Subthalamic nucleus activity dynamics and limb movement prediction in Parkinson's disease

Bimanual coordination and neuromuscular synchronization in Parkinson's disease and healthy adults

Parkinson's disease (PD) is a progressive neurodegenerative disorder that significantly impairs motor function, affecting over 1.5 million people in the U.S. PD is characterized by deficits in movement speed, force timing, and force modulation, particularly during coordinated upper limb actions. These impairments contribute to reduced functional independence and a diminished quality of life in individuals with PD. This study investigated the impact of PD on bimanual coordination, focusing on temporal accuracy, force production, and neuromuscular synchronization. The goal was to compare these parameters across individuals with PD, healthy older adults (HOA), and healthy young adults (HYA) during a stable force coordination task. Thirteen individuals with PD (median age [min-max] = 73 [60-83] years; 6 males), 13 HOAs median age [min-max] = 74 [60-84] years; 7 males), and 15 HYAs (median age [min-max] = 21 [18-23] years; 7 males) performed a 1:1 in-phase (0°) bimanual coordination task, requiring participants to rhythmically produce isometric forces with their left and right index fingers. Muscle activity from the First Dorsal Interosseus (FDI) muscles were recorded using electromyography (EMG). Each participant completed 21, 30-second trials. Temporal accuracy and stability were assessed using frequency ratio, absolute error (AE), and variability (VE) of relative phase. Force production was evaluated in terms of force harmonicity, force asymmetry, and peak force. Neuromuscular synchronization was analyzed using force-force and EMG-EMG coherence across different frequency bands. All groups achieved the target frequency ratio of 1.0, with no significant differences in AE or VE, suggesting comparable temporal accuracy and stability across groups. However, the PD group demonstrated significantly lower harmonicity, indicating less smooth force production, and greater force asymmetry compared to HOA and HYA groups. Reduced force-force coherence, especially in the 1-4 Hz and 4-8 Hz frequency bands, further highlighted challenges in bilateral force synchronization for the PD group. EMG-EMG coherence analysis revealed that the HYA group exhibited higher muscle activation synchronization, particularly in the alpha band, compared to the PD group. These findings suggest that while basic temporal coordination remains intact in PD, the disease impairs the smoothness and symmetry of force production, likely due to disrupted neural synchronization. The observed correlations between higher force coherence, greater harmonicity, and lower force asymmetry underscore the critical role of neural drive coherence in achieving smooth and symmetrical force production. However, it is important to consider the impact of medication state, since all participants were tested around their "ON" medication state. Understanding these impairments can inform the development of targeted interventions and rehabilitation strategies aimed at improving motor function and quality of life in individuals with PD.

Structural covariance alterations reveal motor damage in periventricular leukomalacia

Periventricular leukomalacia is a common neuroimaging finding in patients with spastic cerebral palsy. Myelin damage disrupts neuronal connectivity. However, specific alterations in the grey matter structure and their impact on the whole brain remain unclear, particularly when differentiating between preterm and full-term periventricular leukomalacia. This study investigated the grey matter network alterations following early white matter injury in infants and young children. High-resolution T1-weighted 3 T brain magnetic resonance imaging, clinical data and motor function scores were collected from 42 children with periventricular leukomalacia and 38 age- and sex-matched healthy controls. Based on gestational age, the periventricular leukomalacia group was stratified into preterm (n = 27) and full-term (n = 15) groups. Voxel-based morphometry was used to analyse whole-brain structural metrics, and motor-related regions were selected as nodes for network construction. Structural covariance analysis was used to quantify the strength of the structural connections between grey matter regions, and graph theory metrics were used to assess network properties. Motor assessments included gross and fine motor skills, and their associations with brain regions were analysed. Both preterm and full-term periventricular leukomalacia groups exhibited abnormal motor networks. Preterm periventricular leukomalacia showed more extensive central grey matter nuclei atrophy, whereas full-term periventricular leukomalacia was predominantly localized to the motor cortex. Children with periventricular leukomalacia displayed decreased connectivity between the central grey matter nuclei and other regions, coupled with increased connectivity between the motor cortex and cerebellar hemispheres. Thalamic volume correlated with gross motor scores in preterm infants. These findings suggest that ischaemic-hypoxic injury disrupts motor grey matter networks, with preterm infants being more severely affected. This study highlights the potential of structural covariance patterns for monitoring brain development and advancing our understanding of aberrant brain development in children with periventricular leukomalacia.

