Cross session cell alignment

Please use the online documentation website going forward: https://bahanonu.github.io/ciatah/

Cross-day or -session cell alignment alignment

Find the main function at https://github.com/bahanonu/calciumImagingAnalysis/blob/master/classification/matchObjBtwnTrials.m

Algorithm overview

For details, see Cross-day analysis of BLA neuronal activity methods section in the associated Science paper: http://science.sciencemag.org/content/sci/suppl/2019/01/16/363.6424.276.DC1/aap8586_Corder_SM.pdf#page=10.

• Example output on several mPFC animals across multiple sessions. Color is used to indicate a global ID cell (e.g. the same cell matched across multiple days).

Usage

The below commands in MATLAB can be used to align sessions across days.

% Create input images, cell array of [x y nCells] matrices
inputImages = {day1Images,day2Images,day3Images};

% options to change
opts.maxDistance = 5; % distance in pixels between centroids for them to be grouped
opts.trialToAlign = 1; % which session to start alignment on
opts.nCorrections = 1; %number of rounds to register session cell maps.
opts.RegisTypeFinal = 2 % 3 = rotation/translation and iso scaling; 2 = rotation/translation, no iso scaling

% Run alignment code
[alignmentStruct] = matchObjBtwnTrials(inputImages,'options',opts);

% Global IDs is a matrix of [globalID sessionID]
% Each (globalID, sessionID) pair gives the within session ID for that particular global ID
globalIDs = alignmentStruct.globalIDs;

% View the cross-session matched cells, saved to `private\_tmpFiles` sub-folder.
[success] = createMatchObjBtwnTrialsMaps(inputImages,alignmentStruct);

In certain cases, you want to run analysis on the registered images, see below.

% Get registered images, cell array of [x y nCells] matrices
registeredImagesCell = alignmentStruct.inputImages;
% Get registered cell maps, cell array of [x y] matrices
registeredCellmaps = alignmentStruct.objectMapTurboreg;

% OR another method below.

% Get the registration coordinates
globalRegCoords = alignmentStruct.registrationCoords;
globalRegCoords = globalRegCoords{folderNo};

% Re-register input images for particular imaging session for later analysis.
for iterationNo = 1:length(globalRegCoords)
	fn=fieldnames(globalRegCoords{iterationNo});
	for i=1:length(fn)
		localCoords = globalRegCoords{iterationNo}.(fn{i});
		[inputImages, localCoords] = turboregMovie(inputImages,'precomputedRegistrationCooords',localCoords);
	end
end

Algorithm results

Cross-session metrics and results on cross-session amygdala response to pain

PFC cross-session alignment.

Dorsal striatum cross-session algorithm comparison

• Original dorsal striatum cell maps from ICA with no motion correction applied.

• calciumImagingAnalysis (Biafra's) registration algorithm
  • Color is used to indicate a global ID cell (e.g. the same cell matched across multiple days). Thus, same color cell = same cell across sessions under the quick iteration parameters used in the below run.

• CellReg registration algorithm Using older code at https://github.com/zivlab/CellReg.