An Event-Based Neural Compressive Telemetry With >11× Loss-Less Data Reduction for High-Bandwidth Intracortical Brain
IEEE Transactions on Biomedical Circuits and Systems
|March 18, 2024
Summary
This study presents a new neural compressive telemetry (NCT) system for brain-computer interfaces. The NCT significantly reduces data transfer power consumption and improves efficiency for high-resolution neural recordings.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Intracortical brain-computer interfaces (BCIs) provide high spatial and temporal resolution for neural recording.
- Increasing channel counts in BCIs generate massive data, necessitating power-intensive telemetry and risking tissue damage.
Purpose of the Study:
- To develop a power-efficient data compression and telemetry system for high-density neural recordings.
- To overcome the limitations of current data transfer methods in intracortical BCIs.
Main Methods:
- Introduction of an event-based neural compressive telemetry (NCT) system.
- The NCT comprises 8 channel-rotating Δ-ADCs, an event-driven serializer with a ternary address event representation protocol, and an event-based LVDS driver.
- Exploitation of extracellular spike sparsity and high-density recording spatial correlation for data compression.
Main Results:
- Achieved a compression ratio greater than 11.4×.
- Demonstrated exceptional power efficiency of only 1 µW per channel, a 127× improvement over existing technologies.
- Maintained spike waveform fidelity with a normalized RMS error below 23% for low-amplitude spikes (down to 31 µV).
Conclusions:
- The proposed NCT system offers a significant advancement in power efficiency for brain-computer interfaces.
- This technology enables high-density neural recordings with reduced power consumption and minimal impact on signal integrity.
- NCT addresses critical challenges in BCI data transmission, paving the way for more sophisticated and less invasive neural implants.
