Where Real-Time Data Fits in the Future of Deep Brain Stimulation

Deep Brain Stimulation (DBS) has been transformative for people with movement disorders such as Parkinson's, providing substantially better symptom control than medication alone. Despite this, the way clinicians optimise therapy from these devices remains surprisingly intensive. It relies heavily on visual observation and patient feedback to adjust settings, and is traditionally a manual, iterative process, often requiring multiple clinic sessions to adjust stimulation parameters whilst observing the patient’s symptom response (1). As a result, the process can be slow, imprecise, and variable in its outcome, with success highly dependent on clinician expertise and accurate patient feedback (2).

In response, each DBS device manufacturer is developing specialised hardware and software tools to facilitate this tuning across conditions like Parkinson’s (PD), dystonia, and Essential Tremor (ET). These range from advanced programmer interfaces to sensor-driven algorithms that aim to improve efficiency and precision. Below, we explore the current generation of DBS programming tools (including those from Medtronic, Abbott, and Boston Scientific) and evaluate how real-time objective symptom feedback could enhance their effectiveness.

Medtronic DBS Programming Tools

Medtronic Clinician Programmers: Medtronic’s DBS systems (e.g., Activa series, Percept™ PC) are calibrated using a clinician programmer device historically featuring a handheld console (Model 8840 N’Vision) and now a tablet-based interface. The Clinician Programmer Application communicates wirelessly with the implanted pulse generator to adjust parameters such as contact selection, amplitude (voltage/current), pulse width, and frequency. The Programmer supports all Medtronic DBS indications (PD, ET, dystonia, etc.) and is used to perform initial “monopolar reviews” (systematic testing of each contact) and fine-tune settings over follow-up visits. Medtronic’s latest software integrates advanced programming features: for example, the FDA-cleared SureTune™ 4 visualisation software which can import patient-specific MRI/CT data. By overlaying the imaging with a reference map of subcortical structures, the clinician can ‘see’ the precise DBS lead location in a patient’s brain. They can also model the shape of the stimulation fields and inspect how these intersect with the target structure. This methodology is referred to in the literature as image-guided programming (IgP) (3). By directly visualising the brain structures affected by a given setting in real-time, SureTune™ 4 constrains the programming process to stimulation configurations that are physiologically sound. In practice, these tools provide clinicians with a more intuitive, 3D map for steering current and avoiding known side-effect regions, thereby reducing the need for trial-and-error programming.

Medtronic Percept™ PC Neurostimulator (BrainSense™): The Percept PC is an implantable pulse generator introduced by Medtronic that includes brain signal sensing capabilities (the first commercially available DBS system of this kind). The implantable captures Local Field Potentials (LFPs) from the implanted DBS leads in real-time (a feature called BrainSense), during programming sessions and even outside of the clinic. Clinicians can use the paired programmer to view these neural signals. For example, Percept’s Electrode/BurstID feature provides a real-time snapshot of pathological beta-band oscillatory activity at each contact, helping identify a “sensing sweet spot”. In PD, a hyperactive beta source is correlated with bradykinesia and rigidity symptoms. Using the contact nearest to the beta source can help alleviate these symptoms. Neurophysiological data points can help with initial contact selection and parameter choices. Moreover, Medtronic have now deployed an Adaptive DBS (aDBS) algorithm in Percept. Using a specified neural biomarker threshold (e.g., hyperactive beta oscillations), the device can automatically adjust the stimulation timing in a closed-loop manner. By limiting the amount of time the stimulation is turned on, battery life can be extended and resistive ‘habituation’ to the effect of stimulation can be delayed. Although aDBS operates autonomously outside of the clinic, new thresholding parameters still need to be calibrated during a programming session.

Benefits of Real-Time Symptom Feedback for Medtronic Tools

Medtronic’s programming process greatly benefits from objective, real-time symptom measures. Currently, clinicians rely on patient reports and visual observation of tremor, rigidity, bradykinesia, etc., which can be subjective and episodic (2). Real-time sensors could quantitatively track symptom severity as settings are adjusted:

• Standard Clinician Programming: During a typical session, a clinician increases voltage until a tremor stops, or side effect(s) appear. With a wearable accelerometer or gyroscope on the tremulous limb, the tremor amplitude could be continuously measured, providing an objective curve of tremor severity versus stimulation intensity. This would improve accuracy in identifying the exact threshold at which tremor is suppressed without inducing side effects, rather than relying on coarse clinical scoring. Short, subtle improvements that might be missed by the naked eye would be captured by sensor data. Objective feedback could thus shorten the trial-and-error process and ensure optimal settings are found more efficiently. Indeed, experts highlight the need for objective markers: the subjective nature of clinical observation makes DBS response assessment challenging, motivating the design of “objective markers of therapeutic response” to improve programming.

