Abstract
Fifteen individuals with Parkinson’s disease (PD) and impulsive compulsive behaviours (PD+ICB) were compared to 15 PD patients without ICBs (PD-ICB) and 15 healthy controls (HC) on a pro-saccades and an anti-saccades task to assess if ICBs are associated with distinct saccadic abnormalities. PD+ICB made shorter saccades than HC and more direction errors in the anti-saccades task than PD-ICB and HC, suggesting that patients with ICBs have greater difficulty in suppressing automatic saccades towards a given target. Saccadic assessment has the potential to evolve into a marker to guide therapeutic decisions in patients at risk of developing ICBs.
Keywords
INTRODUCTION
The motor abnormalities in Parkinson’s disease (PD) are frequently reflected in saccadic movement in the form of hypometria and increased reaction time (latency) [1, 2]. While automatic saccades appear to be preserved in PD [3], voluntary saccades are impaired [4–6]. Although treatment with levodopa can improve some of these abnormalities [5], anti-saccadic reaction time and direction errors worsen as PD progresses.
Impulsive compulsive behaviours (ICBs) such as the dopaminergic dysregulation syndrome, hypersexuality, pathological gambling, compulsive shopping and punding, affect 16% of patients with PD [7]. Anti-saccadic error rate has been associated with impulsivity in healthy controls (HC) [8], but to date no studies have assessed eye movements of individuals with PD and ICBs (PD+ICB). The presence of motor and reflection impulsivity in PD+ICB [9] would predict that premature saccades and anti-saccade direction errors are increased in these individuals. To answer this question, we studied pro and anti-saccades of PD+ICB and compared the results with individuals with PD without ICBs (PD-ICB) and HC.
ICBs are usually underreported by PD patients [10] and distinct saccadic abnormalities in PD+ICB may represent a novel way of identifying these behavioural abnormalities, which would be of clinical value.
MATERIAL AND METHODS
Fifteen PD+ICB were matched with 15 PD-ICBs and 15 HC. Diagnosis of ICBs was based on the Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease - Rating Scale (QUIP-RS) [11] and confirmed with an interview. Patients with Montreal Cognitive Assessment (MoCA) score <26 were excluded. For details on other scales/questionnaires used see Supplementary materials. PD patients using levodopa were tested one hour after intake. Eyetracking was conducted with an e(ye)BRAINcopyrightT2 device (SuriCogcopyright). The research was approved by the local ethics committee. Two paradigms were created:
Pro-saccade task
Participants started focusing on a central grey dot and made a saccade to a blue dot appearing eccentrically (15 degrees horizontally) on the screen, displayed for 1500 ms.
Anti-saccade task
Participants were instructed to focus on a central grey dot. Immediately after it disappeared, a red dot appeared randomly on either side of the screen 15 degrees horizontally. Participants were instructed to look to the opposite side of the screen and return their eyes to the central fixation point after the red dot disappeared, 1500 ms later.
Forty randomized trials were conducted to each side, totalling 80 trials. Direction errors, saccadic movements towards the opposite direction of the visual stimulus, were computed as total values and proportion of errors in relation to the total number of detected saccades. Data points outside the interquartile ranges were excluded. Due to the possibility that saccades with latency >900 ms or <120 ms were not related to stimulus presentation, these were excluded from the analysis. Saccadic parameteres were corrected for multiple comparisons using the Benjamini-Hochberg [12] method. A p value <0.05 was considered significant. For statistical analysis details, see the Supplementary Material.
RESULTS
There were more males in the PD+ICB (87%) and PD-ICB (80%) groups compared with HCs (47%). PD+ICB scored higher on the QUIP-RS, AIMS and UPDRS III. None of PD+ICB were receiving drugs for ICBs and no PD patients were using anticholinergics. For demographic/clinical data see Table 1, for types of ICBs see Supplementary materials.