Beta: bursts of cognition

Beta oscillations are linked to the control of goal-directed processing of sensory information and the timing of motor output. Recent evidence demonstrates they are not sustained but organized into intermittent high-power bursts mediating timely functional inhibition. This implies there is a considerable moment-to-moment variation in the neural dynamics supporting cognition. Beta bursts thus offer new opportunities for studying how sensory inputs are selectively processed, reshaped by inhibitory cognitive operations and ultimately result in motor actions. Recent method advances reveal diversity in beta bursts that provide deeper insights into their function and the underlying neural circuit activity motifs. We propose that brain-wide, spatiotemporal patterns of beta bursting reflect various cognitive operations and that their dynamics reveal nonlinear aspects of cortical processing.

Subthalamic stimulation modulates context-dependent effects of beta bursts during fine motor control

Increasing evidence suggests a considerable role of pre-movement beta bursts for motor control and its impairment in Parkinson's disease. However, whether beta bursts occur during precise and prolonged movements and if they affect fine motor control remains unclear. To investigate the role of within-movement beta bursts for fine motor control, we here combine invasive electrophysiological recordings and clinical deep brain stimulation in the subthalamic nucleus in 19 patients with Parkinson's disease performing a context-varying task that comprised template-guided and free spiral drawing. We determined beta bursts in narrow frequency bands around patient-specific peaks and assessed burst amplitude, duration, and their immediate impact on drawing speed. We reveal that beta bursts occur during the execution of drawing movements with reduced duration and amplitude in comparison to rest. Exclusively when drawing freely, they parallel reductions in acceleration. Deep brain stimulation increases the acceleration around beta bursts in addition to a general increase in drawing velocity and improvements of clinical function. These results provide evidence for a diverse and task-specific role of subthalamic beta bursts for fine motor control in Parkinson's disease; suggesting that pathological beta bursts act in a context dependent manner, which can be targeted by clinical deep brain stimulation.

Oscillatory and non-oscillatory features of the magnetoencephalic sensorimotor rhythm in Parkinson's disease

Parkinson's disease (PD) is associated with changes in neural activity in the sensorimotor alpha and beta bands. Using magnetoencephalography (MEG), we investigated the role of spontaneous neuronal activity within the somatosensory cortex in a large cohort of early- to mid-stage PD patients (N = 78) on Parkinsonian medication and age- and sex-matched healthy controls (N = 60) using source reconstructed resting-state MEG. We quantified features of the time series data in terms of oscillatory alpha power and central alpha frequency, beta power and central beta frequency, and 1/f broadband characteristics using power spectral density. Furthermore, we characterised transient oscillatory burst events in the mu-beta band time-domain signals. We examined the relationship between these signal features and the patients' disease state, symptom severity, age, sex, and cortical thickness. PD patients and healthy controls differed on PSD broadband characteristics, with PD patients showing a steeper 1/f exponential slope and higher 1/f offset. PD patients further showed a steeper age-related decrease in the burst rate. Out of all the signal features of the sensorimotor activity, the burst rate was associated with increased severity of bradykinesia, whereas the burst duration was associated with axial symptoms. Our study shows that general non-oscillatory features (broadband 1/f exponent and offset) of the sensorimotor signals are related to disease state and oscillatory burst rate scales with symptom severity in PD.

Beta bursts question the ruling power for brain-computer interfaces

Objective: Current efforts to build reliable brain-computer interfaces (BCI) span multiple axes from hardware, to software, to more sophisticated experimental protocols, and personalized approaches. However, despite these abundant efforts, there is still room for significant improvement. We argue that a rather overlooked direction lies in linking BCI protocols with recent advances in fundamental neuroscience.Approach: In light of these advances, and particularly the characterization of the burst-like nature of beta frequency band activity and the diversity of beta bursts, we revisit the role of beta activity in 'left vs. right hand' motor imagery (MI) tasks. Current decoding approaches for such tasks take advantage of the fact that MI generates time-locked changes in induced power in the sensorimotor cortex and rely on band-passed power changes in single or multiple channels. Although little is known about the dynamics of beta burst activity during MI, we hypothesized that beta bursts should be modulated in a way analogous to their activity during performance of real upper limb movements.Main results and Significance: We show that classification features based on patterns of beta burst modulations yield decoding results that are equivalent to or better than typically used beta power across multiple open electroencephalography datasets, thus providing insights into the specificity of these bio-markers.

