The acquisition and optimization of movement sequences required, for example, for driving a car or brushing one’s teeth is essential in daily life. It has been suggested that such action sequences are arranged into subsequences ( Sakai, Kitaguchi, & Hikosaka, 2003). Progression through these subsequences might occur by a switching operation, by which, as one subsequence is completed, the representation of this sequence is inhibited and the next one activated ( Hayes, Davidson, Keele, & Rafal, 1998).
In PD patients, both sequence learning and switching between different tasks has been shown to be impaired (Cools et al., 2001 and Woodward et al., 2002). A progressive degeneration of nigrostriatal and, to a lesser extent, of mesocortical dopaminergic neurons is the main pathological feature of PD and leads to a lack of dopamine in the basal ganglia and the prefrontal cortex. This dopaminergic deficit causes not only the classical motor manifestations resting tremor, bradykinesia, rigidity, and postural instability (Grahn, Parkinson, & Owen, 2009) but also other deficits, such as in reinforcement learning, planning, sequence learning and set-switching (Carbon et al., 2003 and Moustafa et al., 2008). Set-switching refers to the changing from one set of rules that guides behavior to another set and is often investigated with the “Wisconsin Card Sorting Task” (Eling et al., 2008 and Hayes et al., 1998). However, Cools, van den Bercken, Horstink, van Spaendonck, and Berger (1984) found evidence for set-switching deficits in PD patients not only in sorting compound stimuli as in the Wisconsin Card Sorting Task, but also in the domain of verbal fluency and motor sequencing. Accordingly, Robertson and Flowers (1990) noted that PD patients made substantially more errors than control subjects when they had to switch between motor sequences.
On the other hand, findings regarding motor sequence learning in PD are mixed, with some studies showing profound impairment in PD patients (Jackson et al., 1995 and Stefanova et al., 2000), and others showing only minor impairment (Ferraro et al., 1993, Pascual-Leone et al., 1993 and Sommer et al., 1999), or none (Smith, Siegert, & McDowall, 2001). This suggests that PD patients can still learn sequences, but less efficiently than normal. It remains unclear, however, which pathophysiological factors influence the sequence learning performance of PD patients. Although it has been suggested that learning performance in PD may be related to the stage of disease, clear evidence for this association is still missing. For example, Muslimovic, Post, Speelman, and Schmand (2007) found a significant, but only weak correlation between the degree of axial disorders and implicit learning impairment by using a one-tailed Spearman’s rho test. Moreover, patients with a higher Hoehn and Yahr stage of disease score showed only a trend towards worse sequence learning (Muslimovic et al., 2007).
Furthermore, in the learning paradigms of Hayes et al., 1998 and Robertson and Flowers, 1990 the motor sequences were prelearned and subjects were aware of the sequence switching. Similarly, in common set-switching tasks such as the Wisconsin Card Sorting Task the subjects try intentionally to identify the rule for stimulus classification even though the set-switching usually occurs unbeknownst to the subjects. In contrast, it is unclear whether PD patients reveal also deficits in switching between two different motor sequences, when they learn these sequences incidentally and are not aware of the sequence switching (Grahn et al., 2009, Hayes et al., 1998 and Woodward et al., 2002). Such a deficit in incidental sequence switching might have a great impact on motor function of PD patients in daily life, where many learning processes occur unconsciously and frequent switching between different action sequences is required.
The aim of this study was thus to clarify the relationship between disease parameters and motor sequence learning in PD and to test whether learning of two different motor sequences in short succession is impaired. By using the same task as in the present study, we have recently shown that healthy subjects can implicitly learn two motor sequences in short succession without significant interference between the sequences (Stephan, Meier, Orosz, Cattapan-Ludewig, & Kaelin-Lang, 2009). We hypothesized that PD patients show more impairment than healthy subjects in learning two sequences in short succession, and that their sequence learning performance correlates with the stage of disease.
