mne_denoise.viz.plot_signal_overlay#

mne_denoise.viz.plot_signal_overlay(inst_before, inst_after, times, pick=None, start=None, stop=None, scale_after=True, before_label='Before', after_label='After', x_label='Time', y_label='Amplitude', title=None, reference=None, reference_label='Reference', highlight_mask=None, highlight_label='Artifact', highlight_spans=None, show=True, fname=None)[source]#

Overlay one before/after trace to inspect reconstruction quality.

Parameters:

inst_beforeMNE object | ndarray: Signal before denoising. Accepted array shapes are (n_times,), (n_channels, n_times), and (n_epochs, n_channels, n_times).
inst_afterMNE object | ndarray: Signal after denoising with matching time and channel dimensions.
timesarray-like of shape (n_times,): Explicit time axis for both traces after length alignment.
pickint | str | None: Channel to display. Required when input has more than one channel. String picks require MNE channel names.
startfloat | None, default=None: Optional lower bound on the time axis.
stopfloat | None, default=None: Optional upper bound on the time axis.
scale_afterbool: If True, scale the after-trace to the before-trace standard deviation.
before_labelstr: Legend label for the first input.
after_labelstr: Legend label for the second input.
x_labelstr: X-axis label.
y_labelstr: Y-axis label.
titlestr | None: Optional custom title.
referencearray-like of shape (n_times,) | None: Optional ground-truth/reference trace overlaid on the same axis (e.g. the clean signal in a simulation). Aligned to the same length and time window as the before/after traces.
reference_labelstr: Legend label for the reference trace.
highlight_maskarray-like of bool of shape (n_times,) | None: Optional boolean mask; samples where True are shaded with fill_between to mark artifact regions.
highlight_labelstr: Legend label for the highlight_mask shading.
highlight_spanssequence[mapping] | None: Optional list of spans to shade with axvspan. Each item is a mapping with onset and duration (in the same units as times) and optional color, alpha, and label keys. Only the first span of each distinct label receives a legend entry. Useful for annotation-based repair/rejection spans.
showbool: If True, display the figure.
fnamepath-like | None: Optional output path used to save the figure.

Returns:

figmatplotlib.figure.Figure: Figure handle.

Raises:

ValueError: If input shapes are invalid, multi-channel data is used without pick, or times length is inconsistent.