Breaking EEG signal into epochs: why should we multipy windows-time length by frequency?

I am new to EEG data, Could you please explain why we need to multiply Fs by window_time_length ? I thought by defining window_time_length, we define the length of epoch! I am working on P300 BCI speller and try to understand how its classifier works. I am using a code from Internet. a part of code for per-processing data is as below:

params = io.loadmat(other_path)
mean = params['mean']
std = params['std']
if standard_before:
signal = (signal-mean)/std

epoch_length = int(Fs * window_time_length / 1000.)

• we have no idea what you're referring to – we're not looking at the same code or algorithm as you. You'll have to please describe what you're considering, please, and in much more breadth – there's not exactly one thing that you can do to EEG data. – Marcus Müller Jun 3 at 20:52
• Because you are converting your window duration from a time unit (s) to a number of samples. And I guess for following processing you need this duration in samples. If the frequency Fs is in Hz, I guess window_time_length is expressed in ms. – Mathieu Jun 3 at 21:49
• Dear Mathieu, Thank you for your answer. could you please tell me what do you mean by number of samples? for example we have a epoch array like this (1500, 64, 144). I understood it that 1500 is number of our epochs, 64 is number of channels AND 144 is length of each epoch! Am I wrong? – user57724 Jun 18 at 7:18