I'm a PhD student trying to make sense of some extracellular recordings from spinal cord slices. I'm interested in detecting rhythmicity and finding the major frequency components in my traces using power spectra and autocorrelation analysis but there is something I am unsure of. I find that my autocorrelations will often show a rhythmic component in my traces only after I have performed DC offset removal using my software (Spike2). What I don't know is a) whether this is to be expected and b) what time constant I should be using to remove the offset (i.e. do I compare each data point with those in the second either side, in the millisecond either side, 10 ms...?) such that I preserve the frequencies I'm looking for? My oscillations are in the range 5-30Hz. Any clues at all would be most welcome.