Motion Impact Score for Detecting Spurious Brain-Behavior Associations

Abstract

AbstractBetween-participant differences in head motion introduce systematic bias to resting state fMRI brain-wide association studies (BWAS) that is not completely removed by denoising algorithms. Researchers who study traits, or phenotypes associated with in-scanner head motion (e.g. psychiatric disorders) need to know if trait-functional connectivity (FC) correlations are biased by residual motion artifact in order to avoid reporting false positive results. We devised an adaptable method, Split Half Analysis of Motion Associated Networks (SHAMAN), to assign a motion impact score to specific trait-FC correlations. The SHAMAN approach distinguishes between motion artifact causing false positive vs false negative bias. SHAMAN was > 95% specific at sample sizes of n = 100 and above. SHAMAN was 95% sensitive to detection of false positive motion impact score at sample sizes of n = 3,000 but only 59% sensitive to detection of false negative motion impact score, making it most useful for large BWAS. We computed motion impact scores for trait-FC correlations with 45 demographic, biophysical, cognitive, and personality traits from n = 7,270 participants in the Adolescent Brain Cognitive Development (ABCD) Study. After standard denoising with ABCD-BIDS and without motion censoring, 60% (27/45) of traits had significant (p < 0.05) false positive motion impact scores and 36% (16/45) of traits had false negative motion impact scores. Censoring at framewise displacement (FD) < 0.2 mm reduced the proportion of traits with significant false positive motion impact scores from 60% to 2% (1/45) but did not decrease the number of traits with significant false negative motion impact scores.