Translational Connectomics: overview of machine learning in macroscale Connectomics for clinical insights.

Connectomics is a neuroscience paradigm focused on noninvasively mapping highly intricate and organized networks of neurons. The advent of neuroimaging has led to extensive mapping of the brain functional and structural connectome on a macroscale level through modalities such as functional and diffusion MRI. In parallel, the healthcare field has witnessed a surge in the application of machine learning and artificial intelligence for diagnostics, especially in imaging. While reviews covering machine learn ing and macroscale connectomics exist for specific disorders, none provide an overview that captures their evolving role, especially through the lens of clinical application and translation. The applications include understanding disorders, classification, identifying neuroimaging biomarkers, assessing severity, predicting outcomes and intervention response, identifying potential targets for brain stimulation, and evaluating the effects of stimulation intervention on the brain and connectome mapping in patients before neurosurgery. The covered studies span neurodegenerative, neurodevelopmental, neuropsychiatric, and neurological disorders. Along with applications, the review provides a brief of common ML methods to set context. Conjointly, limitations in ML studies within connectomics and strategies to mitigate them have been covered.

Prediction of Obsessive-Compulsive Disorder: Importance of Neurobiology-Aided Feature Design and Cross-Diagnosis Transfer Learning.

BACKGROUND: Machine learning applications using neuroimaging provide a multidimensional, data-driven approach that captures the level of complexity necessary for objectively aiding diagnosis and prognosis in psychiatry. However, models learned from small training samples often have limited generalizability, which continues to be a problem with automated diagnosis of mental illnesses such as obsessive-compulsive disorder (OCD). Earlier studies have shown that features incorporating prior neurobiological knowledge of brain function and combining brain parcellations from various sources can potentially improve the overall prediction. However, it is unknown whether such knowledge-driven methods can provide a performance that is comparable to state-of-the-art approaches based on neural networks. METHODS: In this study, we apply a transparent and explainable multiparcellation ensemble learning framework EMPaSchiz (Ensemble algorithm with Multiple Parcellations for Schizophrenia prediction) to the task of predicting OCD, based on a resting-state functional magnetic resonance imaging dataset of 350 subjects. Furthermore, we apply transfer learning using the features found effective for schizophrenia to OCD to leverage the commonality in brain alterations across these psychiatric diagnoses. RESULTS: We show that our knowledge-based approach leads to a prediction performance of 80.3% accuracy for OCD diagnosis that is better than domain-agnostic and automated feature design using neural networks. Furthermore, we show that a selection of reduced feature sets can be transferred from schizophrenia to the OCD prediction model without significant loss in prediction performance. CONCLUSIONS: This study presents a machine learning framework for OCD prediction with neurobiology-aided feature design using resting-state functional magnetic resonance imaging that is generalizable and reasonably interpretable.

Extending schizophrenia diagnostic model to predict schizotypy in first-degree relatives.

Recently, we developed a machine-learning algorithm "EMPaSchiz" that learns, from a training set of schizophrenia patients and healthy individuals, a model that predicts if a novel individual has schizophrenia, based on features extracted from his/her resting-state functional magnetic resonance imaging. In this study, we apply this learned model to first-degree relatives of schizophrenia patients, who were found to not have active psychosis or schizophrenia. We observe that the participants that this model classified as schizophrenia patients had significantly higher "schizotypal personality scores" than those who were not. Further, the "EMPaSchiz probability score" for schizophrenia status was significantly correlated with schizotypal personality score. This demonstrates the potential of machine-learned diagnostic models to predict state-independent vulnerability, even when symptoms do not meet the full criteria for clinical diagnosis.

Leukocyte mitochondrial DNA copy number in schizophrenia.

