Linear-phase delay filters for ultra-low-power signal processing in neural recording implants

Area of Science:

• Electrical Engineering
• Biomedical Engineering
• Signal Processing

Background:

• Neural recording implants require efficient signal processing and power management.
• Traditional digital signal processing in implants incurs significant overhead.
• Analog delay elements offer a power-efficient alternative for neural waveform processing.

Purpose of the Study:

• To design and implement linear-phase delay filters for ultra-low-power signal processing in neural recording implants.
• To utilize these filters for integral waveform extraction and efficient power management.
• To compare different filter realizations for optimal performance.

Main Methods:

• Implementation of continuous-time OTA-C filters with 9th-order equiripple transfer functions.
• Utilizing allpass transfer functions for wider constant-delay bandwidth.
• Comparison of Cascaded and Inverse follow-the-leader feedback filter structures.
• Modeling of parasitics and non-idealities, and transistor-level simulations.
• Experimental validation of the chosen filter topology.

Main Results:

• Demonstration of linear-phase delay filters operating at ultra-low power (200 nA budget).
• Achieved low-distortion delay elements suitable for neural waveform processing.
• Comparison of two filter topologies, assessing their strengths and weaknesses.
• Successful experimental validation of the proposed analog delay filter approach.

Conclusions:

• Linear-phase delay filters are effective for ultra-low-power signal processing in neural implants.
• Analog delay elements offer advantages over digital methods for neural waveform extraction.
• The presented OTA-C filter design provides an efficient solution for power-constrained neural recording devices.