 This paper proposes a novel method called multi-scale temporal self-attention and dynamical graph convolution hybrid network, MTSDGCHN. It uses electroencephalography, e.g., signals to detect strabismus patients, even without real-time communication. The method first uses a multi-scale temporal self-attention module to identify time continuity in e.g. signals. This is followed by a dynamical graph convolution module to capture spatial functional relationships between different e.g. electrodes. Finally, the temporal and spatial features are fed into a classification module to predict the outcome. Experimental results show that the proposed MTSDGCHN achieves better classification accuracy than other methods. This article was authored by Lily Shen, Minyang Sun, Chiwen Xiaoli, and others.