Lauren H. Hammer

One-third of the world’s 70 million people with epilepsy have seizures uncontrolled by medication, and implantable devices are a promising treatment. These devices can detect impending seizures, missed medication and impaired cognition, but, like most medical devices, they cannot share this information with patients in real time. This is a missed opportunity for devices to learn from clinical events and teach patients how their behaviors affect their health. Here we present a device platform that converses with patients to help them manage their epilepsy. The inpatient prototype links scalp and intracranial electroencephalography (EEG) to a secure, large language model-based artificial intelligence (AI) that communicates bidirectionally through a smartphone app. The AI processes biomarkers of sleep, medication, cognition and seizure probability, notifies patients of clinical events and trends, records symptoms and responses, and uses their feedback to retrain its seizure detector, reducing false alarms. We validated this platform in 13 patients who are undergoing video-EEG monitoring for epilepsy (scalp, n = 11; intracranial, n = 2). In a pseudoprospective simulation, seizure detectors improved precision from patient feedback alone, without expert intervention. We propose this system as a scalable, safeguarded platform for medical devices that can rapidly adapt to patient and provider needs across epilepsy and other medical conditions.

Objective: Adaptive deep brain stimulation (aDBS) for Parkinson’s disease is a recently-approved therapy that adjusts stimulation in response to neurophysiologic biomarkers of motor-symptom state. Most real-time implementations of aDBS rely on instantaneous, noise-susceptible classifiers that apply simple thresholds to neurophysiologic biomarkers. We examined whether incorporating temporal history through Bayesian state-space modeling improved motor-state classification compared to instantaneous discriminant classifiers. Approach: We analyzed naturalistic neural data from three patients with Parkinson’s disease chronically implanted with investigational sensing-enabled DBS systems, recording from both the subthalamic nucleus (STN) and sensorimotor cortex. Biomarkers were extracted across multiple window lengths and labeled using wearable-derived bradykinesia and dyskinesia scores. Classifier behavior was evaluated using two biomarkers (cortical stimulation-entrained gamma and STN beta oscillations) across a factorial combination of two conditions: (1) instantaneous discriminant analysis vs. Bayesian time-history modeling via hidden Markov models (HMMs), and (2) single Gaussian vs. Gaussian mixture modeling of each motor state’s biomarker distribution. Performance metrics included F1 scores, accuracy, prediction smoothness, latency, and computational load. Main Results: Using entrained-gamma biomarkers, incorporating time history via HMMs significantly improved hyperkinetic-state detection (F1: +12.9 ± 1.8%; accuracy: +30.0 ± 2.7%; both padj < 0.001) with modest decreases in hypokinetic-state performance, yielding a net increase in average F1 (+4.7 ± 0.9%, p < 0.001). HMMs also yielded smoother and more accurate predictions for a given latency compared to simply increasing the window length used to extract neurophysiologic biomarkers. Entrained-gamma biomarkers outperformed STN beta biomarkers across all classifiers (average F1: +12.9% ± 0.5%, p < 0.001). All methods operated within sub-millisecond prediction times and demonstrated sublinear empirical computational scaling. Significance: Bayesian time-history modeling enhanced motor-state classification while preserving the low latency and computational efficiency required for real-time aDBS. These findings, derived from chronic at-home recordings, support the translational potential of Bayesian state-space models for next-generation aDBS systems.

Precise and synchronized multimodal data capture in neurosurgical environments is essential for further understanding brain function and will be crucial to advancing the development of brain-computer interface technology. We have developed an open-source software platform named Thalamus, for multimodal data capture integrated with existing sensors and hardware commonly utilized in the operating room and other clinical environments such as pulse oximeters, inertial sensors, electromyography and neural electrophysiology. Thalamus facilitates synchronous recording of neural and behavioral data, enabling real-time computation for closed-loop experiments and detailed analysis of complex motor functions and neural activity. Thalamus uses a modular, configurable node-based pipeline with a tiered Python and C + + architecture. These design elements allow Thalamus to support a wide range of high-resolution sensors for diverse behavioral data types and enable robust closed-loop synchronization of various data streams. Validation experiments demonstrate that Thalamus is capable of data integration and concurrent analysis with up to sub-millisecond precision, offering great potential for enhancing neurosurgical research and clinical applications. By leveraging conventional sensors and hardware already in use, Thalamus supports adoption into the clinical environment, paving the way for more comprehensive, data-driven approaches to neurological care and improving the personalization and rigor of treatment strategies.

