Quality assessment of high angular resolution diffusion imaging data using bootstrap on Q‐ball reconstruction

Abstract

To develop a bootstrap method to assess the quality of High Angular Resolution Diffusion Imaging (HARDI) data using Q-Ball imaging (QBI) reconstruction. HARDI data were re-shuffled using regular bootstrap with jackknife sampling. For each bootstrap dataset, the diffusion orientation distribution function (ODF) was estimated voxel-wise using QBI reconstruction based on spherical harmonics functions. The reproducibility of the ODF was assessed using the Jensen-Shannon divergence (JSD) and the angular confidence interval was derived for the first and the second ODF maxima. The sensitivity of the bootstrap method was evaluated on a human subject by adding synthetic noise to the data, by acquiring a map of image signal-to-noise ratio (SNR) and by varying the echo time and the b-value. The JSD was directly linked to the image SNR. The impact of echo times and b-values was reflected by both the JSD and the angular confidence interval, proving the usefulness of the bootstrap method to evaluate specific features of HARDI data. The bootstrap method can effectively assess the quality of HARDI data and can be used to evaluate new hardware and pulse sequences, perform multifiber probabilistic tractography, and provide reliability metrics to support clinical studies.