mne.time_frequency.psd_welch#

mne.time_frequency.psd_welch(inst, fmin=0, fmax=inf, tmin=None, tmax=None, n_fft=256, n_overlap=0, n_per_seg=None, picks=None, proj=False, n_jobs=None, reject_by_annotation=True, average='mean', window='hamming', *, verbose=None)[source]#

Compute the power spectral density (PSD) using Welch’s method.

Calculates periodograms for a sliding window over the time dimension, then averages them together for each channel/epoch.

Parameters:

instinstance of Epochs or Raw or Evoked

The data for PSD calculation.

fminfloat

Min frequency of interest.

fmaxfloat

Max frequency of interest.

tminfloat | None

Min time of interest.

tmaxfloat | None

Max time of interest.

n_fftint

The length of FFT used, must be >= n_per_seg (default: 256). The segments will be zero-padded if n_fft > n_per_seg. If n_per_seg is None, n_fft must be <= number of time points in the data.

n_overlapint

The number of points of overlap between segments. Will be adjusted to be <= n_per_seg. The default value is 0.

n_per_segint | None

Length of each Welch segment (windowed with a Hamming window). Defaults to None, which sets n_per_seg equal to n_fft.

picksstr | list | slice | None

Channels to include. Slices and lists of integers will be interpreted as channel indices. In lists, channel type strings (e.g., ['meg', 'eeg']) will pick channels of those types, channel name strings (e.g., ['MEG0111', 'MEG2623'] will pick the given channels. Can also be the string values “all” to pick all channels, or “data” to pick data channels. None (default) will pick good data channels (excluding reference MEG channels). Note that channels in info['bads'] will be included if their names or indices are explicitly provided.

projbool

Apply SSP projection vectors. If inst is ndarray this is not used.

n_jobsint | None

The number of jobs to run in parallel. If -1, it is set to the number of CPU cores. Requires the joblib package. None (default) is a marker for ‘unset’ that will be interpreted as n_jobs=1 (sequential execution) unless the call is performed under a joblib.parallel_backend() context manager that sets another value for n_jobs.

reject_by_annotationbool

Whether to omit bad segments from the data before fitting. If True (default), annotated segments whose description begins with 'bad' are omitted. If False, no rejection based on annotations is performed.

Has no effect if inst is not a mne.io.Raw object.

New in version 0.15.0.

averagestr | None

How to average the segments. If mean (default), calculate the arithmetic mean. If median, calculate the median, corrected for its bias relative to the mean. If None, returns the unaggregated segments.

New in version 0.19.0.

windowstr | float | tuple

Windowing function to use. See scipy.signal.get_window().

New in version 0.22.0.

verbosebool | str | int | None

Control verbosity of the logging output. If None, use the default verbosity level. See the logging documentation and mne.verbose() for details. Should only be passed as a keyword argument.

Returns:

psdsndarray, shape (…, n_freqs) or (…, n_freqs, n_segments): The power spectral densities. If average='mean or average='median' and input is of type Raw or Evoked, then psds will be of shape (n_channels, n_freqs); if input is of type Epochs, then psds will be of shape (n_epochs, n_channels, n_freqs). If average=None, the returned array will have an additional dimension corresponding to the unaggregated segments.
freqsndarray, shape (n_freqs,): The frequencies.