An accumarray call can do that relatively easily:
D = load('test.mat');
task = D.task;
time_s = D.time_s;
TFmin = islocalmin(task, 'MinProminence',10);
TFmax = islocalmax(task, 'MinProminence',10);
MinIdx = find(TFmin);
MaxIdx = find(TFmax);
Idx = sort([MinIdx; MaxIdx]);
IdxCell = accumarray(Idx(1:numel(Idx)-1), (1:numel(Idx)-1).', [], @(x){Idx(x):Idx(x+1)}); % Cell Array Of Index Ranges
IdxRng = find(cellfun(@(x)~isempty(x), IdxCell)); % Index Ranges
GetData = @(x) [time_s(IdxCell{IdxRng(x)}), task(IdxCell{IdxRng(x)})]; % Function To Retrieve Data Easily
Get1 = GetData(1); % Get First Set
X1 = Get1(:,1);
Y1 = Get1(:,2);
Get3 = GetData(3); % Get Third Set
X3 = Get3(:,1);
Y3 = Get3(:,2);
figure
plot(time_s,task, time_s (TFmin),task(TFmin),'r*')
hold on
plot(time_s,task,time_s(TFmax),task(TFmax),'b*')
plot(X1, Y1, '.r') % Plot First Set
plot(X3, Y3, '.g') % Plot Third Set
hold off
Producing this example plot:
Thius code recovers all of the intervals, not only the ascending ones, so choose the odd-numbered arguments to ‘GetData’.
