Graph-Theory-Based Degree Centrality Combined with Machine Learning Algorithms Can Predict Response to Treatment with Antipsychotic Medications in Patients with First-Episode Schizophrenia

Network-based artificial intelligence approaches for advancing personalized psychiatry

Psychiatric disorders have a complex biological underpinning likely involving an interplay of genetic and environmental risk contributions. Substantial efforts are being made to use artificial intelligence approaches to integrate features within and across data types to increase our etiological understanding and advance personalized psychiatry. Network science offers a conceptual framework for exploring the often complex relationships across different levels of biological organization, from cellular mechanistic to brain-functional and phenotypic networks. Utilizing such network information effectively as part of artificial intelligence approaches is a promising route toward a more in-depth understanding of illness biology, the deciphering of patient heterogeneity, and the identification of signatures that may be sufficiently predictive to be clinically useful. Here, we present examples of how network information has been used as part of artificial intelligence within psychiatry and beyond and outline future perspectives on how personalized psychiatry approaches may profit from a closer integration of psychiatric research, artificial intelligence development, and network science.

Association between degree centrality and neurocognitive impairments in patients with Schizophrenia: A Longitudinal rs-fMRI Study

BACKGROUND

Evidence indicates that patients with schizophrenia (SZ) experience significant changes in their functional connectivity during antipsychotic treatment. Despite previous reports of changes in brain network degree centrality (DC) in patients with schizophrenia, the relationship between brain DC changes and neurocognitive improvement in patients with SZ after antipsychotic treatment remains elusive.

METHODS

A total of 74 patients with acute episodes of chronic SZ and 53 age- and sex-matched healthy controls were recruited. The Positive and Negative Syndrome Scale (PANSS), Symbol Digit Modalities Test, digital span test (DST), and verbal fluency test were used to evaluate the clinical symptoms and cognitive performance of the patients with SZ. Patients with SZ were treated with antipsychotics for six weeks starting at baseline and underwent MRI and clinical interviews at baseline and after six weeks, respectively. We then divided the patients with SZ into responding (RS) and non-responding (NRS) groups based on the PANSS scores (reduction rate of PANSS ≥50%). DC was calculated and analyzed to determine its correlation with clinical symptoms and cognitive performance.

RESULTS

After antipsychotic treatment, the patients with SZ showed significant improvements in clinical symptoms, semantic fluency performance. Correlation analysis revealed that the degree of DC increase in the left anterior inferior parietal lobe (aIPL) after treatment was negatively correlated with changes in the excitement score (r = -0.256, p = 0.048, adjusted p = 0.080), but this correlation failed the multiple test correction. Patients with SZ showed a significant negative correlation between DC values in the left aIPL and DST scores after treatment, which was not observed at the baseline (r = -0.359, p = 0.005, adjusted p = 0.047). In addition, we did not find a significant difference in DC between the RS and NRS groups, neither at baseline nor after treatment.

CONCLUSIONS

The results suggested that DC changes in patients with SZ after antipsychotic treatment are correlated with neurocognitive performance. Our findings provide new insights into the neuropathological mechanisms underlying antipsychotic treatment of SZ.

Shared and distinctive dysconnectivity patterns underlying pure generalized anxiety disorder (GAD) and comorbid GAD and depressive symptoms

The resting-state connectivity features underlying pure generalized anxiety disorder (GAD, G1) and comorbid GAD and depressive symptoms (G2) have not been directly compared. Furthermore, it is unclear whether these features might serve as potential prognostic biomarkers and change with treatment. Degree centrality (DC) in G1 (40 subjects), G2 (58 subjects), and healthy controls (HCs, 54 subjects) was compared before treatment, and the DC of G1 or G2 at baseline was compared with that after 4 weeks of paroxetine treatment. Using support vector regression (SVR), voxel-wise DC across the entire brain and abnormal DC at baseline were employed to predict treatment response. At baseline, G1 and G2 exhibited lower DC in the left mid-cingulate cortex and vermis IV/V compared to HCs. Additionally, compared to HCs, G1 had lower DC in the left middle temporal gyrus, while G2 showed higher DC in the right inferior temporal/fusiform gyrus. However, there was no significant difference in DC between G1 and G2. The SVR based on abnormal DC at baseline could successfully predict treatment response in responders in G2 or in G1 and G2. Notably, the predictive performance based on abnormal DC at baseline surpassed that based on DC across the entire brain. After treatment, G2 responders showed lower DC in the right medial orbital frontal gyrus, while no change in DC was identified in G1 responders. The G1 and G2 showed common and distinct dysconnectivity patterns and they could potentially serve as prognostic biomarkers. Furthermore, DC in patients with GAD could change with treatment.

Shared and Disorder-Specific Alterations of Brain Temporal Dynamics in Obsessive-Compulsive Disorder and Schizophrenia

BACKGROUND

Obsessive-compulsive disorder (OCD) and schizophrenia have distinct but also overlapping symptoms. Few studies have examined the shared and disorder-specific disturbances in dynamic brain function in the 2 disorders.

STUDY DESIGN

Resting-state functional magnetic resonance imaging data of 31 patients with OCD and 49 patients with schizophrenia, all untreated, and 45 healthy controls (HCs) were analyzed using spatial group independent component (IC) analysis. Time-varying degree centrality patterns across the whole brain were clustered into 3 reoccurring states, and state transition metrics were obtained. We further explored regional temporal variability of degree centrality for each IC across all time windows.

STUDY RESULTS

Patients with OCD and patients with schizophrenia both showed decreased occurrence of a state having the highest centrality in the sensorimotor and auditory networks. Additionally, patients with OCD and patients with schizophrenia both exhibited reduced dynamics of degree centrality in the superior frontal gyrus than controls, while dynamic degree centrality of the cerebellum was lower in patients with schizophrenia than with OCD and HCs. Altered dynamics of degree centrality nominally correlated with symptom severity in both patient groups.

CONCLUSIONS

Our study provides evidence of transdiagnostic and clinically relevant functional brain abnormalities across OCD and schizophrenia in neocortex, as well as functional dynamic alterations in the cerebellum specific to schizophrenia. These findings add to the recognition of overlap in neocortical alterations in the 2 disorders, and indicate that cerebellar alterations in schizophrenia may be specifically important in schizophrenia pathophysiology via impact on cerebellar thalamocortical circuitry.

Artificial intelligence in psychiatry research, diagnosis, and therapy

Psychiatric disorders are now responsible for the largest proportion of the global burden of disease, and even more challenges have been seen during the COVID-19 pandemic. Artificial intelligence (AI) is commonly used to facilitate the early detection of disease, understand disease progression, and discover new treatments in the fields of both physical and mental health. The present review provides a broad overview of AI methodology and its applications in data acquisition and processing, feature extraction and characterization, psychiatric disorder classification, potential biomarker detection, real-time monitoring, and interventions in psychiatric disorders. We also comprehensively summarize AI applications with regard to the early warning, diagnosis, prognosis, and treatment of specific psychiatric disorders, including depression, schizophrenia, autism spectrum disorder, attention-deficit/hyperactivity disorder, addiction, sleep disorders, and Alzheimer's disease. The advantages and disadvantages of AI in psychiatry are clarified. We foresee a new wave of research opportunities to facilitate and improve AI technology and its long-term implications in psychiatry during and after the COVID-19 era.