Bayesian predictive modeling based on multidimensional connectivity profiling

The neurobiology of functional neurological disorders characterised by impaired awareness

We review the neurobiology of Functional Neurological Disorders (FND), i.e., neurological disorders not explained by currently identifiable histopathological processes, in order to focus on those characterised by impaired awareness (functionally impaired awareness disorders, FIAD), and especially, on the paradigmatic case of Resignation Syndrome (RS). We thus provide an improved more integrated theory of FIAD, able to guide both research priorities and the diagnostic formulation of FIAD. We systematically address the diverse spectrum of clinical presentations of FND with impaired awareness, and offer a new framework for understanding FIAD. We find that unraveling the historical development of neurobiological theory of FIAD is of paramount importance for its current understanding. Then, we integrate contemporary clinical material in order to contextualise the neurobiology of FIAD within social, cultural, and psychological perspectives. We thus review neuro-computational insights in FND in general, to arrive at a more coherent account of FIAD. FIAD may be based on maladaptive predictive coding, shaped by stress, attention, uncertainty, and, ultimately, neurally encoded beliefs and their updates. We also critically appraise arguments in support of and against such Bayesian models. Finally, we discuss implications of our theoretical account and provide pointers towards an improved clinical diagnostic formulation of FIAD. We suggest directions for future research towards a more unified theory on which future interventions and management strategies could be based, as effective treatments and clinical trial evidence remain limited.

Age Related Functional Connectivity Signature Extraction Using Energy-Based Machine Learning Techniques

The study of brain connectivity plays an important role in understanding the functional organizations of the brain. It also helps to identify connectivity signatures that can be used for evaluating neural disorders and monitoring treatment efficacy. In this work, age-related changes in brain connectivity are studied to obtain aging signatures based on various modeling techniques. These include an energy-based machine learning technique to identify brain network interaction differences between two age groups with a large (30 years) age gap between them. Disconnectivity graphs and activation maps of the seven prominent resting-state networks (RSN) were obtained from functional MRI data of old and young adult subjects. Two-sample t-tests were performed on the local minimums with Bonferroni correction to control the family-wise error rate. These local minimums are connectivity states showing not only which brain regions but also how strong they are working together. They work as aging signatures that can be used to differentiate young and old groups. We found that the attention network's connectivity signature is a state with all the regions working together and young subjects have a stronger average connectivity among these regions. We have also found a common pattern between young and old subjects where the left and right brain regions of the frontal network are sometimes working separately instead of together. In summary, in this work, we combined machine learning and statistical approaches to extract connectivity signatures, which can be utilized to distinguish aging brains and monitor possible treatment efficacy.

Neural Decoding on Imbalanced Calcium Imaging Data with a Network of Support Vector Machines

We present a novel neural decoding system for calcium imaging data. Miniature calcium imaging is of great utility for examining population neural activity of animals. Our neural decoding system is developed using a carefully-designed support vector machine subsystem together with dataflow-based techniques for system design, which capture the high-level structure of the application and enable powerful system-level analysis and optimization. Also, we introduce a framework for handling imbalanced data. This addresses a problem of imbalanced datasets, which arises commonly in neural decoding applications, as well as in a wide variety of other applications in biomedical engineering and advanced robotics. We developed an ensemble learning based method to tackle this problem. The proposed framework systemically incorporates two heterogeneous model characteristics into a combined model. Through extensive experiments, we evaluate the proposed system using calcium imaging datasets in which neural activities of D1 medium spiny neurons in the dorsal striatum were recorded. The results show that the F 1 score of the proposed system is significantly better than those of previously developed neural decoding systems for calcium imaging.