BioWolf: A Sub-10-mW 8-Channel Advanced Brain-Computer Interface Platform With a Nine-Core Processor and BLE...

Area of Science:

• Neuroscience and Biomedical Engineering
• Embedded Systems and Computer Architecture
• Signal Processing and Machine Learning

Background:

• Brain-computer interfaces (BCIs) are advancing human-machine interaction through electroencephalography (EEG).
• Current BCIs are limited by the lack of powerful, energy-efficient, and compact embedded platforms.
• Existing systems often require bulky hardware, hindering wearable applications.

Purpose of the Study:

• To develop a highly wearable and energy-efficient BCI platform.
• To address the limitations of current embedded systems for real-time BCI processing.
• To demonstrate the feasibility of a compact BCI for practical applications.

Main Methods:

• Developed BioWolf, a compact BCI platform featuring the ultra-low-power Mr. Wolf System-on-Chip with nine RISC-V cores.
• Integrated an 8-channel medical-grade analog-to-digital converter and an ARM-Cortex M4 microcontroller with Bluetooth Low Energy.
• Implemented and tested a BCI using canonical correlation analysis (CCA) for steady-state visual evoked potentials.

Main Results:

• Achieved an average information transfer rate of 1.46 bits/s, comparable to bench-top systems.
• The digital computational platform consumed less power than the analog front-end.
• The total power consumption was 6.31 mW, enabling up to 38 hours of operation with a 65 mAh battery.

Conclusions:

• BioWolf is the first BCI platform to leverage a parallel multi-core architecture for significant energy efficiency in CCA-based BCIs.
• The developed platform offers a wearable, low-power solution for advanced human-machine interaction.
• This design paves the way for more accessible and practical BCI applications.