`tedana.decomposition`.tedpca

tedpca(data_cat, data_optcom, mask, adaptive_mask, io_generator, tes, n_independent_echos=None, algorithm='aic', kdaw=10.0, rdaw=1.0, low_mem=False)[source]

Use principal components analysis (PCA) to identify and remove thermal noise from data.

Parameters:

data_cat ((Mb x E x T) array_like) – Input functional data
data_optcom ((Mb x T) array_like) – Optimally combined time series data
mask ((Mb,) array_like) – Boolean mask array. This is more restrictive than the base mask (M), so additional masking and unmasking will take place within the function.
adaptive_mask ((Mb,) array_like) – Array where each value indicates the number of echoes with good signal for that voxel. This mask may be thresholded; for example, with values less than 3 set to 0. For more information on thresholding, see make_adaptive_mask.
io_generator (tedana.io.OutputGenerator) – The output generation object for this workflow
tes (list) – List of echo times associated with data_cat, in seconds
n_independent_echos (int, optional) – Number of independent echoes to use in goodness of fit metrics (fstat). Primarily used for EPTI acquisitions. If None, number of echoes will be used. Default is None.
algorithm ({‘kundu’, ‘kundu-stabilize’, ‘mdl’, ‘aic’, ‘kic’, float}, optional) – Method with which to select components in TEDPCA. PCA decomposition with the mdl, kic and aic options are based on a Moving Average (stationary Gaussian) process and are ordered from most to least aggressive (see [1]). If a float is provided, then it is assumed to represent percentage of variance explained (0.0-1.0) to retain from PCA. If an int is provided, then it is assumed to be the number of components to select Default is ‘aic’.
kdaw (float, optional) – Dimensionality augmentation weight for Kappa calculations when algorithm is ‘kundu’. Must be a non-negative float, or -1 (a special value). Default is 10.
rdaw (float, optional) – Dimensionality augmentation weight for Rho calculations when algorithm is ‘kundu’. Must be a non-negative float, or -1 (a special value). Default is 1.
low_mem (bool, optional) – Whether to use incremental PCA (for low-memory systems) or not. This is only compatible with the “kundu” or “kundu-stabilize” algorithms. Default: False

Returns:

kept_data ((Mc x T) numpy.ndarray) – Dimensionally reduced optimally combined functional data, where Mc is samples in classification mask, T is time.
n_components (int) – Number of components retained from PCA decomposition

Notes

Notation	Meaning
$\kappa$	Component pseudo-F statistic for TE-dependent (BOLD) model.
$\rho$	Component pseudo-F statistic for TE-independent (artifact) model.
$v$	Voxel
$V$	Total number of voxels in mask
$\zeta$	Something
$c$	Component
$p$	Something else

Steps:

Variance normalize either multi-echo or optimally combined data, depending on settings.
Decompose normalized data using PCA or SVD.
Compute ${\kappa}$ and ${\rho}$ :

${\kappa}_c = \frac{\sum_{v}^V {\zeta}_{c,v}^p * \ F_{c,v,R_2^*}}{\sum {\zeta}_{c,v}^p} {\rho}_c = \frac{\sum_{v}^V {\zeta}_{c,v}^p * \ F_{c,v,S_0}}{\sum {\zeta}_{c,v}^p}$
Some other stuff. Something about elbows.
Classify components as thermal noise if they meet both of the following criteria:
- Nonsignificant ${\kappa}$ and ${\rho}$ .
- Nonsignificant variance explained.

Generated Files

Default Filename	Content
desc-PCA_metrics.tsv	PCA component table
desc-PCA_metrics.json	Metadata sidecar file describing the component table
desc-PCA_mixing.tsv	PCA mixing matrix
desc-PCA_components.nii.gz	Component weight maps
desc-PCA_decomposition.json	Metadata sidecar file describing the PCA decomposition

References

tedana.decomposition.tedpca

`tedana.decomposition`.tedpca