Implicit sequence learning despite multitasking: the role of across-task predictability

One often replicated finding is that implicit sequence learning is hampered in dual-task situations. Thus, one crucial question has been whether implicit learning processes require attentional resources. Meanwhile, focusing exclusively on limited attentional resources might be considered as too unspecific. Overall, the focus lies now rather on the possibility that the impairment is due to interference coming along with (a) task integration (e.g., Schmidtke and Heuer in Psychol Res 60(1–2):53–71, 1997)—or with (b) parallel response selection (Schumacher and Schwarb in J Exp Psychol Gen 138(2):270–290, 2009). Yet, other explanations have also been put forward—and there is still no agreement. Our goal here is to contribute to this debate by testing several constraints that have been suggested in the literature within one single paradigm, originating by Schumacher and Schwarb (J Exp Psychol Gen 138(2):270–290, 2009). Therefore, we paired the same visual-manual serial reaction time task (SRTT; Nissen and Bullemer in Cogn Psychol 19(1):1–32, 1987) with different auditory-vocal tone-discrimination tasks across seven dual-task conditions. We manipulated (a) its relation to the SRTT and/or (b) the difficulty of response selection. The results suggest that task integration is indeed a crucial factor for implicit sequence learning: since the tone-task is a potential source of noisy patterns of covariation in a complex arrangement of task components, sequence learning is disrupted. In line with Rah, Reber, and Hsiao (Psychon Bull Rev 7(2):309–313, 2000), the usefulness (in terms of sequence learning) of task integration seems to depend on the predictive value of across-task stimulus and/or response events.