how to color part of a rasterplot using a color triplets array

Hi,

i'm trying to color with two different colors a rasterplot using a binary array that represents the presence/absence of a stimulus.

the rasterplot is made this way:

% TIME_SERIES is a cell array with the times of each event in each cell,
% representing different electrodes
% STIM_TIMES is the stimulation binary signal which is shared between series
% both are passed to the algorithm via the arguments of a function
for i = 1:numel(time_series)
  xdata = repmat(time_series{i}', 3, 1);
  ydata = nan(size(xdata));
  ydata(1, :) = i-1;
  ydata(2, :) = i;
  colors = [stim_times zeros(size(stim_times)) zeros(size(stim_times))];
  plot(hndl, xdata, ydata, "color", colors)
end

Unrecognized function or variable 'time_series'.

the issue is (in my opinion) the size of the stim_times array but the error reported is

Error using plot

Invalid RGB triplet. Specify a three-element vector of values between 0 and 1.

[DGM]

If colors is intentionally supposed to be a (binarized?) color table, then no, plot() can't accept that. You might try using scatter() instead.

[Robert Daly]

By "binary" do you mean "boolean" i.e. true or false?

If you want to make a raster style plot I suggest you should try

imagesc

or

pcolor

something like

figure
pcolor(xdata, ydata, colors)
shading flat

Answer 1

[ag]

Hi,

I understand that you are getting an error while using the "plot" function in your script.

The error is in the following line of code:

colors = [stim_times zeros(size(stim_times)) zeros(size(stim_times))]; 

The error is caused due to the values being passed to the "color" input parameter of the "plot" function. To resolve this issue, make sure that the values assigned to the "colors" variable are triplets with each value ranging from 0 to 1.

For example, you can plot a figure with blue lines by using the following code:

plot(hndl, xdata, ydata, "color", [0 0 1]); 

For more detailed information, please refer to the following documentation page:

https://in.mathworks.com/help/releases/R2022b/matlab/ref/plot.html#btzitot-Color

Hope this helps!

Answer 2

[DGM]

There were multiple ways to solve this:

% this emulates the inputs as described
N = 10;
time_series = linspace(0,10,N);
time_series = num2cell(time_series);
stim_times = rand(1,N)>0.5;

% either do it in a loop, generating one tuple at a time
for i = 1:numel(time_series)
    xdata = repmat(time_series{i}', 3, 1);
    ydata = nan(size(xdata));
    ydata(1, :) = i-1;
    ydata(2, :) = i;
    colors = [stim_times(i) 0 0]; % a color tuple, not a table
    plot(xdata, ydata, 'color', colors,'linewidth',2); hold on
end

... or instead of generating each tuple one at a time, you can generate the color table as was attempted originally:

figure
% or generate a color table and then use it in a loop
CT = [stim_times(:) zeros(numel(stim_times),2)]; % a color table
for i = 1:numel(time_series)
    xdata = repmat(time_series{i}', 3, 1);
    ydata = nan(size(xdata));
    ydata(1, :) = i-1;
    ydata(2, :) = i;
    plot(xdata, ydata, 'color', CT(i,:),'linewidth',2); hold on
end

... but the loop isn't necessary. the whole thing can be simplified:

figure
% assemble x,y data
xdata = repmat(cell2mat(time_series),[3 1]);
ydata = (1:numel(time_series)) - [1; 0; NaN];

% set colororder, plot
CT = [stim_times(:) zeros(numel(stim_times),2)]; % a color table
colororder(CT)
plot(xdata, ydata,'linewidth',2)

The question remains whether this barely-readable staircase of tiny line segments was actually the appropriate output. This could probably be done simpler yet with a scatter plot or something else.

figure
% assemble x,y data
xdata = cell2mat(time_series);
ydata = 1:numel(time_series);

% plot the data using a short color table
scatter(xdata,ydata,100,stim_times,'filled');
CT = [0 0 0; 1 0 0]; % one color for each class
colormap(CT)

... but OP never clarified, so we can guess all we want now.