I'm considering an automated method for removal of spontaneous activity. I spend a lot of time manually weeding out cells which we can't analyze because of spontaneous activity. I would consider maybe a method by which we just look at how frequently an event is detected in the windows of time between stimulations, e.g. if a brush occurs between frames 100 and 200 and then a second brush occurs at 1000 and 1100, then consider how many 'event-starts' there are between 200 and 1000. Then do that for every stimulus pair (i.e. the gaps between stimuli). Alternatively we could set up a csv which would include a range of 'events' to specifically examine for spontaneous activity. These cells could then be flagged automatically and we could just go through them manually? Or maybe we validated them and implement some thresholds and just remove the really obvious ones automatically and then have it only ask for checks on the questionable ones?
I'm considering an automated method for removal of spontaneous activity. I spend a lot of time manually weeding out cells which we can't analyze because of spontaneous activity. I would consider maybe a method by which we just look at how frequently an event is detected in the windows of time between stimulations, e.g. if a brush occurs between frames 100 and 200 and then a second brush occurs at 1000 and 1100, then consider how many 'event-starts' there are between 200 and 1000. Then do that for every stimulus pair (i.e. the gaps between stimuli). Alternatively we could set up a csv which would include a range of 'events' to specifically examine for spontaneous activity. These cells could then be flagged automatically and we could just go through them manually? Or maybe we validated them and implement some thresholds and just remove the really obvious ones automatically and then have it only ask for checks on the questionable ones?