If you want to divide the signal into non-overlapping segments and then analyse each segment, use the buffer function. That would go something like this for your vector that I will call ‘signal’:
Fs = ...
ts = 1;
n = Fs*ts;
signal_buffer = buffer(signal, n);
Each column of ‘signal_buffer’ now has a 1-second segment of ‘signal’. If ‘signal’ has more than one channel, do this for each channel. It just makes the code easier to work with and interpret.
EDIT — (21 Nov 2023 at 06:43)
Determining the frequency of the clicks in each second would require counting them. They are likely not simple square-wave pulses, and instead are probably decaying exponential sine curves. Modeling one of them and then using the findsignal function (or one of its friends) would likely make this easier.
EDIT — (4 Dec 2023 at 18:16)
If all the pulses are like the plot image (and relatively clean with no noise), use either findpeaks or islocalmax with a prominence level to detect each peak. Convert the returned indices (my preference) either directly from findpeaks or using find with islocalmax to times by using them to index into the time vector. It might be necessary to plot a time-frequency spectrum. For that, I suggest the pspectrum funciton with the 'spectrogram' option.