Data for Festa et al 2021 - Neuronal variability reflects probabilistic inference tuned to natural image statistics

This data is associated to the following paper: Festa D., Aschner A, Davila A, Kohn A, Coen-Cagli R. Neuronal variability reflects probabilistic inference tuned to natural image statistics. The data consists of: 1) photographic natural images from the BSD500 dataset https://github.com/BIDS/BSDS, used to train the Gaussian Scale Mixture model. Model equations and implementation details are fully described in the associated paper. 2) multi-electrode recordings from V1 in anesthetized and awake macaque monkeys, while natural images and gratings were flashed on the screen. Recordings were performed using “Utah” electrode arrays. Images were presented at different sizes and orientations, to quantify surround modulation of response strength and variability in single neurons. Experimental procedures and stimuli are fully described in the associated paper. Code to read in and process this dataset is provided at https://github.com/rubencoencagli/festa-et-al-2020 . The code reproduces the main figures of the associated paper.