  
  

• BrainSense™ Guided Programming: While the Percept’s internal LFP sensing is itself an objective tool, it measures a neural biomarker rather than the clinical symptom directly. Real-time symptom measures can complement BrainSense. For example, a wearable sensor measuring bradykinesia or tremor could be used in conjunction with LFP readings to correlate changes in beta power with actual motor improvement (2,5). If a certain contact exhibits high beta suppression and the wearable sensor indicates a net reduction in bradykinesia scores, the clinician gains confidence that this contact and setting are truly optimal. Conversely, if the sensed beta signal is ambiguous, the external symptom data provides direct confirmation of benefit. Neural biomarkers are still an active area of research, and many relationships with the various symptoms in the many pathologies of interest need to be firmly validated. Additionally, the quality and reliability of their capture slowly improve with the maturity of the implantable hardware. In essence, real-time symptom feedback would help translate the Percept’s rich neural data into verified clinical outcomes.

  
  

• Adaptive DBS: The closed-loop aDBS feature in Percept utilises an objective feedback signal to automatically adjust therapy. Early trials of adaptive DBS in PD cohorts have shown that patients experience longer periods of good symptom control (e.g., more time without tremor or bradykinesia) compared to traditional continuous stimulation. For instance, Little et al. (2013) demonstrated that beta-band-triggered aDBS could reduce bradykinesia and improve motor states with less overall stimulation time, illustrating the promise of real-time biomarker feedback in dynamic therapy adjustment (15). Those results underscore the importance of feedback-driven adjustment in improving therapy. Future adaptive systems could also integrate external symptom monitors (for instance, a wrist accelerometer detecting tremor onset), and the device could directly respond to actual symptom fluctuations (not just a proxy, such as LFP). Such multi-modal feedback could further refine stimulation control. Furthermore, during in-clinic programming, having an “adaptive” preview, where a wearable indicates symptom severity in real-time, would enable the clinician to find optimal settings more quickly and help assess the effects of the additional adaptive threshold settings. Overall, objective symptom monitoring would enhance Medtronic’s programming by increasing the precision of parameter selection and potentially reducing programming time and “DBS failures” due to suboptimal settings (1).

  
  

Abbott DBS Programming Tools

Abbott Infinity™ DBS Programmer (iOS-based): Abbott’s DBS platform (formerly St. Jude Medical) introduced a novel programmer design using consumer-grade mobile devices (e.g., Apple iPad/iPod) running a dedicated programming app. The clinician programmer is a wireless, iOS-based tablet interface that connects via Bluetooth to the implanted pulse generator. This system was the first DBS platform to adopt a touchscreen mobile device for programming, aiming to make the interface more intuitive and modern. The programmer application provides an easy-to-use GUI for setting stimulation parameters, which was particularly helpful to introduce users to innovative device features such as Abbott’s directional lead technology (Infinity leads). The programmer interface (with Informity software) is designed for “simplified functional directional programming,” supporting clinicians in steering current by distributing the relative current amplitudes of the different contact segments. In practice, Informity streamlines the process of testing multiple orientations by recommending or automatically calculating optimal current splits, thereby reducing the programming burden on the clinician. Abbott also offers expanded therapy options with complex field shaping from MultiStim™ capabilities, which the programmer can configure. All these features are accessible through the wireless tablet, giving the clinician mobility and flexibility in how they program (e.g., standing with the patient doing gait trials, unencumbered by cables).

NeuroSphere™ Virtual Clinic (Remote Programming): A standout feature of Abbott’s DBS ecosystem is its remote programming capability. In 2021, Abbott launched NeuroSphere Virtual Clinic, the first FDA-approved telehealth platform for DBS. This system enables clinicians to perform DBS adjustments in real-time over the internet, allowing patients to remain at home. The patient uses their personal controller device (an app on a smartphone or dedicated iPod) to receive updated settings from the clinician, who controls the session via a secure cloud connection. An integrated video call enables the clinician to observe the patient and converse during the tuning session (7). NeuroSphere was recognised as one of TIME’s Best Inventions of 2021 (6) for its ability to improve patient access to care. This tool was especially valuable during the COVID-19 pandemic and for patients who live far from DBS centres  (9 out of 10 patients reported improved access to care with remote programming). Functionally, the remote session offers the same controls as an in-person programmer (the clinician adjusts settings on their tablet, which then updates the patient’s device via the cloud). The limitation is that the clinician must rely on what they can see on video and what the patient reports to assess symptom changes.

