Joint principles of motor and cognitive dysfunction in Parkinson’s disease

Highlights

• PD-patients show impaired internal but relatively preserved external motor control.

• We tested PD-patients with a cognitive task (SPT: Serial Prediction Task).

• SPT activates the lateral premotor cortex and SMA.

• PD-patients are impaired in off-state with increasing internal load in SPT.

• Cognitive performance in SPT correlates with motor abilities (UPDRS III).

Abstract

Traditionally, the lateral premotor cortex (PM) is assigned a role in stimulus-driven rather than memory-driven motor control, whereas the opposite holds for the mesial premotor cortex (supplementary motor area, SMA). Consistently, patients with Parkinson’s Disease (PD), in which a specific functional degradation of the mesial loop (i.e., SMA-Striatum) occurs, show impaired memory-driven but relatively preserved stimulus-driven motor control. However, both parts of the premotor cortex are involved in perceptual prediction tasks as well. Here we tested whether the functional bias described on the motor level (i.e., memory-driven/mesial versus stimulus-driven/lateral) can also be detected in perceptual prediction tasks thereby suggesting that PD patients exhibit the same pattern of impaired memory-driven and preserved stimulus-driven control in the cognitive domain. To this end, we investigated 20 male PD-patients “on” and “off” dopaminergic medication while performing a serial prediction task (SPT). A specific modification was implemented to the classical SPT (SPT0) that caused shifts from stimulus- to memory-based prediction (SPT+). As a result, PD patients showed a significantly impaired performance “off” compared to “on” medication for SPT+, whereas no significant “on”/“off”-effects were found for SPT0. Descriptively, the “off”-performance decreased gradually with increasing demands on memory-based prediction. Furthermore, the severity of motor deficits according to the UPDRS III correlated significantly with impaired performance in SPT0 “on” medication. Importantly, an even stronger dependency was found for UPDRS III and SPT+. These findings point to a role of the SMA-striatal loop in memory-driven serial prediction beyond the motor domain.

Introduction

Apart from motor deficits, cognitive impairments have a major influence on the quality of life in Parkinson’s disease (PD) (Schrag et al., 2000, Ziemssen and Reichmann, 2007). Characteristic neuropsychological symptoms of PD such as deficits in attention, working-memory, concept formation, planning, and set-shifting are reminiscent of those detected in patients with prefrontal cortex lesions (Brown and Marsden, 1988, Kulisevsky, 2000, Muslimovic et al., 2005, van Spaendonck et al., 1996) and are therefore often subsumed under the notion of a “dysexecutive syndrome” (Martinez-Horta & Kulisevsky, 2011). In PD, frontal dysfunction is most probably caused by deficient input from the caudate nucleus (Dubois and Pillon, 1997, Saint-Cyr et al., 1988, Taylor et al., 1990) which receives no longer sufficient dopamine projections from the degenerating substantia nigra (Alexander et al., 1986, Dubois and Pillon, 1997, Taylor et al., 1986). Frontal functions may be further deteriorated due to degeneration of the dopaminergic mesocortical pathway emanating from ventral tegmental area (Javoy-Agid & Agid, 1980). In contrast to the caudate-prefrontal loops, the so-called “motor loop” (Alexander et al., 1986) that connects the putamen to the lateral premotor cortex (PM) and the supplementary motor area (SMA), is hardly ever considered as potential origin of cognitive dysfunction in PD. However, evidence has accumulated that some cognitive functions draw particularly on the premotor loops (Jeannerod, 2001, Schubotz, 2007).

In a review addressing PD-associated cognitive impairment, Brown and Marsden (1990) argued that cognitive impairment in PD is present when patients have to rely on internal strategies, whereas performance is preserved when external cues or guidance are provided (e.g., Dubois and Pillon, 1997, Flowers et al., 1984, Flowers and Robertson, 1985). Notably, difficulties in internal guidance and relatively preserved external guidance of behaviour are well-known features of motor control in PD. A striking example of this bias is provided by the phenomenon of “paradoxical kinesis”: Patients who suffer from hypokinesia or akinesia are able to improve their gait with help of external cues like rhythmic auditory stimulation (McIntosh, Brown, Rice, & Thaut, 1997) or visual stimuli such as transversely oriented lines on the walking surface (Azulay et al., 1999, Hanakawa et al., 1999, Martin, 1967).

