Meeting Banner
Abstract #3650

TBSS Result Variations: Is the Analysis Dependent on the Fitting Algorithm?

Ivan I. Maximov1, Heike Thoennessen1, 2, Kerstin Konrad2, 3, Laura Amort1, 4, Irene Neuner1, 4, Nadim Jon Shah1, 5

1Institute of Neuroscience and Medicine 4, Forschungszentrum Juelich GmbH, Juelich, Germany; 2Department of Child and Adolescent Psychiatry and Psychotherapy, RWTH Aachen University, Aachen, Germany; 3Institute of Neuroscience and Medicine 3, Forschungszentrum Juelich GmbH, Juelich, Germany; 4Department of Psychiatry, Psychotherapy and Psychosomatics, RWTH Aachen University, Aachen, Germany; 5Department of Neurology, RWTH Aachen University, Aachen, Germany

Voxelwise analysis is a powerful and useful technique that allows one to detect white/grey matter changes and to perform inter-subject group comparisons. However, conventional voxelwise approaches suffer from multiple artefacts originating from the absence of a gold standard in the data processing pipeline. TBSS is a promising framework that reduces the voxelwise comparison in skeleton space. However, this approach is still under debate: does a transition to the FA skeleton space decrease the result variability? We demonstrate that application of the developed robust post-processing framework allows one to reduce the variability of TBSS results. We also found that the TBSS analysis with the developed robust framework exhibits higher reproducibility compared to other DTI algorithms.

Keywords

absence address adolescent allows analyses applied arise array arrive assessed automatic available basis blue brain channel child clinical coil coincidence common contradictory control corrections corresponds criteria criterion decreased dependent described desirable detail detect detected detection diffusion eddy estimation estimations estimator estimators example except exhibits expected exploiting field fitting framework furthermore gold good gradient great green head highly house hypothesize implemented important improve improvement institute inter laura least linear localization localize localized machine matrix medical medicine metrics might minimal modified motion neurology noise normalization ordinary orientation overcoming overlapping patient patients pipeline planning plots potential potentially preparation preprocessing principal problems procedure processing produce producing psychiatry psychosomatics psychotherapy reduction reliability reliable repetition reproduce reproducibility robust rotation scalar scatter scripts selected selection shah slice slices smith smoothing sonata spatial squares stable statistics strength studies subject submitted substantial surgery systems tracking tract transverse treated trimmed turn typical variation variations view vision visualization watch white whole