Revolutionizing brain-computer interfaces: Compact and high-speed wireless neural signal acquisition.

Abstract

A brain-computer interface (BCI) facilitates the connection between the human brain and external devices by decoding neurophysiological signals, thereby enabling seamless interaction between humans and machines. However, existing neural signal acquisition systems often suffer from limited channel counts, low sampling rates, and challenges in miniaturization and wireless bandwidth, which restrict their ability to support large-scale and real-time neural recordings. Given the rapid advancements in BCI technologies and the increasing demand for high-resolution neural data, there is an imperative need for BCI systems that are high-throughput, high-speed, and miniaturized. This paper presents a wireless neural signal acquisition system based on FPGA technology, supporting 1024 channels at 32 kSPS and employing a stacked architecture for compact, low-power wireless transmission. Following the creation of the functional prototype, laboratory electrical performance tests were conducted. The system exhibited a noise voltage of 8.56 μVrms, which is in close proximity to the 6 μVrms specified by the chip. In addition, the system accurately captured weak sine wave inputs in both time and frequency domains, confirming its ability to record weak bioelectrical signals. Subsequent animal experiments involving mice implanted with EEG electrodes demonstrated that the system could reliably acquire brain neural signals in real time. The maximum and minimum values of signal-to-noise ratios among the channels were measured at 28.66 and 30.56 dB, thereby providing additional validation for the system's signal quality and consistency.