rsatoolbox.io.spm module

Handeling SPM (Statistical Parametric Mapping) fMRI data

Utility object that helps to extract time series data, beta coefficients, and residuals from a GLM stored in a SPM.mat file.

## Usage ` spm = SpmGlm('path/to/spm') spm.get_info_from_spm_mat() [residuals, beta, info] = spm.get_residuals('my_ROI_Mask.nii') `

class rsatoolbox.io.spm.SpmGlm(path: str, nitoolsMock=None)

Bases: object

class for handling first-levels GLMs estimated in SPM

Variables:

path (str) – paths to directory containing SPM files

get_betas(mask: Nifti1Image | NDArray | str) → Tuple[NDArray, NDArray, Dict]

Samples the beta images of an estimated SPM GLM at the mask locations also returns the ResMS values, and the obseration descriptors (run and condition) name

Parameters:

mask (ndarray or nibabel nifti1image) – Indicates which voxels to extract Could be a binary 3d-array or a nifti image of the same size as the data Or a 3xP array of coordinates to extract (in mm space)

Returns:

N x P array of beta coefficients resms (ndarray): 1d array of ResMS values obs_descriptors (dict): with lists reg_name and run_number (N long)

Return type:

data (ndarray)

get_info_from_spm_mat() → None

Initializes information for SPM.mat file

Parameters:

spm_mat_path (str) – _description_

get_residuals(mask: Nifti1Image | NDArray | str) → Tuple[NDArray, NDArray, Dict]

Collects 3d images of a range of GLM residuals (typical SPM GLM results) and corresponding metadata (scan number + run info) into respective lists

Parameters:

res_range (range) – range of to be saved residual images per run

relocate_file(fpath: str) → str

SPM file entries to current project directory and OS.

These are file paths suffixed with two spaces and an index number. They are stored as absolute paths on the computer where it is generated, meaning the project directory, as well as OS-dependent path separator may have changed. This method fixes these.

Parameters:

fpath (str) – SPM-style file path and number from unknown OS

Returns:

SPM-style file path

Return type:

str

spm_filter(data: NDArray) → NDArray

Does high pass-filtering and temporal weighting of the data (indentical to spm_filter)

Parameters:

data (ndarray) – 2d array of time series data (TxP)

Returns:

2d array of time series data (TxP)

Return type:

data (ndarray)