Towards guided and automated programming of subthalamic area stimulation in Parkinson’s disease

Are we ready for automated deep brain stimulation programming?

Deep brain stimulation (DBS) requires individualized programming of stimulation parameters, a time-consuming process performed manually by clinicians with specialized training. This process limits DBS accessibility, delays treatment, and constrains the potential for next-generation technology to improve patient outcomes. This review describes technological advancements that could automate DBS programming, focusing on Parkinson's disease biomarkers that can provide objective outcome measurement and algorithms that can quickly and automatically identify optimal DBS settings. We first define key performance criteria for an automated programming system, including effectiveness, efficiency, and ease of use, and then describe and evaluate each component with respect to these criteria. We conclude that the state of current research provides a strong foundation for developing automated DBS programming. The primary remaining obstacle lies in identifying and deploying the best combination of available techniques that will overcome the high entry barrier needed for wide-scale clinical deployment and adoption.

Clinically Implemented Sensing-Based Initial Programming of Deep Brain Stimulation for Parkinson's Disease: A Retrospective Study

OBJECTIVES

Initial deep brain stimulation (DBS) programming using a monopolar review is time-consuming, subjective, and burdensome. Incorporating neurophysiology has the potential to expedite, objectify, and automatize initial DBS programming. We aimed to assess the feasibility and performance of clinically implemented sensing-based initial DBS programming for Parkinson's disease (PD).

MATERIALS AND METHODS

We conducted a single-center retrospective study in 15 patients with PD (25 hemispheres) implanted with a sensing-enabled neurostimulator in whom initial subthalamic nucleus/globus pallidus pars interna DBS programming was guided by beta power in real-time local field potential recordings, instead of a monopolar review.

RESULTS

The initial sensing-based programming visit lasted on average 42.2 minutes (SD 18) per hemisphere. During the DBS optimization phase, a conventional monopolar clinical review was not required in any patients. The initial stimulation contact level remained the same at the final follow-up visit in all hemispheres except three. The final amplitude was on average 0.8 mA (SD 0.9) higher than initially set after the original sensing-based programming visit. One year after surgery, off-medication International Parkinson and Movement Disorder Society Unified Parkinson's Disease Rating Scale (MDS-UPDRS) III total score, tremor subscore, MDS-UPDRS IV, and levodopa-equivalent dose improved by 47.0% (p < 0.001), 77.7% (p = 0.001), 51.1% (p = 0.006), and 44.8% (p = 0.011) compared with preoperatively using this approach.

CONCLUSIONS

This study shows that sensing-based initial DBS programming for PD is feasible and rapid, and selected clinically effective contacts in most patients, including those with tremor. Technologic innovations and practical developments could improve sensing-based programming.

Long-Term Stability of Spatial Distribution and Peak Dynamics of Subthalamic Beta Power in Parkinson's Disease Patients

BACKGROUND

Subthalamic beta oscillations are a biomarker for bradykinesia and rigidity in Parkinson's disease (PD), incorporated as a feedback signal in adaptive deep brain stimulation with potential for guiding electrode contact selection. Understanding their longitudinal stability is essential for successful clinical implementation.

OBJECTIVES

We aimed to analyze the long-term dynamics of beta peak parameters and beta power distribution along electrodes.

METHODS

We recorded local field potentials from 12 channels per hemisphere of 33 PD patients at rest, in a therapy-off state at two to four sessions (0, 3, 12, 18-44 months) post-surgery. We analyzed bipolar beta power (13-35 Hz) and estimated monopolar beta power in subgroups with consistent recordings.

RESULTS

During the initial 3 months, beta peak power increased (P < 0.0001). While detection of high-beta peaks was more consistent, low- and high-beta peak frequencies shifted substantially in some hemispheres during all periods. Spatial distribution of beta power correlated over time. Maximal beta power across segmented contact levels and directions was significantly stable compared with chance and increased in stability over time. Active contacts for therapeutic stimulation showed consistently higher normalized beta power than inactive contacts (P < 0.0001).

CONCLUSIONS

Our findings indicate that beta power is a stable chronic biomarker usable for beta-guided programming. For adaptive stimulation, high-beta peaks might be more reliable over time. Greater stability of beta power, center frequency, and spatial distribution beyond an initial stabilization period suggests that the microlesional effect significantly impacts neuronal oscillations, which should be considered in routine clinical practice when using beta activity for automated programming algorithms. © 2025 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Stimulation artefact removal: review and evaluation of applications in evoked responses

Objective.This study investigated software methods for removing stimulation artefacts in recordings undertaken during deep brain stimulation (DBS). We aimed to evaluate artefact attenuation using sample recordings of evoked resonant neural activity (ERNA), as well as a synthetic ground-truth waveform that emulated observed ERNA characteristics.Approach.The synthetic waveform and eight raw DBS recordings were processed by fourteen algorithms spanning the following categories: signal modification, signal decomposition, and template subtraction. For the synthetic waveform, performance was quantified by comparing each reconstructed signal against the ground-truth waveform. For DBS recordings, performance was contrasted amongst each other. The stimulation artefact was quantified by its amplitude and subsequent decay to baseline by the time to first zero-crossing. Each reconstructed ERNA signal was characterised by peak-to-peak-amplitude, root-mean-square amplitude, latency, and number of zero-crossings.Main results.None of the methods performed overall as well as the Backward Filter. Signal decomposition techniques were able to attenuate stimulation artefact albeit with unacceptable ERNA distortion.Significance.Upon evaluation of common software methods for DBS artefact attenuation, we advocate the use of the Backward Filter for reducing such artefacts while reconstructing ERNA.

