ImagingPreProcess

2P imaging preprocess functions I used in RinbergLab

This preprocessing packages is written to analyze odor and light stimulation responses of mitral/tufted cells and glomeruli. I have inherited some parts from Hirofumi Nakayama's imaging analysis packages (check https://github.com/olfa-lab/2P_Pipeline and https://github.com/hirofuminakayama/Imaging).

We used followings for data collection

Imaging: ScanImage 2019. Data are in int16 and saved as tif files.

Behavior and Stimulus Control: Voyeur ( https://github.com/olfa-lab/Voyeur ) and olfactometers (https://github.com/olfa-lab/olfactometry ) developed in rinberglab. Data (sniff signal,licks, frametrigger, metadata for stimulus and timing information are in) saved as H5 file.

optional: As our behavior box during its communication with PC can drop frames and has low sampling rate, we occasionally use Wavesurfer to record critical parameters. Critical parameters are sniff signal and frame trigger generated by ScanImage.

Most of analysis relies on the following packages:

NORMCorre: This package is used for motion correction. https://github.com/flatironinstitute/NoRMCorre

Wavesurfer: This package is required for blanked odor recording. Each trial has its own TIF file. For continuous recording, this package is not necessary. https://github.com/JaneliaSciComp/Wavesurfer

Breathmetrics: This package is optional and used for sniff processing. It is useful if you want to realign inhalation onset. Voyeur's inhalation onset can be slightly inaccurate. https://github.com/zelanolab/breathmetrics

Steps for preprocessing:

1. Motion correction to register the field to reference field
2. Drawing ROIs
3. Calculating dFF and sniff responses for each trials

At the end you will get 'fieldname_S_v73' .MAT file

1. Motion Correction

Do necessary changes in bp_motioncorrection_mk.s file. Example batch file to create slurm jobs for motion correction. Critical parameters are current directory and array range (this number should be higher than maximum tiffs need to be corrected)

#SBATCH --array=0-250
module purge
module load matlab/R2018a
cd /gpfs/scratch/yourID/2Pdata/SC18/220908/
tifs=(/gpfs/scratch/yourID/2Pdata/SC18/220908/SC18_220908_field1stim*.tif)

tif=${tifs[$SLURM_ARRAY_TASK_ID]}
{
    echo $tif
    matlab -nodisplay -r "normcorremotioncorrection_single('$tif','Ref.tif'); exit"
} > $tif.log 2>&

It relies on normcorremotioncorrection_single.m file to be in the path. Do necessaary changes in following lines so that it can reach to NoRMCorre.

addpath(genpath(fullfile('/gpfs/scratch/yourID','NoRMCorre-master')));

Ref.tif is your best tif file that imaging session can be aligned. You can acquire 200 frames before session starting odor/stim trials to create your own single Ref.tif file.

2 . Draw ROI

I use FIJI to draw ROIs

https://imagej.net/software/fiji/

1. Be cautious when drawing and avoid including dendrites from other cells.

2. After drawing all the regions of interest (ROIs), use the "combine" and then "split" functions (under properties in ROI manager window). All ROIs will be labeled from top-to-bottom order.

3. Save the ROIs with the following naming convention: mouseID_YYMMDD_session_ROISet.zip

3. Odor preprocessing

Odor Imaging : Blanked Recording

Acquisition is discrete (this is our regular way of data acqusition )

Must have: ScanImage Tiff Files and Voyeur h5 File and Wavesurfer H5

Optional : Breathmetric ('inh_realign' get_prepocessed_odordata/get_prepocessed_stimdata parameter rely on this package)

folderpath = '/gpfs/scratch/karadm01/2Pdata/OMP02/230207';
roifile ='OMP02_230207';

fieldname = 'OMP02_230207_HighConc';
VoyeurH5_file = 'OMP02_230207_HighConc_1_01';
WSfieldname = 'OMP02_230207_HighConcWS';
get_prepocessed_odordata('tiffpath',folderpath, 'fieldname', fieldname, 'roiname', roifile,'VoyeurH5_name', VoyeurH5_file, 'WSfieldname',WSfieldname,'isOdor',true)

Only Stim

addpath(genpath('/gpfs/scratch/yourID/ImagingPreProcess/'))
%%
tiffpath = '/gpfs/data/rinberglab/Jon/JG44524/230614';
roifile ='JG44524_230614_glomfield4_stim';

fieldname = 'JG44524_230614_glomfield4_stim_00001';
VoyeurH5_file = 'JG44524_230614_glomfield4_stim_1_01';
get_prepocessed_stimdata('tiffpath',folderpath, 'fieldname', fieldname, 'roiname', roifile,'VoyeurH5_name', VoyeurH5_file,'isOdor',false,'calculate_diff_image',true)

Odor Imaging : Continuous acquisition

Must have: ScanImage Tiff Files and Voyeur h5 File Optional: Breathmetric

All above code will generate .MAT files

%% Data format, I used it to share data with Jon and Saeed. My Python scripts are also written based on this format
Session.OdorResponse = {};
Session.F= Fluo_cell_Kalman';
Session.blockTrials = {};
Session.fieldname = fieldname;

for i = 1: size(OdorInfo.odors,1)
    Session.OdorResponse{i} = permute(dffKalman(:,:,OdorInfo.odorTrials{i}),[2,1,3]);
    Session.blockTrials{i} = ones(length(OdorInfo.odorTrials{i}),1);
end
% df images [-0.5s 1s] wrt inhalation onset
if calculate_diff_image
   Session.diff_image = img_df_percond;
end
% Odor/stim conditions
Session.UniqueConds = cellstr(OdorInfo.odors);
% Odor/stim trial numbers
Session.OdorTrials = OdorInfo.odorTrials;
%sniff per trial
Session.Sniffs = Sniff_trial';
Session.CellMask = cellMask_vec;

Session.Infos.fps = fps;
Session.Infos.ImgFormat = img_format;
Session.Infos.imgwithROIs = img;
Session.Infos.pre_inh = pre_inh;
Session.Infos.post_inh = post_inh;
Session.Infos.TrialsRead = trials_read;
%  Voyeur data
Session.VoyeurData = Data;
save(strcat(fieldname,'_S_v73.mat'), 'Session','-v7.3')

MAT file has following parameters

Parameter name	Description
OdorResponse	dF/F responses, 1X N unique odor/stim condition cell array, every cell is 3D array, Ntime x Nroi x Ntrial
F	Fluorescence data, NtotalFrame x Nroi
overlap_pre	size of overlapping region in each direction before upsampling
fieldname	fieldname is retrieved from tiffs
UniqueConds	Unique trial condition presented
OdorTrials	Odor trials for each unique conditions
Sniffs	Ntime x Ntrials sniff data, [-pre post] wrt inhalation onset
CellMask	Cell masks for each ROI, NxNy x Nroi
VoyeurData	Voyeur data with prepocessed infos
Infos	it has extra information regarding pre, post inhalation duration, fps, image format and trial read infos