mne_denoise.qa.compute_all_qa_metrics#

mne_denoise.qa.compute_all_qa_metrics(raw_before: mne.io.BaseRaw, raw_after: mne.io.BaseRaw, line_freq: float = 50.0, n_harmonics: int = 0, fmax: float = 125.0) → dict[source]#

Compute all QA metrics for a line-noise removal benchmark.

Parameters:

  • raw_before (mne.io.BaseRaw) – Raw recordings before and after cleaning.

  • raw_after (mne.io.BaseRaw) – Raw recordings before and after cleaning.

  • line_freq (float) – Fundamental line-noise frequency (Hz).

  • n_harmonics (int) – Number of harmonics above the fundamental to evaluate.

  • fmax (float) – Maximum frequency for PSD computation.

Returns:

metrics – Dictionary with scalar summary metrics and per-harmonic vectors. Scalar keys: peak_attenuation_db, R_f0, below_noise_distortion_db, overclean_proportion, underclean_proportion, geometric_mean_psd_ratio.

Return type:

dict

Notes

peak_attenuation_db and R_f0 scalar outputs correspond to the first evaluated harmonic (fundamental line frequency).

Examples

>>> # metrics = compute_all_qa_metrics(raw_before, raw_after, line_freq=50.0)
>>> # float(metrics["peak_attenuation_db"])