1/1
Zatka-Haas et al 2020 dataset
dataset
posted on 2020-09-25, 16:55 authored by Peter Zatka-HaasPeter Zatka-Haas, Nick SteinmetzNick Steinmetz, Matteo CarandiniMatteo Carandini, Kenneth HarrisKenneth HarrisDataset from Zatka-Haas et al, 2020, bioRxiv
https://www.biorxiv.org/content/10.1101/501627v3
Widefield
The Widefield folder contains the following:
-Two CSV files detailing information on the mice and sessions
-ctxOutlines.mat a set of coordinates which define the Allen Common Coordinate Framework atlas, useful for overlaying onto the widefield data.
-behavDataMatchedRTs: behavioural data for each session. For each individual mouse the trials are subsampled so that the reaction times for Left and Right choices are similar.
-SVD: Widefield data for each session, which has undergone singular value decomposition denoising. The variables contained within the file are as follows:
U_dff – The U component of SVD, which has been normalised to the mean image activity.
dV – The V component of SVD, which has undergone derivative filtering.
meanImg_registered – the mean image of the mouse cortex across the entire session, registered to match the mean image for the first session for each mouse (contained in the folder reference_image)
wfTime_dt – The timestamps associated with dV
You can re-generate the activity from the U and V components using this toolbox’s function https://github.com/cortex-lab/widefield/blob/master/generalUtils/svdFrameReconstruct.m
-ROIs: Set of ROI coordinates used for each mouse.
-decoding: contains the result of the main decoding analyses. Variables are:
coordSet – the set of stereotaxic coordinates used for extracting widefield activity for later processing
ROImove – the movement-aligned activity extracted at each coordinate from coordSet, across each trial and timestamp defined in moveT.
ROIstim – the stimulus-aligned activity extracted at each coordinate from coordSet, across each trial and timestamp defined in stimT.
decoding_probability_moveAligned – the decoder performance for each timestamp (movement-aligned), coordinate and decoder type
decoding_probability_stimAligned - the decoder performance for each timestamp (stimulus-aligned), coordinate and decoder type
labels – labels for the separate decoders employed in the dataset. ‘Detect prob’ refers to action decoding. ‘stimL prob’ refers to stimulus-left decoding. ‘choice prob’ refers to choice decoding.
moveT - timestamps for all movement-aligned activity and decoding
stimT – timestamps for all stimulus-aligned activity and decoding
The ROI folder contains 1 file for each mouse (named ‘Chain’,’Cori’, etc). Within each file is:
-mask_xy = the set of image coordinates which define the boundaries of the widefield imaging implant. This is used to crop out the edges of the widefield image which are not part of the exposed brain.
-px = the set of ROIs used for that mouse. The fields ‘row’ and ‘col’ refer to the row and column of that coordinate in the whole brain image (row and column of the U_dff variable (from the SVD folder), and the meanImg_registered variable, and also any images you generate after running svdFrameReconstruct).
Stan model fits
The StanModel folder contains files relating to the neurometric model fit:
-Two-Level-Mechanistic-SYMMETRICAL.stan = the Stan code used to define the main neurometric model
-neurometric_wf_original.mat = saved fit of the neurometric model. This is fit to widefield ROI data at VISp and MOs. The copy of the data used in the fitting is in the variable ‘data’. The posterior distribution of the parameters (defined in the .stan file) are in the variable ‘posterior’.
-The other two files are the .stan code and .mat fit for the neurometric model fit to the non-laser trials of the inactivation sessions. This model is the one used to predict the effect of laser inactivation.
The ROI folder contains 1 file for each mouse (named ‘Chain’,’Cori’, etc). Within each file is:
-mask_xy = the set of image coordinates which define the boundaries of the widefield imaging implant. This is used to crop out the edges of the widefield image which are not part of the exposed brain.
-px = the set of ROIs used for that mouse. The fields ‘row’ and ‘col’ refer to the row and column of that coordinate in the whole brain image (row and column of the U_dff variable (from the SVD folder), and the meanImg_registered variable, and also any images you generate after running svdFrameReconstruct).
Optogenetic inactivation:
This folder contains behavioural data as tables, where each row is a single trial. Included in the table is all information about behavioural trials.
Code:
Main code for generating Figs 1-4, as well as associated utility files.