Characterization of subthalamic nucleus deep brain stimulation evoked resonant neural activity in a large animal model: A pilot study

Recent reports have described stimulation evoked resonant neural activity (ERNA) recorded in the subthalamic nucleus (STN) and globus pallidus internus (GPi) of patients during Deep Brain Stimulation (DBS) surgery. The constraints imposed during intraoperative recordings in patients limit the opportunity for in-depth study of new findings such as ERNA. In this pilot study, we leverage a large animal model to focus on detailed characterization of ERNA. Bilateral DBS leads were implanted in the STN in three ovine subjects and externalized for chronic use with custom stimulation and recording circuitry. ERNA was reliably recorded from the STN region in all three subjects with distinct specificity to recording and stimulation sites/contacts. Basic neural response characteristics such as input/output behavior, frequency response and strength/duration curves were evaluated. ERNA amplitude was highly dependent upon stimulation frequency, due to the interaction of the underlying resonant activity and the evoked response from each stimulus pulse. The results could be predicted by a mathematical model of constructive/destructive phase interference, and importantly, the evoked response latency. Significant time dependent dynamics in these evoked potentials were observed, which will be critically important to understand for future clinical applications. Based upon these recordings from leads in the STN region of healthy ovine subjects, these data confirm that DBS evokes high frequency resonant activity in the basal ganglia network. The clinical utility of ERNA remains to be demonstrated, but its direct association with DBS therapy makes it an interesting biomarker for potential use in contact selection and closed loop therapy.

Detection of evoked resonant neural activity in Parkinson's disease

Objective. This study investigated a machine-learning approach to detect the presence of evoked resonant neural activity (ERNA) recorded during deep brain stimulation (DBS) of the subthalamic nucleus (STN) in people with Parkinson's disease.Approach. Seven binary classifiers were trained to distinguish ERNA from the background neural activity using eight different time-domain signal features.Main results. Nested cross-validation revealed a strong classification performance of 99.1% accuracy, with 99.6% specificity and 98.7% sensitivity to detect ERNA. Using a semi-simulated ERNA dataset, the results show that a signal-to-noise ratio of 15 dB is required to maintain a 90% classifier sensitivity. ERNA detection is feasible with an appropriate combination of signal processing, feature extraction and classifier. Future work should consider reducing the computational complexity for use in real-time applications.Significance. The presence of ERNA can be used to indicate the location of a DBS electrode array during implantation surgery. The confidence score of the detector could be useful for assisting clinicians to adjust the position of the DBS electrode array inside/outside the STN.

Multiple input algorithm-guided Deep Brain stimulation-programming for Parkinson's disease patients

Technological advances of Deep Brain Stimulation (DBS) within the subthalamic nucleus (STN) for Parkinson's disease (PD) provide increased programming options with higher programming burden. Reducing the effort of DBS optimization requires novel programming strategies. The objective of this study was to evaluate the feasibility of a semi-automatic algorithm-guided-programming (AgP) approach to obtain beneficial stimulation settings for PD patients with directional DBS systems. The AgP evaluates iteratively the weighted combination of sensor and clinician assessed responses of multiple PD symptoms to suggested DBS settings until it converges to a final solution. Acute clinical effectiveness of AgP DBS settings and DBS settings that were found following a standard of care (SoC) procedure were compared in a randomized, crossover and double-blind fashion in 10 PD subjects from a single center. Compared to therapy absence, AgP and SoC DBS settings significantly improved (p = 0.002) total Unified Parkinson's Disease Rating Scale III scores (median 69.8 interquartile range (IQR) 64.6|71.9% and 66.2 IQR 58.1|68.2%, respectively). Despite their similar clinical results, AgP and SoC DBS settings differed substantially. Per subject, AgP tested 37.0 IQR 34.0|37 settings before convergence, resulting in 1.7 IQR 1.6|2.0 h, which is comparable to previous reports. Although AgP long-term clinical results still need to be investigated, this approach constitutes an alternative for DBS programming and represents an important step for future closed-loop DBS optimization systems.

Can brain signals and anatomy refine contact choice for deep brain stimulation in Parkinson's disease?

INTRODUCTION

Selecting the ideal contact to apply subthalamic nucleus deep brain stimulation (STN-DBS) in Parkinson's disease is time-consuming and reliant on clinical expertise. The aim of this cohort study was to assess whether neuronal signals (beta oscillations and evoked resonant neural activity (ERNA)), and the anatomical location of electrodes, can predict the contacts selected by long-term, expert-clinician programming of STN-DBS.

METHODS

We evaluated 92 hemispheres of 47 patients with Parkinson's disease receiving chronic monopolar and bipolar STN-DBS. At each contact, beta oscillations and ERNA were recorded intraoperatively, and anatomical locations were assessed. How these factors, alone and in combination, predicted the contacts clinically selected for chronic deep brain stimulation at 6 months postoperatively was evaluated using a simple-ranking method and machine learning algorithms.

RESULTS

The probability that each factor individually predicted the clinician-chosen contact was as follows: ERNA 80%, anatomy 67%, beta oscillations 50%. ERNA performed significantly better than anatomy and beta oscillations. Combining neuronal signal and anatomical data did not improve predictive performance.

CONCLUSION

This work supports the development of probability-based algorithms using neuronal signals and anatomical data to assist programming of deep brain stimulation.