Using Minimal-Redundant and Maximal-Relevant Whole-Brain Functional Connectivity to Classify Bipolar Disorder

TSP-GNN: a novel neuropsychiatric disorder classification framework based on task-specific prior knowledge and graph neural network

Neuropsychiatric disorder (ND) is often accompanied by abnormal functional connectivity (FC) patterns in specific task contexts. The distinctive task-specific FC patterns can provide valuable features for ND classification models using deep learning. However, most previous studies rely solely on the whole-brain FC matrix without considering the prior knowledge of task-specific FC patterns. Insight by the decoding studies on brain-behavior relationship, we develop TSP-GNN, which extracts task-specific prior (TSP) connectome patterns and employs graph neural network (GNN) for disease classification. TSP-GNN was validated using publicly available datasets. Our results demonstrate that different ND types show distinct task-specific connectivity patterns. Compared with the whole-brain node characteristics, utilizing task-specific nodes enhances the accuracy of ND classification. TSP-GNN comprises the first attempt to incorporate prior task-specific connectome patterns and the power of deep learning. This study elucidates the association between brain dysfunction and specific cognitive processes, offering valuable insights into the cognitive mechanism of neuropsychiatric disease.

Characterizing the spatiotemporal features of functional connectivity across the white matter and gray matter during the naturalistic condition

Introduction

The naturalistic stimuli due to its ease of operability has attracted many researchers in recent years. However, the influence of the naturalistic stimuli for whole-brain functions compared with the resting state is still unclear.

Methods

In this study, we clustered gray matter (GM) and white matter (WM) masks both at the ROI- and network-levels. Functional connectivity (FC) and inter-subject functional connectivity (ISFC) were calculated in GM, WM, and between GM and WM under the movie-watching and the resting-state conditions. Furthermore, intra-class correlation coefficients (ICC) of FC and ISFC were estimated on different runs of fMRI data to denote the reliability of them during the two conditions. In addition, static and dynamic connectivity indices were calculated with Pearson correlation coefficient to demonstrate the associations between the movie-watching and the resting-state.

Results

As the results, we found that the movie-watching significantly affected FC in whole-brain compared with the resting-state, but ISFC did not show significant connectivity induced by the naturalistic condition. ICC of FC and ISFC was generally higher during movie-watching compared with the resting-state, demonstrating that naturalistic stimuli could promote the reliability of connectivity. The associations between static and dynamic ISFC were weakly negative correlations in the naturalistic stimuli while there is no correlation between them under resting-state condition.

Discussion

Our findings confirmed that compared to resting-state condition, the connectivity indices under the naturalistic stimuli were more reliable and stable to investigate the normal functional activities of the human brain, and might promote the applications of FC in the cerebral dysfunction in various mental disorders.

Sampling inequalities affect generalization of neuroimaging-based diagnostic classifiers in psychiatry

BACKGROUND

The development of machine learning models for aiding in the diagnosis of mental disorder is recognized as a significant breakthrough in the field of psychiatry. However, clinical practice of such models remains a challenge, with poor generalizability being a major limitation.

METHODS

Here, we conducted a pre-registered meta-research assessment on neuroimaging-based models in the psychiatric literature, quantitatively examining global and regional sampling issues over recent decades, from a view that has been relatively underexplored. A total of 476 studies (n = 118,137) were included in the current assessment. Based on these findings, we built a comprehensive 5-star rating system to quantitatively evaluate the quality of existing machine learning models for psychiatric diagnoses.

RESULTS

A global sampling inequality in these models was revealed quantitatively (sampling Gini coefficient (G) = 0.81, p < .01), varying across different countries (regions) (e.g., China, G = 0.47; the USA, G = 0.58; Germany, G = 0.78; the UK, G = 0.87). Furthermore, the severity of this sampling inequality was significantly predicted by national economic levels (β =  - 2.75, p < .001, R²_adj = 0.40; r =  - .84, 95% CI: - .41 to - .97), and was plausibly predictable for model performance, with higher sampling inequality for reporting higher classification accuracy. Further analyses showed that lack of independent testing (84.24% of models, 95% CI: 81.0-87.5%), improper cross-validation (51.68% of models, 95% CI: 47.2-56.2%), and poor technical transparency (87.8% of models, 95% CI: 84.9-90.8%)/availability (80.88% of models, 95% CI: 77.3-84.4%) are prevailing in current diagnostic classifiers despite improvements over time. Relating to these observations, model performances were found decreased in studies with independent cross-country sampling validations (all p < .001, BF₁₀ > 15). In light of this, we proposed a purpose-built quantitative assessment checklist, which demonstrated that the overall ratings of these models increased by publication year but were negatively associated with model performance.

CONCLUSIONS

Together, improving sampling economic equality and hence the quality of machine learning models may be a crucial facet to plausibly translating neuroimaging-based diagnostic classifiers into clinical practice.

Evaluation of Risk of Bias in Neuroimaging-Based Artificial Intelligence Models for Psychiatric Diagnosis: A Systematic Review

IMPORTANCE

Neuroimaging-based artificial intelligence (AI) diagnostic models have proliferated in psychiatry. However, their clinical applicability and reporting quality (ie, feasibility) for clinical practice have not been systematically evaluated.

