Dataset Information

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

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

SUBMITTER: Ghiassian S

PROVIDER: S-EPMC5193362 | biostudies-literature | 2016

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

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

Ghiassian Sina S Greiner Russell R Jin Ping P Brown Matthew R G MR

PloS one 20161228 12

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 the goal of automated diagnosis, we propose an approach for automated classification of ADHD and autism based on histogram of oriented gradients (HOG) features extracted from MR brain images, as well as personal characteristic data features. We describe a learning algorithm that can ...[more]

PMID: 28030565

Similar Datasets

Project description:BackgroundMorton neuroma is a common cause of forefoot pain and sensory disturbances, but it is difficult to identify on magnetic resonance imaging (MRI). The aim of this study was to verify the usefulness of a characteristic MRI finding (slug sign) for identifying Morton neuroma and to clarify the relationship between excised neuroma characteristics and preoperative MRI findings.MethodsTwenty-two web spaces were retrospectively assessed from the second and third intermetatarsal spaces of 11 feet of 10 patients (7 women and 3 men, aged average 59.5 years) who underwent surgical excision of Morton neuroma between 2017 and 2022. Asymptomatic web spaces were used as control. Neuromas with 2 branches of the plantar digital nerves on axial T1-weighted MRI (MRI-T1WI) were considered the slug sign. We investigated the preoperative presence of the slug sign in Morton neuroma and asymptomatic control web spaces. We also investigated the relationship between the maximum transverse diameter of the excised specimen and that estimated on coronal MRI-T1WI.ResultsA total of 15 Morton neuromas were excised and assessed. The slug signs were present in 10 intermetatarsal spaces in 15 web spaces with Morton neuroma whereas the sign was found in 1 intermetatarsal space in 7 asymptomatic web spaces. The sensitivity and specificity for the slug sign to diagnose Morton neuroma was 66.7% and 85.7%, respectively. The positive and negative predictive values were 90.9% and 54.5%, respectively. The mean maximum transverse diameter of excised neuromas was 4.7 mm. The mean maximum transverse diameter of neuromas on coronal MRI-T1WI was 3.4 mm. A significant positive correlation was found between the maximum transverse diameters of excised specimens and diameters estimated on coronal MRI-T1WI (r = 0.799, P < .001).ConclusionThe slug sign may be a useful indicator of Morton neuroma on MRI to confirm nerve involvement after bifurcation.Level of evidenceLevel IV, retrospective series.

Project description:Group differences in resting state functional magnetic resonance imaging connectivity between individuals with autism and typically developing controls have been widely replicated for a small number of discrete brain regions, yet the whole-brain distribution of connectivity abnormalities in autism is not well characterized. It is also unclear whether functional connectivity is sufficiently robust to be used as a diagnostic or prognostic metric in individual patients with autism. We obtained pairwise functional connectivity measurements from a lattice of 7266 regions of interest covering the entire grey matter (26.4 million connections) in a well-characterized set of 40 male adolescents and young adults with autism and 40 age-, sex- and IQ-matched typically developing subjects. A single resting state blood oxygen level-dependent scan of 8 min was used for the classification in each subject. A leave-one-out classifier successfully distinguished autism from control subjects with 83% sensitivity and 75% specificity for a total accuracy of 79% (P = 1.1 × 10(-7)). In subjects <20 years of age, the classifier performed at 89% accuracy (P = 5.4 × 10(-7)). In a replication dataset consisting of 21 individuals from six families with both affected and unaffected siblings, the classifier performed at 71% accuracy (91% accuracy for subjects <20 years of age). Classification scores in subjects with autism were significantly correlated with the Social Responsiveness Scale (P = 0.05), verbal IQ (P = 0.02) and the Autism Diagnostic Observation Schedule-Generic's combined social and communication subscores (P = 0.05). An analysis of informative connections demonstrated that region of interest pairs with strongest correlation values were most abnormal in autism. Negatively correlated region of interest pairs showed higher correlation in autism (less anticorrelation), possibly representing weaker inhibitory connections, particularly for long connections (Euclidean distance >10 cm). Brain regions showing greatest differences included regions of the default mode network, superior parietal lobule, fusiform gyrus and anterior insula. Overall, classification accuracy was better for younger subjects, with differences between autism and control subjects diminishing after 19 years of age. Classification scores of unaffected siblings of individuals with autism were more similar to those of the control subjects than to those of the subjects with autism. These findings indicate feasibility of a functional connectivity magnetic resonance imaging diagnostic assay for autism.