Benefits of Real-Time Symptom Feedback for Abbott Tools

Abbott’s programming tools, particularly due to their wireless and remote capabilities, would strongly benefit from objective symptom monitoring integrated into the workflow:

• In-Clinic Directional Programming: Tuning a directional lead (with many possible current steering combinations) is complex. Although Abbott’s Informity software simplifies this process, clinicians still need to evaluate which configuration yields the best symptom improvement. Real-time kinematic sensors on the patient can provide quantitative outcome measures for each configuration tested. For example, if programming a PD patient’s wearable motion sensors on the legs could measure stride length or shuffling objectively as different directional settings are tried. Sensing could reveal subtle side-effect degradation in gait with a certain current steering that might not be immediately obvious otherwise. Similarly, for tremor or dystonia, accelerometers or muscle activity sensing could measure tremor amplitude or muscle overactivity in real time. By pairing these measurements with the programmer’s ability to quickly switch configurations, the clinician can more confidently identify the optimal setting. Objective feedback would thus enhance the “simplified functional programming” claim by adding data-driven decision support, reducing reliance on subjective scoring.

  
  

• Remote Programming Sessions: During teleprogramming, objective measures are even more crucial. Over a video call, certain symptoms can be challenging to evaluate, for instance, mild rigidity or subtle changes in bradykinesia are difficult to discern through a screen. If patients at home use connected devices (such as a smartwatch or a sensor-enabled wearable) that continuously send symptom metrics to the clinician, it would bridge the assessment gap in remote care. For example, a patient with ET during a remote session could wear a smartwatch that streams tremor amplitude data to the clinician’s dashboard. The clinician could watch the tremor power drop or rise as they adjust DBS settings in NeuroSphere, rather than relying solely on asking “Is your tremor better now?” or trying to visually gauge it over video. This objective telemonitoring could make remote tuning almost as informative as an in-person visit. It improves confidence that the changes made are truly helping. Given that Abbott’s platform already leverages personal smart devices, integrating other mobile health apps that, for instance, track motion or even voice (for speech changes) is feasible. In summary, objective symptom feedback would significantly improve the efficacy and safety of remote DBS programming, ensuring that optimal settings are achieved despite the physical distance.

  
  

• Efficiency and Accuracy: Whether in-person or remote, having quantitative symptom data speeds up programming by allowing rapid comparison of settings. Instead of lengthy clinical examinations at each step, a stream of data (e.g., a numeric tremor severity score) instantly shows the effect of a new setting. This can shorten sessions and potentially reduce the total number of sessions needed to reach therapeutic settings. It also aids less experienced programmers, for instance, an advanced nurse practitioner in a community setting using NeuroSphere could rely on sensor data to guide decisions, which might make them feel more confident. Thus, real-time symptom monitoring could democratise DBS programming expertise, aligning with Abbott’s goal of making therapy more accessible.

  
  

Boston Scientific DBS Programming Tools

Boston Scientific Vercise™ Programmer and Image-Guided Programming: Boston Scientific’s DBS systems (Vercise PC, Vercise Genus, etc.) come with a clinician programmer that is distinct for its Multiple Independent Current Control (MICC) capability. The Vercise platform allows each contact on the lead to be driven with its own current source, enabling flexible current steering and “current fractionalisation” across contacts (2). The clinician programmer (often a laptop-based software called Vercise Neural Navigator, with the latest version being Neural Navigator 5) provides an interface to adjust these independent contact currents and configure directional stimulation. To manage the complexity of programming with MICC and directional leads (Cartesia™ directional leads), Boston Scientific has heavily integrated imaging and software guidance into its programming workflow. They partnered with Brainlab to create an Image Guided Programming (IGP) system (commercially known as StimView™ XT or Guide XT when referring to the planning side). This tool imports the patient’s postoperative lead coordinates and anatomical imaging, allowing the clinician to visualise the DBS lead within a 3D brain model on the programmer screen (8). Crucially, the software can display the estimated volume of neural activation for given stimulation settings in real time, effectively showing a map of where current is flowing in the brain as the clinician adjusts parameters.

The impact of this image-guided approach is supported by clinical data: Lange et al. (2021) conducted a trial in Parkinson’s patients comparing standard clinical-based programming to anatomy-driven programming using Guide XT, and found that imaging-guided programming reduced the average programming time by more than half (~20 minutes vs. ~45 minutes) while achieving equivalent motor symptom outcomes (8). Patients programmed with the image-guided method had similar improvements (MDS-UPDRS-III scores) as those with conventional programming, but the process was significantly more efficient. This is a key advantage of Boston Scientific’s programming tools: by seeing exactly where the lead and stimulation effect lie, clinicians can skip much of the blind trial-and-error. The programmer’s interface (StimView) effectively suggests which contacts and orientations to activate to cover the target (e.g., the dorsolateral STN for PD) while avoiding spread to side-effect regions. Boston Scientific’s recent FDA approvals of these software tools emphasise not only precision but also time savings – one report notes a 56% reduction in programming time using image-guided programming (9).

