Exploiting All Programmable SoCs in Neural Signal Analysis: A Closed-Loop Control for Large-Scale CMOS Multielectrode...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computer Engineering

Background:

• Microelectrode array (MEA) systems offer high-resolution recording of neural activity but are limited to open-loop operations.
• Current MEA systems lack real-time feedback capabilities for closed-loop neural interfacing.
• Studying complex neuronal network dynamics requires advanced systems for large-scale neural recording and stimulation.

Purpose of the Study:

• To design and validate a fully reconfigurable FPGA-based processing system for closed-loop, multichannel MEA control.
• To enable low-latency preprocessing and real-time feedback for large-scale neural interfacing.
• To overcome the limitations of open-loop systems in neuroscience research.

Main Methods:

• Developed a processing platform using a Xilinx Zynq all-programmable system on chip.
• Integrated reconfigurable logic and a dual-core ARM processor for low-latency spike detection and filtering.
• Validated the system using ex vivo mouse retina experiments with a 4096-electrode MEA at 18 kHz sampling rate.

Main Results:

• Achieved a total latency below 2 ms from data acquisition to stimulus generation.
• Demonstrated simultaneous recording of light-evoked spikes from thousands of retinal ganglion cells.
• Successfully implemented real-time feedback for closed-loop neural control.

Conclusions:

• The developed FPGA-based system enables low-latency, closed-loop control for large-scale MEA systems.
• This platform facilitates advanced neuroscience research and opens new avenues for biomedical applications.
• Future research can leverage this system for both planar and implantable MEA technologies.