Introduction: Processing ECoG stored in the NWB format

process_nwb is meant to be an experiment-agnostic library for preprocessing ECoG data stored in the Neurodata Without Borders (NWB) format. It contains low-level signal processing routines and wrapper functions to interface with the NWB format.

These functions can be used individually but they are also packaged into an executable script preprocess_folder. This can be used to preprocess all nwb files in a given folder. Preprocessing consists of

  1. downsampling the original signal (to 3200 Hz by default) and storing this new signal in the file,

  2. notch filter for 60 Hz linenoise and calculate, store, and remove a common average reference (mean of 95% by default), the new signal is stored

  3. calculate and potentially resample the amplitudes for a filterbank of wavelets, by default constant-Q and log-spaced center frequencies. By default, only the amplitudes are stored.

Run

preprocess_folder -h

for options.