Machine learning for adaptive deep brain stimulation in Parkinson's disease: closing the loop

Parkinson's disease (PD) is the second most common neurodegenerative disease bearing a severe social and economic impact. So far, there is no known disease modifying therapy and the current available treatments are symptom oriented. Deep Brain Stimulation (DBS) is established as an effective treatment for PD, however current systems lag behind today's technological potential. Adaptive DBS, where stimulation parameters depend on the patient's physiological state, emerges as an important step towards "smart" DBS, a strategy that enables adaptive stimulation and personalized therapy. This new strategy is facilitated by currently available neurotechnologies allowing the simultaneous monitoring of multiple signals, providing relevant physiological information. Advanced computational models and analytical methods are an important tool to explore the richness of the available data and identify signal properties to close the loop in DBS. To tackle this challenge, machine learning (ML) methods applied to DBS have gained popularity due to their ability to make good predictions in the presence of multiple variables and subtle patterns. ML based approaches are being explored at different fronts such as the identification of electrophysiological biomarkers and the development of personalized control systems, leading to effective symptom relief. In this review, we explore how ML can help overcome the challenges in the development of closed-loop DBS, particularly its role in the search for effective electrophysiology biomarkers. Promising results demonstrate ML potential for supporting a new generation of adaptive DBS, with better management of stimulation delivery, resulting in more efficient and patient-tailored treatments.

Wavelet-Based Bracketing, Time-Frequency Beta Burst Detection: New Insights in Parkinson's Disease

Studies have shown that beta band activity is not tonically elevated but comprises exaggerated phasic bursts of varying durations and magnitudes, for Parkinson's disease (PD) patients. Current methods for detecting beta bursts target a single frequency peak in beta band, potentially ignoring bursts in the wider beta band. In this study, we propose a new robust framework for beta burst identification across wide frequency ranges. Chronic local field potential at-rest recordings were obtained from seven PD patients implanted with Medtronic SenSight™ deep brain stimulation (DBS) electrodes. The proposed method uses wavelet decomposition to compute the time-frequency spectrum and identifies bursts spanning multiple frequency bins by thresholding, offering an additional burst measure, ∆f, that captures the width of a burst in the frequency domain. Analysis included calculating burst duration, magnitude, and ∆f and evaluating the distribution and likelihood of bursts between the low beta (13-20 Hz) and high beta (21-35 Hz). Finally, the results of the analysis were correlated to motor impairment (MDS-UPDRS III) med off scores. We found that low beta bursts with longer durations and larger width in the frequency domain (∆f) were positively correlated, while high beta bursts with longer durations and larger ∆f were negatively correlated with motor impairment. The proposed method, finding clear differences between bursting behavior in high and low beta bands, has clearly demonstrated the importance of considering wide frequency bands for beta burst behavior with implications for closed-loop DBS paradigms.

Diverse beta burst waveform motifs characterize movement-related cortical dynamics

Classical analyses of induced, frequency-specific neural activity typically average band-limited power over trials. More recently, it has become widely appreciated that in individual trials, beta band activity occurs as transient bursts rather than amplitude-modulated oscillations. Most studies of beta bursts treat them as unitary, and having a stereotyped waveform. However, we show there is a wide diversity of burst shapes. Using a biophysical model of burst generation, we demonstrate that waveform variability is predicted by variability in the synaptic drives that generate beta bursts. We then use a novel, adaptive burst detection algorithm to identify bursts from human MEG sensor data recorded during a joystick-based reaching task, and apply principal component analysis to burst waveforms to define a set of dimensions, or motifs, that best explain waveform variance. Finally, we show that bursts with a particular range of waveform motifs, ones not fully accounted for by the biophysical model, differentially contribute to movement-related beta dynamics. Sensorimotor beta bursts are therefore not homogeneous events and likely reflect distinct computational processes.

Oscillatory beta dynamics inform biomarker-driven treatment optimization for Parkinson's disease

Parkinson's disease (PD) is a neurodegenerative disorder characterized by loss of dopaminergic neurons and dysregulation of the basal ganglia. Cardinal motor symptoms include bradykinesia, rigidity, and tremor. Deep brain stimulation (DBS) of select subcortical nuclei is standard of care for medication-refractory PD. Conventional open-loop DBS delivers continuous stimulation with fixed parameters that do not account for a patient's dynamic activity state or medication cycle. In comparison, closed-loop DBS, or adaptive DBS (aDBS), adjusts stimulation based on biomarker feedback that correlates with clinical state. Recent work has identified several neurophysiological biomarkers in local field potential recordings from PD patients, the most promising of which are 1) elevated beta (∼13-30 Hz) power in the subthalamic nucleus (STN), 2) increased beta synchrony throughout basal ganglia-thalamocortical circuits, notably observed as coupling between the STN beta phase and cortical broadband gamma (∼50-200 Hz) amplitude, and 3) prolonged beta bursts in the STN and cortex. In this review, we highlight relevant frequency and time domain features of STN beta measured in PD patients and summarize how spectral beta power, oscillatory beta synchrony, phase-amplitude coupling, and temporal beta bursting inform PD pathology, neurosurgical targeting, and DBS therapy. We then review how STN beta dynamics inform predictive, biomarker-driven aDBS approaches for optimizing PD treatment. We therefore provide clinically useful and actionable insight that can be applied toward aDBS implementation for PD.

