Hi, I was hoping you could clarify the following regarding overlapping the actigraphy, psg and sleep stages data.
I think that I have successfully extracted the overlapping data using the csv, edf and xml files for the actigraphy, psg and sleep stages data respectively. However, there are inconsistencies regarding the number of data points that I have extracted from them. For all of the individuals I obtain 2 extra actigraphy datapoints than sleep stages data points. In the case of the psg data, I use a rolling average to obtain data with 30 second epochs - however still obtain a larger number of points.
Is this to be expected and may I assume that, as long as the start times are in line as given by the overlap information, I can discard the extra data points? Or does this inconsistency mean that it is likely that I have made a mistake?
Many thanks!
How are you extracting the actigraphy (CSV rows) and PSG (XML SleepStaging) data points you need?
Here's one way to think about the task of pulling the data that overlap. Let's say I found a 10 hour PSG recording that contained 1200 epochs of staging data (30 seconds in length). For that subject I could then lookup the overlap line number from the epoch-by-epoch CSV file and select the next 1200 actigraphy epochs to set beside the PSG epochs. Perhaps you are choosing extra actigraphy epochs somehow? Maybe it also has to do with the slight time differences (e.g., actigraphy line times always end in :00 or :30, whereas PSG would not normally start at those times)?