Impaired initiation and slowed execution of movements are key impairments in Parkinson's disease (PD) (Hallett and Khoshbin 1980; Stelmach et al. 1986). Clinical and experimental data suggest that movement difficulties are exacerbated when patients are required to program sequences of movements (Benecke et al., 1987); perform two or more movements simultaneously (Benecke et al., 1986); switch between movement directions (Mak and Hui-Chan, 2002); or perform self-paced, repetitive movements (O'Boyle et al., 1996).#

Current opinion acknowledges that degeneration of nigrostriatal regions, the neuropathological hallmark of idiopathic PD, results in a malfunctioning of frontostriatal pathways, particularly those involving the supplementary motor area (Cunnington et al. 1999; Marsden 1989). In particular, the disruption of basal ganglia output is associated with deficits providing an internal cue to terminate sustained activity (that follows movement execution) in the supplementary motor area to allow for a preparatory phase for upcoming movements or components of movements (Brotchie et al., 1991b), resulting in impaired self-paced (Fattapposta et al., 2000), simultaneous (Benecke et al., 1986), or sequential (Benecke et al., 1987) movements.#

The eye movement system is considered a model motor system with fewer degrees of variability than the limb movement system (for a review, see Leigh et al., 1999). Saccades are rapid eye movements that shift the line of sight to enable foveation of images of interest. There is an extensive literature regarding the generation of saccades and their modulation. Much is known regarding the anatomical substrate, thus, evaluation of saccadic function may well provide particular insights into motor impairments in PD.#

The ability to respond to changes in task demand has yielded mixed findings in PD. In an experimental setting, responses to changes in tasks can be assessed by measuring movements toward an occasional, unpredictable (“oddball”) target that occurs during a sequence of well-learned reciprocating movements. Although one study claimed that PD patients do not experience difficulty altering the direction and/or amplitude characteristics of ongoing movement (Jones et al., 1993), other studies have suggested that patients may have particular difficulty specifying the direction but not amplitude of forthcoming movements (Pullman et al. 1988; Pullman et al. 1990).#

In line with limb movement studies (Ackermann et al. 1997; O'Boyle et al. 1996), the ability to self-pace saccades may also be deficient in PD. External timing signals can aid the initiation and structure of movements in PD (Georgiou et al., 1993), but it is unclear whether external pacing signals used in training can improve the rhythm and accuracy of subsequent self-paced, repetitive movements (Freeman et al. 1993; Konczak et al. 1997). Hence, the ability to self-pace saccades and maintain timing established during an externally cued tracking (training) period was explored in patients with PD. In addition, this study sought to clarify the ability of PD patients to respond to unexpected changes in either the amplitude or direction of the target. Saccades to oddball targets have not been previously explored in PD (Fig. 1).#

We explored the ability of mild to moderately affected PD patients to self-pace saccades at a requested speed of one saccade per second between two continuously illuminated targets before and after an externally cued training period. We also explored the ability to generate saccades to unexpected “oddball” targets, requiring a change in saccade direction and amplitude. Key findings in the PD group were (i) hypometric saccades during the external cueing period, but hypermetric self-paced saccades (a similar pattern of results was noted in the control group); (ii) self-paced saccades (in a smaller sample of PD patients) with increased frequency before external cueing, but performed at a speed of one saccade per second (as requested) after external cueing; (iii) increased variability in terms of saccade accuracy when performing ‘oddball direction’ saccades; and (iv) a significant correlation between symptom severity and intrasubject variability of latencies for saccades to oddball direction targets.#