Advanced Programming Features: In addition to imaging, the Vercise system and programmer support other advanced features such as grouped programs and schedules. Clinicians can set up multiple stimulation programs (for different patient conditions or times of day) and use features akin to interleaving pulses or cycling between contacts to manage complex symptoms. The programmer’s software is designed to handle the large parameter space (current on each of 8 contacts, pulse width, frequency, etc.), and Boston Scientific has also explored algorithmic assistance. Notably, some of the researchers and engineers at Boston Scientific have been involved in developing algorithm-guided programming frameworks (like the CLOVER trial, discussed below), indicating the company’s interest in semi-automated tuning solutions that could be integrated into future programmer software.

Benefits of Real-Time Symptom Feedback for Boston Scientific Tools

Boston Scientific’s programming approach is highly technology-driven (image and software), but it still relies on clinical validation of efficacy. Real-time objective symptom measures would provide that validation and further boost the capabilities of their tools:

• Image-Guided Programming: The anatomical models predict where stimulation will go, but patient-specific physiology can vary. Two patients might have leads in a similar position, yet their symptom response to a given setting could differ. By incorporating objective symptom monitoring during image-guided tuning, clinicians could cross-verify the software’s suggestions with actual outcomes. For example, if StimView recommends activating a specific directional segment to avoid the internal capsule (side-effect area) while covering the motor area, the clinician can do so and then immediately see, e.g., the tremor frequency drop on an accelerometer reading. If the tremor doesn’t improve as expected despite coverage, that real-time data might prompt further adjustment (maybe increasing current or shifting direction slightly). Essentially, sensor feedback would add a “ground truth” layer on top of the virtual model, ensuring that the predicted optimal setting is truly optimal for the patient. It could also catch cases where imaging alone is misleading (e.g., if fibrosis or anatomy causes different current spread than expected, the symptom data will reflect that). Thus, the combination of image-guided programming with real-time symptom metrics merges the best of both worlds – anatomical insight and empirical validation.

  
  

• Handling Complex Parameter Spaces: The MICC and directional capabilities in Vercise mean the clinician has many dials to turn. Objective feedback can act as a compass in this high-dimensional space. For instance, programming a dystonia patient might involve balancing improvements in muscle spasms against side effects like speech slurring. If myography (muscle activity) sensors on involved muscles show reduced pathological activity and a voice recorder app measures any speech changes, the clinician can quantitatively find a compromise (e.g., a slightly lower current that still yields 80% myography improvement but avoids crossing the threshold where speech deteriorates). Without such data, the clinician might overshoot or undershoot optimal settings or require more rounds of adjustments to get it right. In PD, where multiple symptoms (tremor, rigidity, bradykinesia) may respond differently to changes, wearable sensors can help track each dimension concurrently (accelerometers for tremor, gyroscope for movement speed, etc.) during programming – guiding the clinician on which symptom is driving the patient’s disability and whether the chosen settings are addressing it. This data-driven approach is especially useful for Boston Scientific’s system because MICC allows simultaneous multi-contact stimulation – one could envision using objective metrics to tune a different contact for each symptom (e.g., one contact’s current optimised for tremor via accelerometer feedback, another contact optimised for bradykinesia via a tapping speed sensor). In short, real-time feedback could unlock the full potential of MICC by mapping each adjustment to a measurable clinical effect in real time, rather than the clinician juggling mental estimations.

  
  

