An integrated system for multichannel neuronal recording with spike/LFP separation, integrated A/D conversion and...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Electrical Engineering

Background:

• Multichannel neuronal recording is crucial for understanding brain activity.
• Existing systems often face challenges with noise, power consumption, and data handling.
• A need exists for integrated, high-performance front-end processors.

Purpose of the Study:

• To develop and validate a mixed-signal front-end processor for multichannel neuronal recording.
• To enable simultaneous acquisition of Local Field Potential (LFP) and spike (SPK) data.
• To achieve high signal-to-noise ratio and efficient data transmission.

Main Methods:

• Designed a mixed-signal processor with 12 differential input channels.
• Implemented programmable High-Pass Filter (HPF) and Low-Pass Filter (LPF) for signal conditioning.
• Utilized 5-bit calibration Digital-to-Analog Converters (DACs) for offset compensation.
• Integrated a threshold filter for spike segment extraction.
• Fabricated a prototype using 0.35-microm CMOS technology.

Main Results:

• Achieved a low noise level of 3 microV.
• Demonstrated variable amplification rates up to 5000 for SPK and 500 for LFP.
• Enabled analog-to-digital conversion to 10-bit resolution.
• Supported data streaming at up to 8 Mbps over a serial digital bus.
• Successfully tested a prototype system.

Conclusions:

• The developed mixed-signal front-end processor is effective for multichannel neuronal recording.
• The system offers robust signal processing capabilities, including filtering and amplification.
• The prototype demonstrates the feasibility of high-performance, low-noise neural data acquisition.