Hi Leon, thanks for pointing out. I tried findpeaks in the signal processing toolbox, and but it returned "No peaks found." Possibly, in the later versions (version I am using now is kinda old), the function has been corrected to locate the peak. Actually, I am not so concerned about the error because it only arises when there are instances of adjacent samples having exactly the same value on the left side of the peak, e.g., findpeaks([1 2 2.001 3 3 2 1]) returns the correct answer. In most cases, at least in the context of HHT/EMD or general signal processing, such occurrences are rare. Sorry I could not be more helpful.