Search

Background Individuals with bipolar disorder are commonly correctly diagnosed a decade after symptom onset. Machine learning techniques may aid in early recognition and reduce the disease burden. As both individuals at risk and those with a manifest disease display structural brain markers, structural magnetic resonance imaging may provide relevant classification features. Methods Following a pre-registered protocol, we trained linear support vector machine (SVM) to classify individuals according to their estimated risk for bipolar disorder using regional cortical thickness of help-seeking individuals from seven study sites (N = 276). We estimated the risk using three state-of-the-art assessment instruments (BPSS-P, BARS, EPIbipolar). Results For BPSS-P, SVM achieved a fair performance of Cohen's κ of 0.235 (95% CI 0.11–0.361) and a balanced accuracy of 63.1% (95% CI 55.9–70.3) in the 10-fold cross-validation. In the leave-one-site-out cross-validation, the model performed with a Cohen's κ of 0.128 (95% CI −0.069 to 0.325) and a balanced accuracy of 56.2% (95% CI 44.6–67.8). BARS and EPIbipolar could not be predicted. In post hoc analyses, regional surface area, subcortical volumes as well as hyperparameter optimization did not improve the performance. Conclusions Individuals at risk for bipolar disorder, as assessed by BPSS-P, display brain structural alterations that can be detected using machine learning. The achieved performance is comparable to previous studies which attempted to classify patients with manifest disease and healthy controls. Unlike previous studies of bipolar risk, our multicenter design permitted a leave-one-site-out cross-validation. Whole-brain cortical thickness seems to be superior to other structural brain features.

Background Smaller hippocampus volume represents a consistent finding in major depression (MDD). Hippocampal neuroplasticity due to chronic stress might have differential effect on hippocampal subfields. We investigated the effects of the rs1360780 polymorphism of the hypothalamic-pituitary-axis related gene FKBP5 in combination with early life stress (ELA) on the structure of hippocampal subfields in MDD. Methods We assessed the hippocampal subfields volumes in 85/67 MDD/healthy controls. We investigated the effects of diagnosis, FKBP5 allelic status and their interaction as predictors of hippocampal subfield volumes as well as the effect of ELA and its interaction with FKBP5. Results MDD patients had smaller hippocampal volumes, in particular within the cornu ammonis (CA) and dentate gyrus (DG) regions. Patients exposed to ELA had larger hippocampi, in particular within the CA and DG. Among the patients exposed to ELA, the T allele carriers displayed lower volumes within the hippocampus-amygdala-transition-area (HATA) as those subjects homozygous for the C allele. Limitations We pooled the subjects from 2 centers in order to increase the sample size. We did not include the cumulative lifetime exposure to medication. Conclusions Hippocampal volume reductions in MDD were present particularly in the CA and DG. MDD with ELA display differential volume changes compared to MDD without ELA. The significant interaction between ELA and the rs1360780 polymorphism in HATA suggests a role of FKBP5 in the pathophysiology of structural alterations in depression.

Background Aberrant amygdala reactivity to affective stimuli represents a candidate factor predisposing patients with bipolar disorder (BD) to relapse, but it is unclear to what extent amygdala reactivity is state-dependent. We evaluated the modulatory influence of mood on amygdala reactivity and functional connectivity in patients with remitted BD and healthy controls. Methods Amygdala response to sad versus neutral faces was investigated using fMRI during periods of normal and sad mood induced by autobiographical scripts. We assessed the functional connectivity of the amygdala to characterize the influence of mood state on the network responsible for the amygdala response. Results We included 20 patients with remitted BD and 20 controls in our study. The sad and normal mood exerted opposite effects on the amygdala response to emotional faces in patients compared with controls (F1,38 = 5.85, p = 0.020). Sad mood amplified the amygdala response to sad facial stimuli in controls but attenuated the amygdala response in patients. The groups differed in functional connectivity between the amygdala and the inferior prefrontal gyrus (p ≤ 0.05, family-wise error-corrected) of ventrolateral prefrontal cortex (vlPFC) corresponding to Brodmann area 47. The sad mood challenge increased connectivity during the period of processing sad faces in patients but decreased connectivity in controls. Limitations Limitations to our study included long-term medication use in the patient group and the fact that we mapped only depressive (not manic) reactivity. Conclusion Our results support the role of the amygdala-vlPFC as the system of dysfunctional contextual affective processing in patients with BD. Opposite amygdala reactivity unmasked by the mood challenge paradigm could represent a trait marker of altered mood regulation in patients with BD.

