Studying depression using imaging and machine learning methods

Do withdrawal symptoms predict depression relapse after antidepressant cessation?

Discontinuing antidepressants after remission poses risks of withdrawal symptoms and relapse. This study addressed four questions: Can withdrawal symptoms be differentiated from relapse? What are their risk factors? Are withdrawal symptoms associated with relapse? Can discontinuation be optimized? 103 patients with a remitted major depressive disorder were randomized to continue or discontinue antidepressants. Withdrawal symptoms were assessed using the Discontinuation Emergent Signs and Symptoms scale (DESS). Withdrawal syndrome was defined as experiencing at least four new or worsened DESS symptoms. Associations between clinical factors and symptom count were examined using linear regressions. After the randomization phase, all patients discontinued treatment and were monitored for six months. The relationship between withdrawal symptoms, clinical factors, and relapse risk was analyzed via logistic regression and a Cox proportional hazards model. Ten symptoms were reported exclusively in the discontinuation group and may aid in distinguishing withdrawal syndrome from relapse. Withdrawal syndrome occurred in 29% (95% PI [8.3%, 72%]) of patients. Women reported more withdrawal symptoms than men (factor 1.67 (95% PI [1.06, 2.56])). None of the other predictors were associated with symptom count. Of 83 patients with outcome data, 54 (65%) remained well and 29 (35%) relapsed. Withdrawal symptoms (0.58, 95% PI [0.07, 1.16]) and early depressive symptoms (0.63, 95% PI [0.16, 1.17]) were associated with a higher relapse risk. Tapering duration was not associated with either withdrawal symptoms or relapse rate. Withdrawal symptoms were common and more frequent in women. Experiencing withdrawal symptoms may increase relapse risk.

Hippocampal Subregions Volume and Texture for the Diagnosis of Mild Cognitive Impairment

The aim was to examine the diagnostic efficacy of hippocampal subregions volume and texture in differentiating amnestic mild cognitive impairment (MCI) from normal aging changes. Ninety MCI subjects and eighty-eight well-matched healthy controls (HCs) were selected. Twelve hippocampal subregions volume and texture features were extracted using Freesurfer and MaZda based on T1 weighted MRI. Then, two-sample t-test and Least Absolute Shrinkage and Selection Operator (LASSO) regression were developed to select a subset of the original features. Support vector machine (SVM) was used to perform the classification task and the area under the curve (AUC), sensitivity, specificity and accuracy were calculated to evaluate the diagnostic efficacy of the model. The volume features with high discriminative power were mainly located in the bilateral CA1 and CA4, while texture feature were gray-level non-uniformity, run length non-uniformity and fraction. Our model based on hippocampal subregions volume and texture features achieved better classification performance with an AUC of 0.90. The volume and texture of hippocampal subregions can be utilized for the diagnosis of MCI. Moreover, we found that the features that contributed most to the model were mainly textural features, followed by volume. These results may guide future studies using structural scans to classify patients with MCI.

An objective quantitative diagnosis of depression using a local-to-global multimodal fusion graph neural network

This study developed an artificial intelligence (AI) system using a local-global multimodal fusion graph neural network (LGMF-GNN) to address the challenge of diagnosing major depressive disorder (MDD), a complex disease influenced by social, psychological, and biological factors. Utilizing functional MRI, structural MRI, and electronic health records, the system offers an objective diagnostic method by integrating individual brain regions and population data. Tested across cohorts from China, Japan, and Russia with 1,182 healthy controls and 1,260 MDD patients from 24 institutions, it achieved a classification accuracy of 78.75%, an area under the receiver operating characteristic curve (AUROC) of 80.64%, and correctly identified MDD subtypes. The system further discovered distinct brain connectivity patterns in MDD, including reduced functional connectivity between the left gyrus rectus and right cerebellar lobule VIIB, and increased connectivity between the left Rolandic operculum and right hippocampus. Anatomically, MDD is associated with thickness changes of the gray and white matter interface, indicating potential neuropathological conditions or brain injuries.

The Untapped Potential of Dimension Reduction in Neuroimaging: Artificial Intelligence-Driven Multimodal Analysis of Long COVID Fatigue

This perspective paper explores the untapped potential of artificial intelligence (AI), particularly machine learning-based dimension reduction techniques in multimodal neuroimaging analysis of Long COVID fatigue. The complexity and high dimensionality of neuroimaging data from modalities such as positron emission tomography (PET) and magnetic resonance imaging (MRI) pose significant analytical challenges. Deep neural networks and other machine learning approaches offer powerful tools for managing this complexity and extracting meaningful patterns. The paper discusses current challenges in neuroimaging data analysis, reviews state-of-the-art AI approaches for dimension reduction and multimodal integration, and examines their potential applications in Long COVID research. Key areas of focus include the development of AI-based biomarkers, AI-informed treatment strategies, and personalized medicine approaches. The authors argue that AI-driven multimodal neuroimaging analysis represents a paradigm shift in studying complex brain disorders like Long COVID. While acknowledging technical and ethical challenges, the paper emphasizes the potential of these advanced techniques to uncover new insights into the condition, which might lead to improved diagnostic and therapeutic strategies for those affected by Long COVID fatigue. The broader implications for understanding and treating other complex neurological and psychiatric conditions are also discussed.

Resting-state functional magnetic resonance imaging and support vector machines for the diagnosis of major depressive disorder in adolescents

BACKGROUND

Research has found that the amygdala plays a significant role in underlying pathology of major depressive disorder (MDD). However, few studies have explored machine learning-assisted diagnostic biomarkers based on amygdala functional connectivity (FC).

AIM

To investigate the analysis of neuroimaging biomarkers as a streamlined approach for the diagnosis of MDD in adolescents.

METHODS

Forty-four adolescents diagnosed with MDD and 43 healthy controls were enrolled in the study. Using resting-state functional magnetic resonance imaging, the FC was compared between the adolescents with MDD and the healthy controls, with the bilateral amygdala serving as the seed point, followed by statistical analysis of the results. The support vector machine (SVM) method was then applied to classify functional connections in various brain regions and to evaluate the neurophysiological characteristics associated with MDD.

RESULTS

Compared to the controls and using the bilateral amygdala as the region of interest, patients with MDD showed significantly lower FC values in the left inferior temporal gyrus, bilateral calcarine, right lingual gyrus, and left superior occipital gyrus. However, there was an increase in the FC value in Vermis-10. The SVM analysis revealed that the reduction in the FC value in the right lingual gyrus could effectively differentiate patients with MDD from healthy controls, achieving a diagnostic accuracy of 83.91%, sensitivity of 79.55%, specificity of 88.37%, and an area under the curve of 67.65%.

CONCLUSION

The results showed that an abnormal FC value in the right lingual gyrus was effective as a neuroimaging biomarker to distinguish patients with MDD from healthy controls.

Neural signatures of emotional biases predict clinical outcomes in difficult-to-treat depression

Background

Neural predictors underlying variability in depression outcomes are poorly understood. Functional MRI measures of subgenual cortex connectivity, self-blaming and negative perceptual biases have shown prognostic potential in treatment-naïve, medication-free and fully remitting forms of major depressive disorder (MDD). However, their role in more chronic, difficult-to-treat forms of MDD is unknown.

Methods

Forty-five participants (n = 38 meeting minimum data quality thresholds) fulfilled criteria for difficult-to-treat MDD. Clinical outcome was determined by computing percentage change at follow-up from baseline (four months) on the self-reported Quick Inventory of Depressive Symptomatology (16-item). Baseline measures included self-blame-selective connectivity of the right superior anterior temporal lobe with an a priori Brodmann Area 25 region-of-interest, blood-oxygen-level-dependent a priori bilateral amygdala activation for subliminal sad vs happy faces, and resting-state connectivity of the subgenual cortex with an a priori defined ventrolateral prefrontal cortex/insula region-of-interest.

Findings

A linear regression model showed that baseline severity of depressive symptoms explained 3% of the variance in outcomes at follow-up (F[3,34] = .33, p = .81). In contrast, our three pre-registered neural measures combined, explained 32% of the variance in clinical outcomes (F[4,33] = 3.86, p = .01).

Conclusion

These findings corroborate the pathophysiological relevance of neural signatures of emotional biases and their potential as predictors of outcomes in difficult-to-treat depression.

