<div dir="ltr">
<div bgcolor="#FFFFFF" text="#000000">
<p class="MsoNormal"><span lang="EN-US">Dear
FT-List,<br>
<br>
in an MEG experiment we tested the power difference (40-100Hz,
10 tapsmofreq)
between two experimental conditions not averaging over time and
frequency. <br>
We found a nice significant cluster in the lower gamma range and
in the first
half of the time range. <br>
Please have a look at the two figures. <br>
<br>
<a href="http://dropcanvas.com/0fnvn" target="_blank">http://dropcanvas.com/0fnvn</a><br>
<br>
Fig a is the multiplotTFR of stat, showing only significant
areas. <br>
Fig b is the sum over channels of the significance mask giving
the number of
significant channels at the corresponding freq/time point.<br>
<br>
Now, we are not sure how to pick the time and freq range for a
subsequent source
analysis. <br>
Should we take the whole time/freq range in which any
channel/time/freq
combination became significant? Or is it better to pick some
center value, also
to increase power in the source analysis?<br>
<br>
</span>Any ideas?<br>
<br>
Greetings, Cornelius<span lang="EN-US"></span></p>
<br>
<br>
<br>
<br>
</div>
</div>