Neurofeedback records and displays neural activity in real time to enable voluntary modulation of brain signals. In Parkinson’s disease (PD), elevated beta oscillations in the basal ganglia have been linked to motor impairment, and prior work has shown that neurofeedback based on beta activity can be effective when using sensing-capable electrodes. This study investigated whether neurofeedback delivered through a fully internalized deep brain stimulation (DBS) system can enable short-term regulation of beta-band power in the subthalamic nucleus (STN) across multiple days.
The authors positioned this work as the first exploration of multi-day neurofeedback training with a fully implanted DBS device, aiming to assess both immediate electrophysiological regulation and whether repeated sessions produce cumulative improvements. The study also performed exploratory analyses of factors that might influence neurofeedback performance, including dopaminergic medication and stimulation state.
Eight patients with Parkinson’s disease took part in the investigation. The neurofeedback signal was derived from STN recordings obtained via implanted DBS electrodes with interleaved sensing capability. Beta frequency bands used for feedback were generally 13–35 Hz, with two exceptions where subject-specific peak frequencies at 12.7 Hz and 8.78 Hz were used.
No additional demographic, clinical severity, or electrode-model details were reported in the source abstract; those specifics were not provided in the summary available here.
Each patient completed three neurofeedback sessions on separate days. Visual feedback reflected ongoing beta oscillatory power from the STN, and participants were instructed to perform both downregulation and upregulation tasks of the targeted beta activity. The sessions were separated in time to permit assessment of within-day regulation and short-term changes across days.
The study design focused on feasibility and short-term learning with a fully internalized device rather than on long-term clinical outcomes. The authors emphasize this exploratory experimental framework.
Neurofeedback produced a statistically significant decline in beta power during the first and third sessions, indicating successful acute downregulation of the beta oscillations used for feedback. The second session showed a less pronounced regulation effect compared with sessions one and three.
Despite the observed declines in beta power during individual sessions, the study did not demonstrate a significant cumulative improvement across the three sessions. In other words, while sessions 1 and 3 showed meaningful modulation of beta-band activity, repeated exposure over these three days did not produce a clear across-session learning effect according to the reported results.
In an exploratory analysis, the investigators examined how two treatment-related factors affected downregulation performance. They reported that increases in dopaminergic medication were associated with deterioration of downregulation ability, suggesting that higher medication states may reduce the capacity to voluntarily suppress beta power in this neurofeedback paradigm.
Conversely, the presence of stimulation improved downregulation ability, indicating that concurrent DBS stimulation may facilitate voluntary modulation of beta oscillations when used as part of a neurofeedback protocol. These observations are presented as exploratory and hypothesis-generating rather than definitive causal findings.
Beyond the targeted beta-band effects, neurofeedback based on beta oscillations also induced changes in the power of gamma oscillations. The authors interpret these gamma changes as further evidence of pro-kinetic electrophysiological modulation — that is, shifts in oscillatory activity that have been associated with improved motor function in PD. The abstract does not provide quantitative gamma results in the summary available here.
This study demonstrated the feasibility of conducting multiple neurofeedback sessions over several days using a fully implanted DBS system and STN beta activity as the feedback signal. However, a statistically significant cumulative improvement across the multi-day protocol was not observed.
The investigation was explicitly explorative. The authors note that potential improvements in the experimental setup could be identified from their findings and suggest that extended-duration training and outpatient implementations of DBS electrode-guided neurofeedback are worthwhile next steps. Specific methodological limitations, patient characteristics, or longer-term clinical outcome data were not detailed in the abstract and therefore are not reported here.
Overall, the results support the technical feasibility of implanted-electrode neurofeedback in PD and identify medication and stimulation state as candidate modulators of neurofeedback performance, while underscoring the need for longer and more comprehensive studies to assess clinical efficacy.