[Eeglablist] Calculating Alpha Asymmetry

[Andero Uusberg]

Dear Kathryn,

all of your pipeline can be done in EEGLAB, but a few steps require 
using Matlab command line interface.

For the segmentation of data EEGLAB has a function called 
pop_importevent which does what the name suggests and thus allows you to 
import event-markers for the 2-second epochs you'll need. As I'm not 
quite sure how the function would cope with overlapping segments, I've 
created non-overlapping ones at this phase and asked the spectral 
analysis function pop_spectopo to apply overlapping during subsequent 
FFT calculations instead.

Here is a piece of code that adds 2-second event-markers to a loaded 
EEGLAB dataset.
dur = 2; % the required duration of the new events
name = 'M'; % name of the new events that appears in EEGLAB scroll plot
newEvent = {name 0 dur}; % initiate the new event list (see help 
pop_importevent for details)
% the following loop will populate the list with required information
for k = 2:round((EEG.xmax-dur)/dur) % the loop runs until the end of the 
loaded dataset
     newEvent = cat(1, newEvent, {name newEvent{k-1,2}+dur dur}); % 
specs of one event are added on each run of the loop
end
EEG = pop_importevent( EEG, 'event','externalEvent','fields',{'type' 
'latency' 'duration'},'timeunit',1,'optimalign','off'); % import the 
created event list to the loaded dataset
eeglab redraw % update the GUI to reflect the command line changes

After running this code you can use the GUI-based tool to actually 
segment the data (Tools -> Extract epochs).

The FFT can be calculated by the pop_spectopo function (Plot -> Plot 
channel spectra and maps). See help pop_spectopo to learn how to tell 
the algorithm to use overlapping in FFT calculations as well as which 
kind of windowing to use.

Finally, you'll need to turn to the command-line again to perform the 
subtraction of right and left power estimates. One way to do this 
involves adding output variables to the pop_spectopo or spectopo 
function (again, see help) and then performing the subtraction  on this 
output. With some struggling the subtracted data can then be plotted 
using EEGLAB or Matlab tools (e.g. plot) and exported (e.g. xlswrite).

Regards,
Andero Uusberg
Tartu, Estonia


On 19.12.2011 3:51, Campbell, Kathryn wrote:
> Hi,
>
> I know there have been lots of questions about this but I have not found
> any that answer all of my questions. I am completely new to using MATLAB
> and EEGLAB and am a bit overwhelmed.
>
> The steps I have taken in the past to calculate alpha asymmetry were
>
> 1)Segment continuous EEG data into 4 one minute blocks based on events
> markers
>
> 2)For each block create a series of 2.048 second epochs overlapping by 50%
>
> 3)Reject epochs with ocular or other artefacts (either manually or
> automatically)
>
> 4)Calculate mean alpha power at each electrode site for each block using
> FFT with Hanning window.
>
> 5)Use these figures to calculate alpha asymmetry (1n (R/L)
>
> Unfortunately the system I now use to record EEG ( a very old version of
> neuroscan) doesn’t seem have all of these functions available.
>
> I am able to complete the first step using EEG lab but can’t find a way
> to create overlapping epochs to move on to the next steps.
>
> Is anyone able to help me? Please bear in mind I’m not that familiar
> with MATLAB.
>
> Thanks
>
> Kathryn
>
>
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to eeglablist-request at sccn.ucsd.edu