• Algorithmic and Automated Programming: Boston Scientific has been at the forefront of exploring automated programming algorithms. The CLOVER-DBS study (Wenzel et al. 2021) evaluated a closed-loop algorithm that automatically adjusted a Boston Scientific DBS system using wearable motion sensor feedback (10). In that trial, the algorithm tested various contact configurations and amplitudes, guided by a sensor measuring motor symptoms, to converge on effective settings. The results showed that in a subset of patients (those using the latest algorithm version), the algorithm-guided settings were as effective as standard expert programming in reducing PD motor symptoms (UPDRS-III scores), with no significant difference (10). This demonstrates that, given reliable objective feedback, an automated or semi-automated tool can identify good settings in a complex device like Vercise. Clinicians could use such algorithmic suggestions as a starting point, greatly reducing the burden of manual search. Notably, the algorithm in CLOVER used a finger-worn motion sensor to score tremor/bradykinesia, reflecting how crucial that objective input was, since without it, the algorithm would have no guidance signal. Beyond CLOVER, a recent study by Ferleger et al. (2022) employed a Bayesian optimisation algorithm with smartwatch tremor measurements to automatically tune DBS for tremor (in PD and ET patients). It successfully reached settings that gave tremor suppression comparable to those found by human clinicians, in on average 15–18 tested iterations (2). These examples reinforce that real-time symptom metrics are not only beneficial but in fact essential for any kind of automated or guided programming system. Boston Scientific’s programmer of the future could incorporate a “smart tuning” mode where it interfaces with wearable sensors and algorithmically steers the parameters – effectively an expert system to assist the clinician. In summary, Boston Scientific’s tools stand to gain immensely from objective feedback, as evidenced by faster programming times with image guidance (8) and the success of sensor-driven algorithms in trials (2). By leveraging these advancements, programming can become more efficient, precise, and even partially automated, all to the patient’s benefit.

  
  

Sensor-Based and Algorithmic Tuning Systems

Beyond the manufacturer-specific hardware/software, there is a growing array of supplemental tools and research prototypes aimed at improving DBS tuning through objective measurements and algorithmic assistance. These tools can be used across devices (regardless of manufacturer) and represent the cutting edge of DBS programming technology:

•  Wearable Sensor Feedback Systems: Wearable motion sensors (accelerometers, gyroscopes, magnetometers) and inertial measurement units can quantify movement disorders symptoms with high sensitivity. In PD and ET, for example, accelerometers have been used to derive tremor power, bradykinesia indexes, dyskinesia scores, and more (11,12). These wearables (often on the wrist or finger for tremor, or on the trunk/feet for gait) can feed data to a computer in real time. Great Lakes NeuroTechnologies’ Kinesia™ system is one such FDA-cleared sensor platform that provides objective tremor and bradykinesia ratings, although it does not capture other symptoms like rigidity. During DBS programming, such a system can track how the symptom severity changes with each new setting. For instance, as a clinician increases voltage, the sensor might show a continuous decrease in tremor amplitude and then plateau or worsen if a side-effect (like dyskinesia) emerges – producing a data curve to pinpoint the optimal intensity. Research has shown that these sensor-based assessments correlate well with clinical exam scores and can even detect subtle changes that clinicians miss (11,12). The use of wearable sensors in programming has been piloted in several studies. One study reported an “automated DBS programming framework” in which a motion sensor on the patient measured tremor and a computer algorithm adjusted settings in a closed-loop manner. This closed-loop system employed Bayesian optimisation to intelligently choose the next parameter set to try, aiming to maximise tremor reduction while respecting safety limits. In tests on 15 patients, the fully automated approach achieved tremor suppression comparable to expert clinician programming and did so efficiently (often within ~15 iterations of testing) (2). Such results underscore that wearable sensor feedback can encapsulate the clinical expertise needed to tune DBS, making it possible to automate or assist the process.

  
  

• EMG (Electromyography), MMG (Mechanomyography) and Physiological Sensors: Surface myography sensors can objectively measure muscle activity, which is useful for symptoms like dystonic muscle contractions or tremors (which produce rhythmic bursts in EMG and MMG). During DBS tuning for dystonia or tremor, myography placed on affected muscles provides a direct, real-time readout of the pathological activity that DBS is meant to alleviate. As stimulation is increased, the myography might show a reduction in the magnitude or frequency of abnormal muscle bursts. This was demonstrated in a study using EMG-driven responsive DBS for tremor: the system would trigger stimulation only when pathological EMG activity (indicating tremor) was detected and turn it off when tremor subsided (16). The outcome was tremor control equivalent to continuous DBS but with far less total stimulation delivered, thanks to the EMG feedback. In a programming context, one could envisage using myography to titrate stimulation: for example, in cervical dystonia (neck muscle spasms), myography from neck muscles could help find the DBS setting that quiets muscle firing without over-suppressing (which could cause weakness). Other physiological signals, like accelerometer-derived gait metrics or even heart rate changes (in cases where autonomic responses are relevant), could provide additional dimensions of feedback.

