activity modulates the moment-to-moment probability of movement initiation. 98%
- Neural circuit mechanisms for steering control in walking Drosophila 97%
Similar papers in this journal
- A cortical circuit mechanism for coding and updating task structural knowledge in inference-based decision-making 98%
- Universal statistics of hippocampal place fields across species and dimensionalities 98%
- Ventral frontostriatal circuitry mediates the computation of reinforcement from symbolic gains and losses 98%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.