[Eeglablist] newtimef.m difference in different versions of EEGLAB

[Yicong Zheng]

Dear all,

I have been doing time frequency analysis in EEGLAB and has been satisfied with it. The thing is that one day I found I have different “tfdata” and “ersp” output values in two versions of EEGLAB (2019_0 and 14_1_2b), using the same code under same MATLAB version(2018b). The function call is:

 [ersp,itc,powbase,times,freqs,erspboot,itcboot,tfdata] = newtimef(EEG.data(jElec,:,trialTypes{kTrialType}), ...
                    EEG.pnts,[EEG.xmin EEG.xmax]*1000, EEG.srate,tfCycles ,'freqs', freqRange, ...
                    'elocs', EEG.chanlocs,'timesout',timeArray,'baseline',[NaN],'chaninfo', EEG.chaninfo,...
                    'nfreqs',freqInterval,'plotphase','off','plotersp','off','plotitc','off');

I want be calculate my own ERSP using this “tfdata” output, but at 14_1_2b I got something like 10/13/18/20 in ERSP whereas at 2019_0 I got something like -10/-12/-17/-22. After examine the output, I found the “itc” remains the same, and tfdata(14_1_2b)./tfdata(2019_0) gives me a constant value matrix. So my guessing here is that there’s some constant that’s multiplied on the “tfdata”. It won’t hurt my result, but the scale of non-baseline corrected power is different in these two versions.

So if anyone has also encountered this issue and has any any answer to what this constant is and what’s the specific change in this function, please let me know and I’d love to discuss with you. Thanks in advance.

-Alan