Dear Kevin,
The simplest thing to do would be to add baseline segments as another
condition and invert them together with activation. This has been
discussed before and there are several people I know of who did so.
You might need to change the time axis with the timeonset method to be
able to merge baseline and activation segments.
Best,
Vladimir
On 23 Dec 2009, at 05:07, Kevin Guise wrote:
> Hello,
>
> I'm wondering if anyone has experience with the above. I'd like to
> do two sets of analyses
> on sources of the induced EEG response: one response locked, where
> the power in the time-
> frequency window of interest is expressed as a change from a pre-
> target-onset baseline, as
> well as a response-locked analysis, where the power is also
> expressed with respect to a
> pre-target-onset baseline. I understand that it's recommended to
> restrict the reconstruction
> time range to the epoch of interest; this is straightforward for
> target locked analyses where
> the time from the start of baseline to the end of the epoch of
> interest is the same across
> trials, but it doesn't seem so easy for response-locked. One thing
> I thought to do was four
> reconstructions (one for each target- and response-locked window of
> interest as well as for
> the corresponding baselines) with four separate time-windows.
> Another was to play with
> the code and use different time windows for each trial of the
> response-locked analysis.
>
> If anyone has additional suggestions or good reasons why I should go
> with one approach
> over the other, it would be greatly appreciated.
>
> Best
> Kevin
|