Recordings of ~20,000 neurons from V1 in response to oriented stimuli

dataset

posted on 2019-11-05, 16:56 authored by Marius PachitariuMarius Pachitariu, Michalis Michaelos, Carsen StringerCarsen Stringer

Recordings from Stringer et al, 2019, now including higher-order visual cortex. The code to reproduce the figures from the paper is here:

https://github.com/MouseLand/stringer-et-al-2019

Please use the "issues" section from this repository for all communications regarding this dataset.

Each npy file is one recording of deconvolved fluorescence traces + all its metadata, with the name of the file indicating the stimulus set.

Each stimulus set is a single image/movie, with each trial rotating that stimulus by a random angle between 0 and 360. Unless otherwise specified, the stimuli cover 90 degrees of visual field at 0.05 cycles/degree and are shown for 750ms, with ~1000ms in-between. This ISI is fixed and determined by the recording rate which changes slightly from dataset to dataset.

1) gratings_static: square static gratings.

2) minnie: complex stimulus (cartoon mouse head).

3) gratings_short: 200ms presentation time

4) local: localized to neural receptive fields (30 deg)

5) static_biased: rotation drawn randomly from interval [43 47] degrees

The next stimulus sets are all drifting at 2Hz:

6) gratings_drifting: drifting gratings

7) gratings_low_contrast: 1/20 the full contrast

8) gratings_noisy: low contrast and per-frame independent noise

Added in revision

9) sinusoidal gratings with randomized phases

10) biased_V2: like #5 but in higher-order areas lateral to V1.

Note that the stimuli are triggered on the microscope frames, so a neuron is always sampled at the same offset from stimulus start times.

If you use these data in a paper, please cite the original research paper as well as this dataset using the figshare doi.