 This research proposes an improved paradigm for continuous decoding of hand movement parameters from EG signals. The new paradigm increases the accuracy of prediction and reduces the gap between the prediction and the actual data. Additionally, the proposed adaptive decoder ensemble framework further improves the prediction accuracy. These results suggest that the proposed paradigm and framework are promising for the development of non-invasive BCI systems. This article was authored by Jerome Wong, Luzhing Bai, Wei Jiafei, and others.