 This paper proposes a novel approach for detecting epileptic seizures based on brain rhythm recurrence biomarkers, BRRM, and an optimized neural network, ONASNet. The authors found that ONASNet outperformed other structures by strong learning capability, high stability, small model size, short latency, and less requirement of computing resources. Additionally, they demonstrated that their method was able to achieve higher accuracy than other existing methods when applied to the same dataset and same detection task. This article was authored by Zhenshi Song, Bin Deng, Jian Wang, and others.