Balance between pallidal neural oscillations correlated with dystonic activity and severity

BACKGROUND AND OBJECTIVE

The balance between neural oscillations provides valuable insights into the organisation of neural oscillations related to brain states, which may play important roles in dystonia. We aim to investigate the relationship of the balance in the globus pallidus internus (GPi) with the dystonic severity under different muscular contraction conditions.

METHODS

Twenty-one patients with dystonia were recruited. All of them underwent bilateral GPi implantation, and local field potentials (LFPs) from the GPi were recorded via simultaneous surface electromyography. The power spectral ratio between neural oscillations was computed as the measure of neural balance. This ratio was calculated under high and low dystonic muscular contraction conditions, and its correlation with the dystonic severity was assessed using clinical scores.

RESULTS

The power spectral of the pallidal LFPs peaked in the theta and alpha bands. Within participant comparison showed that the power spectral of the theta oscillations significantly increased during high muscle contraction compared with that during low contraction. The power spectral ratios between the theta and alpha, theta and low beta, and theta and high gamma oscillations were significantly higher during high contraction than during low contraction. The total score and motor score were associated with the power spectral ratio between the low and high beta oscillations, which was correlated with the dystonic severity both during high and low contractions. The power spectral ratios between the low beta and low gamma and between the low beta and high gamma oscillations showed a significantly positive correlation with the total score during both high and low contractions; a correlation with the motor scale score was found only during high contraction. Meanwhile, the power spectral ratio between the theta and alpha oscillations during low contraction showed a significantly negative correlation with the total score. The power spectral ratios between the alpha and high beta, alpha and low gamma, and alpha and high gamma oscillations were significantly correlated with the dystonic severity only during low contraction.

CONCLUSION

The balance between neural oscillations, as quantified by the power ratio between specific frequency bands, differed between the high and low muscular contraction conditions and was correlated with the dystonic severity. The balance between the low and high beta oscillations was correlated with the dystonic severity during both conditions, making this parameter a new possible biomarker for closed-loop deep brain stimulation in patients with dystonia.

Landscape and future directions of machine learning applications in closed-loop brain stimulation

Brain stimulation (BStim) encompasses multiple modalities (e.g., deep brain stimulation, responsive neurostimulation) that utilize electrodes implanted in deep brain structures to treat neurological disorders. Currently, BStim is primarily used to treat movement disorders such as Parkinson's, though indications are expanding to include neuropsychiatric disorders like depression and schizophrenia. Traditional BStim systems are "open-loop" and deliver constant electrical stimulation based on manually-determined parameters. Advancements in BStim have enabled development of "closed-loop" systems that analyze neural biomarkers (e.g., local field potentials in the sub-thalamic nucleus) and adjust electrical modulation in a dynamic, patient-specific, and energy efficient manner. These closed-loop systems enable real-time, context-specific stimulation adjustment to reduce symptom burden. Machine learning (ML) has emerged as a vital component in designing these closed-loop systems as ML models can predict / identify presence of disease symptoms based on neural activity and adaptively learn to modulate stimulation. We queried the US National Library of Medicine PubMed database to understand the role of ML in developing closed-loop BStim systems to treat epilepsy, movement disorders, and neuropsychiatric disorders. Both neural and non-neural network ML algorithms have successfully been leveraged to create closed-loop systems that perform comparably to open-loop systems. For disorders in which the underlying neural pathophysiology is relatively well understood (e.g., Parkinson's, essential tremor), most work has involved refining ML models that can classify neural signals as aberrant or normal. The same is seen for epilepsy, where most current research has focused on identifying optimal ML model design and integrating closed-loop systems into existing devices. For neuropsychiatric disorders, where the underlying pathologic neural circuitry is still being investigated, research is focused on identifying biomarkers (e.g., local field potentials from brain nuclei) that ML models can use to identify onset of symptoms and stratify severity of disease.

