Poster BB11 - Computational Approach to Identifying Neuropsychiatric Disorder Signatures in Cerebral Organoids using Multi-Electrode Array Analysis

Introduction: Neuropsychiatric disorders, such as schizophrenia (SCZ) and bipolar disorder (BPI), are currently diagnosed using subjective measures, which lack precision and reliability. The complexity of organoid data, characterized by its high-dimensional, dependent, and time-varying nature, presents significant challenges for analysis and interpretation. To overcome these limitations, we propose a novel approach that combines cerebral organoids derived from induced pluripotent stem cells (iPSCs) of SCZ and BPI patients with Multi-Electrode Array (MEA) data analysis using the Sink-Source Index (SSI) method. We hypothesize that distinct electrical patterns or 'signatures' can serve as potential biomarkers for psychiatric disorders and can be identified by applying computational techniques to MEA data from human cerebral organoids.

Materials and

Methods: Cerebral organoids were generated from iPSCs derived from patients with SCZ, BPI, and healthy controls. The organoids were characterized using immunohistochemistry and single-cell RNA sequencing to investigate disease-specific cellular and molecular phenotypes. Functional differences at the mesoscopic level were assessed using MEA recordings, which captured spontaneous electrical activity. The SSI method was employed to analyze the MEA data, distinguishing inhibiting nodes (sources) from inhibited ones (sinks) within the network. Computational techniques, including power spectral density analysis and feature extraction, were applied to identify distinct electrical patterns associated with each disorder.

Results, Conclusions, and Discussions: Our results demonstrate that cerebral organoids derived from SCZ and BPI patients exhibit distinct cellular and molecular phenotypes compared to healthy controls. Immunohistochemistry and single-cell RNA sequencing revealed alterations in neuronal subtypes, glial cells, and synaptic markers, providing insights into the underlying mechanisms of these disorders. Remarkably, our computational analysis of MEA data using the SSI method allowed us to distinguish between organoids derived from control, SCZ, and BPI patients with a minimal prediction accuracy of 89% for SCZ and 88% for BPI. The most informative features for classification included power spectral density, sink index percent, and median of spontaneous activity recorded as spike frequency.Our study presents a groundbreaking approach to identifying electrophysiological signatures of neuropsychiatric disorders using cerebral organoids and MEA data analysis. By combining cutting-edge stem cell technology, electrophysiology, and computational methods, we have demonstrated the potential to uncover novel biomarkers for SCZ and BPI. These biomarkers can enhance diagnostic precision, improve our understanding of disorder mechanisms, and contribute to the development of personalized treatment strategies and predictive cellular models. The ability to distinguish between organoids derived from patients and healthy controls with high accuracy highlights the translational potential of this approach. Future studies should focus on expanding the sample size, investigating additional neuropsychiatric disorders, and validating the identified biomarkers in clinical settings. Our findings pave the way for more accurate diagnosis and targeted treatment strategies, offering hope for improved outcomes for individuals affected by these debilitating disorders.

Acknowledgements (Optional):