How fMRIPrep head-motion regressors are calculated?

Does fMRIPrep’s motion variables: trans_x trans_y trans_z rot_x rot_y rot_z and the derivatives: trans_x_derivative1 trans_y_derivative1 trans_z_derivative1 rot_x_derivative1 rot_y_derivative1 rot_z_derivative1 are derived by
removing the mean and linear trend from each of the above variables and its derivates?

The trans_* and rot_* are the raw values retrieved from MCFLIRT, without detrending. The derivatives are just the discrete derivative, dx[i] = x[i] - x[i-1].

1 Like