Classification of first-episode psychosis using cortical thickness: A large multicenter MRI study.

Machine learning classifications of first-episode psychosis (FEP) using neuroimaging have predominantly analyzed brain volumes. Some studies examined cortical thickness, but most of them have used parcellation approaches with data from single sites, which limits claims of generalizability. To address these limitations, we conducted a large-scale, multi-site analysis of cortical thickness comparing parcellations and vertex-wise approaches. By leveraging the multi-site nature of the study, we further investigated how different demographical and site-dependent variables affected predictions. Finally, we assessed relationships between predictions and clinical variables. 428 subjects (147 females, mean age 27.14) with FEP and 448 (230 females, mean age 27.06) healthy controls were enrolled in 8 centers by the ClassiFEP group. All subjects underwent a structural MRI and were clinically assessed. Cortical thickness parcellation (68 areas) and full cortical maps (20,484 vertices) were extracted. Linear Support Vector Machine was used for classification within a repeated nested cross-validation framework. Vertex-wise thickness maps outperformed parcellation-based methods with a balanced accuracy of 66.2% and an Area Under the Curve of 72%. By stratifying our sample for MRI scanner, we increased generalizability across sites. Temporal brain areas resulted as the most influential in the classification. The predictive decision scores significantly correlated with age at onset, duration of treatment, and positive symptoms. In conclusion, although far from the threshold of clinical relevance, temporal cortical thickness proved to classify between FEP subjects and healthy individuals. The assessment of site-dependent variables permitted an increase in the across-site generalizability, thus attempting to address an important machine learning limitation.

Impact of polygenic schizophrenia-related risk and hippocampal volumes on the onset of psychosis.

Alterations in hippocampal volume are a known marker for first-episode psychosis (FEP) as well as for the clinical high-risk state. The Polygenic Schizophrenia-related Risk Score (PSRS), derived from a large case-control study, indicates the polygenic predisposition for schizophrenia in our clinical sample. A total of 65 at-risk mental state (ARMS) and FEP patients underwent structural magnetic resonance imaging. We used automatic segmentation of hippocampal volumes using the FSL-FIRST software and an odds-ratio-weighted PSRS based on the publicly available top single-nucleotide polymorphisms from the Psychiatric Genomics Consortium genome-wide association study (GWAS). We observed a negative association between the PSRS and hippocampal volumes (ß=-0.42, P=0.01, 95% confidence interval (CI)=(-0.72 to -0.12)) across FEP and ARMS patients. Moreover, a higher PSRS was significantly associated with a higher probability of an individual being assigned to the FEP group relative to the ARMS group (ß=0.64, P=0.03, 95% CI=(0.08-1.29)). These findings provide evidence that a subset of schizophrenia risk variants is negatively associated with hippocampal volumes, and higher values of this PSRS are significantly associated with FEP compared with the ARMS. This implies that FEP patients have a higher genetic risk for schizophrenia than the total cohort of ARMS patients. The identification of associations between genetic risk variants and structural brain alterations will increase our understanding of the neurobiology underlying the transition to psychosis.

BDNF Val66Met polymorphism and hippocampal volume in neuropsychiatric disorders: A systematic review and meta-analysis.

BACKGROUND: Brain-derived neurotrophic factor (BDNF) is a neurotrophin involved in neurogenesis and synaptic plasticity in the central nervous system, especially in the hippocampus, and has been implicated in the pathophysiology of several neuropsychiatric disorders. Its Val66Met polymorphism (refSNP Cluster Report: rs6265) is a functionally relevant single nucleotide polymorphism affecting the secretion of BDNF and is implicated in differences in hippocampal volumes. METHODS: This is a systematic meta-analytical review of findings from imaging genetic studies on the impact of the rs6265 SNP on hippocampal volumes in neuropsychiatric patients with major depressive disorder, anxiety, bipolar disorder or schizophrenia. RESULTS: The overall sample size of 18 independent clinical cohorts comprised 1695 patients. Our results indicated no significant association of left (Hedge's g=0.08, p=0.12), right (g=0.07, p=0.22) or bilateral (g=0.07, p=0.16) hippocampal volumes with BDNF rs6265 in neuropsychiatric patients. There was no evidence for a publication bias or any demographic, clinical, or methodological moderating effects. Both Val/Val homozygotes (g=0.32, p=0.004) and Met-carriers (g=0.20, p=0.004) from the patient sample had significantly smaller hippocampal volumes than the healthy control sample with the same allele. The magnitude of these effects did not differ between the two genotypes. CONCLUSION: This meta-analysis suggests that there is no association between this BDNF polymorphism and hippocampal volumes. For each BDNF genotype, the hippocampal volumes were significantly lower in neuropsychiatric patients than in healthy controls.

The association of the BDNF Val66Met polymorphism and the hippocampal volumes in healthy humans: a joint meta-analysis of published and new data.

BACKGROUND: The brain-derived neurotrophic factor (BDNF) Val66Met polymorphism (refSNP Cluster Report: rs6265) is a common and functionally relevant single nucleotide polymorphism (SNP). The gene itself, as well as the SNP rs6265, have been implicated in hippocampal learning and memory. However, imaging genetic studies have produced controversial results about the impact of this SNP on hippocampal volumes in healthy subjects. METHODS: We examined the association between the rs6265 polymorphism and hippocampal volume in 643 healthy young subjects using automatic segmentation and subsequently included these data in a meta-analysis based on published studies with 5298 healthy subjects in total. RESULTS: We found no significant association between SNP rs6265 and hippocampal volumes in our sample (g=0.05, p=0.58). The meta-analysis revealed a small, albeit significant difference in hippocampal volumes between genotype groups, such that Met-carriers had slightly smaller hippocampal volumes than Val/Val homozygotes (g=0.09, p=0.04), an association that was only evident when manual (g=0.22, p=0.01) but not automatic tracing approaches (g=0.04, p=0.38) were used. Studies using manual tracing showed evidence for publication bias and a significant decrease in effect size over the years with increasing sample sizes. CONCLUSIONS: This study does not support the association between SNP rs6265 and hippocampal volume in healthy individuals. The weakly significant effect observed in the meta-analysis is mainly driven by studies with small sample sizes. In contrast, our original data and the meta-analysis of automatically segmented hippocampal volumes, which was based on studies with large samples sizes, revealed no significant genotype effect. Thus, meta-analyses of the association between rs6265 and hippocampal volumes should consider possible biases related to measuring technique and sample size.