Background: Adaptive deep brain stimulation (aDBS) using subthalamic nucleus (STN) beta‐band oscillations as a biomarker for the akinetic‐rigid symptoms of Parkinson's disease has been proposed to improve the efficacy of stimulation and decrease adverse effects. However, most studies validating STN beta oscillations as a biomarker used perioperative, off medication, and OFF stimulation conditions, limiting their potential relevance to real‐world aDBS deployment. Objective: We evaluated how stimulation and dopaminergic medication affect beta‐range STN oscillatory activity and its ability to predict the hypokinetic parkinsonian state. Methods: In six patients with Parkinson's disease who experienced residual levodopa‐related motor fluctuations despite standard‐of‐care DBS, we studied medication cycle‐induced fluctuations of peak STN beta oscillation power during active DBS. We evaluated the magnitude of beta oscillation fluctuations, the frequency of peak beta activity, and the ability to predict the off medication state. Neural signals were collected in the clinic during a medication challenge and at home during naturalistic medication cycles. Results: Medication‐induced beta oscillation fluctuations were smaller in the presence of therapeutic DBS compared with when the stimulator was off ( P = 0.008). Therapeutic stimulation also decreased the frequency of peak beta activity ( P < 0.001). As a result, the beta frequencies with the highest power in the OFF stimulation/ off medication state were less accurate than other beta‐range bands at predicting the patient's off medication state during active stimulation ( P = 0.02). Conclusions: The spectral characteristics of STN beta oscillations and their performance as a biomarker in aDBS for Parkinson's disease are influenced by the stimulation and medication conditions in which it is implemented. © 2026 International Parkinson and Movement Disorder Society.

One-third of the world’s 70 million people with epilepsy have seizures that are not controlled by medication; and implantable devices are an exciting option for treatment. These devices improve seizure control and can detect impending attacks, missed medication, and impaired cognition. Unfortunately, they have no way to share this information with their hosts in real-time – a limitation common to most medical devices. This is a missed opportunity for implants and wearables to learn from patients, focus on what matters most to them, and teach them how their behavior affects their health. Here, we present a device platform that converses with patients and learns to co-manage epilepsy. The inpatient pro-totype links scalp and intracranial EEG (electroencephalograms) to secure large language models that communicate freely and bidirectionally with their hosts through a smartphone app. An AI agent ingests biomarkers of sleep, medication level, cognition, and seizure risk extracted from brain activity. It con-verses with patients to inform them of clinical events and physiological trends, records their symptoms, responses, and behaviors, and automatically retrains itself to improve performance. Both patients and the AI agent can initiate conversations to teach each other and personalize interactions. We demon-strate this platform in 13 patients undergoing inpatient video-EEG monitoring for epilepsy and validate its performance. Algorithms for detecting seizures optimized their precision over several days without expert intervention – in contrast to the months of iterative, in-person physician programming currently required. Patients responded positively to messages regarding sleep, cognition, and seizure risk while rating the system as highly usable. The platform includes several safeguards, including a system for further algorithm fine-tuning using efficient expert review, and features that ensure data security and regulate communication content. Further work will link other biosensors to measure behavior, improve performance, and optimize therapeutic stimulation. We propose this system as a scalable platform for medical devices that can rapidly adapt to patient and provider needs; one that is broadly adaptable to improving care for many medical conditions.

We recorded over 500 h of neural activity in a patient with cervical dystonia using a sensing-enabled DBS device—the most extensive dataset to date. Multisite recordings were paired with wearable and kinematic analysis of 100+ home videos. Low-frequency and narrowband gamma oscillations correlated with tremor and were suppressed by DBS, which also induced entrained gamma oscillations. These neural changes tracked with clinical and kinematic improvements, suggesting potential biomarkers of DBS effectiveness.

