Smooth pursuit eye movements and deep brain stimulation
Although DBS in the STN is a highly effective therapeutic intervention for PD, its mechanism and effects on smooth pursuit eye movements have to date been rarely investigated . One previous study from Pinkhardt and colleagues has investigated the effect of STN stimulation on smooth pursuit while the participants received anti-PD medication and found then no significant effect of DBS . However, the present study reveals that when STN stimulation was applied alone after withdrawal of PD medication for 12 hours, DBS significantly improved smooth pursuit performance, both in terms of increasing smooth pursuit velocity gain but also by improving the ability to maintain a steady eye movement velocity without using supportive saccade disruptions or velocity deviations. The improvements made with DBS ON were more detectable at faster target velocities both in smooth pursuit gain and smooth pursuit velocity accuracy. Comparing the study by Pinkhardt and colleagues to the present study there were large differences, both methodological (i.e. different test situations) and between the participants (ours had longer disease duration, i.e., 16.6 versus 10.5 years). Discrepancies between study results may also be caused by differences in electrode locations and DBS settings, which are not always reported [22, 24, 25, 28]. In order to gain an increased understanding of how STN stimulation affects oculomotor performance, it may be beneficiary if forthcoming studies report both electrode locations and DBS-parameter settings since these are of importance for the results. Hence, the recorded effectiveness of applied DBS is likely related to how strong the DBS stimulation is set to be and to what proportional degree the PD symptoms are suppressed by DBS stimulation or anti-PD medication respectively.
The observed, consistently positive, effects of STN stimulation on smooth pursuit performance contrasts with the lower coherency in the effects reported for PD medication. Some have reported that dopaminergic therapy improved smooth pursuit velocity gain [11, 14], others have detected no difference , whereas some have even found that dopaminergic therapy worsened smooth pursuit velocity gain . Hence, more research is motivated to investigate whether STN stimulation can provide a better stable improvement of smooth pursuit performance than PD medication or whether induced performance enhancements are associated with other factors such as state of PD progression.
The relationship between PD and smooth pursuit performance has been a subject of debate. There are a number of reports of decreased smooth pursuit performance in persons with PD [11, 16, 40, 41]. Although abnormal smooth pursuit eye movements have been reported in about 75% of the PD patients , previous studies have been unable to conclude why this is as there appeared to be no direct ocular pursuit pathways traversing the BG . However, new findings made by Yoshida and Tanaka  provide important physiological evidence for the existence of an oculomotor feedback circuit involving the putamen and globus pallidus, structures that have previously been associated in neurophysiological studies primarily with the somatomotor system. Although anatomical studies have previously showed projections from the thalamic nuclei, which receive input from the globus pallidus to the smooth pursuit subregion of the frontal eye field [44, 45], the study by Yoshida and Tanaka showed a relationship between neural activity in the globus pallidus and the pursuit eye movements. Hence, rather than being controlled primarily by areas in extrastriate cortex specialized for processing visual motion, smooth pursuit eye movements involve an extended network of cortical areas, including pursuit-related region in the frontal eye fields and newly identified routes involving structures previously associated with the control of saccades, including the basal ganglia, the superior colliculus, and nuclei in the brain stem reticular formation . Additionally, improvements in smooth pursuit with STN stimulation could be associated with immediate increases in attention brought about by re-activation of dopamine pathways . Attention is a well-known actor in the control of smooth pursuit performance [32, 33]. Interestingly, the STN and globus pallidum have direct connections to the attention network .
Saccadic eye movements and deep brain stimulation
Saccade eye movements were clearly improved by STN stimulation alone, though mainly their amplitude accuracy. These improvements were clearer for smaller saccadic amplitudes (i.e., at 20 and 40 degrees), suggesting that STN stimulation improves the capacity to perform controlled small amplitude saccades more than the capacity to perform large amplitude saccades. One previous study did not detect any effect on saccades when the STN stimulation was turned ON, but their values were close to normal already with anti-PD medication and with the DBS turned OFF . The improved saccadic performance found with DBS ON in the present study is in line with other reports [22–24, 27]. Given, the vast improvements in saccadic performance observed in the present study as well as others, STN stimulation may re-establish direct signals involved in saccadic eye movements, such as to the superior colliculus. One strong possibility is that STN stimulation affects the substantia nigra reticulate-superior colliculi pathway since the striatum has a major outflow of signals via the substantia nigra pars reticulata to the superior colliculus . In line with these findings, previous studies show that in untreated PD, poor smooth pursuit function triggers catch-up saccades  of small amplitude .
