Search

Background: Major depressive disorder (MDD) is a common and burdensome condition that has low rates of treatment success for each individual treatment. This means that many patients require several medication switches to achieve remission; selecting an effective antidepressant is typically a sequential trial-and-error process. Machine learning techniques may be able to learn models that can predict whether a specific patient will respond to a given treatment, before it is administered. This study uses baseline clinical data to create a machine-learned model that accurately predicts remission status for a patient after desvenlafaxine (DVS) treatment., Methods: We applied machine learning algorithms to data from 3,399 MDD patients (90% of the 3,776 subjects in 11 phase-III/IV clinical trials, each described using 92 features), to produce a model that uses 26 of these features to predict symptom remission, defined as an 8-week Hamilton Depression Rating Scale score of 7 or below. We evaluated that learned model on the remaining held-out 10% of the data ( n  = 377)., Results: Our resulting classifier, a trained linear support vector machine, had a holdout set accuracy of 69.0%, significantly greater than the probability of classifying a patient correctly by chance. We demonstrate that this learning process is stable by repeatedly sampling part of the training dataset and running the learner on this sample, then evaluating the learned model on the held-out instances of the training set; these runs had an average accuracy of 67.0% ± 1.8%., Conclusions: Our model, based on 26 clinical features, proved sufficient to predict DVS remission significantly better than chance. This may allow more accurate use of DVS without waiting 8 weeks to determine treatment outcome, and may serve as a first step toward changing psychiatric care by incorporating clinical assistive technologies using machine-learned models.

Background: Neurocognitive impairments are frequently observed in schizophrenia and major depressive disorder (MDD). However, it remains unclear whether reported neurocognitive abnormalities could objectively identify an individual as having schizophrenia or MDD., Methods: The current study included 220 first-episode patients with schizophrenia, 110 patients with MDD and 240 demographically matched healthy controls (HC). All participants performed the short version of the Wechsler Adult Intelligence Scale-Revised in China; the immediate and delayed logical memory of the Wechsler Memory Scale-Revised in China; and seven tests from the computerized Cambridge Neurocognitive Test Automated Battery to evaluate neurocognitive performance. The three-class AdaBoost tree-based ensemble algorithm was employed to identify neurocognitive endophenotypes that may distinguish between subjects in the categories of schizophrenia, depression and HC. Hierarchical cluster analysis was applied to further explore the neurocognitive patterns in each group., Results: The AdaBoost algorithm identified individual's diagnostic class with an average accuracy of 77.73% (80.81% for schizophrenia, 53.49% for depression and 86.21% for HC). The average area under ROC curve was 0.92 (0.96 in schizophrenia, 0.86 in depression and 0.92 in HC). Hierarchical cluster analysis revealed for MDD and schizophrenia, convergent altered neurocognition patterns related to shifting, sustained attention, planning, working memory and visual memory. Divergent neurocognition patterns for MDD and schizophrenia related to motor speed, general intelligence, perceptual sensitivity and reversal learning were identified., Conclusions: Neurocognitive abnormalities could predict whether the individual has schizophrenia, depression or neither with relatively high accuracy. Additionally, the neurocognitive features showed promise as endophenotypes for discriminating between schizophrenia and depression., (Copyright © 2017 Elsevier B.V. All rights reserved.)

Purpose: Evaluate brain iron accumulation in alcohol use disorder (AUD) patients compared to controls using quantitative susceptibility mapping (QSM)., Methods: QSM was performed retrospectively by using phase images from resting state functional magnetic resonance imaging (fMRI). 20 male AUD patients and 15 matched healthy controls were examined. Susceptibility values were manually traced in deep grey matter regions including caudate nucleus, combined putamen and globus pallidus, combined substantia nigra and red nucleus, dentate nucleus, and a reference white matter region in the internal capsule. Average susceptibility values from each region were compared between the patients and controls. The relationship between age and susceptibility was also explored., Results: The AUD group exhibited increased susceptibility in caudate nucleus (+8.5%, p=0.034), combined putamen and globus pallidus (+10.8%, p=0.006), and dentate nucleus (+14.9%, p=0.022). Susceptibility increased with age in two of the four measured regions - combined putamen and globus pallidus (p=0.013) and combined substantia nigra and red nucleus (p=0.041). AUD did not significantly modulate the rate of susceptibility increase with age in our data., Conclusion: Retrospective QSM computed from standard fMRI datasets provides new opportunities for brain iron studies in psychiatry. Substantially elevated brain iron was found in AUD subjects in the basal ganglia and dentate nucleus. This was the first human AUD brain iron study and the first retrospective clinical fMRI QSM study., (Copyright © 2017. Published by Elsevier Inc.)

Background: Growing evidence documents the potential of machine learning for developing brain based diagnostic methods for major depressive disorder (MDD). As symptom severity may influence brain activity, we investigated whether the severity of MDD affected the accuracies of machine learned MDD-vs-Control diagnostic classifiers., Methods: Forty-five medication-free patients with DSM-IV defined MDD and 19 healthy controls participated in the study. Based on depression severity as determined by the Hamilton Rating Scale for Depression (HRSD), MDD patients were sorted into three groups: mild to moderate depression (HRSD 14-19), severe depression (HRSD 20-23), and very severe depression (HRSD ≥ 24). We collected functional magnetic resonance imaging (fMRI) data during both resting-state and an emotional-face matching task. Patients in each of the three severity groups were compared against controls in separate analyses, using either the resting-state or task-based fMRI data. We use each of these six datasets with linear support vector machine (SVM) binary classifiers for identifying individuals as patients or controls., Results: The resting-state fMRI data showed statistically significant classification accuracy only for the very severe depression group (accuracy 66%, p = 0.012 corrected), while mild to moderate (accuracy 58%, p = 1.0 corrected) and severe depression (accuracy 52%, p = 1.0 corrected) were only at chance. With task-based fMRI data, the automated classifier performed at chance in all three severity groups., Conclusions: Binary linear SVM classifiers achieved significant classification of very severe depression with resting-state fMRI, but the contribution of brain measurements may have limited potential in differentiating patients with less severe depression from healthy controls.