Evaluating the Clinical Validity and Reliability of Artificial Intelligence-Enabled Diagnostic Tools in Neuropsychiatric Disorders

Neuropsychiatric disorders (NPDs) pose a substantial burden on the healthcare system. The major challenge in diagnosing NPDs is the subjective assessment by the physician which can lead to inaccurate and delayed diagnosis. Recent studies have depicted that the integration of artificial intelligence (AI) in neuropsychiatry could potentially revolutionize the field by precisely diagnosing complex neurological and mental health disorders in a timely fashion and providing individualized management strategies. In this narrative review, the authors have examined the current status of AI tools in assessing neuropsychiatric disorders and evaluated their validity and reliability in the existing literature. The analysis of various datasets including MRI scans, EEG, facial expressions, social media posts, texts, and laboratory samples in the accurate diagnosis of neuropsychiatric conditions using machine learning has been profoundly explored in this article. The recent trials and tribulations in various neuropsychiatric disorders encouraging future scope in the utility and application of AI have been discussed. Overall machine learning has proved to be feasible and applicable in the field of neuropsychiatry and it is about time that research translates to clinical settings for favorable patient outcomes. Future trials should focus on presenting higher quality evidence for superior adaptability and establish guidelines for healthcare providers to maintain standards.

MDDBranchNet: A Deep Learning Model for Detecting Major Depressive Disorder Using ECG Signal

Major depressive disorder (MDD) is a chronic mental illness which affects people's well-being and is often detected at a later stage of depression with a likelihood of suicidal ideation. Early detection of MDD is thus necessary to reduce the impact, however, it requires monitoring vitals in daily living conditions. EEG is generally multi-channel and due to difficulty in signal acquisition, it is unsuitable for home-based monitoring, whereas, wearable sensors can collect single-channel ECG. Classical machine-learning based MDD detection studies commonly use various heart rate variability features. Feature generation, which requires domain knowledge, is often challenging, and requires computation power, often unsuitable for real time processing, MDDBranchNet is a proposed parallel-branch deep learning model for MDD binary classification from a single channel ECG which uses additional ECG-derived signals such as R-R signal and degree distribution time series of horizontal visibility graph. The use of derived branches was able to increase the model's accuracy by around 7%. An optimal 20-second overlapped segmentation of ECG recording was found to be beneficial with a 70% prediction threshold for maximum MDD detection with a minimum false positive rate. The proposed model evaluated MDD prediction from signal excerpts, irrespective of location (first, middle or last one-third of the recording), instead of considering the entire ECG signal with minimal performance variation stressing the idea that MDD phenomena are likely to manifest uniformly throughout the recording.

One size does not fit all: notable individual variation in brain activity correlates of antidepressant treatment response

Introduction

To date, no robust electroencephalography (EEG) markers of antidepressant treatment response have been identified. Variable findings may arise from the use of group analyses, which neglect individual variation. Using a combination of group and single-participant analyses, we explored individual variability in EEG characteristics of treatment response.

Methods

Resting-state EEG data and Montgomery-Åsberg Depression Rating Scale (MADRS) symptom scores were collected from 43 patients with depression before, at 1 and 12 weeks of pharmacotherapy. Partial least squares (PLS) was used to: 1) identify group differences in EEG connectivity (weighted phase lag index) and complexity (multiscale entropy) between eventual medication responders and non-responders, and 2) determine whether group patterns could be identified in individual patients.

Results

Responders showed decreased alpha and increased beta connectivity, and early, widespread decreases in complexity over treatment. Non-responders showed an opposite connectivity pattern, and later, spatially confined decreases in complexity. Thus, as in previous studies, our group analyses identified significant differences between groups of patients with different treatment outcomes. These group-level EEG characteristics were only identified in ~40-60% of individual patients, as assessed quantitatively by correlating the spatiotemporal brain patterns between groups and individual results, and by independent raters through visualization.

Discussion

Our single-participant analyses suggest that substantial individual variation exists, and needs to be considered when investigating characteristics of antidepressant treatment response for potential clinical applicability.

Clinical trial registration

https://clinicaltrials.gov, identifier NCT00519428.

Identification of central symptoms of children depression and development of two short version of Children's Depression Inventory: Based on network analysis and machine learning

BACKGROUND

Using network analysis to study the central symptoms is important for understanding the mechanism of depression symptoms and selecting items for the short version depression screening scale. This study aimed to identify the central symptoms of depression and develop the short and effective depression screening tools for Chinese rural children.

METHODS

Firstly, the 2458 individuals (Mage = 10.74; SDage = 1.64; 51.2 % were female) were recruited from the rural children's mental health database. Children's Depression Inventory (CDI) was used to assess depression symptoms. Then, network analysis was used to identify the central symptoms of depression. The accuracy, stability, and gender invariance of the depression symptoms network were tested. Finally, a short version of CDI with central symptoms (CDI-SC) and a new CDI-10 (CDI-10-N) were developed by network analysis and feature selection techniques to optimize the existing CDI-10. Their performances in screening depression symptoms were validated by the cutoff threshold and machine learning.

RESULTS

The central symptoms of Chinese rural children's depression were sadness, self-hatred, loneliness and self-deprecation. This result was accurate and stable and depression symptoms network has gender invariance. The AUC values of CDI-10-N and CDI-SC are over 0.9. The CDI-10-N has a higher AUC than CDI-10. The optimal cutoff thresholds for CDI-10-N and CDI-SC are 6 and 1. The performance of machine learning on AUC generally outperforms those of cutoff threshold.

CONCLUSIONS

The central symptoms identified in this study should be highlighted in screening depression symptoms, and CDI-10-N and CDI-SC are effective tools for screening depression symptoms in Chinese rural children.

Neuroanatomy of post-stroke depression: the association between symptom clusters and lesion location

Post-stroke depression affects about 30% of stroke patients and often hampers functional recovery. The diagnosis of depression encompasses heterogeneous symptoms at emotional, motivational, cognitive, behavioural or somatic levels. Evidence indicates that depression is caused by disruption of bio-aminergic fibre tracts between prefrontal and limbic or striatal brain regions comprising different functional networks. Voxel-based lesion-symptom mapping studies reported discrepant findings regarding the association between infarct locations and depression. Inconsistencies may be due to the usage of sum scores, thereby mixing different symptoms of depression. In this cross-sectional study, we used multivariate support vector regression for lesion-symptom mapping to identify regions significantly involved in distinct depressive symptom domains and global depression. MRI lesion data were included from 200 patients with acute first-ever ischaemic stroke (mean 0.9 ± 1.5 days of post-stroke). The Montgomery-Åsberg Depression Rating interview assessed depression severity in five symptom domains encompassing motivational, emotional and cognitive symptoms deficits, anxiety and somatic symptoms and was examined 8.4 days of post-stroke (±4.3). We found that global depression severity, irrespective of individual symptom domains, was primarily linked to right hemispheric lesions in the dorsolateral prefrontal cortex and inferior frontal gyrus. In contrast, when considering distinct symptom domains individually, the analyses yielded much more sensitive results in regions where the correlations with the global depression score yielded no effects. Accordingly, motivational deficits were associated with lesions in orbitofrontal cortex, dorsolateral prefrontal cortex, pre- and post-central gyri and basal ganglia, including putamen and pallidum. Lesions affecting the dorsal thalamus, anterior insula and somatosensory cortex were significantly associated with emotional symptoms such as sadness. Damage to the dorsolateral prefrontal cortex was associated with concentration deficits, cognitive symptoms of guilt and self-reproach. Furthermore, somatic symptoms, including loss of appetite and sleep disturbances, were linked to the insula, parietal operculum and amygdala lesions. Likewise, anxiety was associated with lesions impacting the central operculum, insula and inferior frontal gyrus. Interestingly, symptoms of anxiety were exclusively left hemispheric, whereas the lesion-symptom associations of the other domains were lateralized to the right hemisphere. In conclusion, this large-scale study shows that in acute stroke patients, differential post-stroke depression symptom domains are associated with specific structural correlates. Our findings extend existing concepts on the neural underpinnings of depressive symptoms, indicating that differential lesion patterns lead to distinct depressive symptoms in the first weeks of post-stroke. These findings may facilitate the development of personalized treatments to improve post-stroke rehabilitation.

Development and validation of a machine learning-based vocal predictive model for major depressive disorder

BACKGROUND

Variations in speech intonation are known to be associated with changes in mental state over time. Behavioral vocal analysis is an algorithmic method of determining individuals' behavioral and emotional characteristics from their vocal patterns. It can provide biomarkers for use in psychiatric assessment and monitoring, especially when remote assessment is needed, such as in the COVID-19 pandemic. The objective of this study was to design and validate an effective prototype of automatic speech analysis based on algorithms for classifying the speech features related to MDD using a remote assessment system combining a mobile app for speech recording and central cloud processing for the prosodic vocal patterns.

METHODS

Machine learning compared the vocal patterns of 40 patients diagnosed with MDD to the patterns of 104 non-clinical participants. The vocal patterns of 40 patients in the acute phase were also compared to 14 of these patients in the remission phase of MDD.

RESULTS

A vocal depression predictive model was successfully generated. The vocal depression scores of MDD patients were significantly higher than the scores of the non-patient participants (p < 0.0001). The vocal depression scores of the MDD patients in the acute phase were significantly higher than in remission (p < 0.02).