OBJECTIVE

To systematically assess the risk of bias (ROB) and reporting quality of neuroimaging-based AI models for psychiatric diagnosis.

EVIDENCE REVIEW

PubMed was searched for peer-reviewed, full-length articles published between January 1, 1990, and March 16, 2022. Studies aimed at developing or validating neuroimaging-based AI models for clinical diagnosis of psychiatric disorders were included. Reference lists were further searched for suitable original studies. Data extraction followed the CHARMS (Checklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modeling Studies) and PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines. A closed-loop cross-sequential design was used for quality control. The PROBAST (Prediction Model Risk of Bias Assessment Tool) and modified CLEAR (Checklist for Evaluation of Image-Based Artificial Intelligence Reports) benchmarks were used to systematically evaluate ROB and reporting quality.

FINDINGS

A total of 517 studies presenting 555 AI models were included and evaluated. Of these models, 461 (83.1%; 95% CI, 80.0%-86.2%) were rated as having a high overall ROB based on the PROBAST. The ROB was particular high in the analysis domain, including inadequate sample size (398 of 555 models [71.7%; 95% CI, 68.0%-75.6%]), poor model performance examination (with 100% of models lacking calibration examination), and lack of handling data complexity (550 of 555 models [99.1%; 95% CI, 98.3%-99.9%]). None of the AI models was perceived to be applicable to clinical practices. Overall reporting completeness (ie, number of reported items/number of total items) for the AI models was 61.2% (95% CI, 60.6%-61.8%), and the completeness was poorest for the technical assessment domain with 39.9% (95% CI, 38.8%-41.1%).

CONCLUSIONS AND RELEVANCE

This systematic review found that the clinical applicability and feasibility of neuroimaging-based AI models for psychiatric diagnosis were challenged by a high ROB and poor reporting quality. Particularly in the analysis domain, ROB in AI diagnostic models should be addressed before clinical application.

Pattern of frustration formation in the functional brain network

The brain is a frustrated system that contains conflictual link arrangements named frustration. The frustration as a source of disorder prevents the system from settling into low-energy states and provides flexibility for brain network organization. In this research, we tried to identify the pattern of frustration formation in the brain at the levels of region, connection, canonical network, and hemisphere. We found that frustration formation has no uniform pattern. Some subcortical elements have an active role in frustration formation, despite low contributions from many cortical elements. Frustrating connections are mostly between-network connections, and triadic frustrations are mainly formed between three regions from three distinct canonical networks. We did not find any significant differences between brain hemispheres or any robust differences between the frustration formation patterns of various life-span stages. Our results may be interesting for those who study the organization of brain links and promising for those who want to manipulate brain networks.

Aberrant Functional Connectivity of Sensorimotor Network and Its Relationship With Executive Dysfunction in Bipolar Disorder Type I

Background

The key pathophysiological mechanism of executive dysfunction in patients with bipolar disorder type I (BD-I) is still unclear. Previous studies have demonstrated that it may be related to the disbalance of the sensory motor network (SMN).

Objective

This study was designed to explore the aberrant functional connectivity (FC) of SMN in BD-I patients and its potential associations with executive dysfunction.

Methods

Eighteen BD-I patients and 20 healthy controls (HCs) underwent resting-state fMRI scans. The intranetwork and internetwork functional connectivities of SMN were extracted by independent component analysis (ICA). Clinical symptoms were assessed by the Bech–Rafaelsen Mania Rating Scale (BRMS) and Positive and Negative Syndrome Scale (PANSS). Executive function was measured by digit span tasks and a verbal fluency test. Finally, linear regression and correlation analyses were applied to measure the potential associations between clinical symptoms, intranetwork and internetwork functional connectivities, and executive function performance.

Results

(1) Patients with BD-I showed increased connectivity in the right paracentral lobule and the right postcentral gyrus within the SMN, and the increased connectivity value was positively correlated with the BRMS score (P < 0.05) but negatively correlated with digit span forward scores (P < 0.05). (2) Compared with HC, the connectivity value increased between the SMN and dorsal attention network (DAN) (P < 0.01) and between the default mode network (DMN) and DAN (P < 0.05) but decreased between the DAN and auditory network (AN) (P < 0.05) and between the SMN and DMN (P < 0.01) in patients with BD-I. (3) Digit span forward scores and education of all participants were negatively correlated with FC between SMN and DAN. Age of all subjects was positively correlated with FC between SMN and DMN.

Conclusion

Our findings suggest that the sensorimotor network of BD-I has abnormal functional connections within and between networks, and the abnormal FC value correlated with clinical symptoms and executive function, which provide new information for exploring the neural physiopathology of executive dysfunction in BD-I patients.

Evaluating the Performance of Machine Learning Models for Automatic Diagnosis of Patients with Schizophrenia Based on a Single Site Dataset of 440 Participants

Background

Methods

Results

Conclusions