Brain activity during pre-stimulus period

[Priya Balasubramanian]

Hi,

My research involves analyzing changes in brain activity using 64 electrode EEG data. I use my pre-stimulus data of 2s for baseline normalization as one of my pre-processing steps. Then I calculate percentage change from baseline to observe brain activity during a motor task. If I would like to observe level of brain activity prior to task ( i.e. pre-stimulus), will it make sense to z-score normalize the pre-stimulus period of 2s so I can compare brain activity across all electrodes? I was also considering the min-max normalization for each electrode. 

Would normalization become a moot point if I perform CSD?

Thank you so much for taking the time to read my query.

with regards,

Priya

[Mike X Cohen]

Hi Priya. CSD is a spatial filter that sharpens the activity (by attenuating low-spatial-frequency components) and tends to improve data quality. That's a separate issue from what you are talking about, so you could apply CSD if you want to.

What do you want to compare the pre-stim activity to? If it's across conditions, then a condition-average baseline should be OK, as it will preserve any possible condition differences. Otherwise, it might just be easiest to go with the raw power, at least for the pre-stim analyses. Baseline normalization is often useful, but it's definitely not necessary.

Mike

--
You received this message because you are subscribed to the Google Groups "AnalyzingNeuralTimeSeriesData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to analyzingneuraltimes...@googlegroups.com.
Visit this group at https://groups.google.com/group/analyzingneuraltimeseriesdata.
To view this discussion on the web visit https://groups.google.com/d/msgid/analyzingneuraltimeseriesdata/db9958b0-e185-49c9-8df7-39e6d723b0df%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Mike X Cohen, PhD
Fresh look: mikexcohen.com

[Priya Balasubramanian]

Dear Mike,

Thank you so much for your explanation about CSD. I will definitely be using CSD as I would appreciate spatial filtered data for my analysis. 

I would like to compare pre-stimulus data between controls and stroke population before and after performing a motor task. My hypothesis is that the pre-stimulus power in the beta band frequency range in the two populations will be different. So I would like to visualize the beta power across all electrodes in the two population for the before and after motor task conditions. I was considering normalizing using z-score. Another option would be, for each electrode divide by the total power across all frequencies at that electrode. My objective is to remove electrode impedance as a confounding factor.

with regards,

Priya