The swift advancement of machine learning and artificial intelligence has spurred the evolution of a new generation of edge human-computer interaction systems endowed with cognitive abilities. These systems have found widespread applications across diverse domains, such as virtual reality, augmented reality, disease monitoring, intelligent prosthetics, and collaborative operations. To address the requirements for low power consumption and rapid response times in edge human-computer interaction systems, the neuromorphic computing paradigm has surfaced as an optimal solution. Notably, single-spike coding, which condenses information into the precise timing of a solitary spike, exhibits superior energy efficiency and reduced latency when compared to conventional frequency coding. This characteristic renders it exceptionally well-suited for edge computing scenarios. Nevertheless, current implementation strategies relying on CMOS or memristors are plagued by challenges including elevated hardware costs and substantial fluctuations in coding time, thereby constraining the deployment of single-spike edge human-computer interaction systems in environments with limited resources.