I want to ask everybody how to do the alignment of audio text and meeg when doing trf analysis or some analysis that needs to align the text time points with the time points of the meeg?
My current approach is like this
1 Make up the text time points into a sequence of time points starting at 0 and ending at the total timepoint, e.g. if 1-2 seconds is ba, and 3-4 seconds is da, I'll make up the 0 at 2-3 seconds, or if there is no text at 0-1 seconds, and the text starts at 1 second, I'll make up the 0 at the beginning of 1 second, then I'll make sure that the whole text timepoints from 0 to the end are all complete.
2 Repeat the line based on the duration of each line of text, for example, 1-2.5 seconds is ba and the duration is 1.5 seconds, if my meeg sampling rate is 1000, then I would repeat the line 1500 times, that is from sampling point 1000-2500, which are all 1500 samples are ba. similarly, repeat the line based on the duration x the sampling rate for all lines of text.
But the result of this method of processing seems to be imprecise because trf results are very poor.
But with the same data and preprocessing, the results with naturally aligned features like envelope are ideal.