Demographic and clinical characteristics divided by group
Demographic and clinical characteristics divided by group. SD, standard deviation; PD, Parkinson’s disease; DA, dopamine agonist; LEDD, levodopa equivalent daily dose; MAOI, monoamine oxidase inhibitor; QUIP-RS, Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease - Rating Scale; UPDRS III, Unified Parkinson’s Disease Rating Scale part III; AIMS, Abnormal Involuntary Movements Scale; MoCA, Montreal Cognitive Assessment; FAB, Frontal Assessment Battery; ms, milliseconds; IQR, interquartile range. *Chi-square test; λ ANOVA; ‡Independent samples t-test; φ Kruskal-Wallis test; β Fisher’s exact test. Significant results in bold. Results expressed in mean values and standard deviation and total values and proportions.
The number of outlying data points excluded and premature saccades was similar between groups in both tasks. We found no difference between centripetal and centrifugal saccadic amplitudes in the pro-saccade task, and no differences in saccadic errors in the anti-saccade task between centripetal and centrifugal saccades, across all groups. As centripetal saccades are more influenced by visual determinants [13], we henceforth report centrifugal saccade data only. There were no differences in the latency or amplitude of pro-saccades, nor anti-saccadic latency between groups. Anti-saccade amplitude was similar between groups (Table 2), but a post-hoc analysis showed decreased amplitude in PD+ICB compared to HC (p = 0.021, Mann-Whitney), but no difference between PD+ICB and PD-ICB (p = 0.110) or between PD-ICBs and HC (p = 0.097).
Results of the pro-saccades and anti-saccades tasks
Eye movements characteristics divided by group. Direction errors are displayed as average number of direction errors per group (N) and as proportion of direction errors in relation to the total number of detected saccades (%). SD, standard deviation; PD, Parkinson’s disease; ms, milliseconds. λ ANOVA; φ Kruskal-Wallis test; Significant results in bold. Benjamini Hochberg correction for multiple comparisons used for all results. Results expressed in mean values and standard deviation (SD) for variables with normal distribution or mean values and median and interquartile range (IQR) for variables with non-normal distribution.
Anti-saccadic error rate differed between groups (Table 2). Post-hoc analysis revealed that PD+ICB made more errors than PD-ICB (p = 0.006) and HC (p = 0.001), but there was no difference between PD-ICB and HC (p = 0.802). Anti-saccadic reaction time for accurate saccades and direction errors are found in Supplementary materials.
We also analysed fixation data as this is relevant to interpretation of saccadic function. The number and duration of fixation deviations from the target (fixation instability) were greater in PD+ICB compared to PD-ICB, but this was not statistically significant (p > 0.1). There were no correlations between fixation instability and anti-saccadic errors, age, disease duration, UPDRSIII, AIMS or QUIP-RS.
DISCUSSION
This is the first study of saccades in PD+ICB. Corroborating previous findings showing preserved automatic saccades in PD [3–6], ICB does not influence pro-saccadic behaviour. The superior colliculus (SC) is the point of convergence for cortical and subcortical structures that influence saccadic control [1] and is modulated by the basal ganglia [14]. This structure is under tonic inhibiton from the substantia nigra pars reticulata (SNr). Cortical visual signals are initially directed to the caudate, which connects to the SNr via direct and indirect pathways. The former inhibits the SNr and release the SC to perform a saccade, whereas the latter increases SC inhibition preventing the generation of saccades towards ‘valueless’ objects [14, 15]. The SNr is affected later in PD [16], therefore integrity of such pathways in early PD could explain the preservation of pro-saccades. It is unlikely that dopaminergic medication contributed to this as levodopa slows reaction time of pro-saccades [5, 17]. However, it is possible that the sample size was insufficient to detect subtle differences.
Contradictory data on anti-saccadic reaction time in PD has been published [3]. Here, all patients were tested during ON which may explain the lack of differences in anti-saccadic latencies. The amplitude of anti-saccades, however, was lower in PD+ICB compared to HC. We were unable to replicate findings of previous reports which found that PD-ICB have hypometric saccades [3]. As previously described in a PD population, saccadic hypometria is unaffected by levodopa use [5].