LIMITATIONS

The main limitation of this study is its relatively small sample size, since machine learning validity improves with big data.

CONCLUSIONS

The computerized analysis of prosodic changes may be used to generate biomarkers for the early detection of MDD, remote monitoring, and the evaluation of responses to treatment.

Classification of major depression disorder via using minimum spanning tree of individual high-order morphological brain network

BACKGROUND

Major depressive disorder (MDD) is an overbroad and heterogeneous diagnosis with no reliable or quantifiable markers. We aim to combine machine-learning techniques with the individual minimum spanning tree of the morphological brain network (MST-MBN) to determine whether the network properties can provide neuroimaging biomarkers to identify patients with MDD.

METHOD

Eight morphometric features of each region of interest (ROI) were extracted from 3D T1 structural images of 106 patients with MDD and 97 healthy controls. Six feature distances of the eight morphometric features were calculated to generate a feature distance matrix, which was defined as low-order MBN. Further linear correlations of feature distances between ROIs were calculated on the basis of low-order MBN to generate individual high-order MBN. The Kruskal's algorithm was used to generate the MST to obtain the core framework of individual low-order and high-order MBN. The regional and global properties of the individual MSTs were defined as the feature. The support vector machine and back-propagation neural network was used to diagnose MDD and assess its severity, respectively.

RESULT

The low-order and high-order MST-MBN constructed by cityblock distance had the excellent classification performance. The high-order MST-MBN significantly improved almost 20 % diagnostic accuracy compared with the low-order MST-MBN, and had a maximum R² value of 0.939 between the predictive and true Hamilton Depression Scale score. The different group-level connectivity strength mainly involves the central executive network and default mode network (no statistical significance after FDR correction).

CONCLUSION

We proposed an innovative individual high-order MST-MBN to capture the cortical high-order morphological correlation and make an excellent performance for individualized diagnosis and assessment of MDD.

Evaluating the utility of daily speech assessments for monitoring depression symptoms

Objective

Depression is a common mental health disorder and a major public health concern, significantly interfering with the lives of those affected. The complex clinical presentation of depression complicates symptom assessments. Day-to-day fluctuations of depression symptoms within an individual bring an additional barrier, since infrequent testing may not reveal symptom fluctuation. Digital measures such as speech can facilitate daily objective symptom evaluation. Here, we evaluated the effectiveness of daily speech assessment in characterizing speech fluctuations in the context of depression symptoms, which can be completed remotely, at a low cost and with relatively low administrative resources.

Methods

Community volunteers (N = 16) completed a daily speech assessment, using the Winterlight Speech App, and Patient Health Questionnaire-9 (PHQ-9) for 30 consecutive business days. We calculated 230 acoustic and 290 linguistic features from individual's speech and investigated their relationship to depression symptoms at the intra-individual level through repeated measures analyses.

Results

We observed that depression symptoms were linked to linguistic features, such as less frequent use of dominant and positive words. Greater depression symptomatology was also significantly correlated with acoustic features: reduced variability in speech intensity and increased jitter.

Conclusions

Our findings support the feasibility of using acoustic and linguistic features as a measure of depression symptoms and propose daily speech assessment as a tool for better characterization of symptom fluctuations.

Altered spatio-temporal state patterns for functional dynamics estimation in first-episode drug-naive major depression

Patients with major depressive disorder (MDD) display affective and cognitive impairments. Although MDD-associated abnormalities of brain function and structure have been explored in depth, the relationships between MDD and spatio-temporal large-scale functional networks have not been evaluated in large-sample datasets. We employed data from International Big-Data Center for Depression Research (IBCDR), and comparable 543 healthy controls (HC) and 314 first-episode drug-naive (FEDN) MDD patients were included. We used a multivariate pattern classification method to learn informative spatio-temporal functional states. Brain states of each participant were extracted for functional dynamic estimation using an independent component analysis. Then, a multi-kernel pattern classification method was developed to identify discriminative spatio-temporal states associated with FEDN MDD. Finally, statistical analysis was applied to intrinsic and clinical brain characteristics. Compared with HC, FEDN MDD patients exhibited altered spatio-temporal functional states of the default mode network (DMN), the salience network, a hub network (centered on the dorsolateral prefrontal cortex), and a relatively complex coupling network (visual, DMN, motor-somatosensory and subcortical networks). Multi-kernel classification models to distinguish patients from HC obtained areas under the receiver operating characteristic curves up to 0.80. Classification scores correlated with Hamilton Depression Rating Scale scores and age at MDD onset. FEDN MDD patients had multiple abnormal spatio-temporal functional states. Classification scores derived from these states were related to symptom severity. The assessment of spatio-temporal states may represent a powerful clinical and research tool to distinguish between neuropsychiatric patients and controls.

Predictive factors and nomogram to evaluate the risk of below-ankle re-amputation in patients with diabetic foot

BACKGROUND

Diabetes mellitus, as the most common metabolic disease, is common worldwide and represents a crucial global health concern. The purpose of this research was to investigate the related risk factors and to develop a re-amputation risk nomogram in diabetic patients who have undergone an amputation.

METHODS

A observational analysis was performed on 459 patients who have underwent amputation for diabetic foot from January 2014 through December 2019 at the First Affiliated Hospital of Wenzhou Medical University. The least absolute shrinkage and selection operator regression and stepwise regression methods were implemented to determine risk selection for the re-amputation risk model, and the predictive nomogram was established with these features. Calibration curve, receiver operating characteristic curve, and decision curve analysis of this re-amputation nomogram were assessed.

RESULTS

Predictors contained in this predictive model included smoking, glycated hemoglobin A1c (HbA1c), ankle-brachial index (ABI) and C-reactive protein (CRP). Good discrimination with a C-index of 0.725 (95% CI, 0.6624-0.7876) and good calibration were displayed with this predictive model. The decision curve analysis showed that this re-amputation nomogram predicting risk adds more benefit than none strategy if the threshold probability of a patient was >6% and <59%.

CONCLUSIONS

This novel re-amputation nomogram incorporating smoking, glycated hemoglobin A1c (HbA1c), ankle-brachial index (ABI), C-reactive protein (CRP), and smoking could be easily used to predict individual re-amputation risk prediction in diabetic foot patients who have undergone an amputation. In the future, further analysis and external testing will be needed as much as possible to reconfirm that this new Nomogram can accurately predict the risk of toe re-amputation.

Resting-state neural signal variability in women with depressive disorders

Aberrant activity and connectivity in default mode (DMN), frontoparietal (FPN), and salience (SN) network regions is well-documented in depression. Recent neuroimaging research suggests that altered variability in the blood oxygen level-dependent (BOLD) signal may disrupt normal network integration and be an important novel predictor of psychopathology. However, no studies have yet determined the relationship between resting-state BOLD signal variability and depressive disorders nor applied BOLD signal variability features to the classification of depression history using machine learning (ML). We collected resting-state fMRI data for 79 women with different depression histories: no history, past history, and current depressive disorder. We tested voxelwise differences in BOLD signal variability related to depression group and severity. We also investigated whether BOLD signal variability of DMN, FPN, and SN regions could predict depression history group using a supervised random forest ML model. Results indicated that individuals with any history of depression had significantly decreased BOLD signal variability in the left and right cerebellum and right parietal cortex (pFWE <0.05). Furthermore, greater depression severity was also associated with reduced BOLD signal variability in the cerebellum. A random forest model classified participant depression history with 74% accuracy, with the ventral anterior cingulate cortex of the DMN as the most important variable in the model. These findings provide novel support for resting-state BOLD signal variability as a marker of neural dysfunction in depression and implicate decreased neural signal variability in the pathophysiology of depression.

A deep tensor-based approach for automatic depression recognition from speech utterances

Depression is one of the significant mental health issues affecting all age groups globally. While it has been widely recognized to be one of the major disease burdens in populations, complexities in definitive diagnosis present a major challenge. Usually, trained psychologists utilize conventional methods including individualized interview assessment and manually administered PHQ-8 scoring. However, heterogeneity in symptomatic presentations, which span somatic to affective complaints, impart substantial subjectivity in its diagnosis. Diagnostic accuracy is further compounded by the cross-sectional nature of sporadic assessment methods during physician-office visits, especially since depressive symptoms/severity may evolve over time. With widespread acceptance of smart wearable devices and smartphones, passive monitoring of depression traits using behavioral signals such as speech presents a unique opportunity as companion diagnostics to assist the trained clinicians in objective assessment over time. Therefore, we propose a framework for automated depression classification leveraging alterations in speech patterns in the well documented and extensively studied DAIC-WOZ depression dataset. This novel tensor-based approach requires a substantially simpler implementation architecture and extracts discriminative features for depression recognition with high f1 score and accuracy. We posit that such algorithms, which use significantly less compute load would allow effective onboard deployment in wearables for improve diagnostics accuracy and real-time monitoring of depressive disorders.

