De Vries, S. E. J. et al. Nat. Neurosci. 23, 138–151 (2020).

Numerous studies have reported recordings from neurons in the visual cortex, but such studies have typically been limited in the number of neurons being recorded and have used a variety of different stimuli. De Vries et al. have acquired a large dataset under standardized experimental conditions to address this limitation. The researchers performed calcium imaging using two-photon microscopy in awake, behaving mice. They imaged activity in about 60,000 excitatory and inhibitory neurons in the visual cortex while the mice were presented with a battery of visual stimuli ranging from drifting gratings to natural movies. The researchers could classify many neurons into several functional response classes and model their responses by combining linear filters and nonlinearities. Nevertheless, many of the recorded neurons could not be modeled, and these may be driven by highly specific stimuli not represented in the battery of stimuli presented here or by non-visual features of the mouse behavior. The dataset is available at http://observatory.brain-map.org/visualcoding.