A System-on-Chip for Closed-loop Optogenetic Sleep Modulation.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Electrical Engineering

Background:

• Optogenetics allows targeted neuronal stimulation during specific sleep stages to study sleep mechanisms.
• Closed-loop systems are needed for real-time, stage-specific stimulation in freely moving animals.
• Existing systems often face limitations in power consumption and integration for untethered applications.

Purpose of the Study:

• To design and validate a fully integrated, low-power system-on-chip (SoC) for closed-loop optogenetic sleep studies.
• To enable real-time sleep stage classification and stage-specific optical stimulation in untethered animals.
• To minimize power consumption while maintaining high classification performance.

Main Methods:

• Developed a system-on-chip (SoC) integrating a 4-channel analog front-end for polysomnography signal recording, a mixed-signal machine-learning (ML) core, and a 16-channel optical stimulation back-end.
• Implemented a novel ML algorithm and innovative circuit design techniques for efficient online classification and low power consumption.
• Designed and simulated the SoC using 180 nm CMOS technology.

Main Results:

• The SoC achieved high performance in discriminating 5 sleep stages, with a sensitivity of 0.806 and specificity of 0.947, evaluated on an expert-labeled sleep database.
• Achieved significantly reduced power consumption, with an overall continuous operation power draw of only 97 µW.
• Demonstrated the feasibility of real-time, stage-specific optogenetic control in a low-power, integrated system.

Conclusions:

• The developed SoC provides a powerful tool for advancing sleep research through precise, closed-loop optogenetic manipulation.
• This technology facilitates the study of sleep's mechanisms and effects in naturalistic, untethered conditions.
• The system's low power and high performance pave the way for more sophisticated in-vivo neuroscience experiments.