Using Functional or Structural Magnetic Resonance Images and Personal Characteristic Data to Identify ADHD and Autism

摘要: A clinical tool that can diagnose psychiatric illness using functional or structural magnetic resonance (MR) brain images has the potential to greatly assist physicians and improve treatment efficacy. Working toward goal of automated diagnosis, we propose an approach for classification ADHD autism based on histogram oriented gradients (HOG) features extracted from MR images, as well personal characteristic data features. We describe a learning algorithm produce effective classifiers when run two large public datasets. The is able distinguish control with hold-out accuracy 69.6% (over baseline 55.0%) characteristics scan trained ADHD-200 dataset (769 participants in training set, 171 test set). It 65.0% 51.6%) Autism Brain Imaging Data Exchange (ABIDE) (889 222 These results outperform all previously presented methods both To our knowledge, this first demonstration single process distinguishing patients vs. controls imaging above-chance datasets different illnesses (ADHD autism). applications requires robustness against real-world conditions, including substantial variability often exists among collected at institutions. therefore important was successful ABIDE datasets, which include hundreds multiple While resulting are not yet clinically relevant, work shows there signal (f)MRI find. anticipate will lead more accurate classifiers, over these other disorders, working high differential diagnosis.