Background: Hyperkinetic movement disorders involve excessive, involuntary movements such as ataxia, chorea, dystonia, myoclonus, tics, and tremor. Recent advances in artificial intelligence (AI) allow investigators to integrate multimodal instrumented movement measurements and imaging techniques and to analyze these data together at scale. In this systematic review, we aim to characterize AI’s performance in diagnosing and quantitatively phenotyping these disorders. Methods: We searched PubMed and Embase using a semi-automated article-screening pipeline. Results: Fifty-five studies met the inclusion criteria (n = 11,946 subjects). Thirty-five studies used machine learning, sixteen used deep learning, and four used both. Thirty-eight studies reported disease diagnosis, twenty-three reported quantitative phenotyping, and six reported both. Diagnostic accuracy was reported in 36 of 38 and correlation coefficients in 10 of 23 studies. Kinematics (e.g., accelerometers and inertial measurement units) were the most used dataset. Diagnostic accuracy was reported in 36 studies and ranged from 56 to 100% compared to clinical diagnoses to differentiate them from healthy controls. The correlation coefficient was reported in 10 studies and ranged from 0.54 to 0.99 compared to clinical ratings for quantitative phenotyping. Five studies had an overall judgment of “low risk of bias” and three had external validation. Conclusion: There is a need to adopt AI-based research guidelines to minimize reporting heterogeneity and bolster clinical interpretability.

Introduction Deep brain stimulation (DBS) of the subthalamic nucleus (STN) or globus pallidus (GP) is an established therapy for Parkinson’s disease (PD). Novel DBS devices can record local field potential (LFP) physiomarkers from the STN or GP. While beta (13-30 Hz) and gamma (40-90 Hz) STN and GP LFP oscillations correlate with PD motor severity and with therapeutic effects of treatments, STN-GP interactions in electrophysiology in patients with PD are not well characterized. Methods Simultaneous bilateral STN and GP LFPs were recorded in a patient with PD who received bilateral STN-DBS and GP-DBS. Power spectra in each target and STN-GP coherence were assessed in various ON- and OFF-levodopa and DBS states, both at rest and with voluntary movement. Results OFF-levodopa and OFF-DBS, beta peaks were present at bilateral STN and GP, coincident with prominent STN-GP beta coherence. Levodopa and dual-target-DBS (simultaneous STN-DBS and GP-DBS) completely suppressed STN-GP coherence. Finely-tuned gamma (FTG) activity at half the stimulation frequency (62.5Hz) was seen in the STN during GP-DBS at rest. To assess the effects of movement on FTG activity, we recorded LFPs during instructed movement. We observed FTG activity in bilateral GP and bilateral STN during contralateral body movements while on GP-DBS and ON-levodopa. No FTG was seen with STN-DBS or dual-target-DBS. Conclusion Dual-target-DBS and levodopa suppressed STN-GP coherence. FTG throughout the basal ganglia was induced by GP-DBS in the presence of levodopa and movement. This bilateral STN-FTG and GP-FTG corresponded with the least severe bradykinesia state, suggesting a pro-kinetic role for FTG.

1. Deep brain stimulation is a widely used therapy for Parkinson’s disease (PD) but currently lacks dynamic responsiveness to changing clinical and neural states. Feedback control has the potential to improve therapeutic effectiveness, but optimal control strategy and additional benefits of “adaptive” neurostimulation are unclear. We implemented adaptive subthalamic nucleus stimulation, controlled by subthalamic or cortical signals, in three PD patients (five hemispheres) during normal daily life. We identified neurophysiological biomarkers of residual motor fluctuations using data-driven analyses of field potentials over a wide frequency range and varying stimulation amplitudes. Narrowband gamma oscillations (65-70 Hz) at either site emerged as the best control signal for sensing during stimulation. A blinded, randomized trial demonstrated improved motor symptoms and quality of life compared to clinically optimized standard stimulation. Our approach highlights the promise of personalized adaptive neurostimulation based on data-driven selection of control signals and may be applied to other neurological disorders.