Advice needed on salvaging data with unusual evoked baseline activity

Oh that sounds great! Yeah I’ll drop by that session. Thanks!

Ahh yes I did not specify. Due to the nature of my tasks, it is impossible to have same number of trials across conditions per participant. This can scale in the order of 2-5 times difference in number of trials for the outlier participants. Typically, the sponblinks and volblinks conditions will have more trials so averaging should work well to suppress trials with unusually high activity. I suppose it will be of concern if I have <10 trials for the blanks control. I have at least 20-30 blanks trial for the outliers as opposed to 60-150 sponblink trials for the problematic data sets. Would 20-30 trials not be sufficient enough to average out the GFP?

Well the big difference could be a potential answer to what you are seeing. Maybe you can see if the difference in “sponblink” vs “blank” trials is especially pronounced in the “weird” data sets.

1 Like

You could also try and only average a subset of 20–30 epochs of the two “good” conditions and see how the GFP behaves

1 Like