How do I cut data into segments?

I have eeg data in a .bdf file and I need to segment the data using the information in a .txt file.

The text file contains start and stop times in milliseconds .

I must cut fromt1 to t2 and then t2 to t3 and so on..

[Ameer Hamza]

How is the data from bdf file available in MATLAB? Do you have the data a matrix? How do you want to partition the data?

[Nikki Larsen]

yes it is a matrix and there are 80 rows and 512 columns. I am not sure how to cut the data according to the times in the text file. I hope I made it clear.

[Ameer Hamza]

Does rows represent time samples or columns. Have you loaded the data from txt file? What is the sample rate?

[Nikki Larsen]

Sampling frequency is 512

[Image Analyst]

Nikki - I think people are still waiting for you to upload the file. How else can we effectively work with it if we don't have it?

[dpb]

Well, the only .bdf file extension I was aware of was the Adobe one that didn't seem to make any sense...looks as though this is a proprietary EEG file format using 24-bit/3-bytes per sample. That's going to take some processing to read...just a simple fread won't cut it.

<biosemi.com/faq/file_format.htm>

There's a very simple and somewhat questionable for general use submittal at the FileExchange

<fileexchange/13070-eeg-bdf-reader>

How did you import the data into MATLAB?

As far as the time, there isn't a timestamp in the file; either the duration and number of samples is given or the duration is encoded as "1" in which case the number of samples is to be interpreted as the sample rate.

Either way, you'll need to determine that and then compute how many records are needed to match up. Easy enough if the two have common divisor; maybe not quite so much if time and sample rates don't match precisely.

[Nikki Larsen]

yes i basically modified the eeg-bdf-reader to read the bdf file.

Answer 1

[Ameer Hamza]

try something like this

data = rand(80, 512); % random data, replace with data from bdf file
t_intervals = [50 100 300 500 700 900]; % values from your txt file in milliseconds

% solution
fs = 512; % sampling frequency
T = 1/fs*1000; % sampling time in milliseconds
t = 0:T:T*fs-1; % sample time instants

t_intervals = [0 t_intervals t(end)];
data_partition = splitapply(@(x) {x}, data, discretize(t, t_intervals));

[Nikki Larsen]

Thank you :)

[Ameer Hamza]

Glad to be of help.