Deep Brain Stimulation-Withdrawal Syndrome in Parkinson's Disease: Risk Factors and Pathophysiological Hypotheses of a Life-Threatening Emergency

BACKGROUND AND OBJECTIVES

Subthalamic nucleus deep brain stimulation (DBS) is the most common therapeutic surgical procedure for patients with Parkinson's disease with motor fluctuations, dyskinesia, or tremor. Routine follow-up of patients allows clinicians to anticipate replacement of the DBS battery reaching the end of its life. Patients who experience a sudden stop of the DBS battery experience a rapid worsening of symptoms unresponsive to high dose of levodopa, in a life-threatening phenomenon called "DBS-withdrawal syndrome." In the current context of the COVID-19 pandemic, in which many surgeries are being deprogrammed, it is of utmost importance to determine to what extent DBS battery replacement surgeries should be considered an emergency. In this study, we attempt to identify risk factors of DBS-withdrawal syndrome and provide new insights about pathophysiological hypotheses. We then elaborate on the optimal approach to avoid and manage such a situation.

MATERIALS AND METHODS

We conducted a systematic review of the literature on the subject and reported the cases of 20 patients (including five from our experience) with DBS-withdrawal syndrome, comparing them with 15 undisturbed patients (including three from our experience), all having undergone neurostimulation discontinuation.

RESULTS

A long disease duration at battery removal and many years of DBS therapy are the main potential identified risk factors (p < 0.005). In addition, a trend for older age at the event and higher Unified Parkinson's Disease Rating Scale motor score before initial DBS implantation (evaluated in OFF-drug condition) was found (p < 0.05). We discuss several hypotheses that might explain this phenomenon, including discontinued functioning of the thalamic-basal ganglia loop due to DBS-stimulation cessation in a context in which cortical-basal ganglia loop had lost its cortical input, and possible onset of a severe bradykinesia through the simultaneous occurrence of an alpha and high-beta synchronized state.

CONCLUSIONS

The patients' clinical condition may deteriorate rapidly, be unresponsive to high dose of levodopa, and become life-threatening. Hospitalization is suggested for clinical monitoring. In the context of the current COVID-19 pandemic, it is important to widely communicate the replacement of DBS batteries reaching the end of their life. More importantly, in cases in which the battery has stopped, there should be no delay in performing replacement as an emergent surgery.

When do bursts matter in the primary motor cortex? Investigating changes in the intermittencies of beta rhythms associated with movement states

Brain activity exhibits significant temporal structure that is not well captured in the power spectrum. Recently, attention has shifted to characterising the properties of intermittencies in rhythmic neural activity (i.e. bursts), yet the mechanisms that regulate them are unknown. Here, we present evidence from electrocorticography recordings made over the motor cortex to show that the statistics of bursts, such as duration or amplitude, in the beta frequency (14-30 Hz) band, significantly aid the classification of motor states such as rest, movement preparation, execution, and imagery. These features reflect nonlinearities not detectable in the power spectrum, with states increasing in nonlinearity from movement execution to preparation to rest. Further, we show using a computational model of the cortical microcircuit, constrained to account for burst features, that modulations of laminar specific inhibitory interneurons are responsible for the temporal organisation of activity. Finally, we show that the temporal characteristics of spontaneous activity can be used to infer the balance of cortical integration between incoming sensory information and endogenous activity. Critically, we contribute to the understanding of how transient brain rhythms may underwrite cortical processing, which in turn, could inform novel approaches for brain state classification, and modulation with novel brain-computer interfaces.

A systematic review of local field potential physiomarkers in Parkinson's disease: from clinical correlations to adaptive deep brain stimulation algorithms

Deep brain stimulation (DBS) treatment has proven effective in suppressing symptoms of rigidity, bradykinesia, and tremor in Parkinson's disease. Still, patients may suffer from disabling fluctuations in motor and non-motor symptom severity during the day. Conventional DBS treatment consists of continuous stimulation but can potentially be further optimised by adapting stimulation settings to the presence or absence of symptoms through closed-loop control. This critically relies on the use of 'physiomarkers' extracted from (neuro)physiological signals. Ideal physiomarkers for adaptive DBS (aDBS) are indicative of symptom severity, detectable in every patient, and technically suitable for implementation. In the last decades, much effort has been put into the detection of local field potential (LFP) physiomarkers and in their use in clinical practice. We conducted a research synthesis of the correlations that have been reported between LFP signal features and one or more specific PD motor symptoms. Features based on the spectral beta band (~ 13 to 30 Hz) explained ~ 17% of individual variability in bradykinesia and rigidity symptom severity. Limitations of beta band oscillations as physiomarker are discussed, and strategies for further improvement of aDBS are explored.