The saccade ratio shows that the changes in relationship between peak saccade velocity and saccade amplitude were markedly different in DBS OFF and ON states. As illustrated in Figure 2, although the peak saccade velocity was maintained fairly high, the saccade amplitude was strikingly lower with DBS OFF compared with DBS ON. Activation of DBS partly restored the peak saccade velocity versus saccade amplitude relationship from 19.9 (DBS OFF) to 16.1 (DBS ON), which can be compared with 12.8 in healthy subjects . Our findings confirm previous observations by Rascol and colleagues  of abnormal relationships between saccade velocity and saccade amplitude in persons with PD. This study also revealed that STN stimulation can partly restore a more normal relationship by enhancing the ability to perform controlled larger saccade amplitudes.
Saccade and smooth pursuit latency
Eye movements have sometimes been found hyper-reflexive in patients with PD as compared to healthy controls, possibly due to PD-induced changes in both the peripheral perceptual processing and in central executive mechanisms involving the BG . Noteworthy, the average smooth pursuit latency value found with DBS ON (218 ms) was almost identical to that found in healthy subjects (221 ms) , whereas the latency with DBS OFF was 171 ms. A potential explanation for this could be that STN stimulation might suppress a hyperactive start of the smooth pursuit eye movements and thereby allow the visual target to be captured better at movement onset. However, it is interesting that patients with PD show impaired predictive smooth pursuit eye movements in the early stage , indicating multi-component dysfunction.
In our participants, STN stimulation also caused a small, but non-significant, shortening of saccadic latencies from 177 ms with DBS OFF to 156 ms with DBS ON, which is quite similar to that found in healthy subjects (153 ms) . Similarly, Yugeta and colleagues found that STN stimulation improved saccade latency of memory guided and visually guided saccades .
Relationship between UPDRS motor scores and oculomotor findings
The correlation analysis revealed that many of the oculomotor parameters were significantly related to the UPDRS motor scores (i.e. part III) or showed trends of such relationship. Saccadic latency reduction has previously been shown to correlate with improvements of motor symptoms with the STN stimulation turned ON . This was however not shown in the study by Pinkhardt and colleagues  possibly because of the continued use of anti-PD medications. Lohnes and Earhart evaluated the effects of STN stimulation alone, and their UPDRS III scores correlated with the percentage improvement in saccade amplitude, but not with saccade velocity . Baseline postural instability and gait scores of the UPDRS part III did however correlate with improvements in both saccade amplitude and velocity. Their findings may indicate that oculomotor performance in people with PD relates more to gait and balance problems than motor symptoms per se (i.e. as assessed by the UPDRS part III). This further highlights the importance of taking oculomotor performance into account while simultaneously investigating balance problems in PD.
In the present study, the statistical relationships found between smooth pursuit and saccade performance and the UPDRS motor scores were only present with DBS ON. To speculate, one factor that could substantially influence many scores assessed by UPDRS in DBS OFF state, without influencing oculomotor performance, is tremor. Notably, many subjects had clear tremor with DBS OFF but the STN stimulation almost completely eliminated this symptom in DBS ON in nearly all subjects assessed. Hence, the UPDRS scores may not always properly reflect the oculomotor functions in certain states, as illustrated in this study particularly with DBS OFF and without PD medication. This observation suggests that important functions, such as oculomotor performance, might have to be assessed separately in each state to properly provide detailed supplementary information to the general UPDRS motor scoring, congruent to other reports [22, 23].
This study was performed on a limited number of subjects. The main reason for this was the restricted inclusion criteria applied to avoid interacting and confounding factors. For example, the subjects investigated were within a limited age range and were assessed within a limited time after DBS surgery. Possible confounding or interacting factors such as time after surgery, age at surgery, sickness duration and PD medication at the time of investigation were investigated using correlation analysis. In the material investigated, none of these factors were found to have a systematic influence on oculomotor performance, though such influences cannot and should not be excluded on general basis. Hence, this study was conducted on a well-defined group to ensure that the effects observed were most likely from STN stimulation alone. The statistical evaluation of the assessments revealed consistent effects among the subjects, evidencing that both saccadic and smooth pursuit eye movements were significantly improved by STN stimulation in persons with PD.
A factor difficult to account for is whether STN stimulation also influences attention and cognitive functions. The STN receives projections involved in emotional and cognitive activities from the anterior cingulate, inferior frontal cortex and medial and dorsolateral pre-frontal cortices , and an integrative function has been suggested . Indeed, recent evidence suggests that DBS stimulation of the STN may have some involvement in modifying attentional cerebral networks . Noteworthy, attention is also important for human motor control and for oculomotor performance [52, 53]. Hence, one cannot rule out that improvements in attention and cognition partly account for the improvements recorded in oculomotor performance with STN stimulation. However, given the revealed relationships between several oculomotor parameters and UPDRS motor scores assessed with DBS ON, it is likely that not only is oculomotor function changed by any such attention enhancement induced by STN stimulation, but also motor functions assessed by the UPDRS.
