Consider the problem on sequential change-point detection on multiple data streams. We provide the asymptotic lower bounds of the detection delays at all levels of change-point sparsity and we derive a smaller asymptotic lower bound of the detection delays for the case of extreme sparsity. A sparsity likelihood stopping rule based on sparsity likelihood scores is designed to achieve the optimal detections. A numerical study is also performed to show that the sparsity likelihood stopping rule performs well at all levels of sparsity. We also illustrate its applications on non-normal models.