Effect of phase-encoding direction on group analysis of resting-state functional magnetic resonance imaging

Echo-planar imaging is a common technique used in functional magnetic resonance imaging (fMRI); however, it suffers from image distortion and signal loss because of large susceptibility effects that are related to the phase-encoding direction of the scan. Despite this relation, the majority of neuroimaging studies has not considered the influence of phase-encoding direction. Here, we aimed to clarify how phase-encoding direction can affect the outcome of an fMRI connectivity study of schizophrenia (SCZ).

Resting-state fMRI using anterior to posterior (A-P) and posterior to anterior (P-A) directions was used to examine 25 patients with SCZ and 37 matched healthy controls (HC). We conducted a functional connectivity (FC) analysis using independent component analysis and performed three group comparisons: (i) A-P versus P-A (all participants); (ii) SCZ versus HC for the A-P and P-A datasets; and (iii) the interaction between phase-encoding direction and participant group.

The estimated FC differed between the two phase-encoding directions in areas that were more extensive than those where signal loss has been reported. Although FC in the SCZ group was lower than that in the HC group for both directions, the A-P and P-A conditions did not exhibit the same specific pattern of differences. Further, we observed an interaction between participant group and the phase-encoding direction in the left temporoparietal junction and left fusiform gyrus.

Phase-encoding direction can influence the results of FC studies. Thus, appropriate selection and documentation of phase-encoding direction will be important in future resting-state fMRI studies.