epoch time window using EEGLAB

[Qing He]

Dear Mike

I try to compare the time-frequency analysis methods introduced by you with the methods derived from EEGLAB toolbox.
Here the epoch time in my dataset is 2500 ms, but the output time is 1385ms after using newtimef function?
However, it can show us a full length time window when I using the methods you provided in Chapter 13.
So, could you tell me the main reason to to this difference?
--------------------------------------------------------------------------------------------
[ersp, itc,powbase, times, freqs,erspboot, itcboot]=newtimef(EEG.data(1,:,:), 2500, [EEG.xmin EEG. xmax]*1000, 1000, [3, 0.5], 'maxfreq',45, freqs,[3, 45], 'winsize',256, 'padratio',8,'timesout',2500);


Thank you very much!

Qing He

[Mike X Cohen]

Hi Qing. This is a question about the inner computations of eeglab, and I'm really not qualified to answer that question, because I don't know that much about the "meat" of eeglab. But a few things come immediately to mind:

- newtimef can do wavelet convolution or STFFT. I'm not sure what it's doing from the inputs you specify. If you really want to compare them directly, you'd have to try to match the sets of parameters as closely as possible.

- I guess eeglab is doing some post-analysis downsampling of roughly 1/2. That's probably in the "times" output.

- How do the results generally look? Any time you change parameters or use a different algorithm, the results will necessarily change. The question is whether the difference is more qualitative (i.e., the two methods looks completely different) or more quantitative (i.e., the plots generally look similar but there are some differences). If it's the former case, then something is off and it deserves investigation. If the latter, then I would say don't worry about it, unless you are investigating this as a learning experience. And in that case, you will need to go through the newtimef function line by line to see what it's doing.

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 analyzingneuraltimeseriesdata+unsub...@googlegroups.com.
Visit this group at https://groups.google.com/group/analyzingneuraltimeseriesdata.
For more options, visit https://groups.google.com/d/optout.

--

Mike X Cohen, PhD
mikexcohen.com - inverse