OBJECTIVE: Schizophrenia is a complex neuropsychiatric disorder with significant genetic predisposition. In a subset of schizophrenia patients, mitochondrial dysfunction could be explained by the genomic defects like mitochondrial DNA Copy Number Variations, which are considered as a sensitive index of cellular oxidative stress. Given the high energy demands for neuronal functions, altered Mitochondrial DNA copy number (mtDNAcn) and consequent impaired mitochondrial physiology would significantly influence schizophrenia pathogenesis. In this context, we have made an attempt to study mitochondrial dysfunction in schizophrenia by assessing mtDNAcn in antipsychotic-naïve/free schizophrenia patients. METHOD: mtDNAcn was measured in 90 antipsychotic-naïve / free schizophrenia (SCZ) patients and 147 Healthy Controls (HC). The relative mtDNAcn was determined by quantitative real-time polymerase chain reaction (qPCR) using TaqMan® multiplex assay method. RESULT: A statistically significant difference between groups [t = 5.22, P < 0.001] was observed, with significantly lower mtDNAcn in SCZ compared to HC. The group differences persisted even after controlling for age and sex [F (4, 232) = 22.68, P < 0.001, η2 = 0.09]. CONCLUSION: Lower mtDNAcn in SCZ compared to HC suggests that mtDNAcn may hold potential to serve as an important proxy marker of mitochondrial function in antipsychotic-naïve/free SCZ patients.

Towards artificial intelligence in mental health by improving schizophrenia prediction with multiple brain parcellation ensemble-learning.

In the literature, there are substantial machine learning attempts to classify schizophrenia based on alterations in resting-state (RS) brain patterns using functional magnetic resonance imaging (fMRI). Most earlier studies modelled patients undergoing treatment, entailing confounding with drug effects on brain activity, and making them less applicable to real-world diagnosis at the point of first medical contact. Further, most studies with classification accuracies >80% are based on small sample datasets, which may be insufficient to capture the heterogeneity of schizophrenia, limiting generalization to unseen cases. In this study, we used RS fMRI data collected from a cohort of antipsychotic drug treatment-naive patients meeting DSM IV criteria for schizophrenia (N = 81) as well as age- and sex-matched healthy controls (N = 93). We present an ensemble model -- EMPaSchiz (read as 'Emphasis'; standing for 'Ensemble algorithm with Multiple Parcellations for Schizophrenia prediction') that stacks predictions from several 'single-source' models, each based on features of regional activity and functional connectivity, over a range of different a priori parcellation schemes. EMPaSchiz yielded a classification accuracy of 87% (vs. chance accuracy of 53%), which out-performs earlier machine learning models built for diagnosing schizophrenia using RS fMRI measures modelled on large samples (N > 100). To our knowledge, EMPaSchiz is first to be reported that has been trained and validated exclusively on data from drug-naive patients diagnosed with schizophrenia. The method relies on a single modality of MRI acquisition and can be readily scaled-up without needing to rebuild parcellation maps from incoming training images.

Construction of population-specific Indian MRI brain template: Morphometric comparison with Chinese and Caucasian templates.

OBJECTIVE: Spatial normalization of brain MR images is highly dependent on the choice of target brain template. Morphological differences caused by factors like genetic and environmental exposures, generates a necessity to construct population specific brain templates. Brain image analysis performed using brain templates from Caucasian population may not be appropriate for non-Caucasian population. In this study, our objective was to construct an Indian brain template from a large population (N = 157 subjects) and compare the morphometric parameters of this template with that of Chinese-56 and MNI-152 templates. In addition, using an independent MRI data of 15 Indian subjects, we also evaluated the potential registration accuracy differences using these three templates. METHODS: Indian brain template was constructed using iterative routines as per established procedures. We compared our Indian template with standard MNI-152 template and Chinese template by measuring global brain features. We also examined accuracy of registration by aligning 15 new Indian brains to Indian, Chinese and MNI templates. Furthermore, we supported our measurement protocol with inter-rater and intra-rater reliability analysis. RESULTS: Our results showed that there were significant differences in global brain features of Indian template in comparison with Chinese and MNI brain templates. The results of registration accuracy analysis revealed that fewer deformations are required when Indian brains are registered to Indian template as compared to Chinese and MNI templates. CONCLUSION: This study concludes that population specific Indian template is likely to be more appropriate for structural and functional image analysis of Indian population.