Properly functioning saccadic and smooth pursuit eye movements are vital because these functions bring visual objects or areas of interest into visual focus and keep moving objects in focus. Through these abilities, vision provides exteroceptive information that allows us to interact with a highly dynamic environment and supports feedforward motor control which helps us to anticipate change . Hence, the improvements gained by STN stimulation on oculomotor functions are likely to have positive implications for patients’ ability to perform tasks that rely on visual motor control and visual feedback such as postural control  and reduce side-effects from otherwise present poor oculomotor control like visual distortions and dizziness.
Currently, many evaluations concerning persons with PD (e.g., effectiveness of therapeutic interventions including DBS and progression of the disease) are based on conventional observer-dependent rating scales such as the UPDRS. This study illustrates that analysis of oculomotor functions might provide an additional information source when evaluating the effects of interventions like STN stimulation. The largely automatic procedures used when performing the assessments and analyzing the data ensures that the patients can be objectively assessed repeatedly using a high resolution evaluation scale.
The positive influence of STN stimulation on the ability to perform accurate saccadic and smooth pursuit eye movements was most clearly illustrated by the novel smooth pursuit velocity accuracy analysis and the saccadic ratio parameter, the latter commonly called main sequence [36, 37].
The novel smooth pursuit velocity accuracy parameter provides information about to what degree a steady smooth pursuit velocity can be maintained without supportive saccades or velocity inaccuracies beyond a 20% velocity error limit. As evidenced in two independent studies [32, 33] and the phenomena itself previously described in other studies [56, 57], one of the first signs of an affected oculomotor function might be an inability to maintain steady and accurate control of smooth pursuit movements within acceptable velocity boundaries over longer periods of time. Furthermore, Armstrong suggests that this method might prove successful for evaluating PD . In the present study and in previous studies [11, 16, 40, 41], the traditional smooth pursuit gain analysis also provided significant statistical evidence for an altered smooth pursuit function, but not with the same high statistical sensitivity as the velocity accuracy parameter. Hence, analysis of smooth pursuit velocity accuracy may provide better means to detect oculomotor deficits at an earlier stage in a number of disorders and in persons with PD where other functional parameters may not.
The saccade ratio parameter illustrates whether there is an imbalance between saccade velocity and the saccade amplitude. Previous reports have shown a relatively fixed relationship between the saccade amplitude and saccade velocity up to about 15–20 degrees amplitude. However, above these saccade amplitudes the relationship changes in a non-linear manner [32, 36, 37]. Recent reports have shown that this relationship can be changed by various factors with central effects such as alcohol intoxication and sleep deprivation [32, 33]. Such saccade relationship changes between saccade velocity and the saccade amplitude were also detected in persons with PD when using the saccade ratio parameter to assess the effects of DBS in the STN.
The possible relationship between UPDRS scores and oculomotor findings were investigated by performing correlations to each of the oculomotor parameters. One should be aware of that multiple analyses can increase the risk of statistical Type 1 errors but also that the Bonferroni correcting factor under some conditions could increase the risk of making Type II errors . Specifically, it is difficult to present multiple systematic relationships in the p-value range of <0.05 to >0.01 without all of them being down-ranked to insignificant level by Bonferroni. Contradictory, if publishing one finding at a time, as singled-out observations, then they are individually regarded as significant evidence for the hypothesis. Moreover, as described by Perneger , if Bonferroni was applied in clinical practice the effects would be absurd. Hence, it is important that systematic investigations can be published, i.e., both significant and insignificant findings, because both kinds of findings are often of direct clinical relevance. Accordingly, we present the complete statistical correlation analysis of all oculomotor parameters so the systematic patterns are displayed, though we do so with a reminder of being aware of the problems associated with Type I and Type II errors.
STN stimulation from DBS alone significantly improves smooth pursuit and saccade performance in patients with PD. The improved oculomotor functions with DBS in the STN may have positive implications for patients’ ability to perform tasks that rely on visual motor control and visual feedback (such as postural control) and reduce side-effects from poor visual control like visual distortions and dizziness. The findings further indicate a critical but complex role of the basal ganglia and STN on oculomotor functions. Moreover, the new oculomotor analysis methods used provide a sensitive vehicle to detect functional enhancements produced by STN stimulation from DBS alone. These methods could be useful in detecting subtle pathological changes from PD. The largely automatic procedures used when performing the assessments and analyzing the data ensures that the patients can be objectively assessed repeatedly using a high resolution evaluation scale.