SCC-LSR: From Technology to Humans: Protecting Users of Neural and Medical Implant Technologies Through Resilience and Safety Engineering

This research will advance community priorities in the areas of safety, security, and human health and wellness pertaining to existing and future neural implant devices. The team includes computer scientists, electrical engineers, MDs, neuroscientists, neural implant community groups, and manufacturers. The partner community groups include patients and their supporting families and caregivers from whom the team will understand the personal impact of neural implant technologies. Expanding engagement with partnered companies who manufacture and medical doctors who use neural implants, the team will address these community challenges and collaboratively deliver a neural implant hardware/software co-design solution that empowers the use and facilitates the safe adoption of emerging neural implant technologies. The research techniques developed for chip security will have applicability to other security related chip-based devices. This project will advance the scientific and technical security and chip hardware design by modeling the operations from a secure and dependable control perspective and developing innovative defense mechanisms that can be applied to emerging smart healthcare devices. Finally, by adding low-resource chip design, the team will advance manufacturing techniques for chip-based devices adding additional security features without impact to overall performance or lifetime. Researchers have demonstrated the impact of “brain-jacking” in mouse models but have not provided solutions. This team’s multipronged project resolves these challenges by leveraging expertise in computer security, sensing, microelectronic design, and strong affiliations with clinical settings. The approaches include the design of (1) verifiably resilient control system and simulations and design upgrades that build upon models of neural sensing and stimulation and explainable AI techniques, (2) automated cybersecurity and resilience testbeds that host physical neural implant devices for fault-injection, side-channel information leakage, and remote connection security analysis, and (3) low-resource (i.e., computational time, power, footprint) intrusion detection with on-chip sensing to continuously monitor anomalies and deter adversarial manipulations. The team will improve the resilience and security of neural implants by protecting system hardware and AI-in-the-loop control software from malicious information injection, disruption in operation, and privacy leakage. A test bed will be developed to validate and test physical devices, see the results from modeling and microelectronic design changes, with real world data collected in lab and translated back to the model for updating and control of realistic interactions. Testing the insertion of malicious intent or unintended interference will allow the team to discern impact and identify potential design changes on the device hardware to prevent future incidents. 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.