Multimodal fusion of brain signals for robust prediction of psychosis transition.

The prospective study of youths at clinical high risk (CHR) for psychosis, including neuroimaging, can identify neural signatures predictive of psychosis outcomes using algorithms that integrate complex information. Here, to identify risk and psychosis conversion, we implemented multiple kernel learning (MKL), a multimodal machine learning approach allowing patterns from each modality to inform each other. Baseline multimodal scans (n = 74, 11 converters) included structural, resting-state functional imaging, and diffusion-weighted data. Multimodal MKL outperformed unimodal models (AUC = 0.73 vs. 0.66 in predicting conversion). Moreover, patterns learned by MKL were robust to training set variations, suggesting it can identify cross-modality redundancies and synergies to stabilize the predictive pattern. We identified many predictors consistent with the literature, including frontal cortices, cingulate, thalamus, and striatum. This highlights the advantage of methods that leverage the complex pathophysiology of psychosis.

Fusion of biomedical imaging studies for increased sample size and diversity: a case study of brain MRI.

Data collection, curation, and cleaning constitute a crucial phase in Machine Learning (ML) projects. In biomedical ML, it is often desirable to leverage multiple datasets to increase sample size and diversity, but this poses unique challenges, which arise from heterogeneity in study design, data descriptors, file system organization, and metadata. In this study, we present an approach to the integration of multiple brain MRI datasets with a focus on homogenization of their organization and preprocessing for ML. We use our own fusion example (approximately 84,000 images from 54,000 subjects, 12 studies, and 88 individual scanners) to illustrate and discuss the issues faced by study fusion efforts, and we examine key decisions necessary during dataset homogenization, presenting in detail a database structure flexible enough to accommodate multiple observational MRI datasets. We believe our approach can provide a basis for future similarly-minded biomedical ML projects.

The roles, challenges, and merits of the p value.

Since the 18th century, the p value has been an important part of hypothesis-based scientific investigation. As statistical and data science engines accelerate, questions emerge: to what extent are scientific discoveries based on p values reliable and reproducible? Should one adjust the significance level or find alternatives for the p value? Inspired by these questions and everlasting attempts to address them, here, we provide a systematic examination of the p value from its roles and merits to its misuses and misinterpretations. For the latter, we summarize modest recommendations to handle them. In parallel, we present the Bayesian alternatives for seeking evidence and discuss the pooling of p values from multiple studies and datasets. Overall, we argue that the p value and hypothesis testing form a useful probabilistic decision-making mechanism, facilitating causal inference, feature selection, and predictive modeling, but that the interpretation of the p value must be contextual, considering the scientific question, experimental design, and statistical principles.

Differential reinforcement learning responses to positive and negative information in unmedicated individuals with depression.

Major depressive disorder (MDD) is characterized by behavioral and neural abnormalities in processing both rewarding and aversive stimuli, which may impact motivational and affective symptoms. Learning paradigms have been used to assess reinforcement encoding abnormalities in MDD and their association with dysfunctional incentive-based behavior, but how the valence and context of information modulate this learning is not well understood. To address these gaps, we examined responses to positive and negative reinforcement across multiple temporal phases of information processing. While undergoing functional magnetic resonance imaging (fMRI), 47 participants (23 unmedicated, predominantly medication-naïve participants with MDD and 24 demographically-matched HC participants) completed a probabilistic, feedback-based reinforcement learning task that allowed us to separate neural activation during motor response (choice) from reinforcement feedback and monetary outcome across two independent conditions: pursuing gains and avoiding losses. In the gain condition, MDD participants showed overall blunted learning responses (prediction error) in the dorsal striatum when receiving monetary outcome, and reduced responses in ventral striatum for positive, but not negative, prediction error. The MDD group showed enhanced sensitivity to negative information, and symptom severity was associated with better behavioral performance in the loss condition. These findings suggest that striatal responses during learning are abnormal in individuals with MDD but vary with the valence of information.

Identifying neural signatures mediating behavioral symptoms and psychosis onset: High-dimensional whole brain functional mediation analysis.

Along the pathway from behavioral symptoms to the development of psychotic disorders sits the multivariate mediating brain. The functional organization and structural topography of large-scale multivariate neural mediators among patients with brain disorders, however, are not well understood. Here, we design a high-dimensional brain-wide functional mediation framework to investigate brain regions that intermediate between baseline behavioral symptoms and future conversion to full psychosis among individuals at clinical high risk (CHR). Using resting-state functional magnetic resonance imaging (fMRI) data from 263 CHR subjects, we extract an α brain atlas and a ß brain atlas: the former underlines brain areas associated with prodromal symptoms and the latter highlights brain areas associated with disease onset. In parallel, we identify and separate mediators that potentially positively and negatively mediate symptoms and psychosis, respectively, and quantify the effect of each neural mediator on disease development. Taken together, these results paint a brain-wide picture of neural markers that are potentially mediating behavioral symptoms and the development of psychotic disorders; additionally, they underscore a statistical framework that is useful to uncover large-scale intermediating variables in a regulatory biological system.

Baseline reward processing and ventrostriatal dopamine function are associated with pramipexole response in depression.

