Working with neuralynx data, and have found a few datasets where there is missing data and the segment times don't line up. In the past, I would fill in data using some form of interpolation and on a previous thread polyphasic resampling was proposed and incorporated into Neo.
However, I have now observed multiple datasets where there is a negative amount of missing data.
To explain further, consider this example. The t_stop of the first segment is after the t_start of the second segment. I had set datasets with this phenomena aside, but now I am coming up against these. This has appeared in several datasets, but in the one I am processing now, I am looking at a total of about 1018 samples that are now ambiguous, or about 32 msec of overlap across many segments.
I don't think this will amount to a crazy amount of differences in downstream spike sorting / LFP analysis but I want to have a formal method for dealing with this particular issue. I checked the overlapping data in question and they are different, i.e. not just duplicate samples.
Working with neuralynx data, and have found a few datasets where there is missing data and the segment times don't line up. In the past, I would fill in data using some form of interpolation and on a previous thread polyphasic resampling was proposed and incorporated into Neo.
However, I have now observed multiple datasets where there is a negative amount of missing data.
To explain further, consider this example. The t_stop of the first segment is after the t_start of the second segment. I had set datasets with this phenomena aside, but now I am coming up against these. This has appeared in several datasets, but in the one I am processing now, I am looking at a total of about 1018 samples that are now ambiguous, or about 32 msec of overlap across many segments.
I don't think this will amount to a crazy amount of differences in downstream spike sorting / LFP analysis but I want to have a formal method for dealing with this particular issue. I checked the overlapping data in question and they are different, i.e. not just duplicate samples.