[Eeglablist] Reg: Conversion of EEG

[Robert Oostenveld]

Dear Krishnaveni

> If I plot this using plot command in Matlab or eegplot command using 
> EEGLAB toolbox, the waveform I am getting doesn't look like eeg 
> signal.

I also had a look at the two samples you provided and they indeed do 
not look like EEG.

>   Is the sampling rate correct ?

I cannot tell. But regardless of the sampling rate these two signals 
never will look like EEG.

Channel 1 only shows some "bit steps", i.e. it seems to be constant, 
but only contains some noise that is in the order of magnitude of the 
least significant bit of your AD converter. Channel 2 contains a 
different type of "noise" which is larger, but looking at the power 
spectrum (try the matlab function psd) I would say that it does not 
show any features that resemble an electrophysiological signal. I would 
at least expect the power spectrum to show that it has been filtered 
near the Nyquist frequency, but it is not.

> To what accuracy (how many decimal places) the data need to be 
> converted. As shown above the digital data is in integer format. Is it 
> correct ? Is this the problem.

Normally the data that you would work with in Matlab would not be in 
integer format (since you are measuring microvolts), but an AD 
converter does converts the potential into an integer after amplifying 
it. I.e., a "16 bit A/D converter" results in integers between -32768 
and 32768. To convert the integers back into microvolts, you would have 
to know the amplification. The amplification of each channel can even 
be slightly different (depends on the calibration). Therefore it is not 
completely surprising that the signal is in integers, i.e. it indicates 
that it has not been scaled yet to microvolts.

But the scaling-factor needed to convert the signal from integers into 
a signal expressed in uV will only change its magnitude and not the 
overall shape. Therefore scaling your two signals will not make them 
look more like EEG.
 
> If we have the data in analog format, how to convert to digital 
> format, so that it can be used by the EEGLAB toolbox. (Because this 
> toolbox is very useful for my work)

I assume that you have an EEG amplifier that is connected to the 
computer. In that case the EEG is recorded on the computer, and hence 
it is in a digital format already. It is hardly impossible to store EEG 
data in an analog format: in principle it could be done using magnetic 
tapes, but I never heard of anyone doing that. Therefore I would assume 
that you don't have the data in an analog format, but already in a 
digital file.

Typically, each EEG manufacturer has his own fileformat, and the 
problem that you have to solve is to find (or create) a Matlab import 
routine for your particular file format. Alternatively, you can try 
around in your EEG software to see whether it can also export the data 
into another file format. EDF is a common file format, and it is 
supported by EEGLAB.

Best regards,
Robert


----------------------------------------------------------------------
Robert Oostenveld, PhD
F.C. Donders Centre for Cognitive Neuroimaging
Radboud University Nijmegen
phone: +31-24-3619695
http://www.ru.nl/fcdonders/
----------------------------------------------------------------------
N.B. Starting from 1 September 2004, the University of Nijmegen has 
changed its name to Radboud University Nijmegen. All web- and 
email-addresses ending in ".kun.nl" should therefore be changed into 
".ru.nl". Please update your address book and links.