  
  

• Video and Machine Learning-Based Assessment: Video recording of patients during programming can be analysed to produce objective scores. Machine learning models have been developed to grade tremor severity, bradykinesia (through finger-tapping or gait analysis), and even speech changes from video or audio inputs. For example, in remote assessments, patients have performed motor tasks on camera which were later rated by blinded clinicians or analysed by software. One pilot study used a smartphone app to capture videos of PD patients post DBS and was able to remotely quantify their motor outcomes with promising reliability (13). In the programming scenario, one can imagine a clinician’s tablet running a live video analysis: as the patient performs a hand rotation task at each setting, the software instantly computes a bradykinesia score (or a tremor amplitude from the video). This would be especially useful for quantifying complex symptoms like axial gait impairment, where a clinician typically has the patient walk and then subjectively notes improvement. A depth-sensor camera or wearable motion capture system (like Microsoft Kinect or Vicon) could measure gait parameters objectively during the tuning, for example, stride length, speed, balance metrics, providing data to decide which DBS settings yield the best gait. While such video/ML tools are still emerging, they hold promise for standardising outcome measurement during DBS programming, both in-clinic and via telemedicine.

  
  

• Software Algorithms for Parameter Optimisation: Several algorithmic approaches independent of specific manufacturer hardware have been proposed to assist DBS programming. These include:

  • Adaptive search algorithms (e.g. Bayesian optimisation, genetic algorithms) that treat programming as an optimisation problem – the algorithm selects a set of parameters to try, observes the result (via an objective measure), and then updates its strategy to try a better set, iteratively converging to an optimum.

  • Closed-loop programming algorithms (like CLOVER-DBS (10)) which systematically map out the contact and amplitude space using sensor feedback. These often use a “stop when good” criterion, identifying when further increases in stimulation yield no additional benefit to the sensor-defined symptom metric.

  • Imaging-based algorithmic programming: A recent concept is using connectomic models and patient-specific simulations to predict optimal settings. For instance, Roediger et al. (2023) developed a data-driven algorithm that suggests DBS contacts and settings based on the electrode’s location relative to a probabilistic atlas of effective stimulation sites (14). While primarily image-based, such algorithms could incorporate symptom feedback for fine-tuning.

  • Teleprogramming decision support: In remote scenarios, cloud-based algorithms might process sensor or patient-reported data between visits to recommend parameter adjustments, which the clinician can then implement via teleprogramming.

These algorithmic tools universally depend on objective input data, and cannot function on subjective observations that are not quantified. The more accurate and real-time the symptom measurement, the better these algorithms can perform in finding therapeutic settings. Importantly, early studies caution that some algorithm-derived programs had contacts slightly outside the classical “optimal” zone, highlighting the need for iterative refinement and maybe multimodal input (10). Nevertheless, they showed the potential to reduce programming burden and even allow non-expert clinicians to achieve good results with guidance. In the coming years, we can expect these algorithms to be integrated into manufacturer programmers or external apps, making data-driven programming a standard part of DBS therapy.

In all, the advent of wearable sensors and intelligent algorithms represents a paradigm shift in DBS tuning. They introduce a level of objectivity, consistency, and automation that complements the clinician’s expertise. Instead of purely empirical, trial-and-error tuning, programming can become a more standardised and efficient process, where technology augments clinical decision-making.

Comparison of DBS Tuning Tools and the Impact of Real-Time Symptom Monitoring

To highlight the role of objective feedback, the table below lists major clinician-used DBS tuning tools (hardware and software) and assesses the potential benefit of integrating real-time symptom severity measures into each tool’s workflow:

