Lindsay Walker1, Lin-Ching Chang1,2, Efstathios Kanterakis3, Luke Bloy3, Kristina Simonyan4, Ragini Verma3, Carlo Pierpaoli1
1NICHD, NIH, Bethesda, MD, USA; 2Dept of EE & CS, Catholic University of America, Washington, DC, USA; 3Dept of Radiology, University of Pennsylvania, Philadelphia, PA, USA; 4NINDS, NIH, Bethesda, MD, USA
Corrupted DWI data has an effect on DTI derived quantities. An analysis of the effect of physiological noise on statistical analysis of a population is presented. Comparison of non-robust and RESTORE robust tensor fitting shows significant differences in the mean and variance of anisotropy and mean diffusivity. The effects are regionally varying across the brain, and not the same for different tensor derived metrics. When considering a statistical analysis of a population, the effect of outliers may not be the same for both patient and control groups. Statistically significant results may originate from the presence of outliers instead of pathology.