SiNAPS: An implantable active pixel sensor CMOS-probe for simultaneous large-scale neural recordings.

Area of Science:

• Neuroscience
• Electrical Engineering
• Biomedical Engineering

Background:

• Understanding brain function requires large-scale neural recordings with high spatial and temporal accuracy.
• Current complementary metal-oxide semiconductor (CMOS) multi-electrode array probes face scaling limitations, increasing chip area with more recording channels.
• Active Pixel Sensor (APS) technology offers a modular approach with front-end circuits beneath each electrode, but faces constraints in area, power, and noise.

Purpose of the Study:

• To develop and validate an APS CMOS-probe technology for simultaneous neural recording that overcomes scaling issues.
• To achieve high channel counts with reduced chip area, low power consumption, and minimal noise.
• To demonstrate the probe's capability for high-resolution neural activity detection in vivo.

Main Methods:

• Designed and fabricated an implantable single-shaft probe using 0.18 μm CMOS technology with 512 electrode-pixels (28 μm pitch).
• Implemented an APS architecture with in-pixel circuits for simultaneous whole-array read-outs at 25 kHz/channel.
• Conducted extensive bench tests for circuit performance evaluation and in vivo acute recordings.

Main Results:

• Achieved whole-array read-outs at 25 kHz/channel from up to 1024 electrode-pixels.
• Demonstrated an in-pixel gain of 45.4 ± 0.4 dB, input-referred noise of 7.5 ± 0.67 μV_RMS, and power consumption < 6 μW/pixel.
• Successfully resolved and discriminated activity from multiple packed neurons in vivo, both spatially and temporally.

Conclusions:

• The developed APS CMOS-probe technology successfully addresses scaling, area, power, and noise challenges for neural recording.
• This technology paves the way for next-generation compact and scalable active brain recording systems.
• The probe enables high-fidelity sampling of bioelectrical signals, advancing neural circuit analysis.