Hi all, I've recently been using EMD as part of a pre-processing step but seem to have hit a bit of a problem. Generally the decomposition runs fine and the signals yield some very nice results. However, when doing this at the single trial level I have found that sometimes the EMD just gets into an endless loop (bad trials?). This is a bit problematic for trying to implement it as an automated step.
Has anybody else had this issue? Does anybody know of an optimum algorithm to use with eeg?
I am no expert on EMD or HHT, but with unfiltered data (or single trials) the number of intrinsic mode functions increases, so my advice would be to low-pass filter your data (cut-off set to 40 Hz, for example, if possible) and run EMD to see what is happening.
I do not know of any optimum algorithm for EEG or similar signals, but maybe we can start an interesting discussion here.
Thanks Joan. I've been running it on filtered data, but ended up changing the filter settings slightly. Managed to find a slightly faster algorithm than the one I'd been using as well. Thanks
- Badges
- Science topic
- Similar topics
- Biological Science
- Physiology
- Physiological Processes
- Stress