Machine learning-based predictive modeling of depression in hypertensive populations

We aimed to develop prediction models for depression among U.S. adults with hypertension using various machine learning (ML) approaches. Moreover, we analyzed the mechanisms of the developed models. This cross-sectional study included 8,628 adults with hypertension (11.3% with depression) from the National Health and Nutrition Examination Survey (2011–2020). We selected several significant features using feature selection methods to build the models. Data imbalance was managed with random down-sampling. Six different ML classification methods implemented in the R package caret—artificial neural network, random forest, AdaBoost, stochastic gradient boosting, XGBoost, and support vector machine—were employed with 10-fold cross-validation for predictions. Model performance was assessed by examining the area under the receiver operating characteristic curve (AUC), accuracy, precision, sensitivity, specificity, and F1-score. For an interpretable algorithm, we used the variable importance evaluation function in caret. Of all classification models, artificial neural network trained with selected features (n = 30) achieved the highest AUC (0.813) and specificity (0.780) in predicting depression. Support vector machine predicted depression with the highest accuracy (0.771), precision (0.969), sensitivity (0.774), and F1-score (0.860). The most frequent and important features contributing to the models included the ratio of family income to poverty, triglyceride level, white blood cell count, age, sleep disorder status, the presence of arthritis, hemoglobin level, marital status, and education level. In conclusion, ML algorithms performed comparably in predicting depression among hypertensive populations. Furthermore, the developed models shed light on variables’ relative importance, paving the way for further clinical research.

Treatment Response Prediction for Major Depressive Disorder Patients via Multivariate Pattern Analysis of Thalamic Features

Antidepressant treatment, as an important method in clinical practice, is not suitable for all major depressive disorder (MDD) patients. Although magnetic resonance imaging (MRI) studies have found thalamic abnormalities in MDD patients, it is not clear whether the features of the thalamus are suitable to serve as predictive aids for treatment responses at the individual level. Here, we tested the predictive value of gray matter density (GMD), gray matter volume (GMV), amplitude of low-frequency fluctuations (ALFF), and fractional ALFF (fALFF) of the thalamus using multivariate pattern analysis (MVPA). A total of 74 MDD patients and 44 healthy control (HC) subjects were recruited. Thirty-nine MDD patients and 35 HC subjects underwent scanning twice. Between the two scanning sessions, patients in the MDD group received selective serotonin reuptake inhibitor (SSRI) treatment for 3-month, and HC group did not receive any treatment. Gaussian process regression (GPR) was trained to predict the percentage decrease in the Hamilton Depression Scale (HAMD) score after treatment. The percentage decrease in HAMD score after SSRI treatment was predicted by building GPRs trained with baseline thalamic data. The results showed significant correlations between the true percentage of HAMD score decreases and predictions (p < 0.01, r² = 0.11) in GPRs trained with GMD. We did not find significant correlations between the true percentage of HAMD score decreases and predictions in GMV (p = 0.16, r² = 0.00), ALFF (p = 0.125, r² = 0.00), and fALFF (p = 0.485, r² = 0.10). Our results suggest that GMD of the thalamus has good potential as an aid in individualized treatment response predictions of MDD patients.

Emotional faces processing in major depressive disorder and prediction of antidepressant treatment response: A NeuroPharm study

BACKGROUND

Major depressive disorder (MDD) is a prevalent neuropsychiatric illness for which it is important to resolve underlying brain mechanisms. Current treatments are often unsuccessful, precipitating a need to identify predictive markers.

AIM

We evaluated (1) alterations in brain responses to an emotional faces functional magnetic resonance imaging (fMRI) paradigm in individuals with MDD, compared to controls, (2) whether pretreatment brain responses predicted antidepressant treatment response, and (3) pre-post change in brain responses following treatment.

METHODS

Eighty-nine medication-free, depressed individuals and 115 healthy controls completed the fMRI paradigm. Depressed individuals completed a nonrandomized, open-label, 8-week treatment with escitalopram, including the option to switch to duloxetine after 4 weeks. We examined patient-control group differences in regional fMRI responses at baseline, whether baseline fMRI responses predicted treatment response at 8 weeks, including early life stress moderating effects, and change in fMRI responses in 36 depressed individuals rescanned following 8 weeks of treatment.

RESULTS

Task reaction time was 5% slower in patients. Multiple brain regions showed significant task-related responses, but we observed no statistically significant patient-control group differences (Cohen's d < 0.35). Patient pretreatment brain responses did not predict antidepressant treatment response (area under the curve of the receiver operator characteristic (AUC-ROC) < 0.6) and brain responses were not statistically significantly changed after treatment (Cohen's d < 0.33).

CONCLUSION

This represents the largest prediction study to date examining emotional faces fMRI features as predictors of antidepressant treatment response. Brain response to this fMRI emotional faces paradigm did not distinguish depressed individuals from healthy controls, nor was it predictive of antidepressant treatment response.Clinical Trial Registration: Site: https://clinicaltrials.gov, Trial Number: NCT02869035, Trial Title: Treatment Outcome in Major Depressive Disorder.

Machine Learning Algorithms for Depression: Diagnosis, Insights, and Research Directions

Over the years, stress, anxiety, and modern-day fast-paced lifestyles have had immense psychological effects on people’s minds worldwide. The global technological development in healthcare digitizes the scopious data, enabling the map of the various forms of human biology more accurately than traditional measuring techniques. Machine learning (ML) has been accredited as an efficient approach for analyzing the massive amount of data in the healthcare domain. ML methodologies are being utilized in mental health to predict the probabilities of mental disorders and, therefore, execute potential treatment outcomes. This review paper enlists different machine learning algorithms used to detect and diagnose depression. The ML-based depression detection algorithms are categorized into three classes, classification, deep learning, and ensemble. A general model for depression diagnosis involving data extraction, pre-processing, training ML classifier, detection classification, and performance evaluation is presented. Moreover, it presents an overview to identify the objectives and limitations of different research studies presented in the domain of depression detection. Furthermore, it discussed future research possibilities in the field of depression diagnosis.

Improving the diagnostic accuracy for major depressive disorder using machine learning algorithms integrating clinical and near-infrared spectroscopy data

BACKGROUND

Given that major depressive disorder (MDD) is both biologically and clinically heterogeneous, a diagnostic system integrating neurobiological markers and clinical characteristics would allow for better diagnostic accuracy and, consequently, treatment efficacy.

OBJECTIVE

Our study aimed to evaluate the discriminative and predictive ability of unimodal, bimodal, and multimodal approaches in a total of seven machine learning (ML) models-clinical, demographic, functional near-infrared spectroscopy (fNIRS), combinations of two unimodal models, as well as a combination of all three-for MDD.

METHODS

We recruited 65 adults with MDD and 68 matched healthy controls, who provided both sociodemographic and clinical information, and completed the HAM-D questionnaire. They were also subject to fNIRS measurement when participating in the verbal fluency task. Using the nested cross validation procedure, the classification performance of each ML model was evaluated based on the area under the receiver operating characteristic curve (ROC), balanced accuracy, sensitivity, and specificity.

RESULTS

The multimodal ML model was able to distinguish between depressed patients and healthy controls with the highest balanced accuracy of 87.98 ± 8.84% (AUC = 0.92; 95% CI (0.84-0.99) when compared with the uni- and bi-modal models.

CONCLUSIONS

Our multimodal ML model demonstrated the highest diagnostic accuracy for MDD. This reinforces the biological and clinical heterogeneity of MDD and highlights the potential of this model to improve MDD diagnosis rates. Furthermore, this model is cost-effective and clinically applicable enough to be established as a robust diagnostic system for MDD based on patients' biosignatures.

Longitudinal clustering analysis and prediction of Parkinson's disease progression using radiomics and hybrid machine learning

BACKGROUND

We employed machine learning approaches to (I) determine distinct progression trajectories in Parkinson's disease (PD) (unsupervised clustering task), and (II) predict progression trajectories (supervised prediction task), from early (years 0 and 1) data, making use of clinical and imaging features.

METHODS

We studied PD-subjects derived from longitudinal datasets (years 0, 1, 2 & 4; Parkinson's Progressive Marker Initiative). We extracted and analyzed 981 features, including motor, non-motor, and radiomics features extracted for each region-of-interest (ROIs: left/right caudate and putamen) using our standardized standardized environment for radiomics analysis (SERA) radiomics software. Segmentation of ROIs on dopamine transposer - single photon emission computed tomography (DAT SPECT) images were performed via magnetic resonance images (MRI). After performing cross-sectional clustering on 885 subjects (original dataset) to identify disease subtypes, we identified optimal longitudinal trajectories using hybrid machine learning systems (HMLS), including principal component analysis (PCA) + K-Means algorithms (KMA) followed by Bayesian information criterion (BIC), Calinski-Harabatz criterion (CHC), and elbow criterion (EC). Subsequently, prediction of the identified trajectories from early year data was performed using multiple HMLSs including 16 Dimension Reduction Algorithms (DRA) and 10 classification algorithms.

