Quality Control for AIND ephys pipeline

aind-ephys-processing-qc

Description

This capsule is designed to compute Metrics following the QualityControl schema form aind-data-schema.

The following metrics are computed:

• Raw data: snippets of AP and LFP traces at multiple points over the session
• PSD: Power Spectrum Density of wide band, high and low freq also over depth
• Noise: RMS channel values of raw data and preprocessed data
• Drift: raster map and estimated drift
• Saturation: how many positive/negative saturation events and trace snippets around them
• Event triggers: average response to selected HARP events (defaults: licks and optogenetics stimulation). This metric needs HARP behavior events.
• Unit Yield: distribution of ISI violation ratios, amplitude cutoff, presence ratio, unit depths, amplitudes, firing rate and their classification (SUA, MUA, noise)
• Firing rate: population firing rate over the session (to detect seizure-like or drops in activity)

Inputs

The data/ folder must include the output of the raw data, the output of the result-collector caspule, and at least one JSON file generated by the aind-ephys-job-dispatch capsule is required.

Parameters

The code/run script takes no arguments.

Output

The capsule produces the following output for each input JSON file:

• results/quality_control_{recording_name}.json file, representing a valid aind_data_schema.core.QualityControl object.
• results/quality_control_{recording_name} folder, with reference figures for the quality_control_{recording_name}.json file.