Our findings indicated that saccades during the externally cued tracking period were hypometric, relative to the self-paced saccades. Although the tracking saccades may be considered externally guided, it is likely that hypometric primary saccade gain, particularly in the PD group, was the result of anticipatory processes that engage striatal–dorsolateral prefrontal cortex pathways (see Crawford et al., 1989; O'Sullivan et al., 1997). In contrast, self-paced saccades are one of the few saccade paradigms that involve purely endogenous saccades that must be internally generated in the absence of any reflexive trigger. Considering associations between hypometria and internally cued movements are well reported in PD (Georgiou et al., 1993), the tendency toward hypermetric self-paced saccades in the PD group was an unexpected finding. One previous study has reported hypermetria in PD patients when they self-paced saccades in the absence of visual targets (O'Sullivan et al., 1997). The authors related overshoot in the PD group to the absence of target visibility. In contrast, our study has demonstrated hypermetria in the presence of a visual target in controls as well as PD patients.#

Martin et al. (1994) also reported overshoot in PD patients in the presence of visual targets during a drawing task that involved repetitive, continuous pen strokes to join dots. The authors interpreted overshoot as a result of deficient synchronizing of cues between basal ganglia output and the supplementary motor area involved in sequential movement preparation (Brotchie et al. 1991ab). If the signal to terminate set-related activity from one movement is defective, this leads to poor preparation for the following movement and results in an endpoint that differs from what was initially planned.#

Interestingly, in our study, both PD patients and controls generated hypermetric self-paced saccades. In accord with our findings, Lemij and Collewijn (1989) reported that when tracking jumping targets, healthy controls generated hypometric saccades, but self-paced saccades between two stationary targets resulted in increased saccade amplitudes. They proposed that the increased time exposed to the target during the self-paced task allowed for greater preparation time to encode the location of the oncoming saccade and, hence, increased accuracy. The fact that both controls and PD patients attained similar primary saccade and final eye position gains for self-paced saccades suggests that the hypermetria may not be a PD impairment. Rather, this finding indicates that self-paced saccade pathways may in fact by-pass the basal ganglia and instead utilize direct frontocollicular pathways (Ferraina et al., 2001). Indeed, positron emission tomography (PET) studies indicated that the supplementary eye fields play a prominent role during self-paced saccades and following repetition or practice, activity in this region increases (Dejardin et al., 1998). Hence, hypermetric self-paced saccades in both the PD and control groups indicate that self-paced saccade accuracy may be determined by direct frontocollicular pathways, which exclude the basal ganglia.#

In their repetitive, continuous drawing task, Martin et al. reported that in addition to hypermetria, PD patients performed movements with an increased speed or ‘hastening’ (Martin et al., 1994). Interestingly, the smaller subset of PD patients who completed our self-paced saccade task before external cueing were also unable to self-pace saccades at the requested rhythm of one saccade per second and instead generated more than one saccade per second. In contrast, the timing of self-paced saccades immediately following the externally cued period (which encouraged saccade tracking at a rhythm of one saccade per second) was similar between the PD and control groups. This finding is consistent with limb studies in PD that have reported beneficial effects of training or practice on the speed and reaction time of upper limb movements during sequential targeting tasks (Behrman et al., 2000) but contrasts with a previous study reporting that withdrawal of external auditory timing cues resulted in marked impairment of patients' rhythm generation (Freeman et al., 1993). Freeman et al.'s (1993) study utilized cross-modality cueing (i.e., synchronizing finger tapping to an auditory cue), whereas cueing in the current study involved a practice period with visual cues. Hence, it is possible that external cues involving the same modality as the task are of greater benefit to PD patients in terms of regulating timing.#

As previously mentioned, when self-paced saccades are performed following practice and repetition, PET studies suggest that activity within the supplementary eye fields increases, while activity in the cerebellum, left superior parietal lobule and left occipital cortex decreases (Dejardin et al., 1998). Although the parameters of the task in the PET study were different to the current study (i.e., saccades were self-paced in darkness, whereas in the current study, saccades were self-paced to continuously illuminated targets) similar principles may apply. That is, while the timing of self-paced saccades in the absence of external cueing may be monitored by the basal ganglia and supplementary eye field connections, the timing of self-paced saccades immediately following an externally cued training period may be mediated primarily through cortical regions (including the supplementary eye fields) and place less emphasis on already compromised striatal regions.#

