Luciano Dyballa

The paper uses encoding manifolds to analyze how mouse visual cortical areas represent natural scenes and static gratings, revealing layer-specific and cell-type-specific response patterns.

This study compares how retinal and V1 neural populations encode visual features, finding discrete encoding in retina and continuous in V1, with CNNs resembling retinal encoding.

The paper uses flow stimuli to show that mouse V1 neurons respond to higher spatial frequencies than previously thought, challenging conventional linear models of visual processing.