Background Early diagnosis of schizophrenia could improve the outcome of the illness. Unlike classical between-group comparisons, machine learning can identify subtle disease patterns on a single subject level, which could help realize the potential of MRI in establishing a psychiatric diagnosis. Machine learning has previously been predominantly tested on gray-matter structural or functional MRI data. In this paper we used a machine learning classifier to differentiate patients with a first episode of schizophrenia-spectrum disorder (FES) from healthy controls using diffusion tensor imaging. Methods We applied linear support-vector machine (SVM) and traditional tract based spatial statistics between group analyses to brain fractional anisotropy (FA) data from 77 FES and 77 age and sex matched healthy controls. We also evaluated the effects of medication and symptoms on the SVM classification. Results The SVM distinguished between patients and controls with significant accuracy of 62.34% (p = 0.005). Participants with FES showed widespread FA reductions relative to controls in a large cluster (N = 56,647 voxels, corrected p = 0.002). The white matter regions, which contributed to the correct identification of participants with FES, overlapped with the regions, which showed lower FA in patients relative to controls. There was no association between the classification performance and medication or symptoms. Conclusions Our results provide a proof of concept that SVM might help differentiate FES patients early in the course of illness from healthy controls using white-matter fractional anisotropy. As there was no effect of medications or symptoms, the SVM classification seemed to be based on trait rather than state markers and appeared to capture the lower FA in FES participants relative to controls.

Introduction Hippocampus (HC) volume reductions have been a consistent finding in major depression (MDD). One possible mechanism that contribute to these reductions might be the inhibition of hippocampal neurogenesis through increased glucocorticoid stimulation following the chronic activation of the hypothalamo-pituitary stress axis (HPA) [1,2]. This is also supported by the fact that the resulting volume reductions can be partially reversed with antidepressant treatment [1,2] Furthermore, preclinical studies suggest that the effect of HPA dysregulation on hippocampal volume reduction, neurogenesis and their reversal by antidepressant response vary across hippocampal regions and are especially present in CA1, CA3 (atrophy of dendrites) and dentate gyrus (neurogenesis) [2]. Compatibly with this hypothesis, genetic variants influencing the endocrine stress axis in MDD, such as the FKBP5 glucocorticoid receptor regulator, were shown to be associated with brain structural changes, especially in the presence of environmental stressors such as early life stress (ELS) [3,4]. In accordance with the above mentioned findings we hypothesized that there might be a differential effect of early life stress on hippocampal subfields volumes in MDD patients with different allelic variants of FKBP5. To answer this question, we investigated whether patients exposed to childhood adversity and carrying the high-risk T allele of the rs1360780 polymorphism of FKBP5 showed smaller hippocampal volumes compared to C homozygous patients. Since CA1, CA3 and the dentate gyrus are especially affected by HPA axis alterations, we decided to focus on these hippocampal subfields. Methods We processed structural MRI data from 65 healthy controls (HC) and 91 MDD patients with automated hippocampal subfields segmentation using Freesurfer 5.3 for whole brain segmentation, followed by hippocampal subfields segmentation using Freesurfer 6 developer version. A quality check was performed using the ENIGMA hippocampal subfield protocols. We selected the following subfields for analysis: granule cells/molecular layer/dentate gyrus (GCMLDG), CA1, CA3, whole hippocampus (WH). ELS was evaluated with the Childhood Trauma Questionnaire Short Form. At least one mild expression of sexual abuse (subscale score > 7), emotional abuse (subscale score > 12) or physical abuse (subscale score > 9) was seen as having had a history of childhood abuse. We predicted the volume of each subfield using generalized linear scale response models, testing for the significance of between-group (diagnosis, ELS, allele status, age, sex, site, total intractranial volume) and within-group (side) factors. Results MDD patients showed reduced volume in all the subfields and in the total HC (p Conclusion In MDD, volume of CA1, CA3, DG and total HC was reduced. Patients with a high-risk variant in a glucocorticoid receptor regulator gene display smaller CA1 volumes in presence of ELS.

