Approximate Computing-Based Processing of MEA Signals on FPGA

Microelectrode arrays (MEAs) are essential equipment in neuroscience for studying the nervous system’s behavior and organization. MEAs of parallel electrodes that work by sensing extracellular potential neurons their proximity. Processing data streams acquired from is a computationally intensive task requiring parallelization. It performed using complex signal processing algorithms architectural templates. In this paper, we propose approximate computing-based on Field Programmable Gate Arrays (FPGAs), which can be very useful custom implementations neural signals MEAs. The motivation to provide better performance gains system area, power consumption, latency associated with real-time at cost reduced output accuracy within certain bounds. Three types adders explored different configurations develop algorithms. used build systems FPGA then compare them accurate system. All tested real biological same settings. Results show an enhancement speed up 37.6% some without loss accuracy. other systems, area reduction 14.3%. Other trade between speed, accuracy, area.