RESULTS

We identified 3 distinct progression trajectories. Hotelling's t squared test (HTST) showed that the identified trajectories were distinct. The trajectories included those with (I, II) disease escalation (2 trajectories, 27% and 38% of patients) and (III) stable disease (1 trajectory, 35% of patients). For trajectory prediction from early year data, HMLSs including the stochastic neighbor embedding algorithm (SNEA, as a DRA) as well as locally linear embedding algorithm (LLEA, as a DRA), linked with the new probabilistic neural network classifier (NPNNC, as a classifier), resulted in accuracies of 78.4% and 79.2% respectively, while other HMLSs such as SNEA + Lib_SVM (library for support vector machines) and t_SNE (t-distributed stochastic neighbor embedding) + NPNNC resulted in 76.5% and 76.1% respectively.

CONCLUSIONS

This study moves beyond cross-sectional PD subtyping to clustering of longitudinal disease trajectories. We conclude that combining medical information with SPECT-based radiomics features, and optimal utilization of HMLSs, can identify distinct disease trajectories in PD patients, and enable effective prediction of disease trajectories from early year data.

Biological Diagnosis of Depression: A Biomarker Panel from Several Nonspecial Indicators Instead of the Specific Biomarker(s)

It is a consensus that the diagnosis efficiency of depression is rather low in clinic. The traditional way of diagnosing depression by symptomatology is flawed. Recent years, a growing body of evidence has underlined the importance of physiological indicators in the diagnosis of depression. However, the diagnosis of depression is difficult to be like some common clinical diseases, which have clear physiological indicators. A single physiological index provides limited information to clinicians and is of little help in the diagnosis of depression. Thus, it is more rational and practical to diagnose depression with a biomarker panel, which covers a few non-specific indicators, such as hormones, cytokines, and neurotrophins. This open review suggested that biomarker panel had a bright future in creating a new model of depression diagnosis or at least providing a reference to the existing depression criteria. The viewpoint is also the future of other psychiatric diagnosis.

Gender differences in major depressive disorders: A resting state fMRI study

BACKGROUND

Major depressive disorder (MDD) has a high disability rate and brings a large disease burden to patients and the country. Significant sex differences exist in both the epidemiological and clinical features in MDD. The effect of sex on brain function in MDD is not clear now. Regional homogeneity (ReHo) and ALFF are widely used research method in the study of brain function. This research aimed to use ReHo and ALFF to explore gender differences in brain function images in MDD.

METHODS

Eighty first-episode drug-naive patients (47 women and 30 men) with MDD and 85 age, education matched healthy volunteers (47 women and 31 men) were recruited in our study and participated in resting-state functional magnetic resonance imaging scans. ReHo and ALFF were used to assess brain activity, two-way ANOVA and post hoc analysis was conducted to explore the sex difference in MDD. Correlation analysis was used to explore the relationship between abnormal brain functioning and clinical symptoms.

RESULTS

We observed sex-specific patterns and diagnostic differences in MDD Patients, further post hoc comparisons indicated that women with MDD showed decreased ALFF value in the right superior occipital gyrus and decreased ReHo value in the left calcarine and left dorsolateral superior frontal gyrus compared with HC females and men with MDD. Men with MDD showed decreased ReHo value in the right median cingulate gyrus compared with HC males and increased ReHo value in the left dorsolateral superior frontal gyrus compared with HC males, we also found that HC males showed higher ReHo value in the right median cingulate gyrus than HC females.

CONCLUSIONS

Men and women do have sex differences in brain function, the occipital lobe, calcarine, DLPFC, and DCG were the main different brain regions found between male and female in MDD, which may be the biomarker brain regions that can help diagnose and treat MDD in men and women.

Performances of whole-brain dynamic and static functional connectivity fingerprinting in machine learning-based classification of major depressive disorder

BACKGROUND

Alterations in static and dynamic functional connectivity during resting state have been widely reported in major depressive disorder (MDD). The objective of this study was to compare the performances of whole-brain dynamic and static functional connectivity combined with machine learning approach in differentiating MDD patients from healthy controls at the individual subject level. Given the dynamic nature of brain activity, we hypothesized that dynamic connectivity would outperform static connectivity in the classification.

METHODS

Seventy-one MDD patients and seventy-one well-matched healthy controls underwent resting-state functional magnetic resonance imaging scans. Whole-brain dynamic and static functional connectivity patterns were calculated and utilized as classification features. Linear kernel support vector machine was employed to design the classifier and a leave-one-out cross-validation strategy was used to assess classifier performance.

RESULTS

Experimental results of dynamic functional connectivity-based classification showed that MDD patients could be discriminated from healthy controls with an excellent accuracy of 100% irrespective of whether or not global signal regression (GSR) was performed (permutation test with P < 0.0002). Brain regions with the most discriminating dynamic connectivity were mainly and reliably located within the default mode network, cerebellum, and subcortical network. In contrast, the static functional connectivity-based classifiers exhibited unstable classification performances, i.e., a low accuracy of 38.0% without GSR (P = 0.9926) while a high accuracy of 96.5% with GSR (P < 0.0002); moreover, there was a considerable variability in the distribution of brain regions with static connectivity most informative for classification.

CONCLUSION

These findings suggest the superiority of dynamic functional connectivity in machine learning-based classification of depression, which may be helpful for a better understanding of the neural basis of MDD as well as for the development of effective computer-aided diagnosis tools in clinical settings.

Machine Learning and Deep Learning for the Pharmacogenomics of Antidepressant Treatments

A growing body of evidence now proposes that machine learning and deep learning techniques can serve as a vital foundation for the pharmacogenomics of antidepressant treatments in patients with major depressive disorder (MDD). In this review, we focus on the latest developments for pharmacogenomics research using machine learning and deep learning approaches together with neuroimaging and multi-omics data. First, we review relevant pharmacogenomics studies that leverage numerous machine learning and deep learning techniques to determine treatment prediction and potential biomarkers for antidepressant treatments in MDD. In addition, we depict some neuroimaging pharmacogenomics studies that utilize various machine learning approaches to predict antidepressant treatment outcomes in MDD based on the integration of research on pharmacogenomics and neuroimaging. Moreover, we summarize the limitations in regard to the past pharmacogenomics studies of antidepressant treatments in MDD. Finally, we outline a discussion of challenges and directions for future research. In light of latest advancements in neuroimaging and multi-omics, various genomic variants and biomarkers associated with antidepressant treatments in MDD are being identified in pharmacogenomics research by employing machine learning and deep learning algorithms.

Machine Learning Methods for Predicting Postpartum Depression: Scoping Review

BACKGROUND

Machine learning (ML) offers vigorous statistical and probabilistic techniques that can successfully predict certain clinical conditions using large volumes of data. A review of ML and big data research analytics in maternal depression is pertinent and timely, given the rapid technological developments in recent years.

OBJECTIVE

This study aims to synthesize the literature on ML and big data analytics for maternal mental health, particularly the prediction of postpartum depression (PPD).

METHODS

We used a scoping review methodology using the Arksey and O'Malley framework to rapidly map research activity in ML for predicting PPD. Two independent researchers searched PsycINFO, PubMed, IEEE Xplore, and the ACM Digital Library in September 2020 to identify relevant publications in the past 12 years. Data were extracted from the articles' ML model, data type, and study results.

RESULTS

A total of 14 studies were identified. All studies reported the use of supervised learning techniques to predict PPD. Support vector machine and random forest were the most commonly used algorithms in addition to Naive Bayes, regression, artificial neural network, decision trees, and XGBoost (Extreme Gradient Boosting). There was considerable heterogeneity in the best-performing ML algorithm across the selected studies. The area under the receiver operating characteristic curve values reported for different algorithms were support vector machine (range 0.78-0.86), random forest method (0.88), XGBoost (0.80), and logistic regression (0.93).

CONCLUSIONS

ML algorithms can analyze larger data sets and perform more advanced computations, which can significantly improve the detection of PPD at an early stage. Further clinical research collaborations are required to fine-tune ML algorithms for prediction and treatment. ML might become part of evidence-based practice in addition to clinical knowledge and existing research evidence.