This study also explored the generation of saccades to unexpected oddball targets. As expected, saccade latencies increased for both groups when they generated saccades to oddball direction targets, as compared to latencies for saccades to expected targets. The variability of latencies for saccades to oddball direction targets was significantly and positively correlated with disease severity (UPDRS scores), suggesting that as the disease progresses, the time taken to respond to changes in target direction becomes more variable. In our previous study, we noted increased errors (i.e., an increased number of inappropriate saccades to the no longer appropriate, but expected, location) in the PD group when required to respond to oddball direction targets (manuscript submitted for publication). Hence, it is likely that increased variability in the latencies of saccades to oddball targets is related to inhibitory deficits within the motor system that probably compromise the selection between competing responses (see Praamstra and Plat, 2001).#

Individual variability of both primary saccade and final eye position gain was significantly increased in PD patients for saccades corresponding to changes in target direction (oddball direction). The ability to accurately respond to unexpected changes in saccade direction is likely to involve a number of regions including the: frontal eye fields (coding saccade metrics) (Gagnon et al., 2002); basal ganglia (facilitate preparation of movement direction) (Postle and D'Esposito, 1999); superior colliculus (directional drive for forthcoming saccades) (Quaia et al., 1999); and, cerebellum (saccade accuracy) (Robinson and Fuchs, 2001). These regions are also involved in target expectation and target selection (Carello and Krauzlis 2004; Gagnon et al. 2002; Hikosaka et al. 2000), processes that play a role during the programming of saccades to oddball targets. In addition, the supplementary eye fields are involved in ocular motor set shifting (i.e., altering saccadic responses to correspond to changes in environmental cues) (Husain et al., 2003). Thus, it is possible that for PD patients, compromised striatal regions (including connections with frontal and collicular areas) may underpin the increased variability in accuracy for saccades corresponding to changes in target direction.#

The PD group was, however, able to generate accurate saccades to oddball amplitude targets. This selective impairment of generating saccades to oddball direction targets is consistent with early limb movement studies that suggested the preparation of movement direction, but not amplitude, may be impaired in PD (Pullman et al. 1988 1990), and a more recent study suggesting deficits switching between movement directions (Mak and Hui-Chan, 2002). Previous studies have also demonstrated that PD patients have difficulty changing set quickly (Chong et al., 2000) and unlike amplitude changes, direction changes may be related to a more general deficit changing set. Taken together, these findings suggest that the basal ganglia are involved in responding to unexpected changes in forthcoming movement direction but not movement amplitude.#

One limitation of this study is that only a small subset of patients completed the self-paced saccade task before external cueing. Nevertheless, our results are consistent with limb studies suggesting that the timing of self-paced movements (without external cueing) is impaired in PD (O'Boyle et al., 1996). Future studies should not only replicate these findings in a larger sample but extend these results to further explore the basis of hypermetric self-paced saccades and determine whether improvements in the timing of self-paced saccades are maintained over time.#

In conclusion, this study has demonstrated that controls and particularly PD patients generated hypermetric self-paced saccades but hypometric saccades during the externally cued tracking period, suggesting different pathways are involved in these different saccade types. Although patients had difficulty regulating the timing of self-paced saccades, they were able to benefit from an externally cued tracking period. It is possible that these beneficial effects on timing of self-paced saccades are mediated through cortical regions (such as the supplementary eye fields), thus placing less emphasis on the already compromised basal ganglia. In addition, while patients were able to respond to unexpected changes in target amplitude, saccades generated in response to changes in target direction were more variable (in terms of latency and accuracy). This finding suggests that the basal ganglia and its connections with the superior colliculus, supplementary and frontal eye fields (regions involved in ocular motor-set shifting, target expectation and target selection), are involved in saccade programming to unexpected changes in target direction, but not amplitude.#