Discretized representations in V1 predict suboptimal orientation discrimination

Calcium Imaging datasets and code for the paper titled "Discretized representations in V1 predict suboptimal orientation discrimination".

This description has 3 sections:

Data tables description

Query example

Code description

1. Data tables description

The data is presented in CSV files that were exported from a SQL database. There are 3 datasets: Trained, Audiovisual and Naive. The archive contains 11 tables, 3 for each dataset used in the study (Trial, Cell and Cell_Trial tables) and 2 containing quartile information about the locomotion and pupil area for the second dataset. 

The content of the tables is as follows:

Trial_Table contains all information about the individual trials and the tracked behavioral variables. The columns are as follows:

Trial - the trial number

Experiment - the recording session identifier

Behav_Cond - the protocol identifier or recording day

Block - block identifier (task or tuning curve)

Visual_Stim - the orientation in degrees of the visual stimulus

Auditory_Stim - the type of auditory stimulus (1 = 5KHz, 2 = 10KHz, other = no stim)

Contrast - the contrast of the visual stimulus

Response - the mouse’s response (0 = no lick, 1 = lick)

Outcome - the outcome of the trial (Hit, Miss, FA, CR)

Lick_trace* - a blob containing the licking time series  

Lick_Trial - same as trace but only during stimulus presentation

Lick_dt - the sampling period of the licking recording

Locomotion_trace* - the mouse’s locomotion for this trial

Locomotion_mean_01 - the average locomotion during the first second of stimulus presentation

Pupil_Area_trace - pupil area during the trial

Pupil_Area_mean_01 - pupil area average during the first second of stimulus presentation

Pupil_Position_trace - pupil position during the trial

Pupil_Position_meanX - pupil position average on the horizontal axis during stimulus presentation

Pupil_Position_meanY - pupil position average on the vertical axis during stimulus presentation   

Inclusion*** - Whether the trial is included or not (due to some signal issues, some trials can be unusable)

Cell_Table contains all information about the segmented cells. The columns are as follows:

Cell - Cell identifier

Experiment - Recording session identifier

Zscore - Zscored response average during stimulus presentation

DeconvA_Tau - Fitted deconvolution parameters (amplitude and decay time)

nSpikeTC - Number of deconvoluted action potential related events (APrEs) during the tuning curve block

DeconvCorr - Correlation coefficient between dF/F trace and convoluted APrEs trace   

Tuning_Curve_spikes_x**** - cell tuning curve fit from APrE data, encoded in bytes (see *)

Best_Fit_spikes_x - The type of fit that was selected

Pref_Orientation_spikes_x - The preferred orientation of the cell

Tuning_Width_spikes_x - The tuning width (at half max)

DSI_spikes_x - Direction selectivity index

OSI_spikes_x -  Orientation selectivity index

All following columns are same than 7 to 12 but using either the mean dF/F or the AUC during stimulus presentation

isCell** -  the cell probability statistic from suite2p

nPix** - the number of pixels in the segment

spikeProb_TCtrials - the probability of having at least one APrE for the orientation closest to the preferred orientation in the tuning block 

Roundness** -  the roundness of the segment (see paper)

Radius** - radius in pixels of the segment

magFactor** -  the number of microns per pixel

weightPO** - the preferred orientation (orientation of the average weight vector) derived from the SNN

weightNorm** - the norm of the average weight vector

Cell_Trial_Table contains information about the cells’ activity during each trial. Columns are as follows:

Experiment - recording session identifier

Trial - Trial number

Cell - Cell identifier

Trace* - dF/F trace

TraceSpikes* - deconvolved APrE trace 

nSpikes - Number of APrEs during stim presentation

Mean_dFoF - Mean dF/F during stim presentation

AuC_dFoF - AuC of dF/F during stim presentation     

Locomotion and Pupil (only in the audiovisual dataset) contain the average locomotion or pupil value during the first second of stimulus presentation (Loc01mean/Pupil01mean) and the corresponding quartile (QuartMean) for every trial identified with Experiment and Trial.

*All traces contain time points from -3 to +6s around the stimulus, they are encoded in bytes and can be unpacked using the getArrayFromByteStream matlab function   

** this column is specific to the trained and naive

*** this column is specific to the audiovisual dataset 

**** x can be 1 or 2, if 2 has data it contains the tuning block recorded before the task. If no number is specified in the column title, then the tuning block was recorded after the task  

In the audiovisual dataset, the following sessions were not included in the calcium imaging analysis

 {'Ang001_0512','Ani5_111918','Ang003_0513','Ani2_092618','Ani4_111918','Ani1_092718','Ang002_0515','Ang003_0515','Ang004_0515','Ang001_0516','Ang002_0516' ,'Ang005_0516' }

—--------------------------------------------------------------------------------------------------------------------------------------------------------

2. Query example

Example of a query used to download data using the criteria used in the article :

SELECT ct.Experiment, ct.Cell, ctt.Trial,tt.Behav_Cond, tt.Visual_Stim, ctt.TraceSpikes, ctt.Trace,  ct.Pref_Orientation_spikes_2, ct.Tuning_Curve_spikes_2, ct.Best_Fit_spikes_2, ct.spikeProb_TCTrials

FROM Cell_Trial_Table as ctt

INNER JOIN Trial_Table as tt ON ctt.Experiment = tt.Experiment AND ctt.Trial = tt.Trial

INNER JOIN Cell_Table as ct ON ctt.Experiment = ct.Experiment AND ctt.Cell = ct.Cell

WHERE  tt.Behav_Cond = ''D1''

AND tt.Block = ''Orientation Tuning''

AND ct.Pref_Orientation_spikes_2 IS NOT NULL

AND ct.DeconvCorr > 0.8

AND ct.isCell > 0.8

AND ct.spikeProb_TCTrials > 0.1

AND ct.roundness > 0.2

AND ct.nPix*ct.magFactor > 15

—--------------------------------------------------------------------------------------------------------------------------------------------------------------------

3. Code description

We provide the scripts allowing to analyze the data and plot the figures of the paper. The "Main" scripts folder contains the scripts generating the figures, the "Dependencies" folder contains the necessary custon functions to run the main scripts, and the "DataFiles" folder contain matfiles containing data pulled from the database in order to generate a specific plot.

All plots are indexed in the “Index” pdf file, along with the file name of the script generating the given plot from the database, and if applicable, a data file corresponding to the corresponding data pulled from the database, allowing to bypass querying the necessary data.