BackgroundEarly diagnosis of schizophrenia could improve the outcomes and limit the negative effects of untreated illness. Although participants with schizophrenia show aberrant functional connectivity in brain networks, these between-group differences have a limited diagnostic utility. Novel methods of magnetic resonance imaging (MRI) analyses, such as machine learning (ML), may help bring neuroimaging from the bench to the bedside. Here, we used ML to differentiate participants with a first episode of schizophrenia-spectrum disorder (FES) from healthy controls based on resting-state functional connectivity (rsFC).MethodWe acquired resting-state functional MRI data from 63 patients with FES who were individually matched by age and sex to 63 healthy controls. We applied linear kernel support vector machines (SVM) to rsFC within the default mode network, the salience network and the central executive network.ResultsThe SVM applied to the rsFC within the salience network distinguished the FES from the control participants with an accuracy of 73.0% (p = 0.001), specificity of 71.4% and sensitivity of 74.6%. The classification accuracy was not significantly affected by medication dose, or by the presence of psychotic symptoms. The functional connectivity within the default mode or the central executive networks did not yield classification accuracies above chance level.ConclusionsSeed-based functional connectivity maps can be utilized for diagnostic classification, even early in the course of schizophrenia. The classification was probably based on trait rather than state markers, as symptoms or medications were not significantly associated with classification accuracy. Our results support the role of the anterior insula/salience network in the pathophysiology of FES.

Early diagnosis of schizophrenia might reduce the negative impact of the untreated disease. Progressive functional/structural changes were repeatedly detected using classical between-group statistics. However, these findings have been due to their low sensitivity and specificity not clinically useful. Machine learning methods are able to learn from the data and make predictions on the individual level, which might have a diagnostic potential. We performed a classification of patients with the first episode of schizophrenia (FES) and healthy controls (HC) from the resting state functional connectivity (rsFC) and fractional anisotropy (FA) using machine learning on 1:1 age and sex matched samples of 63/63 patients/HC (rsFC) and 77/77 (DTI). Support vector machine distinguished between patients and controls with an accuracy 73.0% (p = 0.001) (rsFC) and 62.34% (p = 0.005) (DTI). These results were not influenced by symptoms or medication. Our work shows that rsFC and FA might be used to classify patients with FES and HC on the individual level. The classification reflects ‘’trait” markers of FES, as the symptoms and medication had no significant effect. Additionally the results of the analysis of rsFC show the significance of anterior insula in the pathophysiology of FES.

Background: The phenomenology of the clinical symptoms indicates that disturbance of the sense of self be a core marker of schizophrenia. Aims: To compare neural activity related to the self/other-agency judgment in patients with first-episode schizophrenia-spectrum disorders (FES, n = 35) and healthy controls (HC, n = 35). Method: A functional magnetic resonance imaging (fMRI) using motor task with temporal distortion of the visual feedback was employed. A task-related functional connectivity was analyzed with the use of independent component analysis (ICA). Results: (1) During self-agency experience, FES showed a deficit in cortical activation in medial frontal gyrus (BA 10) and posterior cingulate gyrus, (BA 31; P < .05, Family-Wise Error [FWE] corrected). (2) Pooled-sample task-related ICA revealed that the self/other-agency judgment was dependent upon anti-correlated default mode and central-executive networks (DMN/CEN) dynamic switching. This antagonistic mechanism was substantially impaired in FES during the task. Discussion: During self-agency experience, FES demonstrate deficit in engagement of cortical midline structures along with substantial attenuation of anti-correlated DMN/CEN activity underlying normal self/other-agency discriminative processes.