Closed-Loop Adaptive Deep Brain Stimulation in Parkinson's Disease: Procedures to Achieve It and Future Perspectives

Parkinson's disease (PD) is a neurodegenerative disease with a heavy burden on patients, families, and society. Deep brain stimulation (DBS) can improve the symptoms of PD patients for whom medication is insufficient. However, current open-loop uninterrupted conventional DBS (cDBS) has inherent limitations, such as adverse effects, rapid battery consumption, and a need for frequent parameter adjustment. To overcome these shortcomings, adaptive DBS (aDBS) was proposed to provide responsive optimized stimulation for PD. This topic has attracted scientific interest, and a growing body of preclinical and clinical evidence has shown its benefits. However, both achievements and challenges have emerged in this novel field. To date, only limited reviews comprehensively analyzed the full framework and procedures for aDBS implementation. Herein, we review current preclinical and clinical data on aDBS for PD to discuss the full procedures for its achievement and to provide future perspectives on this treatment.

Progress in the development of a fully implantable brain-computer interface: the potential of sensing-enabled neurostimulators

This perspective article investigates the performance of using a sensing-enabled neurostimulator as a motor brain-computer interface.

Electrocorticography is superior to subthalamic local field potentials for movement decoding in Parkinson's disease

Brain signal decoding promises significant advances in the development of clinical brain computer interfaces (BCI). In Parkinson's disease (PD), first bidirectional BCI implants for adaptive deep brain stimulation (DBS) are now available. Brain signal decoding can extend the clinical utility of adaptive DBS but the impact of neural source, computational methods and PD pathophysiology on decoding performance are unknown. This represents an unmet need for the development of future neurotechnology. To address this, we developed an invasive brain-signal decoding approach based on intraoperative sensorimotor electrocorticography (ECoG) and subthalamic LFP to predict grip-force, a representative movement decoding application, in 11 PD patients undergoing DBS. We demonstrate that ECoG is superior to subthalamic LFP for accurate grip-force decoding. Gradient boosted decision trees (XGBOOST) outperformed other model architectures. ECoG based decoding performance negatively correlated with motor impairment, which could be attributed to subthalamic beta bursts in the motor preparation and movement period. This highlights the impact of PD pathophysiology on the neural capacity to encode movement vigor. Finally, we developed a connectomic analysis that could predict grip-force decoding performance of individual ECoG channels across patients by using their connectomic fingerprints. Our study provides a neurophysiological and computational framework for invasive brain signal decoding to aid the development of an individualized precision-medicine approach to intelligent adaptive DBS.

Machine learning based brain signal decoding for intelligent adaptive deep brain stimulation

Sensing enabled implantable devices and next-generation neurotechnology allow real-time adjustments of invasive neuromodulation. The identification of symptom and disease-specific biomarkers in invasive brain signal recordings has inspired the idea of demand dependent adaptive deep brain stimulation (aDBS). Expanding the clinical utility of aDBS with machine learning may hold the potential for the next breakthrough in the therapeutic success of clinical brain computer interfaces. To this end, sophisticated machine learning algorithms optimized for decoding of brain states from neural time-series must be developed. To support this venture, this review summarizes the current state of machine learning studies for invasive neurophysiology. After a brief introduction to the machine learning terminology, the transformation of brain recordings into meaningful features for decoding of symptoms and behavior is described. Commonly used machine learning models are explained and analyzed from the perspective of utility for aDBS. This is followed by a critical review on good practices for training and testing to ensure conceptual and practical generalizability for real-time adaptation in clinical settings. Finally, first studies combining machine learning with aDBS are highlighted. This review takes a glimpse into the promising future of intelligent adaptive DBS (iDBS) and concludes by identifying four key ingredients on the road for successful clinical adoption: i) multidisciplinary research teams, ii) publicly available datasets, iii) open-source algorithmic solutions and iv) strong world-wide research collaborations.

Combining Multimodal Biomarkers to Guide Deep Brain Stimulation Programming in Parkinson Disease

BACKGROUND

Deep brain stimulation (DBS) programming of multicontact DBS leads relies on a very time-consuming manual screening procedure, and strategies to speed up this process are needed. Beta activity in subthalamic nucleus (STN) local field potentials (LFP) has been suggested as a promising marker to index optimal stimulation contacts in patients with Parkinson disease.

OBJECTIVE

In this study, we investigate the advantage of algorithmic selection and combination of multiple resting and movement state features from STN LFPs and imaging markers to predict three relevant clinical DBS parameters (clinical efficacy, therapeutic window, side-effect threshold).

