[Eeglablist] automated detection of slow wave sleep periods in intracranial EEG

[James Desjardins]

Hi Athena and Erin,

You can manually place events in the continuous data using "Edit >  
Visually edit events and identify bad channels". This will bring up  
the eegplot scroll window. If you hold CTRL and left click on the data  
a window will pop up allowing you insert a new event and you can call  
it whatever you want (type the new name into the "Event type" edit  
box). If you place new events in your data that mark the boundaries  
for periods of interest (eg. "slowW_start" & "slowW_stop") and then  
click on the "Update EEG structure" button in the bottom right corner  
of the data scrolling figure you will be able to select the periods of  
time between those events by doing the following in the Matlab command  
line (assuming the event labels "slowW_start" & "slowW_stop" were used):

%create a points vector from the new events in the data.
pnts=[EEG.event(strmatch('slowW_start',{EEG.event.type},'exact)).latency;EEG.event(strmatch('slowW_stop',{EEG.event.type},'exact)).latency]';

%select the data using the points vector.
EEG=pop_select(EEG,'points',pnts);

%redraw eeglab
eeglab redraw



I hope that this is helpful.


James Desjardins
Technician, MA Student
Department of Psychology, Behavioural Neuroscience
Cognitive and Affective Neuroscience Lab
Brock University
500 Glenridge Ave.
St. Catharines, ON, Canada
L2S 3A1
905-688-5550 x4676


Quoting Erin Wamsley <EWamsley at bidmc.harvard.edu>:

> Athena-
>
> Unfortunately, the only strongly reliable way to pull out periods of slow
> wave sleep is to score this data by hand, marking every epoch that you want
> to include.  Although there are some automated sleep scoring algorithms out
> there, reliability is generally not great, and I am not aware of any
> automated scoring method which has been applied to intracranial data. In
> EEGLab, the simplest thing to do would be to manually highlight and delete
> all non slow-wave periods using the artifact rejection tool (Tools > Reject
> continuous data).  Alternatively, you could use an outside program to
> classify each 30sec epoch according to sleep stage, and then use the ³select
> data² function to select a list of time periods that were classified as slow
> wave sleep ‹ but I am not aware of any existing script for doing this type
> of sleep scoring in MatLab/EEGLab ‹ does anyone else know of something?
>
> Best,
> Erin Wamsley
>
>
> Erin J. Wamsley, Ph.D.
> Instructor, Harvard Medical School
> 330 Brookline Avenue E/FD 862
> Boston MA 02215
> 617.667.8486
>
>
>
> On 2/23/11 8:33 AM, "Athena Lemesiou" <a.lemesiou at ucl.ac.uk> wrote:
>
>> Hi All,
>>
>> I have recently started my PhD and I am very new to EEGLab and Matlab in
>> general so please forgive my ignorance.
>>
>> I have a lot of intracranial EEG data from which I only want to look at slow
>> wave sleep periods. Is this something that can be done through   
>> EEGlab? If not,
>> is anyone aware of any scripts written in Matlab for this purpose?
>>
>> Any help/feedback is very much appreciated.
>>
>> Best wishes,
>> Athena
>>
>
>
> Erin J. Wamsley, Ph.D.
> Instructor, Harvard Medical School
> 330 Brookline Avenue E/FD 862
> Boston MA 02215
> 617.667.8486
>
>
>
>