Machine Learning-Based Behavioral Diagnostic Tools for Depression: Advances, Challenges, and Future Directions

The psychiatric diagnostic procedure is currently based on self-reports that are subject to personal biases. Therefore, the diagnostic process would benefit greatly from data-driven tools that can enhance accuracy and specificity. In recent years, many studies have achieved promising results in detecting and diagnosing depression based on machine learning (ML) analysis. Despite these favorable results in depression diagnosis, which are primarily based on ML analysis of neuroimaging data, most patients do not have access to neuroimaging tools. Hence, objective assessment tools are needed that can be easily integrated into the routine psychiatric diagnostic process. One solution is to use behavioral data, which can be easily collected while still maintaining objectivity. The current paper summarizes the main ML-based approaches that use behavioral data in diagnosing depression and other psychiatric disorders. We classified these studies into two main categories: (a) laboratory-based assessments and (b) data mining, the latter of which we further divided into two sub-groups: (i) social media usage and movement sensors data and (ii) demographic and clinical information. The paper discusses the advantages and challenges in this field and suggests future research directions and implementations. The paper's overarching aim is to serve as a first step in synthetizing existing knowledge about ML-based behavioral diagnosis studies in order to develop interventions and individually tailored treatments in the future.

A Novel Approach to Learning Models on EEG Data Using Graph Theory Features—A Comparative Study

Brain connectivity is studied as a functionally connected network using statistical methods such as measuring correlation or covariance. The non-invasive neuroimaging techniques such as Electroencephalography (EEG) signals are converted to networks by transforming the signals into a Correlation Matrix and analyzing the resulting networks. Here, four learning models, namely, Logistic Regression, Random Forest, Support Vector Machine, and Recurrent Neural Networks (RNN), are implemented on two different types of correlation matrices: Correlation Matrix (static connectivity) and Time-resolved Correlation Matrix (dynamic connectivity), to classify them either on their psychometric assessment or the effect of therapy. These correlation matrices are different from traditional learning techniques in the sense that they incorporate theory-based graph features into the learning models, thus providing novelty to this study. The EEG data used in this study is trail-based/event-related from five different experimental paradigms, of which can be broadly classified as working memory tasks and assessment of emotional states (depression, anxiety, and stress). The classifications based on RNN provided higher accuracy (74–88%) than the other three models (50–78%). Instead of using individual graph features, a Correlation Matrix provides an initial test of the data. When compared with the Time-resolved Correlation Matrix, it offered a 4–5% higher accuracy. The Time-resolved Correlation Matrix is better suited for dynamic studies here; it provides lower accuracy when compared to the Correlation Matrix, a static feature.

Individualized diagnosis of major depressive disorder via multivariate pattern analysis of thalamic sMRI features

BACKGROUND

Magnetic resonance imaging (MRI) studies have found thalamic abnormalities in major depressive disorder (MDD). Although there are significant differences in the structure and function of the thalamus between MDD patients and healthy controls (HCs) at the group level, it is not clear whether the structural and functional features of the thalamus are suitable for use as diagnostic prediction aids at the individual level. Here, we were to test the predictive value of gray matter density (GMD), gray matter volume (GMV), amplitude of low-frequency fluctuations (ALFF), and fractional amplitude of low-frequency fluctuations (fALFF) in the thalamus using multivariate pattern analysis (MVPA).

METHODS

Seventy-four MDD patients and 44 HC subjects were recruited. The Gaussian process classifier (GPC) was trained to separate MDD patients from HCs, Gaussian process regression (GPR) was trained to predict depression scores, and Multiple Kernel Learning (MKL) was applied to explore the contribution of each subregion of the thalamus.

RESULTS

The primary findings were as follows: [1] The balanced accuracy of the GPC trained with thalamic GMD was 96.59% (P < 0.001). The accuracy of the GPC trained with thalamic GMV was 93.18% (P < 0.001). The correlation between Hamilton Depression Scale (HAMD) score targets and predictions in the GPR trained with GMD was 0.90 (P < 0.001, r2 = 0.82), and in the GPR trained with GMV, the correlation between HAMD score targets and predictions was 0.89 (P < 0.001, r2 = 0.79). [2] The models trained with ALFF and fALFF in the thalamus failed to discriminate MDD patients from HC participants. [3] The MKL model showed that the left lateral prefrontal thalamus, the right caudal temporal thalamus, and the right sensory thalamus contribute more to the diagnostic classification.

CONCLUSIONS

The results suggested that GMD and GMV, but not functional indicators of the thalamus, have good potential for the individualized diagnosis of MDD. Furthermore, the thalamus shows the heterogeneity in the structural features of thalamic subregions for predicting MDD. To our knowledge, this is the first study to focus on the thalamus for the prediction of MDD using machine learning methods at the individual level.

Feature selection and machine learning methods for optimal identification and prediction of subtypes in Parkinson's disease

OBJECTIVES

The present work focuses on assessment of Parkinson's disease (PD), including both PD subtype identification (unsupervised task) and prediction (supervised task). We specifically investigate optimal feature selection and machine learning algorithms for these tasks.

METHODS

We selected 885 PD subjects as derived from longitudinal datasets (years 0-4; Parkinson's Progressive Marker Initiative), and investigated 981 features including motor, non-motor, and imaging features (SPECT-based radiomics features extracted using our standardized SERA software). Two different hybrid machine learning systems (HMLS) were constructed and applied to the data in order to select optimal combinations in both tasks: (i) identification of subtypes in PD (unsupervised-clustering), and (ii) prediction of these subtypes in year 4 (supervised-classification). From the original data based on years 0 (baseline) and 1, we created new datasets as inputs to the prediction task: (i,ii) CSD0 and CSD01: cross-sectional datasets from year 0 only and both years 0 & 1, respectively; (iii) TD01: timeless dataset from both years 0 & 1. In addition, PD subtype in year 4 was considered as outcome. Finally, high score features were derived via ensemble voting based on their prioritizations from feature selector algorithms (FSAs).

RESULTS

In clustering task, the most optimal combinations (out of 981) were selected by individual FSAs to enable high correlation compared to using all features (arriving at 547). In prediction task, we were able to select optimal combinations, resulting in an accuracy >90% only for timeless dataset (TD01); there, we were able to select the most optimal combination using 77 features, directly selected by FSAs. In both tasks, however, using combination of only high score features from ensemble voting did not enable acceptable performances, showing optimal feature selection via individual FSAs to be more effective.

CONCLUSION

Combining non-imaging information with SPECT-based radiomics features, and optimal utilization of HMLSs, can enable robust identification of subtypes as well as appropriate prediction of these subtypes in PD patients. Moreover, use of timeless dataset, beyond cross-sectional datasets, enabled predictive accuracies over 90%. Overall, we showed that radiomics features extracted from SPECT images are important in clustering as well as prediction of PD subtypes.

Neurite orientation dispersion and density imaging reveals white matter microstructural alterations in adults with autism

Background

Evidences suggesting the association between behavioral anomalies in autism and white matter (WM) microstructural alterations are increasing. Diffusion tensor imaging (DTI) is widely used to infer tissue microstructure. However, due to its lack of specificity, the underlying pathology of reported differences in DTI measures in autism remains poorly understood. Herein, we applied neurite orientation dispersion and density imaging (NODDI) to quantify and define more specific causes of WM microstructural changes associated with autism in adults.

Methods

NODDI (neurite density index [NDI], orientation dispersion index, and isotropic volume fraction [ISOVF]) and DTI (fractional anisotropy [FA], mean diffusivity [MD], axial diffusivity, and radial diffusivity [RD]) measures were compared between autism (N = 26; 19 males and 7 females; 32.93 ± 9.24 years old) and age- and sex-matched typically developing (TD; N = 25; 17 males and 8 females; 34.43 ± 9.02 years old) groups using tract-based spatial statistics and region-of-interest analyses. Linear discriminant analysis using leave-one-out cross-validation (LDA-LOOCV) was also performed to assess the discriminative power of diffusion measures in autism and TD.

Results

Significantly lower NDI and higher ISOVF, suggestive of decreased neurite density and increased extracellular free-water, respectively, were demonstrated in the autism group compared with the TD group, mainly in commissural and long-range association tracts, but with distinct predominant sides. Consistent with previous reports, the autism group showed lower FA and higher MD and RD when compared with TD group. Notably, LDA-LOOCV suggests that NDI and ISOVF have relatively higher accuracy (82%) and specificity (NDI, 84%; ISOVF, 88%) compared with that of FA, MD, and RD (accuracy, 67–73%; specificity, 68–80%).

Limitations

The absence of histopathological confirmation limit the interpretation of our findings.

Conclusions

Our results suggest that NODDI measures might be useful as imaging biomarkers to diagnose autism in adults and assess its behavioral characteristics. Furthermore, NODDI allows interpretation of previous findings on changes in WM diffusion tensor metrics in individuals with autism.

