EEG sequencies visualization ideas

I'm looking for ideas on how to visualize my EEG signal data for sequence classification. My dataset consists of 62 channels, and the data is organized in a cell format where each element is a 2D matrix with dimensions 62 by timestep. What are some effective ways to represent this data in a 2D graph to showcase the samples?

* The feature extraction technique is embedded in the netLayer:

netSPN = [sequenceInputLayer(numChannels,"MinLength",800,"Name","input");
          cwtLayer("SignalLength",200,"IncludeLowpass",true,"Wavelet","amor")];

Answer 1

[Star Strider]

Based oin your description, my choice would be tiledlayout. It would be relatively straightforward to put each sequence in a vertical ‘stack’ using that function. Once you decide on the number of plots you want, the rest could be done in a for loop.

[Walter Roberson]

Not stackedplot ?

[Star Strider]

My first thought was to use stackedplot, however it wants everything to have the same independent variable (in this instance, time), and these appear to be sequential records, so every plot has to have its own sequential time vector.

[Ahmed]

Can I visulaize the CWT results on a certain graph to see how smaple points are related together?

[Star Strider]

It depends on what you are plotting. Just now, I do not have any idea what that actually is.

If you are plotting the different wavelet decompositions over the same time range, then stackedplot (or a consecutive series of stackedplot figures) might be best. If you are plotting single traces consecutively, then tiledlayout might be best. There are also Wavelet Toolbox plots that could do what you want.

Answer 2

[Diego Caro]

Hi, if you're familiar with eeglab, I'd suggest you use my function: dcaro_stacked.m (I just published it, lol)

It's just a fancier way to visualize eeg channels/components than using the regular eegplot or pop_eegplot.

Hope it helps!