A Low-Cost and Ultra-Low-Data-Rate Event-Driven Platform for in-Vitro Real-Time Multichannel Neural Signal Acquisition

In vitro multi-electrode array (MEA) recordings are essential tools for neuroscience research, drug screening, and closed-loop neuromodulation. Yet, conventional MEA recording rely on continuous high-speed sampling, generating high data volumes that complicate processing, increase cost, and limit scalability. This work introduces a low-cost, low-data-rate, eventdriven neural acquisition platform that detects neural spikes directly in hardware, encoding them as discrete digital events, retaining only their spatiotemporal information. This architecture removes the need for high-resolution digitization and postprocessing, achieving over 99% data-rate reduction compared to continuous acquisition at 10 kS/s. Under controlled stress-test validation with synthetic signals across 15 simultaneous channels, the event-driven system reached F1-scores above 0.98 at 10 dB signal-to-noise ratio and remained stable up to 300 Hz firing rates, while maintaining sub-millisecond temporal accuracy. The platform can also be reprogrammed for conventional raw waveform acquisition, offering dual-use flexibility. Its modular design allows for straightforward scalability up to 60 channels, front-end upgrades or reconfiguration for other electrochemical measurements without a full hardware redesign. While the event-driven mode is not intended for experiments where spike morphology is critical, it provides an accessible and scalable solution for applications where spatio-temporal spike patterns are the key information, including network synchronization studies, burst detection, pharmacological assays, and long-term closed-loop protocols. By combining affordability, real-time performance, and high data reduction, this system establishes a practical and effective alternative to conventional electrophysiology systems for next-generation neuroscience research.