DBS Tuning Tool	Description & Usage	Benefit from Real-Time Symptom Feedback
Medtronic Clinician Programmers (Tablet-based programmer, incl. older N’Vision)	External programmer devices for Medtronic DBS (PD, ET, dystonia). Provides manual control of voltage/current, pulse width, frequency, and electrode selection. Modern versions (tablet app) support advanced leads (directional). Standard practice involves clinician observing patient response to adjustments.	High: Currently relies on clinician’s subjective observation and patient input. Continuous objective monitoring (e.g. accelerometer for tremor) would improve accuracy and speed of finding optimal settings 3. It would help quantify symptom changes with each tweak, reducing guesswork and potentially shortening programming sessions.
Medtronic Percept™ PC (BrainSense™) Sensing-enabled IPG with SureTune™ visual programming	Implantable neurostimulator with LFP sensing; can display patient’s brain signals during programming. SureTune software provides patient specific 3D lead visualisation and allows “visual” programming on the clinician’s tablet 6. Used for refined programming by identifying biomarker peaks and imaging-defined sweet spots.	Moderate – High: The device’s internal neural feedback is valuable, but adding direct symptom measures would enhance correlation between biomarker changes and actual clinical improvement. For example, linking tremor sensor data with BrainSense readings can validate that a drop in beta power corresponds to tremor reduction in real time. Objective symptom input would also aid adaptive tuning, ensuring the closed-loop adjustments based on LFP truly translate to symptom relief. Overall, external feedback could refine the use of sensing and confirm the optimal setting more precisely.
Abbott Infinity™ DBS Programmer iOS-based wireless tablet and NeuroSphere™ Virtual Clinic	Clinician programmer on an Apple iPad/iPod with Bluetooth link to IPG 13. Intuitive touch interface for setting amplitudes and steering directional leads (via Informity™ software). Also enables remote programming through NeuroSphere Virtual Clinic (cloud-based, with video call), unique to Abbott. Used across PD, ET, dystonia.	High: Both in-clinic and remote tuning would strongly benefit from real-time symptom monitoring. In clinic, objective sensors can guide directional lead adjustments by quantifying improvements that are hard to judge by eye. In remote sessions (teleprogramming), having the patient wear a sensor or use a smartphone app to stream symptom data (tremor amplitude, tapping speed, etc.) would compensate for the clinician’s inability to perform a hands-on exam. This would increase the safety and efficacy of adjustments made over telehealth and improve overall programming precision.
Boston Scientific Vercise™ Programming Neural Navigator software with StimView™ image guidance	Clinician programmer (usually PC-based) for MICC (multiple current) DBS. Allows flexible current steering on 8-contact directional leads. Integrated with Brainlab’s imaging it displays lead location in patient’s MRI and estimated stimulation field in 3D. Greatly speeds up contact selection and setting determination. Used for all approved conditions (PD, ET, dystonia).	Medium: The image-guided system provides a head start by indicating likely optimal settings, but real-time symptom feedback would serve as an important confirmation and fine-tuning tool. While anatomy-based programming drastically cuts down search time, continuous objective measures (e.g. a wearable recording tremor or gait metrics) ensure that the chosen setting truly maximises symptom relief. They can catch any discrepancies between predicted and actual clinical effect and help adjust stimulation intensity or orientation on the fly for best results.
Algorithmic Tuning Systems (e.g. CLOVER iterative algorithm; Bayesian autotuner)	Software algorithms that automate or assist DBS programming by systematically testing settings and using sensor feedback to evaluate outcomes. Examples: CLOVER-DBS (closed-loop optimisation using a motion sensor); Bayesian optimisation that learns the patient’s response curve using smartwatch data. Currently in research/prototype stage.	Essential: These systems fundamentally depend on objective real-time feedback. They treat the tuning process as a data driven optimisation problem; without a quantifiable outcome measure, the algorithms have nothing to optimise. In studies where they were tested, the use of wearable sensors for feedback enabled automated programming to achieve outcomes on par with expert clinicians. As such, objective symptom inputs are the cornerstone of algorithm-guided DBS tuning, and broader adoption of these systems will require reliable real-time symptom monitoring for every patient.
Adaptive (Closed-Loop) DBS Devices (e.g. Medtronic aDBS, investigational closed-loop systems)	Next-generation DBS devices that self-adjust stimulation in real time based on physiological signals. Medtronic’s Adaptive DBS uses the patient’s brain LFP as a trigger to modulate stimulation (on, off, or varying amplitude). Other research systems have used sensors like EMG or accelerometers to trigger stimulation only when symptoms are present 41. These devices blur the line between “programming” and ongoing therapy, but initial setup is a form of tuning (determining the control threshold, etc.).	High: Adaptive systems already illustrate the power of feedback – by responding to a biomarker, they can improve symptom control and efficiency of therapy. Incorporating direct symptom feedback (such as a tremor or movement sensor) could further enhance these systems. For example, an adaptive DBS that ramps up stimulation when a wrist accelerometer detects tremor onset would directly counteract symptoms. During the setup of adaptive DBS, having objective symptom data helps set the correct trigger criteria (e.g., calibrating the LFP threshold to when tremor actually appears). Thus, even in closed-loop systems, symptom monitoring remains valuable to ensure the device’s automatic adjustments align with real clinical needs.

Overview of key DBS tuning tools and the expected benefit from integrating real-time objective symptom monitoring.

Conclusions

DBS therapy has seen major technological advances in the tools clinicians use to program and optimise devices. Traditional programmer consoles have evolved into sophisticated interfaces with imaging guidance, wireless connectivity, and even sensing capabilities. Across the major manufacturers, the trend is toward more personalised and efficient programming, where therapy is personalised accurately and specifically to the person receiving it, whether by Medtronic’s neural sensing and closed-loop features, Abbott’s remote tuning, or Boston Scientific’s precise image-guided steering. Despite these different directions, the core aim in DBS tuning remains the same: find the optimal balance of symptom control vs. side effects. This is where real-time objective measures of symptom severity show immense value. They enable data-driven decision making into a process that used to be subjective.