MATERIALS AND METHODS

STN LFPs were recorded at rest and during voluntary movements from multicontact DBS leads in 27 hemispheres. Resting- and movement-state features from multiple frequency bands (alpha, low beta, high beta, gamma, fast gamma, high frequency oscillations [HFO]) were used to predict the clinical outcome parameters. Subanalyses included an anatomical stimulation sweet spot as an additional feature.

RESULTS

Both resting- and movement-state features contributed to the prediction, with resting (fast) gamma activity, resting/movement-modulated beta activity, and movement-modulated HFO being most predictive. With the proposed algorithm, the best stimulation contact for the three clinical outcome parameters can be identified with a probability of almost 90% after considering half of the DBS lead contacts, and it outperforms the use of beta activity as single marker. The combination of electrophysiological and imaging markers can further improve the prediction.

CONCLUSION

LFP-guided DBS programming based on algorithmic selection and combination of multiple electrophysiological and imaging markers can be an efficient approach to improve the clinical routine and outcome of DBS patients.

An Impending Paradigm Shift in Motor Imagery Based Brain-Computer Interfaces

The development of reliable assistive devices for patients that suffer from motor impairments following central nervous system lesions remains a major challenge in the field of non-invasive Brain-Computer Interfaces (BCIs). These approaches are predominated by electroencephalography and rely on advanced signal processing and machine learning methods to extract neural correlates of motor activity. However, despite tremendous and still ongoing efforts, their value as effective clinical tools remains limited. We advocate that a rather overlooked research avenue lies in efforts to question neurophysiological markers traditionally targeted in non-invasive motor BCIs. We propose an alternative approach grounded by recent fundamental advances in non-invasive neurophysiology, specifically subject-specific feature extraction of sensorimotor bursts of activity recorded via (possibly magnetoencephalography-optimized) electroencephalography. This path holds promise in overcoming a significant proportion of existing limitations, and could foster the wider adoption of online BCIs in rehabilitation protocols.

Controlling Clinical States Governed by Different Temporal Dynamics With Closed-Loop Deep Brain Stimulation: A Principled Framework

Closed-loop strategies for deep brain stimulation (DBS) are paving the way for improving the efficacy of existing neuromodulation therapies across neurological disorders. Unlike continuous DBS, closed-loop DBS approaches (cl-DBS) optimize the delivery of stimulation in the temporal domain. However, clinical and neurophysiological manifestations exhibit highly diverse temporal properties and evolve over multiple time-constants. Moreover, throughout the day, patients are engaged in different activities such as walking, talking, or sleeping that may require specific therapeutic adjustments. This broad range of temporal properties, along with inter-dependencies affecting parallel manifestations, need to be integrated in the development of therapies to achieve a sustained, optimized control of multiple symptoms over time. This requires an extended view on future cl-DBS design. Here we propose a conceptual framework to guide the development of multi-objective therapies embedding parallel control loops. Its modular organization allows to optimize the personalization of cl-DBS therapies to heterogeneous patient profiles. We provide an overview of clinical states and symptoms, as well as putative electrophysiological biomarkers that may be integrated within this structure. This integrative framework may guide future developments and become an integral part of next-generation precision medicine instruments.

Adaptive, personalized closed-loop therapy for Parkinson's disease: biochemical, neurophysiological, and wearable sensing systems

INTRODUCTION

Motor complication management is one of the main unmet needs in Parkinson's disease patients.

AREAS COVERED

Among the most promising emerging approaches for handling motor complications in Parkinson's disease, adaptive deep brain stimulation strategies operating in closed-loop have emerged as pivotal to deliver sustained, near-to-physiological inputs to dysfunctional basal ganglia-cortical circuits over time. Existing sensing systems that can provide feedback signals to close the loop include biochemical-, neurophysiological- or wearable-sensors. Biochemical sensing allows to directly monitor the pharmacokinetic and pharmacodynamic of antiparkinsonian drugs and metabolites. Neurophysiological sensing relies on neurotechnologies to sense cortical or subcortical brain activity and extract real-time correlates of symptom intensity or symptom control during DBS. A more direct representation of the symptom state, particularly the phenomenological differentiation and quantification of motor symptoms, can be realized via wearable sensor technology.

EXPERT OPINION

Biochemical, neurophysiologic, and wearable-based biomarkers are promising technological tools that either individually or in combination could guide adaptive therapy for Parkinson's disease motor symptoms in the future.

