Alzheimer’s disease (AD), the most common form of dementia, is a progressive and degenerative brain disease that causes loss of memory or other cognitive impairment. Brain organoids derived from human induced pluripotent stem cells mimic the 3D structure of the brain allowing researchers to effectively model pathology and analyze disease mechanisms in human neuronal cells. We are investigating how well and how early a separation can be made between healthy organoids and organoids with AD. Multi-Electrode Array (MEA) recordings repeatedly taken from these brain organoids represent the neuronal activities over time, whilst being non-destructive to the organoids. Thus, the analysis of these MEA data would reveal informative observations, strengthening the understanding of the pathogenesis and pathophysiology of Alzheimer’s Disease. This study includes an in-depth exploration of the existing computational models employed to analyse similar data, identifying the existing gaps and finally developing novel algorithms to overcome them. The current attempt is also aimed at identifying the response patterns of drugs in AD organoids, to determine when and whether they would show neuronal activities in contrast to healthy organoids treated with or without drugs. Initial feature engineering steps followed by unsupervised clustering of each electrode channel show promising results by separating drugged vs undrugged organoids, while further separating AD organoid channels from healthy organoid channels. Further analysis will be conducted to identify the change in neuronal activities over time in AD organoids in comparison to control organoids. In conclusion, MEA data derived from brain organoids provide a comprehensive experimental premise to investigate the underlying factors of AD.
This research is partially funded by Dementia Australia Research Foundation and the Yulgilbar Foundation.