Currently, MNE-Python provides a set of functions
allowing to compute spectral analyses in the source space.
Many these functions return `mne.SourceEstimate()`

objects or collections thereof.

Note

The `mne.SourceEstimate()`

object was initially designed for classical time-domain analyses.
In this context, the time axis can actually refer to frequencies. This might be improved
in the future.

The following functions are based on minimum norm estimates (MNE).

`mne.minimum_norm.compute_source_psd_epochs()`

returns single-trial power spectral density (PSD) esitmates using multi-tapers. Here, the time axis actually refers to frequencies, even if labeled as time.`mne.minimum_norm.compute_source_psd()`

returns power spectral density (PSD) esitmates from continuous data usign FFT. Here, the time axis actually refers to frequencies, even if labeled as time.`mne.minimum_norm.source_band_induced_power()`

returns a collection of time-domain`mne.SourceEstimate()`

for each frequency band, based on Morlet-Wavelets.`mne.minimum_norm.source_induced_power()`

returns power and inter-trial-coherence (ITC) as raw numpy arrays, based on Morlet-Wavelets.

Alternatively, the source power spectral density can also be estimated using the DICS beamformer,
see `mne.beamformer.dics_source_power()`

.