Predictive modeling of depression and anxiety using electronic health records and a novel machine learning approach with artificial intelligence

Generalized anxiety disorder (GAD) and major depressive disorder (MDD) are highly prevalent and impairing problems, but frequently go undetected, leading to substantial treatment delays. Electronic health records (EHRs) collect a great deal of biometric markers and patient characteristics that could foster the detection of GAD and MDD in primary care settings. We approached the problem of predicting MDD and GAD using a novel machine learning pipeline to re-analyze data from an observational study. The pipeline constitutes an ensemble of algorithmically distinct machine learning methods, including deep learning. A sample of 4,184 undergraduate students completed the study, undergoing a general health screening and completing a psychiatric assessment for MDD and GAD. After explicitly excluding all psychiatric information, 59 biomedical and demographic features from the general health survey in addition to a set of engineered features were used for model training. We assessed the model's performance on a held-out test set and found an AUC of 0.73 (sensitivity: 0.66, specificity: 0.7) and 0.67 (sensitivity: 0.55, specificity: 0.7) for GAD, and MDD, respectively. Additionally, we used advanced techniques (SHAP values) to illuminate which features had the greatest impact on prediction for each disease. The top predictive features for MDD were being satisfied with living conditions and having public health insurance. The top predictive features for GAD were vaccinations being up to date and marijuana use. Our results indicate moderate predictive performance for the application of machine learning methods in detection of GAD and MDD based on EHR data. By identifying important predictors of GAD and MDD, these results may be used in future research to aid in the early detection of MDD and GAD.

Robust identification of Parkinson's disease subtypes using radiomics and hybrid machine learning

OBJECTIVES

It is important to subdivide Parkinson's disease (PD) into subtypes, enabling potentially earlier disease recognition and tailored treatment strategies. We aimed to identify reproducible PD subtypes robust to variations in the number of patients and features.

METHODS

We applied multiple feature-reduction and cluster-analysis methods to cross-sectional and timeless data, extracted from longitudinal datasets (years 0, 1, 2 & 4; Parkinson's Progressive Marker Initiative; 885 PD/163 healthy-control visits; 35 datasets with combinations of non-imaging, conventional-imaging, and radiomics features from DAT-SPECT images). Hybrid machine-learning systems were constructed invoking 16 feature-reduction algorithms, 8 clustering algorithms, and 16 classifiers (C-index clustering evaluation used on each trajectory). We subsequently performed: i) identification of optimal subtypes, ii) multiple independent tests to assess reproducibility, iii) further confirmation by a statistical approach, iv) test of reproducibility to the size of the samples.

RESULTS

When using no radiomics features, the clusters were not robust to variations in features, whereas, utilizing radiomics information enabled consistent generation of clusters through ensemble analysis of trajectories. We arrived at 3 distinct subtypes, confirmed using the training and testing process of k-means, as well as Hotelling's T2 test. The 3 identified PD subtypes were 1) mild; 2) intermediate; and 3) severe, especially in terms of dopaminergic deficit (imaging), with some escalating motor and non-motor manifestations.

CONCLUSION

Appropriate hybrid systems and independent statistical tests enable robust identification of 3 distinct PD subtypes. This was assisted by utilizing radiomics features from SPECT images (segmented using MRI). The PD subtypes provided were robust to the number of the subjects, and features.

Predicting Antidepressant Citalopram Treatment Response via Changes in Brain Functional Connectivity After Acute Intravenous Challenge

Introduction: The early and therapy-specific prediction of treatment success in major depressive disorder is of paramount importance due to high lifetime prevalence, and heterogeneity of response to standard medication and symptom expression. Hence, this study assessed the predictability of long-term antidepressant effects of escitalopram based on the short-term influence of citalopram on functional connectivity. Methods: Twenty nine subjects suffering from major depression were scanned twice with resting-state functional magnetic resonance imaging under the influence of intravenous citalopram and placebo in a randomized, double-blinded cross-over fashion. Symptom factors were identified for the Hamilton depression rating scale (HAM-D) and Beck's depression inventory (BDI) taken before and after a median of seven weeks of escitalopram therapy. Predictors were calculated from whole-brain functional connectivity, fed into robust regression models, and cross-validated. Results: Significant predictive power could be demonstrated for one HAM-D factor describing insomnia and the total score (r = 0.45-0.55). Remission and response could furthermore be predicted with an area under the receiver operating characteristic curve of 0.73 and 0.68, respectively. Functional regions with high influence on the predictor were located especially in the ventral attention, fronto-parietal, and default mode networks. Conclusion: It was shown that medication-specific antidepressant symptom improvements can be predicted using functional connectivity measured during acute pharmacological challenge as an easily assessable imaging marker. The regions with high influence have previously been related to major depression as well as the response to selective serotonin reuptake inhibitors, corroborating the advantages of the current approach of focusing on treatment-specific symptom improvements.

Automated classification of depression from structural brain measures across two independent community-based cohorts

Major depressive disorder (MDD) has been the subject of many neuroimaging case-control classification studies. Although some studies report accuracies ≥80%, most have investigated relatively small samples of clinically-ascertained, currently symptomatic cases, and did not attempt replication in larger samples. We here first aimed to replicate previously reported classification accuracies in a small, well-phenotyped community-based group of current MDD cases with clinical interview-based diagnoses (from STratifying Resilience and Depression Longitudinally cohort, 'STRADL'). We performed a set of exploratory predictive classification analyses with measures related to brain morphometry and white matter integrity. We applied three classifier types-SVM, penalised logistic regression or decision tree-either with or without optimisation, and with or without feature selection. We then determined whether similar accuracies could be replicated in a larger independent population-based sample with self-reported current depression (UK Biobank cohort). Additional analyses extended to lifetime MDD diagnoses-remitted MDD in STRADL, and lifetime-experienced MDD in UK Biobank. The highest cross-validation accuracy (75%) was achieved in the initial current MDD sample with a decision tree classifier and cortical surface area features. The most frequently selected decision tree split variables included surface areas of bilateral caudal anterior cingulate, left lingual gyrus, left superior frontal, right precentral and paracentral regions. High accuracy was not achieved in the larger samples with self-reported current depression (53.73%), with remitted MDD (57.48%), or with lifetime-experienced MDD (52.68-60.29%). Our results indicate that high predictive classification accuracies may not immediately translate to larger samples with broader criteria for depression, and may not be robust across different classification approaches.

Towards a brain-based predictome of mental illness

Neuroimaging-based approaches have been extensively applied to study mental illness in recent years and have deepened our understanding of both cognitively healthy and disordered brain structure and function. Recent advancements in machine learning techniques have shown promising outcomes for individualized prediction and characterization of patients with psychiatric disorders. Studies have utilized features from a variety of neuroimaging modalities, including structural, functional, and diffusion magnetic resonance imaging data, as well as jointly estimated features from multiple modalities, to assess patients with heterogeneous mental disorders, such as schizophrenia and autism. We use the term "predictome" to describe the use of multivariate brain network features from one or more neuroimaging modalities to predict mental illness. In the predictome, multiple brain network-based features (either from the same modality or multiple modalities) are incorporated into a predictive model to jointly estimate features that are unique to a disorder and predict subjects accordingly. To date, more than 650 studies have been published on subject-level prediction focusing on psychiatric disorders. We have surveyed about 250 studies including schizophrenia, major depression, bipolar disorder, autism spectrum disorder, attention-deficit hyperactivity disorder, obsessive-compulsive disorder, social anxiety disorder, posttraumatic stress disorder, and substance dependence. In this review, we present a comprehensive review of recent neuroimaging-based predictomic approaches, current trends, and common shortcomings and share our vision for future directions.

Brain network topology predicts participant adherence to mental training programs

Adherence determines the success and benefits of mental training (e.g., meditation) programs. It is unclear why some participants engage more actively in programs for mental training than others. Understanding neurobiological factors that predict adherence is necessary for understanding elements of learning and to inform better designs for new learning regimens. Clustering patterns in brain networks have been suggested to predict learning performance, but it is unclear whether these patterns contribute to motivational aspects of learning such as adherence. This study tests whether configurations of brain connections in resting-state fMRI scans can be used to predict adherence to two programs: meditation and creative writing. Results indicate that greater system segregation and clustering predict the number of practice sessions and class participation in both programs at a wide range of network thresholds (corrected p value < 0.05). At a local level, regions in subcortical circuitry such as striatum and accumbens predicted adherence in all subjects. Furthermore, there were also some important distinctions between groups: Adherence to meditation was predicted by connectivity within local network of the anterior insula and default mode network; and in the writing program, adherence was predicted by network neighborhood of frontal and temporal regions. Four machine learning methods were applied to test the robustness of the brain metric for classifying individual capacity for adherence and yielded reasonable accuracy. Overall, these findings underscore the fact that adherence and the ability to perform prescribed exercises is associated with organizational patterns of brain connectivity.