The efficacy of dopamine agonists in treating major depressive disorder has been hypothesized to stem from effects on ventrostriatal dopamine and reward function. However, an important question is whether dopamine agonists are most beneficial for patients with reward-based deficits. This study evaluated whether measures of reward processing and ventrostriatal dopamine function predicted response to the dopamine agonist, pramipexole (ClinicalTrials.gov Identifier: NCT02033369). Individuals with major depressive disorder (n = 26) and healthy controls (n = 26) (mean ± SD age = 26.5 ± 5.9; 50% female) first underwent assessments of reward learning behaviour and ventrostriatal prediction error signalling (measured using functional MRI). 11C-(+)-PHNO PET before and after oral amphetamine was used to assess ventrostriatal dopamine release. The depressed group then received open-label pramipexole treatment for 6 weeks (0.5 mg/day titrated to a maximum daily dose of 2.5 mg). Symptoms were assessed weekly, and reward learning was reassessed post-treatment. At baseline, relative to controls, the depressed group showed lower reward learning (P = 0.02), a trend towards blunted reward-related prediction error signals (P = 0.07), and a trend towards increased amphetamine-induced dopamine release (P = 0.07). Despite symptom improvements following pramipexole (Cohen's d ranging from 0.51 to 2.16 across symptom subscales), reward learning did not change after treatment. At a group level, baseline reward learning (P = 0.001) and prediction error signalling (P = 0.004) were both associated with symptom improvement, albeit in a direction opposite to initial predictions: patients with stronger pretreatment reward learning and reward-related prediction error signalling improved most. Baseline D2/3 receptor availability (P = 0.02) and dopamine release (P = 0.05) also predicted improvements in clinical functioning, with lower D2/3 receptor availability and lower dopamine release predicting greater improvements. Although these findings await replication, they suggest that measures of reward-related mesolimbic dopamine function may hold promise for identifying depressed individuals likely to respond favourably to dopaminergic pharmacotherapy.

Resting-state brain information flow predicts cognitive flexibility in humans.

The human brain is a dynamic system, where communication between spatially distinct areas facilitates complex cognitive functions and behaviors. How information transfers between brain regions and how it gives rise to human cognition, however, are unclear. In this article, using resting-state functional magnetic resonance imaging (fMRI) data from 783 healthy adults in the Human Connectome Project (HCP) dataset, we map the brain's directed information flow architecture through a Granger-Geweke causality prism. We demonstrate that the information flow profiles in the general population primarily involve local exchanges within specialized functional systems, long-distance exchanges from the dorsal brain to the ventral brain, and top-down exchanges from the higher-order systems to the primary systems. Using an information flow map discovered from 550 subjects, the individual directed information flow profiles can significantly predict cognitive flexibility scores in 233 novel individuals. Our results provide evidence for directed information network architecture in the cerebral cortex, and suggest that features of the information flow configuration during rest underpin cognitive ability in humans.

Gene expression links functional networks across cortex and striatum.

The human brain is comprised of a complex web of functional networks that link anatomically distinct regions. However, the biological mechanisms supporting network organization remain elusive, particularly across cortical and subcortical territories with vastly divergent cellular and molecular properties. Here, using human and primate brain transcriptional atlases, we demonstrate that spatial patterns of gene expression show strong correspondence with limbic and somato/motor cortico-striatal functional networks. Network-associated expression is consistent across independent human datasets and evolutionarily conserved in non-human primates. Genes preferentially expressed within the limbic network (encompassing nucleus accumbens, orbital/ventromedial prefrontal cortex, and temporal pole) relate to risk for psychiatric illness, chloride channel complexes, and markers of somatostatin neurons. Somato/motor associated genes are enriched for oligodendrocytes and markers of parvalbumin neurons. These analyses indicate that parallel cortico-striatal processing channels possess dissociable genetic signatures that recapitulate distributed functional networks, and nominate molecular mechanisms supporting cortico-striatal circuitry in health and disease.

The human cortex possesses a reconfigurable dynamic network architecture that is disrupted in psychosis.

Higher-order cognition emerges through the flexible interactions of large-scale brain networks, an aspect of temporal coordination that may be impaired in psychosis. Here, we map the dynamic functional architecture of the cerebral cortex in healthy young adults, leveraging this atlas of transient network configurations (states), to identify state- and network-specific disruptions in patients with schizophrenia and psychotic bipolar disorder. We demonstrate that dynamic connectivity profiles are reliable within participants, and can act as a fingerprint, identifying specific individuals within a larger group. Patients with psychotic illness exhibit intermittent disruptions within cortical networks previously associated with the disease, and the individual connectivity profiles within specific brain states predict the presence of active psychotic symptoms. Taken together, these results provide evidence for a reconfigurable dynamic architecture in the general population and suggest that prior reports of network disruptions in psychosis may reflect symptom-relevant transient abnormalities, rather than a time-invariant global deficit.

Motivational Context Modulates Prediction Error Response in Schizophrenia.

BACKGROUND: Recent findings demonstrate that patients with schizophrenia are worse at learning to predict rewards than losses, suggesting that motivational context modulates learning in this disease. However, these findings derive from studies in patients treated with antipsychotic medications, D2 receptor antagonists that may interfere with the neural systems that underlie motivation and learning. Thus, it remains unknown how motivational context affects learning in schizophrenia, separate from the effects of medication. METHODS: To examine the impact of motivational context on learning in schizophrenia, we tested 16 unmedicated patients with schizophrenia and 23 matched controls on a probabilistic learning task while they underwent functional magnetic resonance imaging (fMRI) under 2 conditions: one in which they pursued rewards, and one in which they avoided losses. Computational models were used to derive trial-by-trial prediction error responses to feedback. RESULTS: Patients performed worse than controls on the learning task overall, but there were no behavioral effects of condition. FMRI revealed an attenuated prediction error response in patients in the medial prefrontal cortex, striatum, and medial temporal lobe when learning to predict rewards, but not when learning to avoid losses. CONCLUSIONS: Patients with schizophrenia showed differences in learning-related brain activity when learning to predict rewards, but not when learning to avoid losses. Together with prior work, these results suggest that motivational deficits related to learning in schizophrenia are characteristic of the disease and not solely a result of antipsychotic treatment.