The Neural Engine: A Reprogrammable Low Power Platform for Closed-Loop Optogenetics.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Implantable Medical Devices

Background:

• Brain-machine interfaces (BMI) are advancing towards closed-loop systems for neurological disorders like epilepsy.
• Existing systems require low-power, thermally stable processing for medical implants with long battery life.
• A shift from open-loop to fully closed-loop neural control necessitates efficient, programmable processing units.

Purpose of the Study:

• To develop a low-power neural engine for closed-loop brain-machine interface applications.
• To demonstrate the system's capability in modulating neural activity in vitro and in vivo.
• To assess the system's reliability, power consumption, and thermal performance for implantable use.

Main Methods:

• Development of a low-power neural engine utilizing a power cycling domain.
• Integration with a custom-designed brain implant chip.
• In vitro and in vivo testing, including freely-moving non-human primate and rodent experiments, to evaluate neural modulation and recording performance.

Main Results:

• The system successfully modulated local field potentials at required central frequency ranges in brain tissues.
• Reliable neural recording performance was demonstrated in 24-hour non-human primate and 1-hour rodent in vivo experiments.
• The system operates at 2.93 mA with a 50 Hz sampling rate, offering a lifespan of approximately 56 hours and minimal heating.

Conclusions:

• The developed low-power neural engine is suitable for closed-loop brain-machine interface applications.
• The system's performance, low power consumption, and extended battery life meet critical requirements for medical implants.
• This technology shows significant potential for both neuroscience research and implantable processing units for neurological disorder treatment.