Meeting Banner
Abstract #2270

Reproducible Differentiation of Individual Subjects with Minimal Acquisition Time Via Resting State FMRI

Raag D. Airan1, Joshua Vogelstein2, 3, Brian S. Caffo4, James J. Pekar1, 5, Haris I. Sair1

1Radiology and Radiological Science, Johns Hopkins Medical Institutions, Baltimore, MD, United States; 2Statistical Sciences, Duke University, Durham, NC, United States; 3Applied Mathematics and Statistics, Johns Hopkins University, Baltimore, MD, United States; 4Biostatistics, Johns Hopkins University, Baltimore, MD, United States; 5F.M. Kirby Research Center for Functional Brain Imaging, Kennedy Krieger Institute, Baltimore, MD, United States

Resting state fMRI (rs-fMRI) has received much recent attention as a complement to task-based fMRI. While reproducibility of rs-fMRI has been well established at the group level, individual subject level analysis is desired for most clinical applications and is less well characterized. Here we establish the reproducibility of subject-level rs-fMRI data and determine that 7 minutes is sufficient to acquire the maximal rs-fMRI information for differentiating individual subjects. We see strong reproducibility of individual subject data with only 1 minute of acquisition time. Additionally, we find that brain regions yielding the most individual differentiation correspond to higher cortical processing centers.

Abstract PDF Poster PDF

Keywords

acquisition adjacency allows although amount analyzed applications assessment attention audience available best biostatistics black body brain calculation capture cerebellum characteristics characterization characterized clinical clinicians clustering coded coefficients color completed content continues contributing correction correlation correlations cortex cortical courses cross custom decrease define defined determinants determine determined deviation differentiability differentiation dissimilarity distance distances distribution documented duke edges either examined excessive excluded extract extraction frequency function functional ganglia gray highest highly identified included indicated indicates individual individuals institutions johns label lowest makes mathematics matrices matrix maximal medical metric minimal minute minutes mode motion near nodes normalization nuisance pair pairings pairs particularly partitioning performance preprocessing processing publicly question radiological random rank ranked ranks received recent regression reproducibility reproducible researchers resolution resting retest saturates schemes science sciences scrambling scripts seeking sense separated series slice squared statistics strong subject subjects sufficient suggesting sums supported taken taking target task template timing trust typically underwent varied whole yielding