NBIO Presents: Marius Pachitariu (Janelia Research Campus)

Linking neural representations to behavior using generalization

Sensory-guided decisions are the result of sensorimotor transformations across many brain areas. Recent studies have localized the motor components of these transformations using correlations between brain-wide neural activity and behavior. It has been more difficult to localize sensory computations with the same approach, as it requires pushing animal performance close to psychophysical thresholds, which is typically not feasible. Here we developed a new approach for linking sensory computations to behavior by training mice to discriminate between two stimuli and testing their responses to new stimuli. We then calculated the similarity of neural representations between train and test stimuli, using recordings of over 50,000 simultaneously-recorded neurons from 9 primary and higher-order visual areas (HVAs) and across cortical depths. We found that neural discrimination on the test images but not the train images correlated with the behavioral discrimination of the animals, across a large number of stimulus pairs and animals. The link between neural and behavioral performance was highest in the medial HVAs, suggesting this region as a critical component of sensory transformations and generalization.