It has been suggested that the neurofunctional mechanisms underlying paradoxical kinesis may be related to a functional dichotomy in the (pre)motor loops: Goldberg (1985) proposed that the supplementary motor area (SMA) is associated with internally or memory guided processing, whereas the lateral premotor cortex supports externally or stimulus driven processing. This view is largely (but not always, cf. Cunnington et al., 2002, Weeks et al., 2001) in keeping with imaging studies comparing internally to externally guided movements (Debaere, Wenderoth, Sunaert, Van Hecke, & Swinnen, 2003; Heuninckx, Wenderoth, & Swinnen, 2010). In Parkinson’s disease, dopamine depletion is worst in the putamen (Brooks et al., 1990), whose main cortical target is the SMA (Alexander et al., 1986). Accordingly, PD patients performing motor tasks show a decreased blood flow in the SMA and putamen compared to age-matched controls (Playford et al., 1992). In contrast, they exhibit an increased blood flow of the lateral premotor cortex during motor tasks (Haslinger et al., 2001, Samuel et al., 1997). Moreover, administration of Levodopa in PD restores SMA-activation at least to a certain amount and decreases lateral hyper-activation (Haslinger et al., 2001). Lateral premotor activity is significantly higher when patients improve their motor abilities by relying on external cues (Hanakawa et al., 1999). Against the background of these observations, it has been suggested that lateral premotor activity may reflect compensatory processes for reduced SMA function in PD (Hanakawa et al., 1999).

We here aimed at investigating whether the known functional dichotomy of the lateral and mesial premotor cortex for motor tasks, i.e., lateral=stimulus−driven, mesial=memory−driven, holds also for tasks drawing on cognitive functions of the motor system. The serial prediction task (SPT) (Schubotz, 1999) has been shown to activate both the lateral premotor cortex and the SMA in the absence of motor demands (Schubotz & von Cramon, 2003). We modified the SPT in order to parametrically increase dependency on sequence memory, and hence internal guidance. Thus our motivation was to test PD patients (1) in a cognitive task that is known to engage the premotor system, which in turn is known to be particularly impaired in PD patients and (2) to vary the degree to which patients can rely on external cues. By this means we tested to what extent PD patients are able to compensate for occasional absence of prediction-triggering and prediction-confirming stimuli. Moreover, in order to uncover the direct role of dopaminergic supply, we examined the modulatory effect of dopaminergic medication on the described task by comparing the patients’ performance “on” and “off” medication to that of healthy age, gender and education matched control subjects.

In the SPT, subjects monitor a repetitive stimulus sequence that accords to the structure 1-2-3-1-2-3-1-2-3; subsequently they have to indicate in a forced choice mode whether the sequence’s last repetition ended orderly (1-2-3) or not (1-3-2 or 2-1-3). Note that the SPT is a purely cognitive task. In this regard, it clearly differs from otherwise related sequential paradigms such as the serial reaction time task (SRT) (Nissen & Bullemer, 1987). The parametric modification we implemented to the classical SPT (SPT+, hereafter) was a masking of a varying number of stimuli in the sequence (0–4 out of 15) during which subjects are forced to keep track of the correct stimulus order on memory basis.

We hypothesized that, due to a functional degradation of the motor system, (i) PD patients show a deficit in serial prediction when compared to healthy controls, (ii) performance correlates with PD-related motor symptoms (according to UPDRS III), and (iii) dopaminergic medication can restore performance significantly. More importantly, due to the particular detriment in the striatal-SMA-loop in PD, we furthermore expected the impairment of PD patients to be even more prominent when prediction is less regularly informed by external stimuli (i.e., in the SPT+ condition).