After initial learning, a one-finger key stroke sequence, defined by a specific relative timing pattern (temporal structure) and absolute total movement time (temporal parameter), was practiced (with KR provided) either under dual-task conditions (experimental group), or under single-task conditions (control group). During dual-task practice, the key stroke sequence (i.e., the primary-task) was always executed in parallel to one of two cognitively demanding secondary tasks (subtracting numbers, or sorting marbles). Secondary tasks were alternated every 20 practice trials. Before (Pre-test) and after practice (Post-test), performance in each group was assessed under single-task and under dual-task conditions (no KR during tests). From Pre- to Post-test, primary-task performance in both groups significantly increased (relative timing in particular). Also, after practice dual-task costs found during Pre-test in both groups were still prevalent in the control group, but completely vanished in the experimental group with respect to those task combinations that were practiced before. However, when a new secondary task (repeating letters) was introduced, dual-task costs fully reappeared in the experimental group with respect to relative timing of the key stroke sequence. These results contradict the notion of readily acquiring automatic control in the course of dual-task practice by "Structural Displacement" (Blischke & Reiter, 2002), but they are well in line with the concept of developing cognitive strategies for "Integrated Task Processing" (Manzey, 1993). Thus, impact of dual-task practice on motor sequence production may be different from that on motor parameter control. In this context, implications of recent findings from neuropsychology on cortical systems engaged in the pursuit of concurrent behavioural goals (cf. Charron & Koechlin, 2010) are discussed.