Prior work has shown that the interleaving of perceptual, motor, and cognitive components results in a considerable speedup in the performance of a simple decision making task (Veksler, Gray, & Schoelles, 2007). The current modeling effort conducted using the ACT-R cognitive architecture (Anderson & Lebiere, 1998) is intended to demonstrate how this interleaving might be learned, and how decision-making in this task might take place. The model learns the interleaving and exhibits a speedup in performance similar to that of human participants (RMSE=4.3sec). Furthermore, the model matches human accuracy by using a simple heuristic to make decisions.