modalities.fmri.realfuncs

Module: modalities.fmri.realfuncs

Helper functions for constructing design regressors

Functions

nipy.modalities.fmri.realfuncs.dct_ii_basis(volume_times, order=None, normcols=False)

DCT II basis up to order order

See: https://en.wikipedia.org/wiki/Discrete_cosine_transform#DCT-II

By default, basis not normalized to length 1, and therefore, basis is not orthogonal. Normalize basis with normcols keyword argument.

Parameters

volume_times : array-like

Times of acquisition of each volume. Must be regular and continuous otherwise we raise an error.

order : None or int, optional

Order of DCT-II basis. If None, return full basis set.

normcols : bool, optional

If True, normalize columns to length 1, so return orthogonal dct_basis.

Returns

dct_basis : array

Shape (len(volume_times), order) array with DCT-II basis up to order order.

Raises

ValueError

If difference between successive volume_times values is not constant over the 1D array.

nipy.modalities.fmri.realfuncs.dct_ii_cut_basis(volume_times, cut_period)

DCT-II regressors with periods >= cut_period

See: http://en.wikipedia.org/wiki/Discrete_cosine_transform#DCT-II

Parameters

volume_times : array-like

Times of acquisition of each volume. Must be regular and continuous otherwise we raise an error.

cut_period: float

Cut period (wavelength) of the low-pass filter (in time units).

Returns

cdrift: array shape (n_scans, n_drifts)

DCT-II drifts plus a constant regressor in the final column. Constant regressor always present, regardless of cut_period.