Hi,
I am working on the spike amplitude in my EEG data and trying to find the channel with the highest amplitude for each spike. However, Brainstorm demonstrates a different amplitude if I look at the same spike and window in the overview with all the channels vs when I look at it on the single channel (see screenshot) Does anyone know why this is? I am looking in average reference montage. Thanks!
Hello – it’s because the y-axis scales are different between plots.
I am not sure I understand what you are meaning. Because it is the same signal being displayed in both cases: to get the second window, we selected the signal from the first recording and did ‘displayed selected channel’
So the amplitudes should be exactly the same in both case. Note that, when we load the file in Matlab, it is the second amplitude that is being displayed ( -61.21 muV in this case)
[Edit] It seems the difference arises if we use the average reference montage in the first plot and is not present if we use [all channels]. But i am not sure it should matter in this case, as we already applied the reference montage to the file (using standardize > apply montage)
In this thread ( Apply montage or re-reference EEG ), @Raymundo.Cassani seemed to indicate that applying the average reference montage twice would have no effect on the signal ?
This is correct, as the in the second time the average across channels is 0.
Could it be that other processing was applied after? ICA/SSP, that would change the weights of the channels, so average across channels may not be 0 any more.
Ah thanks! I have marked some channels as ‘bad’ after the average reference montage. If I define these as ‘good’ again, the issue indeed goes away. Does that mean I should first mark the channels and then apply the average reference montage?
In my case, I already made single spike event files that I use in Matlab and these demonstrate the correct amplitude. So should I redo my steps and first do the marking, or is it good enough if I know that causes the issue and change my view to [all channels]?
Yes, that should be the first step, otherwise the signal of those bad channel will get into the average reference, and from there to all the other channels (when the avg reference is used).
You may want to check if those bad channels have a significant influence on the average referencing, and decide based on that.
