Confounds in Functional Magnetic Resonance Imaging—Part III: Auxiliary Data

In the first part of this series, we reviewed the pathways that bring about confounding nuisance signals in fMRI. Then, in the second part, we considered how human factors can modify these pathways. In this, the third and final part, we review auxiliary data that may allow researchers to identify and perhaps correct for nuisance signals and their modifiers. We consider three broad categories of data: (1) measures that can be collected synchronously with the fMRI, (2) other MRI scans that either capture trait information or attempt to define the state that exists during an fMRI scan in the same session, and (3) non-MRI measures that can be collected before or after the scan and that also capture trait or relevant state factors influencing the fMRI. We conclude with some recommendations on ways to improve fMRI experiments going forward.