Hello everyone,
I am currently working on my Bachelor's thesis for which I am doing a simple EEG analysis. The objective of my thesis is to examine power differences in the theta, alpha and beta frequency band in a pre-stimulus interval (-1s) between later remembered vs forgotten items. I am completely new to EEG analysis as well as python (or any other programming language). Currently, I am using MNE python for the EEG signal analysis. So far, I pre-processed my data and epoched it to the relevant time interval. For the frequency analysis I followed the following tutorial (https://mne.tools/dev/auto_tutorials/time-freq/plot_sensors_time_frequency.html#sphx-glr-auto-tutorials-time-freq-plot-sensors-time-frequency-py). However, I am not interested in a time-frequency analysis and only want to extract the power spectrum. As far as I understood it, I can use psd_welch to retrieve the relevant data, right? I would then like to use this data of several subjects and do my statistical analysis with the retrieved values. Basically, I just want to compare the means of a subset of electrodes (e.g.FT7,T7,TP7) with each other . How do I merge the different data sets so that I have all PSD mean values of all subjects in one dataset? At the moment, I save them separately, but could not figure out how to merge them. In addition, I am not sure where to look for the proper code for an ANOVA given that I usually use SPSS and not python, maybe someone can give me a tip.
Thank you very much for your help. It is a very new topic for me and I am just a little bit lost.
Kind regards,
Hanna
You are right, you can use https://mne.tools/stable/generated/mne.time_frequency.psd_array_welch.html to compute the PSD for every subject, in a loop. If I were you, at each interaction of this loop I would make a Pandas DataFrame out of these PSD values (https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.html) and I would append this DataFrame object to a list (lets call this variable df_lst). At the end of the loop, you can concatenate all the subject-specific data frames of this list into a single DataFrame using something like `merged_df = pandas.concat(df_lst)`. Then you can run statistical analyses on that merge_df using the statsmodels package. I personally find it easier to use the "formula" interface (R-liked style): https://www.statsmodels.org/dev/example_formulas.html
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python