BackgroundSchizophrenia is a chronic disorder with an early onset and high disease burden in terms of life disability. Its early recognition may delay the resulting brain structural/functional alterations and improve treatment outcomes. Unlike conventional group-statistics, machine-learning techniques made it possible to classify patients and controls based on the disease patterns on an individual level. Diagnostic classification in first-episode schizophrenia to date was mostly performed on sMRI or fMRI data. DTI modalities have not gained comparable attention.MethodsWe performed the classification of 77 FES patients and 77 healthy controls matched by age and sex from fractional anisotropy data from using linear support-vector machine (SVM). We further analyzed the effect of medication and symptoms on the classification performance using standard statistical measures (t-test, linear regression) and machine learning (Kernel–Ridge regression).ResultsThe SVM distinguished between patients and controls with significant accuracy of 62.34% (P = 0.005). There was no association between the classification performance and medication nor symptoms. Group level statistical analysis yielded brain-wide significant differences in FA.ConclusionThe SVM in combination with brain white-matter fractional anisotropy might help differentiate FES from HC. The performance of our classification model was not associated with symptoms or medications and therefore reflects trait markers in the early course of the disease.Disclosure of interestThe authors have not supplied their declaration of competing interest.

Background White matter abnormality has been recently proposed as a pathophysiological feature of schizophrenia (SZ). However, most of the data available has been gathered from chronic patients, and was therefore possibly confounded by factors such as duration of the disease, and treatment received. The extent and localization of these changes is also not clear. Methods We examined a population of early stage SZ patients using diffusion tensor imaging (DTI). 77 SZ patients and 60 healthy controls (HCs) were included in the analysis using Tract-Based Spatial Statistics (TBSS). We have also analyzed 250 randomly created subsets of the original cohort, to investigate the relation between the result of TBSS analysis, and the size of the sample studied. Results We have found a significant decrease in fractional anisotropy (FA) in the patient group. This change is present in most major white matter (WM) tracts including the corpus callosum, superior and inferior longitudinal fasciculi, inferior fronto-occipital fasciculus, and posterior thalamic radiation. Furthermore, we identified a clear trend towards an increase in the number and spatial extent of significant voxels reported, with an increasing number of subjects included in the analysis. Conclusion Our study shows that FA is significantly decreased in patients at an early stage of schizophrenia, and that the extent of this finding is dependent on the size of studied sample; therefore underpowered studies might produce results with false spatial localization.

The assessment of cognitive functions represents a crucial step in the diagnostics and therapy of mental disorders such as schizophrenia. To assess spatial learning, mental flexibility and working memory using a test applicable in clinical research, we designed virtual analogue of the spatial animal task: the Morris water maze (MWM). The task of the tested subject was simply to find and remember several hidden goal positions placed in an enclosed circular arena. Data obtained in 30 schizophrenia patients show cognitive deficit in the newly-developed virtual task in comparison to healthy volunteers, indicating its possible application as a tool for diagnostics or remediation of cognitive functions.

NCoR and SMRT are two paralogous vertebrate proteins that function as corepressors with unliganded nuclear receptors. Although C. elegans has a large number of nuclear receptors, orthologues of the corepressors NCoR and SMRT have not unambiguously been identified in Drosophila or C. elegans. Here, we identify GEI-8 as the closest homologue of NCoR and SMRT in C. elegans and demonstrate that GEI-8 is expressed as at least two isoforms throughout development in multiple tissues, including neurons, muscle and intestinal cells. We demonstrate that a homozygous deletion within the gei-8 coding region, which is predicted to encode a truncated protein lacking the predicted NR domain, results in severe mutant phenotypes with developmental defects, slow movement and growth, arrested gonadogenesis and defects in cholinergic neurotransmission. Whole genome expression analysis by microarrays identified sets of de-regulated genes consistent with both the observed mutant phenotypes and a role of GEI-8 in regulating transcription. Interestingly, the upregulated transcripts included a predicted mitochondrial sulfide:quinine reductase encoded by Y9C9A.16. This locus also contains non-coding, 21-U RNAs of the piRNA class. Inhibition of the expression of the region coding for 21-U RNAs leads to irregular gonadogenesis in the homozygous gei-8 mutants, but not in an otherwise wild-type background, suggesting that GEI-8 may function in concert with the 21-U RNAs to regulate gonadogenesis. Our results confirm that GEI-8 is the orthologue of the vertebrate NCoR/SMRT corepressors and demonstrate important roles for this putative transcriptional corepressor in development and neuronal function.