In this project, we extract the temporal signature for different motor tasks and show the correlation and dependency among these tasks using spatio-temporal network. As adaptive sequential behavior such as singing a song, locomotion etc is observed in animal behavior, the existing neural model fails to address them because of storing the space and temporal information on the same networks. Hence, in this project, we consider associated cluster-dependent sequence (ACDC) architecture, which adopt the characteristics of spatio-temporal network through distributing the network's weight and dynamics among three different sub-modules named as primary motor cortex (PMC), basal ganglia (BG), and thalamus (A) to address storing the weight in the same networks. For PMC, we choose recurrent network clusters with Hebbian learning mechanisms for weight update, whereas we use nonlinear activation functions for other networks updates rule. We perform training and testing on ACDC architectures for eight different motor actions and observe network dynamics through weight updates. Finally, we extract temporal signature, perform temporal shifting, re-scaling, and different temporal compositionality, which proves the efficacy of the ACDC architecture as well as dependency and correlation among these eight different motor actions.