29. COMPUTATIONAL APPROACHES TO FACE MULTILEVEL COMPLEXITY IN SCHIZOPHRENIA. (9th April 2019)

• Record Type: Journal Article
• Authors: Do, Kim
• Is Part Of: Schizophrenia bulletin. Volume 45(2019)Supplement 2
• Page Start: S135
• Page End: S136
• Publication Date: 2019-04-09
• Subjects: Schizophrenia -- Periodicals; Schizophrenia -- Research -- Periodicals; 616.898005
• DOI: 10.1093/schbul/sbz022.116
• Languages: English
• ISSNs: 0586-7614

Abstract

Abstract: Schizophrenia faces many challenges due to complexity at multiple levels, from genetic, physiological to symptoms levels. The explosion in genetics information, the development of multimodal functional and structural imaging technics and of ecological momentary assessment, generate large data sets. Computational approaches were proposed to allow the translation of knowledge and understanding between numerous levels of analysis, from subcellular to sociological (Redish and Gordon, 2016). This symposium aims to discuss various computational approaches towards the understanding of psychopathological, neurocognitive as well as predicting functional outcomes of schizophrenia. Dr. Powers will discuss novel findings on specific psychotic symptom dimensions using computational modeling of sensory processing: hallucinations of simple tones produced by audiovisual Pavlovian conditioning occur more frequently in individuals who experience daily hallucinations, engage a network of brain regions previously identified as being important for clinical hallucinations, and occur because of the over-weighting of priors during perceptual inference. Critically, higher-level learning parameters identified using this approach differentiate those with psychosis from those without, regardless of hallucination status. These approaches have relevance to risk stratification in individuals at clinical high risk of developing psychosis and may be useful in identifying differential contributions of neurotransmitter systems to specific computational abnormalities in psychosis. Cognitive deficits are core symptoms of psychotic disorders like schizophrenia, responsible for significant functional impairment and disability. Yet quantifying cognition in schizophrenia remains a challenge in the busy clinical setting. Digital technologies like smartphones represent a novel proxy for relevant cognitive symptoms given the complex social, executive, working memory, psychomotor processing, and attentional domains often involved in using these devices. Dr. Torous will share novel results of smartphone based micro-cognitive measures (MCM) captured in patients with schizophrenia, with MCM defined as a multimodal stream of data variably composed of smartphone metadata, smartphone digital phenotyping data, and smartphone cognitive assessments. Dr. Hess investigated clinical and metabolic features that allow to assess the risk of poor functional outcomes in early psychosis patients (EPP), using state of the art topological analysis. Three clusters of EPP with similar metabolic and/or clinical profiles were identified. The comparison of these clusters allowed to characterize patients' subgroups that displayed distinct social and occupational functioning after 3 years of follow-up. These novel findings suggest that topological analysis of a combination of symptoms and blood metabolic profile might contribute to the prediction of functional outcome at early stages of psychosis. Dr. Koutsouleris will present recent findings from the multi-site European PRONIA project (Personalised Prognostic Tools for Early Psychosis Management). The project's first results highlight the potential of multi-modal machine learning techniques to establish an individualized prediction of clinically relevant outcomes in adolescents and young adults in at-risk or early stages for psychosis or recent-onset depression. He will discuss the challenges of translating these tools to clinical real-world with a focus on how to optimally combine examinations under the premises of accuracy, economic viability, and patient safety.