DetrendTimeCourses.m
2cb18aae5832master

DetrendTimeCourses.m
View Options

			%% This function detrends the time courses to consider for total activation
			%
			% Inputs:
			% - TC is a n_ret_voxels x n_time_points 2D matrix with the data to detrend
			% - param is the structure containing relevant TA parameters; here, we
			% require the 'TR' (TR of the data), 'DCT_TS' (cutoff frequency for the DCT
			% basis), 'Covariates' (list of covariates to include in the
			% detrending/regression process; set to [] if none wished), and 'NbrVoxels'
			% (number of retained voxels for TA)

			% Younes, Jun 2019.
			% Update: Allow for Normalization only.
			% Younes, Dec 2019.
			% Update: Save mean and Std maps from normalization.
			function [TCN,STD_MAP] = DetrendTimeCourses(TC,param,fid)

			% TCN has n_vox x n_TP dimensions
			TCN = zeros(size(TC));
			STD_MAP = zeros(param.NbrVoxels,1);

			if param.doDetrend

			for i=1:param.NbrVoxels;

			% Regresses out low frequency components + possible other
			% covariates
			TCN(i,:) = sol_dct(TC(i,:)',param.TR,param.DCT_TS,param.Covariates);

			% Normalization
			stdval = std(TCN(i,:));
			STD_MAP(i) = stdval;
			TCN(i,:) = TCN(i,:)./stdval;
			end

			WriteInformation(fid,['Detrending and normalizing the data with DCT = ',...
			num2str(param.DCT_TS),' [s] and ',num2str(size(param.Covariates,2)),...
			' covariate(s)']);
			else

			for i=1:param.NbrVoxels;
			% Only Normalization
			stdval = std(TCN(i,:));
			STD_MAP(i) = stdval;
			TCN(i,:) = TCN(i,:)./stdval;
			end

			WriteInformation(fid,['Normalizing the data']);

			end

			end