mne_denoise.dss.denoisers.QuasiPeriodicDenoiser#

class mne_denoise.dss.denoisers.QuasiPeriodicDenoiser(peak_distance: int = 100, peak_height_percentile: float = 75.0, *, warp_length: int | None = None, smooth_template: bool = True)[source]#

Quasi-periodic denoiser via cycle averaging.

For signals with repeating structure (ECG, respiration, periodic artifacts): 1. Detect peaks/cycles in the source 2. Segment into individual cycles 3. Time-warp cycles to common length 4. Average to create template 5. Replace each cycle with time-warped template

Parameters:

  • peak_distance (int) – Minimum distance between peaks in samples. Default 100.

  • peak_height_percentile (float) – Percentile of signal for peak detection threshold. Default 75.

  • warp_length (int, optional) – Length to warp each cycle to. If None, use median cycle length.

  • smooth_template (bool) – If True, smooth the template. Default True.

Examples

>>> # For ECG-like signal at 250 Hz (peaks ~1 sec apart)
>>> from mne_denoise.dss.denoisers import QuasiPeriodicDenoiser
>>> denoiser = QuasiPeriodicDenoiser(peak_distance=200)
>>> denoised = denoiser.denoise(ecg_source)

References

Särelä & Valpola (2005). Section 4.1.4 “DENOISING OF QUASIPERIODIC SIGNALS”

__init__(peak_distance: int = 100, peak_height_percentile: float = 75.0, *, warp_length: int | None = None, smooth_template: bool = True) → None[source]#

Methods

__init__([peak_distance, ...])

denoise(source)

Apply quasi-periodic denoising.