Biology subjects

Dannlowski, U.

Publications and source records attributed to Dannlowski, U..

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↗

bioRxiv2018

Shared vulnerability for connectome alterations across psychiatric and neurological brain disorders

Macroscale white matter pathways form the infrastructure for large-scale communication in the human brain, a prerequisite for healthy brain function. Conversely, disruptions in the brains connectivity architecture are thought to play an important role in a wide range of psychiatric and neurological brain disorders. Here we show that especially connections important for global communication and network integration are involved in a wide range of brain disorders. We report on a meta-analytic connectome study comprising in total 895 patients and 1,016 controls across twelve neurological and psychiatric disorders. We extracted disorder connectome fingerprints for each of these twelve disorders, which were then combined into a cross-disorder disconnectivity involvement map, representing the involvement of each brain pathway across brain disorders. Our findings show connections central to the brains infrastructure are disproportionally involved across a wide range of disorders. Connections critical for global network communication and integration display high disturbance across disorders, suggesting a general cross-disorder involvement and importance of these pathways in normal function. Taken together, our cross-disorder study suggests a convergence of disconnectivity across disorders to a partially shared disconnectivity substrate of central connections.

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↗