We are committed to advancing the state- of-the-art in neural interface technologies through innovative integrated circuit design, edge AI solutions, and biomedical systems engineering.
Our contributions to the SwissChips initiative focus on developing energy-efficient and scalable Systems-on-Chip (SoCs) including:
Analog Front-End Circuits: We aim to design cutting-edge Analog Front-End (AFE) circuits that optimize power efficiency, noise efficiency, area, linearity, and sensitivity. These circuits serve as the backbone for converting real-world signals (e.g., neural activity) into precise electrical representations. By pioneering novel architectures and fabrication techniques, we will enable compact and low-power AFE solutions for neural and biomedical applications.
Flexible Feature Extractors for Edge AI: we are exploring flexible and computationally efficient feature extraction methods tailored for edge AI systems. By focusing on energy-efficient techniques like compressive sensing, signal-adaptive acquisition, on-chip symptom detection and movement decoding, we aim to reduce the data movement bottleneck, improve power efficiency, and support real-time processing in implantable devices. Our designs emphasize low- power operation while maintaining adaptability across diverse biomedical signals and dynamic ranges.
High-Throughput Wireless Readout and Wireless Powering Circuits: To address the challenge of managing high data rates (100+ Mbps) in neural implants, IoT systems, and multi- channel sensor interfaces, we are investigating ultra-low-power IR-UWB transmission systems. We also actively explore advanced powering techniques combining magneto-electric and inductive coupling to enable wireless battery-free devices. These solutions will enable energy- efficient, high-bandwidth wireless data transfer and powering systems, making such devices ideal for battery-free, chronic operation in freely-moving subjects.
AI-Embedded Closed-Loop Stimulation Systems and Brain-Machine Interfaces: We are developing AI-embedded systems for closed-loop neural recording and stimulation and high- performance brain-machine interfacing. Our approach integrates advanced AI algorithms and circuits to create responsive systems capable of real-time neural sensing and stimulation, or movement control and stimulation for sensory feedback. This includes designing low-power, programmable recording and stimulation SoCs for precise therapy delivery in disorders like epilepsy, Parkinson’s disease, depression, and chronic pain.
Related Publications
| A 49.8-mm 2 IR-UWB Transmitter With Co-Designed Power Amplifier and Antenna for Neural Implants With Extended Transmission Range | |||
| Ding, Cong; Gao, Mingxiang; Skrivervik, Anja K.; Shoaran, Mahsa | |||
| 2025-01-29 | IEEE Journal of Solid-State Circuits | ||