Project description:BackgroundThe aim of the study was to perform a feature tracking analysis on cine magnetic resonance (MR) images to elucidate if functional measurements of the motion of the left ventricular wall may detect scar defined with gadolinium enhanced MR.Myocardial contraction can be measured in terms of the velocity, displacement and local deformation (strain) of a particular myocardial segment. Contraction of the myocardial wall will be reduced in the presence of scar and as a consequence of reduced myocardial blood flow.MethodsThirty patients (3 women and 27 men) were selected based on the presence or absence of extensive scar in the anteroseptal area of the left ventricle. The patients were investigated in stable clinical condition, 4-8 weeks post ST-elevation myocardial infarction treated with percutaneous coronary intervention. Seventeen had a scar area >75% in at least one anteroseptal segment (scar) and thirteen had scar area <1% (non-scar). Velocity, displacement and strain were calculated in the longitudinal direction, tangential to the endocardial outline, and in the radial direction, perpendicular to the tangent.ResultsIn the scar patients, segments with scar showed lower functional measurements than remote segments. Radial measurements of velocity, displacement and strain performed better in terms of receiver-operator-characteristic curves (ROC) than the corresponding longitudinal measurements. The best area-under-curve was for radial strain, 0.89, where a cut-off value of 38.8% had 80% sensitivity and 86% specificity for the detection of a segment with scar area >50%. As a percentage of the mean, intraobserver variability was 16-14-26% for radial measurements of displacement-velocity-strain and corresponding interobserver variability was 13-12-18%.ConclusionFeature tracking analysis of cine-MR displays velocity, displacement and strain in the radial and longitudinal direction and may be used for the detection of transmural scar. The accuracy and repeatability of the radial functional measurements is satisfactory and global measures agree.

Project description:BackgroundWhile autism and attention-deficit/hyperactivity disorder (ADHD) are considered distinct conditions from a diagnostic perspective, clinically they share some phenotypic features and have high comorbidity. Regardless, most studies have focused on only one condition, with considerable heterogeneity in their results. Taking a dual-condition approach might help elucidate shared and distinct neural characteristics.MethodGraph theory was used to analyse topological properties of structural covariance networks across both conditions and relative to a neurotypical (NT; n = 87) group using data from the ABIDE (autism; n = 62) and ADHD-200 datasets (ADHD; n = 69). Regional cortical thickness was used to construct the structural covariance networks. This was analysed in a theoretical framework examining potential differences in long and short-range connectivity, with a specific focus on relation between central graph measures and cortical thickness.ResultsWe found convergence between autism and ADHD, where both conditions show an overall decrease in CT covariance with increased Euclidean distance between centroids compared with a NT population. The 2 conditions also show divergence. Namely, there is less modular overlap between the 2 conditions than there is between each condition and the NT group. The ADHD group also showed reduced cortical thickness and lower degree in hub regions than the autism group. Lastly, the ADHD group also showed reduced wiring costs compared with the autism groups.ConclusionsOur results indicate a need for taking an integrated approach when considering highly comorbid conditions such as autism and ADHD. Furthermore, autism and ADHD both showed alterations in the relation between inter-regional covariance and centroid distance, where both groups show a steeper decline in covariance as a function of distance. The 2 groups also diverge on modular organization, cortical thickness of hub regions and wiring cost of the covariance network. Thus, on some network features the groups are distinct, yet on others there is convergence.

Dataset Information

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

Publications

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

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets