Brain Feature Extraction With an Artifact-Tolerant Multiplexed Time-Encoding Neural Frontend for True Real-Time...

Area of Science:

• Biomedical Engineering
• Neuroscience
• Integrated Circuit Design

Background:

• Closed-loop neuromodulation offers targeted neurological treatment but struggles with recording brain activity during stimulation due to artifacts.
• Existing systems cannot continuously monitor treatment effectiveness because stimulation artifacts corrupt neural signals.
• This limitation hinders the development of advanced, adaptive neuromodulation therapies.

Purpose of the Study:

• To develop a novel readout frontend and signal processing technique for artifact removal in neural recordings during closed-loop neuromodulation.
• To enable continuous monitoring of brain activity and treatment effectiveness in real-time.
• To present a power-efficient and compact solution suitable for high-channel-count systems.

Main Methods:

• A rapid-artifact-recovery time-multiplexed neural readout frontend combined with backend linear interpolation was designed.
• A prototype 13-bit incremental ADC was implemented in 180-nm CMOS, focusing on small area and low power consumption.
• The system was tested by measuring artifact-corrupted local field potentials (LFPs), followed by reconstruction and feature extraction.

Main Results:

• The readout frontend achieved a small area of 0.0018 mm²/channel and low power consumption of 4.51 μW/channel.
• It demonstrated a best-in-class total harmonic distortion of -72.6 dB.
• Signal reconstruction using linear interpolation showed relative accuracies above 95% for LFP features compared to artifact-free signals.

Conclusions:

• The proposed hybrid technique effectively reconstructs artifact-corrupted neural signals, overcoming limitations of current neuromodulation systems.
• The compact and power-efficient design is ideal for integration into high-channel-count closed-loop neuromodulation devices.
• This technology facilitates more accurate monitoring and adaptive control in neuromodulation therapies for neurological disorders.