Skip to content. | Skip to navigation

Personal tools
You are here: Home Forum Using data sets HC-5 Dataset - electrode location and missing channels

HC-5 Dataset - electrode location and missing channels

Up to Using data sets

HC-5 Dataset - electrode location and missing channels

Posted by Ralf Mueller at April 03. 2018


we want to use the data set hc-5 (Eva Pastalkova, Gyorgy Buzsáki). In the description of the download it says that all of the data are incorporated in the matlab file (.mat).

In this mat-file I could find data from a single EEG channel.

My questions are:


This EEG-channel does belong to which neural area? Possible positions are "left and right hippocampal areas CA1 and right entorhinal cortex" - but I couldn't find a hint where the data is taken from.


Where are the EEG data sets from the remaining two areas?


Best regards,


Re: HC-5 Dataset - electrode location and missing channels

Posted by Jeff Teeters at April 09. 2018

I think that the matlab file only contains one example eeg channel and the eeg recorded from all the channels is in file with extension ".eeg" (which is a binary file, 2 bytes per sample).  The xml file indicates which channels are in the .eeg file.  The data description indicates which location the channels are recorded from (e.g. Right EC: channels 1-32, ...).


For example, with data set "i01_maze08.001", the matlab file:

load i01_maze08_MS.001_eeg_1250Hz.mat

contains variable sh7 which is the same as variable Track.eeg inside

load i01_maze08_MS.001_BehavElectrData.mat

>> size(Track.eeg)

ans =

     2689125           1

>> isequal(Track.eeg, sh7)

ans =




The number of minutes of data stored in these variables is given by dividing the size by the sampling rate, then 60 (convert seconds to minutes):

>> 2689125 / 1250 / 60

ans =


Which matches the time given in the data description document for data file i01_maze08.001.tar.gz.

Now consider the binary ".eeg" file.  It has size:

ls -l i01_maze08_MS.001.eeg

-rwxr-xr-x@ 1 jt  staff  548581500 Jun  7  2008 i01_maze08_MS.001.eeg


Taking that size (in bytes), dividing by 2 (2 bytes per sample), then dividing by the sampling rate and the time (in seconds) gives the number of channels of data in the file:

>> 548581500 / 2 / 1250 / (35.8550*60)

ans =


So there are 102 channels of data in that file.  The xml file should indicate which channels (of the 128 channels described in the data description document) they are.  btw, the scripts that come with the hc-2 dataset might be useful for loading the data from the binary file into Matlab.


Powered by Ploneboard
Document Actions