Meeting Banner
Abstract #2126

Which Blind Tract Clustering Method Is Most Robust to False Positives?

Mark Drakesmith1, John Evans1, Anthony David2, Derek Jones1

1CUBRIC, Cardiff University, Cardiff, Wales, United Kingdom; 2Institute of Psychiatry, Kings College London, London, United Kingdom

The effect of false positives on the performance of blind tract clustering is unclear. From an idealised dataset of 6 pre-defined bundles, various distance metrics and clustering methods were tested across varying FP-rates by substituting a random proportion of true tracts with FPs. Most methods deteriorate gradually with noise. Affinities computed from maximum (Hausdorff) and endpoint distances were most robust to noise. These methods showed FP misclassification were concentrated on the smallest fibre bundle while other methods showed more diffuse misclassification across adjacent bundles. These two distance metrics are therefore best for clustering noisy tractography datasets.

Abstract PDF Poster PDF

Keywords

accommodate active adequately adjacent affinity although ambiguity analyses anatomy another applying approaches arbitrariness assigned atlas attempt automated beneficial blind body bundle bundles chamfer characterize characterizes cluster clustering clusters coded collected college combined computational computed computing concentrated conform consistently correction correctly crossing dampened dataset deconvolution defined degradation derived distance distinct distinctive distortion distribution downsides eddy embedding endpoint endpoints equal error evaluate every except expressed extremities false focal fraction genuine good gradual inaccuracies increasing increment inevitably initial intra issue kingdom kings known launched lead likely manual mark matrix metric metrics middle minimal misclassification misclassified model motion noise noticeable occur offer opposed orientation pairs pattern percentage performance permutations portions positive positives preferable presence prevalent procedure proportion psychiatry quantified quantify random randomly rather reduce refer replaced resolve responses robust robustness scale segmentation selected smaller spatially specified spectral step still streamline streamlines studies subset taken towards tract tracts trajectories transform true type variance vectors versus wales