We refer to the study design paper [1] as well as the primary outcome of the study [3] for a detailed description.

Briefly, the NMS-Nab Study was a mono-centric phase II, randomized, placebo-controlled, double-blind, parallel-group, enriched enrollment withdrawal study in patients with a variety of NMS in PD. A sample of 47 patients with PD with stable motor disease and disturbing NMS defined by a score of ≥4 points on the Movement Disorder Society-Unified PD Rating Scale-I (MDS-UPDRS-I) was included in this study (Figure 1). At the baseline visit open-label nabilone was started with a dosage of 0.25 milligrams (mg) in the evening. During dose titration period, nabilone was titrated in 0.25-mg increments once or twice daily until a maximum dosage of 1 mg twice daily. After this titration phase (phase 1), subjects were either considered responders according to change in the Clinical Global Impression of Improvement Scale (CGI-I) or discontinued from the study (n = 9). The remaining 38 participants were randomized to either optimal nabilone dose or matching placebo (phase 2). After the 4-week, double-blind, withdrawal phase, a termination visit was performed.

Eye-tracking was added as an exploratory outcome measure after study initiation of the original protocol as an amendment in January 2018. Therefore, only a total of 26 out of the 38 patients could be included in this eye-tracking analysis. All subjects performed a Montreal Cognitive Assessment (MoCA), the MDS-UPDRS parts I-IV, in addition a Mini-Mental State Examination (MMSE,) and an eye-tracking battery at the baseline visit as well as at the termination visit (after a 4-week double-blind phase). Eye-tracking was only performed twice in order to minimize learning effects.

Eye movement data were recorded using the Tobii TX-300 eye tracker and the Tobii Pro Lab Software (v.1.83). A sampling rate of 300 Hz was used. Before every task, a 9-point-calibration was done in order to assure a gaze position accuracy of less than 0.4°. Subjects were seated at a 64 cm distance from the presentation screen. The eye-tracking assessment consisted of a prosaccade task, an antisaccade task, a mixed (Go/-NoGo) task, and a mixed pro-, antisaccade task. The centralized fixation cross and the peripheral cues were 0.4° in size. The cues were placed approximately 10° horizontally to each side. The background of the display was set to black and cues were presented in white or in some cases colored depending on the task, in order to achieve maximum contrast. All tests were pseudorandomized in order to assure that all participants experienced the same order of trials. Prior to every task, subjects were orally and visually instructed and performed four test runs.

The prosaccade task consisted of 80 trials in total (40 on each side). A centralized fixation cross (white) appeared on screen. We used an overlap design. Therefore, the peripheral cue appeared after a delay and patients were instructed to look at the peripheral cue only after the central fixation cross disappeared. In order to minimize habituation, we used different delays (500 ms vs. 700 ms).

Secondly, all participants performed an antisaccade task with a step design. The assessment consisted of 20 trials. Cues used were the same as in the prosaccade task (central fixation cross [1500 ms], peripheral round cue [1000 ms], both approximately 0.4° size).

Then, all patients were asked to perform a Go/NoGo task (stop saccades), consisting of 60 trials in total. In this test the direction of the appearing cue was indicated in advance by a flashing green arrow (100 ms). However, in 30% of trials, a red stop signal (200 ms) appeared after the green arrow. In this case, participants were asked to suppress the prosaccade towards the cue. No gap was used.

Lastly, all participants performed a mixed pro- and antisaccade task. The centralized fixation cross (1000 ms) would change color into either blue or green (500 ms), indicating the saccade that had to be initiated by the subject. The test consisted of 40 trials (28 prosaccades, 12 antisaccades). This trial was then repeated with the colors being switched.

Artifact correction and calculation of saccade parameters were carried out by the use of custom-developed interactive analysis software. Reaction times were calculated by the use of linear regression as well as eye-movement interval classification. The automated detection was manually corrected and adjusted if necessary and exported to Microsoft excel files for further analysis. Saccade initiation was defined as an eye movement greater than 2° and a velocity greater than 30°/s. Trials where evaluation was not possible due to a lack of data quality (blinks, eye recognition errors, etc.) were excluded from analysis. Reaction times in milliseconds (ms) were defined by the latency between stimulus onset and saccadic initiation. Furthermore, error rates (%) were calculated. Errors were defined as premature saccades (latency below 70 ms), direction errors, and failure to initiate a saccade. In the Go/NoGo paradigm, saccades that should have been suppressed were considered errors as well.

The study (NMS-Nab Study, EudraCT 2017-000198-86) was approved by the Research Ethics Committee of Innsbruck Medical University, Austria, before study initiation (reference number 1008/2017). All participants provided written informed consent according to the Declaration of Helsinki.

Statistical analysis was carried out using SPSS Statistics 26. Parametric or nonparametric tests were used for statistical analysis depending on the scale type and data distribution. Quantitative values are given in mean ± SD or median and interquartile range. For evaluation of the eye-tracking parameters, a repeated-measures ANOVA was calculated. For variables not meeting the criterion of normal distribution, logarithmic transformation or the min-max method were applied. The level of significance was set at p < 0.05.