fmri_corregressor

Description

FMRI_CORREGRESSOR Voxelwise correlation of regressor(s) with fMRI data. Correlates one or more regressors with every voxel in the brain, returning a brain map of Pearson r values (or Spearman rho if requested). This is the primary function for univariate GLM-style analysis (e.g. correlating a stimulus time course, a behavioural score, or a seed-region time series with the whole brain).

Usage

c = fmri_corregressor(regr, data)
c = fmri_corregressor(regr, data, method)

Inputs

REGR: Regressor matrix (T x R): T time points, R regressors. Will be normalised internally. For a single regressor, a (T x 1) vector is accepted.
DATA: fMRI time series, vectorized (228453 x T). Each row is a voxel, each column a time point.
METHOD: Correlation method: ‘pearson’ (default) or ‘spearman’. Spearman is rank-based and more robust to outliers but slower.

Outputs

C: Correlation map (228453 x R). Each column is the r-map for one regressor. Values are clipped to [-1, 1].

Examples

r = fmri_corregressor(stim_ts, data);
fmri_showslices(r, 1, 2, [0.2 0.6])
% Spearman (rank-based, outlier-robust)
r = fmri_corregressor(score, group_data, 'spearman');
% Multiple regressors at once
r = fmri_corregressor([regr1, regr2], data);