This package has grown organically and most of the time it lives in Python. Spike detection is still done in Matlab. Analysis has flowed like this:
- Get all the .abf files of the desired type ('GapFree', '10^-6 Pilo', etc); this can be done with getExtFromDF in the main package, drophila.py
- Pass that to batchWriteSpikes.m, which uses a moving window to retrieve all spikes from every file listed
- Every .abf file now has a .csv file that includes spike properties for every spike time
- If burst detection is desired, the main package can also find bursts using inter-spike intervals and Markov chain Monte Carlo simulation -- this is not perfect, but gets about 90% of the spikes. Burst inclusion (1=burst, 0=tonic) is appended to the trace's .abf file
- Specific burst features can be extracted if desired using tools in the main function including burst times, inter-burst intervals, and coefficients of variation for each of these
Some details for the burst detection algorithm can be found in the document burst_tonic.pdf in this repo.