249 Full text: 2804 Number of figures/tables: 4 Number of references: 39 Abstract Background: Beta-based adaptive Deep Brain Stimulation (aDBS) is effective in Parkinson's disease (PD), when assessed in the immediate post-implantation phase. However, the potential benefits of aDBS in chronically implanted patients, in whom the benefits of the microlesion effect have disappeared, are yet to be assessed. Methods: To determine the effectiveness and side-effect profile of aDBS in PD compared to conventional continuous DBS (cDBS) and no stimulation (NoStim) in the chronically implanted state. 13 PD patients undergoing battery replacement were pseudo-randomised in a crossover fashion, into three conditions (NoStim, aDBS or cDBS). Patient videos were blindly evaluated using a short version of the Unified Parkinson's Disease Rating Scale (subUPDRS) and the Speech Intelligibility Test (SIT). Results: Patients had a mean disease duration of 16 years, and the mean time since DBS implantation was 6.9 years. subUPDRS scores (11 patients tested) were significantly lower both in aDBS (p=<.001), and cDBS (p=.001), when compared to NoStim. Bradykinesia subscores were significantly lower in aDBS (p=.002), but not in cDBS (p=.08), when compared to NoStim. SIT scores of patients with stimulation-induced dysarthria (11 patients tested) significantly worsened in cDBS (p=.009), but not in aDBS (p=.407), when compared to NoStim. Overall, stimulation was applied 48.8% of the time during aDBS. Conclusion: Beta oscillations remain informative of the clinical state in patients with advanced PD, after several years of DBS implantation. Beta-based aDBS is as effective as cDBS for PD in chronically implanted patients, but delivers less stimulation, and has a more favourable speech side-effect profile.