High risk behaviors such as narcotic use or physical fighting can be caused by impulsive decision making in emotionally-charged situations. Improved neuroscientific understanding of how emotional context interacts with the control of impulsive behaviors may lead to advances in public policy and/or treatment approaches for high risk groups, including some high-risk adolescents or adults with poor impulse control. Inferior frontal gyrus (IFG) is an important contributor to response inhibition (behavioral impulse control). IFG also has a role in processing emotional stimuli and regulating emotional responses. The mechanism(s) whereby response inhibition processes interact with emotion processing in IFG are poorly understood. We used 4.7 T fMRI in 20 healthy young adults performing a rapid event-related emotional Go/NoGo task. This task combined the Go/NoGo task, which is a classic means of recruiting response inhibition processes, with emotionally neutral and aversive distractor images. In IFG, both response inhibition in an emotionally neutral context (neutral NoGo trials) and aversive emotional picture processing (aversive Go trials) evoked activation greater than the simple response baseline (neutral Go trials). These results are consistent with the literature. Activation for response inhibition in aversive contexts (aversive NoGo-neutral Go trials) was approximately the sum of response inhibition activation (neutral NoGo-neutral Go) and aversive emotional distractor activation (aversive Go-neutral Go). We conclude that response inhibition and aversive emotional stimulus processing activities combine additively (linearly) in IFG, rather than interfering with each other (sub-linearly) or mutually-enhancing each other (super-linearly). We also found previously undocumented interaction effects between response inhibition (NoGo vs. Go) and emotional context (aversive vs. neutral distractor pictures) in bilateral posterior middle temporal gyrus and angular gyrus, right frontal eye field, and other brain regions. These results may reflect the interaction of attention processes driven by emotional stimuli with conflict resolution processes related to Go/NoGo performance., (Copyright © 2012 Elsevier Inc. All rights reserved.)

Previous functional magnetic resonance imaging (fMRI) studies have compared saccade trials and nogo trials, which required subjects to look at peripheral visual stimuli and to inhibit automatic saccades evoked by peripheral stimuli, respectively. These studies surprisingly reported no activation differences in cortical saccade regions between the two tasks, despite their opposite response requirements. Here, we re-examined this issue by comparing saccades and nogo trials using a rapid event-related fMRI design in which saccade trials were presented twice as frequently as nogo trials to make the saccade response more prepotent. We hypothesized that this should increase recruitment of response inhibition processes in the nogo task, thereby increasing fMRI activation on nogo trials. Saccade and nogo trials were presented in whole and half trial versions. Whereas whole trials included a trial type instruction followed by peripheral stimulus presentation and subject response, half trials included only the instruction component, allowing us to measure instruction-related activation separately from response-evoked signals for both saccades and nogo trials. Instruction-related activation was greater for nogo versus saccade trials in right frontal eye field, middle frontal gyrus, intraparietal sulcus, and precuneus, which we attribute to a mixture of preparatory and task switching processes. Response-related activation was greater for nogo trials in supplementary eye field, anterior cingulate cortex, inferior frontal gyrus, and right supramarginal gyrus, and we attribute these results to saccade inhibition and other processes associated with an increased requirement for inhibition of the automatic saccade in nogo trials.

Flexible, adaptive behavior often requires the inhibition of automatic responses in favor of voluntary response generation. The antisaccade task requires active inhibition of the automatic saccade to a peripheral stimulus followed by generation of a voluntary antisaccade to the opposite location. Previous studies demonstrated greater functional magnetic resonance imaging (fMRI) activation for antisaccades than prosaccades in cortical saccade areas but did not distinguish the relative contributions of saccadic inhibition and generation. To address this question, we compared prosaccades, antisaccades, and nogo trials in a rapid event-related fMRI experiment with ten human subjects (6 female, 4 male). Trials were compound, containing a colored fixation point to cue trial type, followed by a 200-ms gap and then peripheral stimulus presentation and response. Required responses for prosaccade, antisaccade, and nogo trials, respectively, were to generate a saccade to the stimulus, look away from it, and inhibit the automatic saccade while maintaining central fixation. Frontal and supplementary eye fields, anterior cingulate cortex, intraparietal sulcus, and precuneus, exhibited surprisingly similar activation patterns for prosaccade and nogo responses, suggesting that BOLD signal in cortical saccade regions might predominantly reflect visual detection and attention processes rather than saccade generation or inhibition. These regions displayed greater activation for antisaccade responses versus prosaccade or nogo responses, possibly due to visuospatial remapping or increased attention levels in the antisaccade task. In right superior frontal sulcus, right supramarginal gyrus, and posterior cingulate sulcus, activation was greater for nogo compared to prosaccade responses, suggesting a role in active saccadic inhibition.