CHRFAM7A gene expression in schizophrenia: clinical correlates and the effect of antipsychotic treatment.

Earlier studies have implicated CHRNA7, coding α-7 nicotinic acetylcholine receptor (α7 nAChR), and its partially duplicated chimeric gene CHRFAM7A in schizophrenia. However, the relationship between the alterations in peripheral gene expression of CHRFAM7A and severity of clinical symptoms has not been examined. Furthermore, potential influence of the antipsychotic medication on CHRFAM7A expression in drug-naive or drug-free schizophrenia is an unexplored area. CHRFAM7A gene expression in lymphocytes was analyzed in 90 antipsychotic-naïve or free schizophrenia patients using TaqMan-based quantitative RT-PCR. Psychotic symptoms were assessed using Scale for Assessment of Positive Symptoms and Scale for Assessment of Negative Symptoms (SANS). The relationship between psychopathology and CHRFAM7A expression was examined. In addition, measurement of CHRFAM7A gene expression was repeated during follow-up after short-term antipsychotic treatment in 38 patients. There was significant inverse correlation between CHRFAM7A expression and total negative psychopathology score-SANS, and this relationship persisted after accounting for possible confounders such as age, sex and smoking. On exploration of the factor structure of psychopathology using principal component analysis, all the negative symptoms-affective flattening, alogia, apathy, anhedonia and inattention were found to be inversely associated with CHRFAM7A expression. Furthermore, analysis of repeated measures revealed a significant increase in CHRFAM7A expression in patients after short-term administration of antipsychotic medication. Our study observations support the role for CHRFAM7A gene in schizophrenia pathogenesis and suggest a potential novel link between deficient CHRFAM7A expression and negative psychopathology. Furthermore, up-regulation of CHRFAM7A gene expression by antipsychotics suggests that it could be a potential state marker for clinical severity.

In SilicoModel-driven Assessment of the Effects of Brain-derived Neurotrophic Factor Deficiency on Glutamate and Gamma-Aminobutyric Acid: Implications for Understanding Schizophrenia Pathophysiology.

OBJECTIVE: Deficient brain-derived neurotrophic factor (BDNF) is one of the important mechanisms underlying the neuroplasticity abnormalities in schizophrenia. Aberration in BDNF signaling pathways directly or circuitously influences neurotransmitters like glutamate and gamma-aminobutyric acid (GABA). For the first time, this study attempts to construct and simulate the BDNF-neurotransmitter network in order to assess the effects of BDNF deficiency on glutamate and GABA. METHODS: Using CellDesigner, we modeled BDNF interactions with calcium influx via N-methyl-D-aspartate receptor (NMDAR)- Calmodulin activation; synthesis of GABA via cell cycle regulators protein kinase B, glycogen synthase kinase and ß-catenin; transportation of glutamate and GABA. Steady state stability, perturbation time-course simulation and sensitivity analysis were performed in COPASI after assigning the kinetic functions, optimizing the unknown parameters using random search and genetic algorithm. RESULTS: Study observations suggest that increased glutamate in hippocampus, similar to that seen in schizophrenia, could potentially be contributed by indirect pathway originated from BDNF. Deficient BDNF could suppress Glutamate decarboxylase 67-mediated GABA synthesis. Further, deficient BDNF corresponded to impaired transport via vesicular glutamate transporter, thereby further increasing the intracellular glutamate in GABAergic and glutamatergic cells. BDNF also altered calcium dependent neuroplasticity via NMDAR modulation. Sensitivity analysis showed that Calmodulin, cAMP response element-binding protein (CREB) and CREB regulated transcription coactivator-1 played significant role in this network. CONCLUSION: The study presents in silicoquantitative model of biochemical network constituting the key signaling molecules implicated in schizophrenia pathogenesis. It provides mechanistic insights into putative contribution of deficient BNDF towards alterations in neurotransmitters and neuroplasticity that are consistent with current understanding of the disorder.