EFRI BEGIN OI: Organoid-Directed Muscle Control

This award will support a research project to explore how brain-like structures grown in the lab, brain organoids, can be trained to control muscles, seeking to offer a new and efficient way to perform tasks that challenge today’s computers. While artificial intelligence (AI) systems can perform impressive feats, they require substantial amounts of energy and computing power—so much so that some companies aim to repurpose nuclear power plants to support their AI systems. By contrast, the human brain performs complex tasks quickly and efficiently, using only as much energy as a light bulb. This project seeks to harness the power of biological neural networks to create low-power, adaptable computing systems. The research team looks to train organoids to control muscle contractions and compare the organoids’ abilities to traditional AI models. In tandem, the team intends to develop ethical guidelines to ensure this new technology aligns with societal values. If successful, this work could lead to more sustainable computing for use in healthcare, robotics, and national defense, and help prepare society for the challenges and opportunities posed by advanced biotechnologies. This project looks to develop and evaluate a biocomputing platform in which human induced pluripotent stem cell (hiPSC)-derived neuronal organoids are trained to control sustained contraction of ex vivo muscle tissue via a flexible, bi-directional microelectrode interface. A high-density, multi-channel stimulation system seeks to map organoid outputs to muscle activation patterns, enabling real-time control and feedback. Complementary in silico digital twins and machine learning (ML) surrogate models aim to simulate muscle response and assess training efficiency, power use, and scalability compared to synthetic algorithms such as deep and recurrent neural networks. Organoid performance will be benchmarked through metrics including rally length, force-matching accuracy, and ATP-based energy consumption. To evaluate long-term use and adaptability, the project intends to test organoid memory of prior training and assess multi-organoid coordination in dynamic control tasks using simulated joints. Ethical, legal, and social implications (ELSI) will be proactively addressed by developing protocols for sourcing, data ownership, and laboratory practices, and by conducting a Delphi study with a range of societal stakeholders. Together, the technical and ethical components of this work seek to inform the development of biologically based, energy-efficient computing systems that may one day augment or replace synthetic AI in key applications. The project will also include activities that support and retain students and scientists from all career stages through structured mentoring, peer networking opportunities, and the creation of a collaborative, convergent, and transformative research and learning environment. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.