CAREER: FLEX-MEG: Flexible Quantum MEG with Physics-Aware Compressed Sensing for Scalable, Real-Time Neural Imaging

Understanding how the brain functions require instruments that can measure neural activity safely, accurately, and in everyday environments. Existing brain-imaging systems that detect magnetic signals from neural activity are large, expensive, and require cryogenic cooling, limiting their accessibility and preventing studies of natural behavior. This project will develop a new generation of wearable brain-imaging technology based on flexible magnetic sensors that operate at room temperature and conform to the scalp. These lightweight sensor arrays are designed to measure weak magnetic signals produced by the brain while individuals or animals move freely, enabling more realistic studies of cognition, sensory processing, and behavior. Beyond imaging, the project also explores an integrated approach in which magnetic sensing and noninvasive magnetic stimulation operate together in a closed-loop system, opening the possibility of precision neuromodulation for neurological disorders. The research advances new computational approaches that compress and analyze large volumes of neural data directly within the sensing hardware, enabling real-time interpretation of brain activity without overwhelming data-storage requirements. In addition to advancing neuroscience and medical technology, the project provides interdisciplinary education and workforce training in quantum sensing, artificial intelligence (AI), and biomedical instrumentation. Undergraduate research courses, graduate training programs, and K-12 outreach activities will broaden participation in science and engineering, while industry partnerships will help prepare students for careers in emerging sensing and neurotechnology industries. By enabling accessible, wearable brain-imaging systems and training the next generation of engineers and scientists, this work supports national priorities in advanced sensing, health technologies, and STEM workforce development. The project establishes a flexible quantum magnetoencephalography (MEG) platform based on tri-axial magnetic tunnel junction (MTJ) sensor arrays integrated with physics-aware compressed sensing (PACS) and edge artificial intelligence for scalable, real-time neural imaging. The central hypothesis is that dense, scalp-conformal MTJ arrays, combined with hierarchical in-sensor preprocessing and PACS-enabled reconstruction algorithms, can achieve high-resolution, room-temperature magnetic neuroimaging while maintaining manageable power and data throughput. Task 1 develops strain-tolerant MTJ thin-film stacks, tri-axial magnetic flux-guide architectures, and hierarchical sensor electronics that enable vector magnetic-field reconstruction with low-latency analog preprocessing. Task 2 investigates nanoscale magnetic-to-thermal transduction using integrated micro coil systems to enable magnetothermal neural stimulation, establishing a bidirectional magnetic neural interface that combines stimulation and sensing in vitro. Task 3 validates wearable flexible MEG arrays for in vivo neural imaging in freely moving animal models, benchmarking performance against multimodal ground truth measurements, including electrophysiology, behavioral readouts, and complementary imaging modalities. Task 4 develops PACS-AI algorithms that exploit neural signal sparsity and spatiotemporal structure to enable dynamic sampling, adaptive reconstruction, and real-time inference across large-scale sensor arrays. These efforts jointly advance reconfigurable quantum sensing architectures, closed-loop magneto genetic neuromodulation, and AI-integrated cyber-physical bio interfaces. The resulting framework establishes a scalable pathway toward wearable, label-free neural imaging systems capable of real-time operation, providing new tools for studying distributed brain dynamics and enabling future adaptive brain-machine technologies. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Founda NSF Award ID: 2541981 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Kai Wu | Institution: Texas Tech University, LUBBOCK, TX | Award Amount: $500,000 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2541981 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2541981.html