Cognitive state detection using electroen-cephalogram (EEG) signals for various tasks has at-tracted significant research attention.However,it is difficult to further improve the performance of cross-subject cognitive state detection.Further,most of the existing deep learning models will degrade sig-nificantly when limited training samples are given,and the feature hierarchical relationships are ignored.To address the above challenges,we propose an effi-cient interpretation model based on multiple capsule networks for cross-subject EEG cognitive state de-tection,termed as Efficient EEG-based Multi-Capsule Framework (E3GCAPS).Specifically,we use a self-expression module to capture the potential connec-tions between samples,which is beneficial to alle-viate the sensitivity of outliers that are caused by the individual differences of cross-subject EEG.In addition,considering the strong correlation between cognitive states and brain function connection mode,the dynamic subcapsule-based spatial attention mech-anism is introduced to explore the spatial relation-ship of multi-channel 1D EEG data,in which multi-channel 1D data greatly improving the training effi-ciency while preserving the model performance.The effectiveness of the E3GCAPS is validated on the Fatigue-Awake EEG Dataset (FAAD) and the SJTU Emotion EEG Dataset (SEED).Experimental results show E3GCAPS can achieve remarkable results on the EEG-based cross-subject cognitive state detection un-der different tasks.