Machine Learning Models for the Prediction of Postpartum Depression: Application and Comparison Based on a Cohort Study

Background

Postpartum depression (PPD) is a serious public health problem. Building a predictive model for PPD using data during pregnancy can facilitate earlier identification and intervention.

Objective

The aims of this study are to compare the effects of four different machine learning models using data during pregnancy to predict PPD and explore which factors in the model are the most important for PPD prediction.

Methods

Information on the pregnancy period from a cohort of 508 women, including demographics, social environmental factors, and mental health, was used as predictors in the models. The Edinburgh Postnatal Depression Scale score within 42 days after delivery was used as the outcome indicator. Using two feature selection methods (expert consultation and random forest-based filter feature selection [FFS-RF]) and two algorithms (support vector machine [SVM] and random forest [RF]), we developed four different machine learning PPD prediction models and compared their prediction effects.

Results

There was no significant difference in the effectiveness of the two feature selection methods in terms of model prediction performance, but 10 fewer factors were selected with the FFS-RF than with the expert consultation method. The model based on SVM and FFS-RF had the best prediction effects (sensitivity=0.69, area under the curve=0.78). In the feature importance ranking output by the RF algorithm, psychological elasticity, depression during the third trimester, and income level were the most important predictors.

Conclusions

In contrast to the expert consultation method, FFS-RF was important in dimension reduction. When the sample size is small, the SVM algorithm is suitable for predicting PPD. In the prevention of PPD, more attention should be paid to the psychological resilience of mothers.

Automated assessment of psychiatric disorders using speech: A systematic review

OBJECTIVE

There are many barriers to accessing mental health assessments including cost and stigma. Even when individuals receive professional care, assessments are intermittent and may be limited partly due to the episodic nature of psychiatric symptoms. Therefore, machine-learning technology using speech samples obtained in the clinic or remotely could one day be a biomarker to improve diagnosis and treatment. To date, reviews have only focused on using acoustic features from speech to detect depression and schizophrenia. Here, we present the first systematic review of studies using speech for automated assessments across a broader range of psychiatric disorders.

METHODS

We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. We included studies from the last 10 years using speech to identify the presence or severity of disorders within the Diagnostic and Statistical Manual of Mental Disorders (DSM-5). For each study, we describe sample size, clinical evaluation method, speech-eliciting tasks, machine learning methodology, performance, and other relevant findings.

RESULTS

1395 studies were screened of which 127 studies met the inclusion criteria. The majority of studies were on depression, schizophrenia, and bipolar disorder, and the remaining on post-traumatic stress disorder, anxiety disorders, and eating disorders. 63% of studies built machine learning predictive models, and the remaining 37% performed null-hypothesis testing only. We provide an online database with our search results and synthesize how acoustic features appear in each disorder.

CONCLUSION

Speech processing technology could aid mental health assessments, but there are many obstacles to overcome, especially the need for comprehensive transdiagnostic and longitudinal studies. Given the diverse types of data sets, feature extraction, computational methodologies, and evaluation criteria, we provide guidelines for both acquiring data and building machine learning models with a focus on testing hypotheses, open science, reproducibility, and generalizability.

LEVEL OF EVIDENCE

3a.

Prediction of individual responses to electroconvulsive therapy in patients with schizophrenia: Machine learning analysis of resting-state electroencephalography

BACKGROUND

Electroconvulsive therapy (ECT) has strong efficacy in patients with treatment refractory schizophrenia. However, access to ECT has been limited by high costs, professional labor, treatment duration, and significant adverse effects. To provide support for the decision to perform ECT, we aimed to predict individual responses to ECT among patients with schizophrenia using machine learning analysis of resting-state electroencephalography (EEG).

METHODS

Forty-seven patients diagnosed with schizophrenia or schizoaffective disorder with EEG recordings before the course of ECT were treated at Seoul National University Hospital. Among these patients, 29 were responders who showed scores of 3 or less on the Clinical Global Impression Severity scale after the course of ECT. Transfer entropy (TE), which represents information flow, was extracted from baseline EEG data and used as a feature. Feature selection was performed with four methods, including Random Subset Feature Selection (RSFS). The random forest classifier was used to predict individual ECT responses.

RESULTS

The averaged TE, especially in frontal regions, was higher in ECT responders than in nonresponders. A predictive model using the RSFS method classified ECT responders and nonresponders with 85.3% balanced accuracy, 85.2% accuracy, 88.7% sensitivity, and 81.8% specificity. The positive predictive value was 82.6%, and the negative predictive value was 88.2%.

CONCLUSIONS

The results of the current study suggest that higher effective connectivity in frontal areas may be associated with a favorable ECT response. Furthermore, personalized decisions to perform ECT in clinical practice could be augmented by resting-state EEG biomarkers of the ECT response in schizophrenia patients.

Predicting Treatment Outcome in Major Depressive Disorder Using Serotonin 4 Receptor PET Brain Imaging, Functional MRI, Cognitive-, EEG-Based, and Peripheral Biomarkers: A NeuroPharm Open Label Clinical Trial Protocol

BACKGROUND

Between 30 and 50% of patients with major depressive disorder (MDD) do not respond sufficiently to antidepressant regimens. The conventional pharmacological treatments predominantly target serotonergic brain signaling but better tools to predict treatment response and identify relevant subgroups of MDD are needed to support individualized and mechanistically targeted treatment strategies. The aim of this study is to investigate antidepressant-free patients with MDD using neuroimaging, electrophysiological, molecular, cognitive, and clinical examinations and evaluate their ability to predict clinical response to SSRI treatment as individual or combined predictors.

METHODS

We will include 100 untreated patients with moderate to severe depression (>17 on the Hamilton Depression Rating Scale 17) in a non-randomized open clinical trial. We will collect data from serotonin 4 receptor positron emission tomography (PET) brain scans, functional magnetic resonance imaging (fMRI), electroencephalogram (EEG), cognitive tests, psychometry, and peripheral biomarkers, before (at baseline), during, and after 12 weeks of standard antidepressant treatment. Patients will be treated with escitalopram, and in case of non-response at week 4 or intolerable side effects, offered to switch to a second line treatment with duloxetine. Our primary outcome (treatment response) is assessed using the Hamilton depression rating subscale 6-item scores at week 8, compared to baseline. In a subset of the patients (n = ~40), we will re-assess the neurobiological response (using PET, fMRI, and EEG) 8 weeks after initiated pharmacological antidepressant treatment, to map neurobiological signatures of treatment responses. Data from matched controls will either be collected or is already available from other cohorts.

DISCUSSION

The extensive investigational program with follow-up in this large cohort of participants provides a unique possibility to (a) uncover potential biomarkers for antidepressant treatment response, (b) apply the findings for future stratification of MDD, (c) advance the understanding of pathophysiological underpinnings of MDD, and (d) uncover how putative biomarkers change in response to 8 weeks of pharmacological antidepressant treatment. Our data can pave the way for a precision medicine approach for optimized treatment of MDD and also provides a resource for future research and data sharing.

CLINICAL TRIAL REGISTRATION

The study was registered at clinicaltrials.gov prior to initiation (NCT02869035; 08.16.2016, URL: https://clinicaltrials.gov/ct2/results?cond=&term=NCT02869035&cntry=&state=&city=&dist=).

Resting-State Functional Network Scale Effects and Statistical Significance-Based Feature Selection in Machine Learning Classification

In recent years, functional brain network topological features have been widely used as classification features. Previous studies have found that network node scale differences caused by different network parcellation definitions significantly affect the structure of the constructed network and its topological properties. However, we still do not know how network scale differences affect the classification accuracy, performance of classification features, and effectiveness of the feature selection strategy using P values in terms of the machine learning method. This study used five scale parcellations, involving 90, 256, 497, 1003, and 1501 nodes. Three local properties of resting-state functional brain networks were selected (degree, betweenness centrality, and nodal efficiency), and the support vector machine method was used to construct classifiers to identify patients with major depressive disorder. We analyzed the impact of the five scales on classification accuracy. In addition, the effectiveness and redundancy of features obtained by the different scale parcellations were compared. Finally, traditional statistical significance (P value) was verified as a feature selection criterion. The results showed that the feature effectiveness of different scales was similar; in other words, parcellation with more regions did not provide more effective discriminative features. Nevertheless, parcellation with more regions did provide a greater quantity of discriminative features, which led to an improvement in the accuracy of the classification. However, due to the close distance between brain regions, the redundancy of parcellation with more regions was also greater. The traditional P value feature selection strategy is feasible with different scales, but our analysis showed that the traditional P < 0.05 threshold was too strict for feature selection. This study provides an important reference for the selection of network scales when applying topological properties of brain networks to machine learning methods.