changes on cond comparison and MeasListAct (735df8ec) · Commits · github_fork / Homer3

FuncRegistry/UserFunctions/hmrS_CalcPvalue.m

+24 −24

Original line number	Diff line number	Diff line
		@@ -24,15 +24,14 @@
		% baselineRange: [-2.0, 0.0]
		% hrfTimeWindow: [-2.0, 20.0]
		%
		function pValues = hmrS_CalcPvalue(yRuns, stimRuns, mlActAuto, baselineRange, hrfTimeWindow)
		function pValues = hmrS_CalcPvalue(yRuns, stimRuns,baselineRange, hrfTimeWindow)

		pValues = cell(length(yRuns{1}),1);


		% extract fq and number of conditions from the first run:
		snirf = SnirfClass(yRuns, stimRuns);
		% % % % % snirf = SnirfClass(yRuns{iRun}, stimRuns{iRun});
		ml = snirf.GetMeasListSrcDetPairs();
		snirf = SnirfClass(yRuns{1}, stimRuns{1});
		ml = yRuns{1}.GetMeasListSrcDetPairs();
		t = snirf.GetTimeCombined();
		s = snirf.GetStims(t);
		ncond = size(s,2);
		@@ -54,7 +53,7 @@ for cond = 1:ncond % for each condition
		for iBlk = 1:length(yRuns)
		% get active measuremnt list for each run
		% 1) IS THIS HOW WE EXTRACT MEASLISTACT
		if isempty(mlActAuto{iBlk})
		if ~exist('mlActAuto')
		mlActAuto{iBlk} = ones(size(ml,1),1);
		end
		mlAct = mlActAuto{iBlk};
		@@ -112,7 +111,7 @@ for i = 1:ncond
		lst_stim_all{i} = find(s(:,i)==1);
		end

		for comb_inx = 1:size(lst_stim_all{i},1) % for each condition
		for comb_inx = 1:size(cond_2_comb,1) % for each condition
		% get current combin.
		foo = cond_2_comb(comb_inx,:);

		@@ -120,7 +119,7 @@ for comb_inx = 1:size(lst_stim_all{i},1) % for each condition
		for iBlk = 1:length(yRuns)
		% get active measuremnt list for each run
		% 1) IS THIS HOW WE EXTRACT MEASLISTACT
		if isempty(mlActAuto{iBlk})
		if ~exist('mlActAuto')
		mlActAuto{iBlk} = ones(size(ml,1),1);
		end
		mlAct = mlActAuto{iBlk};
		@@ -129,19 +128,19 @@ for comb_inx = 1:size(lst_stim_all{i},1) % for each condition
		for hb = 1:3 % across HbO/HbR/HbT
		for iTrial = 1:size(lst_stim_all{foo(1)},1) % across trials
		% Hb: # of time points X # of trials X # of runs X # of channels X HbO/HbR
		Hb_peak1(:,iTrial,iRun,:,hb,cond) = squeeze(y([lst_stim_all{foo(1)}(iTrial) + round(abs(hrfTimeWindow(1))fq)]:[lst_stim_all{foo(1)}(iTrial) + round(hrfTimeWindow(2)fq)],hb,:)); % get each trial
		Hb_peak1(:,iTrial,iRun,:,hb) = squeeze(y([lst_stim_all{foo(1)}(iTrial) + round(abs(hrfTimeWindow(1))fq)]:[lst_stim_all{foo(1)}(iTrial) + round(hrfTimeWindow(2)fq)],hb,:)); % get each trial
		end
		for iTrial = 1:size(lst_stim_all{foo(2)},1) % across trials
		% Hb: # of time points X # of trials X # of runs X # of channels X HbO/HbR
		Hb_peak2(:,iTrial,iRun,:,hb,cond) = squeeze(y([lst_stim_all{foo(2)}(iTrial) + round(abs(hrfTimeWindow(1))fq)]:[lst_stim_all{foo(2)}(iTrial) + round(hrfTimeWindow(2)fq)],hb,:)); % get each trial
		Hb_peak2(:,iTrial,iRun,:,hb) = squeeze(y([lst_stim_all{foo(2)}(iTrial) + round(abs(hrfTimeWindow(1))fq)]:[lst_stim_all{foo(2)}(iTrial) + round(hrfTimeWindow(2)fq)],hb,:)); % get each trial
		end
		end
		end
		end

		% put together trials from all runs:
		Hb_peak_rs1 = reshape(Hb_peak1, size(Hb_peak1,1), size(Hb_peak1,2)* size(Hb_peak1,3), size(Hb_peak1,4),size(Hb_peak1,5),size(Hb_peak1,6));
		Hb_peak_rs2 = reshape(Hb_peak2, size(Hb_peak2,1), size(Hb_peak2,2)* size(Hb_peak2,3), size(Hb_peak2,4),size(Hb_peak2,5),size(Hb_peak2,6));
		Hb_peak_rs1 = reshape(Hb_peak1, size(Hb_peak1,1), size(Hb_peak1,2)* size(Hb_peak1,3), size(Hb_peak1,4),size(Hb_peak1,5));
		Hb_peak_rs2 = reshape(Hb_peak2, size(Hb_peak2,1), size(Hb_peak2,2)* size(Hb_peak2,3), size(Hb_peak2,4),size(Hb_peak2,5));

		% take the mean in time range hrfTimeWindow
		for hb = 1:3 % HbO/HbR/HbT
		@@ -153,19 +152,20 @@ for comb_inx = 1:size(lst_stim_all{i},1) % for each condition
		end

		% get stats
		if isempty(pValues_cond{iBlk})
		if ~exist('pValuesS_cond')
		for ch = 1:size(MEAN_Hb_peak,2) % channels
		if mlAct(ch) ~=0
		format long
		for hb = 1:3% HbO/HbR/HbT
		[h,p,c,stats] = ttest2(MEAN_Hb_peak1(:,ch,hb),(MEAN_Hb_peak2(:,ch,hb)));
		pValuesS_cond(foo(1),foo(2),hb,ch,cond) = p;
		pValuesS_cond(hb,ch,comb_inx) = p;
		% or pValuesS_cond(foo(1),foo(2),hb,ch) = p;

		% 2) Pvalue SHOULD BE CONVERTED TO DATA CLASS IN THE RIGHT DIM AT THE END

		end
		else
		pValuesS_cond(foo(1),foo(2),hb,ch,cond) = 'NaN';
		pValuesS_cond(hb,ch,comb_inx) = 'NaN';
		end
		end
		end

Admin message