As detailed, virtually all current programming tools would benefit from objective symptom feedback. In clinical literature, there is a consensus that objective assessments can improve programming outcomes and efficiency. For example, using wearable sensor feedback, researchers achieved automated programming that was equally effective as expert tuning, but potentially faster and more standardised. Image-guided programming combined with sensor validation can cut down programming time without sacrificing effectiveness (8).

Remote programming, which is becoming more commonplace, is substantially enabled by reliable objective monitoring of patients at a distance. Overall, real-time symptom feedback stands to enhance clinical effectiveness by ensuring optimal symptom relief, improve efficiency by reducing trial-and-error, and increase accuracy by providing quantitative outcomes for every programming method, both manual and automated.

Looking forward, the integration of wearable sensors and video analysis with DBS programming is likely to become routine. We may see hybrid systems where the clinician’s expertise is augmented by algorithmic suggestions and continuous data readouts, making DBS tuning more precise than ever before. The ultimate vision is a closed-loop DBS ecosystem: the clinician sets broad parameters, and the system fine-tunes itself in real time using multi-modal feedback (brain signals, motion sensors, etc.), with the patient’s symptoms continuously kept in check. Achieving this will require close collaboration between device manufacturers, software developers, and clinicians, but we see the beginnings of this shift already. The currently available DBS programming tools are powerful, but when paired with objective, real-time symptom monitoring, they become markedly more powerful in delivering optimal therapy while achieving high efficiency, ultimately making DBS accessible to more patients with greater ease and success.

1         DBS Programming: An Evolving Approach for Patients with Parkinson's Disease (https://pmc.ncbi.nlm.nih.gov/articles/PMC5632902/)

2         Automated deep brain stimulation programming with safety constraints for tremor suppression in patients with Parkinson’s disease and essential tremor (https://pmc.ncbi.nlm.nih.gov/articles/PMC9614806/)

3         SureTune™ 4 software (https://www.medtronic.com/en-us/healthcare-professionals/products/neurological/deep-brain-stimulation/dbs-therapy-leads/sensight-directional-lead/suretune-software.html) 

4         Percept™ PC neurostimulator with BrainSense™ technology (https://www.medtronic.com/en-us/healthcare-professionals/products/neurological/deep-brain-stimulation/electrical-stimulation-systems/percept-pc-neurostimulator.html)

5         The Place of Local Field Potentials in Deep Brain Stimulation Programming for Parkinson’s Disease: A Review (https://www.mdpi.com/2076-3425/15/2/116)

6         Access to Care at Home - Abbott NeuroSphere Virtual Clinic (https://time.com/collection/best-inventions-2021/6113022/neurosphere-virtual-clinic/)

7         Revolutionizing treatment for Parkinson's disease and essential tremor (https://www.neuromodulation.abbott/us/en/healthcare-professionals/parkinsons.html) 

8         Reduced Programming Time and Strong Symptom Control Even in Chronic Course Through Imaging-Based DBS Programming (https://pubmed.ncbi.nlm.nih.gov/34819915/)

9         Boston Scientific Nets FDA Approval for Image-Guided DBS Programming Software (https://appliedradiology.com/articles/boston-scientific-nets-fda-approval-for-image-guided-dbs-programming-software)

10      CLOVER-DBS: Algorithm-Guided Deep Brain Stimulation-Programming Based on External Sensor Feedback Evaluated in a Prospective, Randomized, Crossover, Double-Blind, Two-Center Study (https://pubmed.ncbi.nlm.nih.gov/34151855/)

11      Overview on wearable sensors for the management of Parkinson’s disease (https://pmc.ncbi.nlm.nih.gov/articles/PMC10622581/)

12      Motion sensor strategies for automated optimization of deep brain stimulation in Parkinson's disease (https://www.prd-journal.com/article/S1353-8020(15)00053-X/abstract)

13      Remote video-based outcome measures of patients with Parkinson’s disease after deep brain stimulation using smartphones: a pilot study (https://thejns.org/focus/view/journals/neurosurg-focus/51/5/article-pE2.xml)

14      Automated deep brain stimulation programming based on electrode location: a randomised, crossover trial using a data-driven algorithm (https://pubmed.ncbi.nlm.nih.gov/36528541/) 

15      Adaptive deep brain stimulation in advanced Parkinson disease (https://pubmed.ncbi.nlm.nih.gov/23852650/) 

16      Wearable sensor-driven responsive deep brain stimulation for essential tremor (https://www.sciencedirect.com/science/article/pii/S1935861X21002321)