07. Memory Usage

Working with microchannels (32k or 30k Hz) requires intensive memory usage. Here is a list of brief memory usage for each step in the pipeline:

Unpack

Unpacking Neuralynx micro channels will utilize approximately 50GB (20GB for macro channels) of memory when running with 10 parallel jobs. Since the longest segment of Neuralynx data is 2 hours and the unpacking process handles each segment independently, memory usage does not increase with the number or duration of experiments.

Spike sort

The signal filtering step and the calculation of cross-channel spike codes are likely the most memory-intensive processes in spike sorting. For a 2-hour Neuralynx dataset, we can run up to 5 parallel jobs with 64GB memory.