Biology subjects

Voineskos, A.

Publications and source records attributed to Voineskos, A..

3 recordsLinked to original sources

bioRxiv2018

Deep Learning for Quality Control of Subcortical Brain 3D Shape Models

We present several deep learning models for assessing the morphometric fidelity of deep grey matter region models extracted from brain MRI. We test three different convolutional neural net architectures (VGGNet, ResNet and Inception) over 2D maps of geometric features. Further, we present a novel geometry feature augmentation technique based on parametric spherical mapping. Finally, we present an approach for model decision visualization, allowing human raters to see the areas of subcortical shapes most likely to be deemed of failing quality by the machine. Our training data is comprised of 5200 subjects from the ENIGMA Schizophrenia MRI cohorts, and our test dataset contains 1500 subjects from the ENIGMA Major Depressive Disorder cohorts. Our final models reduce human rater time by 46-70%. ResNet outperforms VGGNet and Inception for all of our predictive tasks.

neuroscience↗

bioRxiv2017

Machine Learning for Large-Scale Quality Control of 3D Shape Models in Neuroimaging

As very large studies of complex neuroimaging phenotypes become more common, human quality assessment of MRI-derived data remains one of the last major bottlenecks. Few attempts have so far been made to address this issue with machine learning. In this work, we optimize predictive models of quality for meshes representing deep brain structure shapes. We use standard vertex-wise and global shape features computed homologously across 19 cohorts and over 7500 human-rated subjects, training kernelized Support Vector Machine and Gradient Boosted Decision Trees classifiers to detect meshes of failing quality. Our models generalize across datasets and diseases, reducing human workload by 30-70%, or equivalently hundreds of human rater hours for datasets of comparable size, with recall rates approaching inter-rater reliability.

neuroscience↗

bioRxiv2017

Personalized intrinsic network topography mapping and functional connectivity deficits in Autism Spectrum Disorder

BackgroundRecent advances demonstrate individually specific variation in brain architecture in healthy individuals using fMRI data. To our knowledge, the effects of individually specific variation in complex brain disorders have not been previously reported.\n\nMethodsWe developed a novel approach (Personalized Intrinsic Network topography, PINT) for localizing individually specific resting state networks using conventional resting state fMRI scans. Using cross-sectional data from participants with ASD (n=393) and TD controls (n=496) across 15 sites we tested: 1) effect of diagnosis and age on the variability of intrinsic network locations and 2) whether prior findings of functional connectivity differences in ASD as compared to TD remain after PINT application.\n\nResultsWe found greater variability in the spatial locations of resting state networks within individuals with ASD as compared to TD. In TD participants, variability decreased from childhood into adulthood, and increased again in late-life, following a U-shaped pattern, which was not present in those with ASD. Comparison of intrinsic connectivity between groups revealed that the application of PINT decreased the number of hypo-connected regions in ASD.\n\nConclusionsOur results provide a new framework for measuring altered brain functioning in neurodevelopmental disorders that may have implications for tracking developmental course, phenotypic heterogeneity, and ultimately treatment response. We underscore the importance of accounting for individual variation in the study of complex brain disorders.

neuroscience↗