Longitudinal change in antisaccade error rate is associated with TUG performance after STN-DBS in Parkinson’s disease

Introduction: Gait and functional mobility in Parkinson's disease (PD) depend not only on dopaminergic motor circuits, but also on attention, executive control, set-shifting and response inhibition (Xu et al., 2014). Antisaccade error rate is an oculomotor marker of inhibitory control, and increased antisaccade errors have been linked to cognitive control deficits in PD, particularly in patients with gait disturbance or freezing of gait (Walton et al., 2015). Deep brain stimulation (DBS) improves cardinal motor symptoms and reduces medication burden, but gait and axial outcomes remain variable (Brozova et al., 2021). We therefore explored whether antisaccade error rate is associated with longitudinal changes in functional mobility after subthalamic nucleus deep brain stimulation (STN-DBS) in PD patients.Methods: This retrospective exploratory analysis included 11 patients with PD treated with STN-DBS who had complete paired preoperative and 12-month postoperative data for Timed Up and Go (TUG) and antisaccade error rate. Functional mobility was assessed using the DynaPort system (McRoberts). Mobility outcomes included sit-to-stand time (STS), stand-to-sit time (STANDTS) and total transfer time, defined as the combined sit-to-stand and stand-to-sit sequence (STS-total), and TUG. Oculomotor parameters were assessed using the EyeBrain T2 eye-tracking system (SURICOG). Antisaccade error rate was averaged across available eye and direction measurements. Trail Making Test – Part B (TMT-B) was administered by a neuropsychologist and used as a measure of executive function and set-shifting. Change scores were calculated as 12-month follow-up minus baseline. For timed mobility tests and TMT-B, positive change indicated worsening. Spearman’s rank correlations were used to assess associations between changes in antisaccade error rate, functional mobility and TMT-B, due to the small sample size and non-normal data distribution.Results: In the main subgroup (n = 11), TUG worsened from median 8.03 s to 9.83 s (median change +1.96 s; Wilcoxon p = 0.009). Antisaccade error rate did not change significantly at group level (median baseline 32.5 %, median 12-month 28.8 %; Wilcoxon p = 0.700), indicating marked inter-individual variability. However, the 12-month change in antisaccade error rate strongly correlated with 12-month change in TUG (Spearman’s ρ = 0.745, p = 0.008). Baseline antisaccade error rate was not consistently associated with functional change, suggesting that the longitudinal change, rather than baseline error rate alone, was most informative. In the same 11-patient subgroup, baseline TMT-B correlated with baseline antisaccade error rate (ρ = 0.665, p = 0.025), supporting the interpretation of antisaccade error rate as a marker of executive/inhibitory control. TMT-B showed a borderline 12-month worsening (median 79 s to 99 s; median change +6 s; p = 0.050). The association between ΔTMT-B and Δantisaccade error rate showed a non-significant trend (ρ = 0.547, p = 0.082), while ΔTMT-B was not significantly associated with ΔTUG (ρ = 0.387, p = 0.239).Conclusions: Baseline antisaccade error rate did not predict 12-month functional mobility change after DBS. In contrast, worsening antisaccade error rate over 12 months was significantly associated with worsening in TUG performance. Together with the correlation between baseline TMT-B and baseline antisaccade error rate, these preliminary findings suggest that eyetracking may capture executive/inhibitory aspects of complex mobility after DBS that are not fully reflected by conventional timed mobility tests alone. Larger prospective studies are needed to confirm these hypothesis-generating results.Acknowledgements: This research was supported by the Slovak Research and Development Agency (APVV) under contract no. APVV-24-0369.