More premature saccades were made in the pro-saccades than the anti-saccades task. It has been shown that anticipatory saccadic movements can occur with predictable tasks similar to our pro-saccade paradigm [18]. Increased anti-saccadic error rate has been reported in drug-naïve PD [19], but the error rate found here was higher than previously reported, possibly due to different assessment protocols. Whereas PD+ICB made more direction errors than both groups, there were no differences between PD-ICB and HC. This is unlikely to be related to abnormalities of fixation, given the similar rates of fixation instability across both groups. PD+ICBs have reflection impulsivity, temporal discounting and a bias towards risky choices in decision-making tasks [9]. However, considering the short time between target onset and saccadic movement and the low number of premature saccades, it is unlikely that decision-making abnormalities are behind the higher error rate in PD+ICB. Previous studies show that correct performance in the anti-saccadic task requires top-down inhibition of neurons in the SC before target onset [1]. PD-related dopaminergic depletion in the dorsolateral prefrontal cortex coupled with deficits in cortical inhibitory circuits in PD+ICB [20] may explain the failure to suppress an automatic saccade [21].
An important caveat is that PD+ICB had higher UPDRS scores, which could contribute to increased anti-saccadic error rate; both anti-saccadic error rate and reaction time increase as PD progresses [22, 23]. However, there are important differences between our study and previous reports of saccadic abnormalities in advanced PD. Firstly, in our study PD+ICB exhibited a significantly higher error rate (48.62%) compared to the literature (36.2%) [23]. Secondly, in our study the anti-saccadic reaction time was also shorter (318.6 ms) than a previous report (410 ms) [23]. Thirdly, severity of bradykinesia has been correlated with longer anti-saccadic latencies [23] but we have found no differences in reaction time between PD+ICB and PD-ICB. Lastly, direction errors were not correlated with QUIP-RS, AIMS nor UPDRS scores. Therefore, although the UPDRSIII suggests that PD+ICB have more advanced PD, the anti-saccades data do not corroborate this. Although the UPDRSIII score influences performance in PD-ICB [23] it is likely that cognitive impulsivity is contributing to poor performance in PD+ICB.
One limitation of this study is the small sample size. This was addressed by sampling 80 saccades in each task, ensuring patients were offered breaks between tasks to avoid fatigue. There were more female HC; previous data show that saccadic parameters do not differ between sexes [24], although one study reported a higher anti-saccadic error rate in women compared to men [25]. However, the 20% error rate of HC in our study is lower than the findings of that study.
This is the first study of saccades in PD+ICB, who made hypometric anti-saccades and a higher number of anti-saccade direction errors. This finding may have important clinical implications if antisaccadic error rate could be confirmed as a marker for ICBs. Future studies should investigate whether PD+ICB are less able to inhibit saccades to less salient stimuli. Tasks with short preparation times, or which present sensory information before motor choice, could help understanding decision-making in such a short time frame and the health of the frontal inhibitory circuits modulating it [26, 27].
CONFLICT OF INTEREST
Funding sources and conflict of interest regarding this manuscript
All authors have nothing to disclose.
Financial disclosure for the previous 12 months
Dr Barbosa has a consultancy agreement and received a grant from Britannia Pharmaceuticals, and is currently supported by a Grant from Conselho Nacional de Desenvolvimento Cientifico e Tecnologico (Brazilian National Council for Scientific and Technological Development). Dr Kaski is supported by the National Institute for Health Research University College London Hospitals Biomedical Research Centre. Prof Lees is a consultant for Britannia Pharmaceuticals and BIAL Portela, and received honoraria from Britannia, Novartis, Teva, Meda, Boehringer-Ingelheim, GSK, Ipsen, Lundbeck, Allergan, Orion, BIAL, Noscira, Roche, UCB, NordcInfu Care, Windrose Consulting Group. Prof Warner received honoraria from Britannia Pharmaceuticals, TEVA and Lundbeck.
Dr Djamshidian received honoraria from TEVA and UCB. Ms Castro has nothing to disclose.
Footnotes
ACKNOWLEDGMENTS
The authors would like to thank Dr Saiful Islam for the statiscal advice and the Reta Lila Weston Institute of Neurological Studies for the support received during this research project. Pedro Barbosa is supported by a grant from Brazilian Conselho Nacional de Desenvolvimento Científico e Tecnológico(CNPq).