Continuously Decoding Grasping Movements using Stereotactic Depth Electrodes

Brain-Computer Interfaces (BCIs) that decode a patient's movement intention to control a prosthetic device could restore some independence to paralyzed patients. An important step on the road towards naturalistic prosthetic control is to decode movement continuously with low-latency. BCIs based on intracortical micro-arrays provide continuous control of robotic arms, but require a minor craniotomy. Surface recordings of neural activity using EEG have made great advances over the last years, but suffer from high noise levels and large intra-session variance. Here, we investigate the use of minimally invasive recordings using stereotactically implanted EEG (sEEG). These electrodes provide a sparse sampling across many brain regions. So far, promising decoding results have been presented using data measured from the subthalamic nucleus or trial-to-trial based methods using depth electrodes. In this work, we demonstrate that grasping movements can continuously be decoded using sEEG electrodes, as well. Beta and high-gamma activity was extracted from eight participants performing a grasping task. We demonstrate above chance level decoding of movement vs rest and left vs right, from both frequency bands with accuracies up to 0.94 AUC. The vastly different electrode locations between participants lead to large variability. In the future, we hope that sEEG recordings will provide additional information for the decoding process in neuroprostheses.

Subthalamic dynamic neural states correlate with motor symptoms in Parkinson's Disease

OBJECTIVE

This study aims to discriminate the dynamic synchronization states from the subthalamic local field potentials and investigate their correlations with the motor symptoms in Parkinson's Disease (PD).

METHODS

The resting-state local field potentials of 10 patients with PD were recorded from the subthalamic nucleus. The dynamic neural states of multiple oscillations were discriminated and analyzed. The Spearman correlation was used to investigate the correlations between occurrence rate or duration of dynamic neural states and the severity of motor symptoms.

RESULTS

The proportion of long low-beta and theta synchronized state was significantly correlated with the general motor symptom and tremor, respectively. The duration of combined low/high-beta state was significantly correlated with rigidity, and the duration of combined alpha/high-beta state was significantly correlated with bradykinesia.

CONCLUSIONS

This study provides evidence that motor symptoms are associated with the neural states coded with multiple oscillations in PD.

SIGNIFICANCE

This study may advance the understanding of the neurophysiological mechanisms of the motor symptoms and provide potential biomarkers for closed-loop deep brain stimulation in PD.

Average beta burst duration profiles provide a signature of dynamical changes between the ON and OFF medication states in Parkinson's disease

Parkinson's disease motor symptoms are associated with an increase in subthalamic nucleus beta band oscillatory power. However, these oscillations are phasic, and there is a growing body of evidence suggesting that beta burst duration may be of critical importance to motor symptoms. This makes insights into the dynamics of beta bursting generation valuable, in particular to refine closed-loop deep brain stimulation in Parkinson's disease. In this study, we ask the question "Can average burst duration reveal how dynamics change between the ON and OFF medication states?". Our analysis of local field potentials from the subthalamic nucleus demonstrates using linear surrogates that the system generating beta oscillations is more likely to act in a non-linear regime OFF medication and that the change in a non-linearity measure is correlated with motor impairment. In addition, we pinpoint the simplest dynamical changes that could be responsible for changes in the temporal patterning of beta oscillations between medication states by fitting to data biologically inspired models, and simpler beta envelope models. Finally, we show that the non-linearity can be directly extracted from average burst duration profiles under the assumption of constant noise in envelope models. This reveals that average burst duration profiles provide a window into burst dynamics, which may underlie the success of burst duration as a biomarker. In summary, we demonstrate a relationship between average burst duration profiles, dynamics of the system generating beta oscillations, and motor impairment, which puts us in a better position to understand the pathology and improve therapies such as deep brain stimulation.

Impaired reach-to-grasp kinematics in parkinsonian patients relates to dopamine-dependent, subthalamic beta bursts

Excessive beta-band oscillations in the subthalamic nucleus are key neural features of Parkinson’s disease. Yet the distinctive contributions of beta low and high bands, their dependency on striatal dopamine, and their correlates with movement kinematics are unclear. Here, we show that the movement phases of the reach-to-grasp motor task are coded by the subthalamic bursting activity in a maximally-informative beta high range. A strong, three-fold correlation linked beta high range bursts, imbalanced inter-hemispheric striatal dopaminergic tone, and impaired inter-joint movement coordination. These results provide new insight into the neural correlates of motor control in parkinsonian patients, paving the way for more informative use of beta-band features for adaptive deep brain stimulation devices.

Local field potentials in Parkinson's disease: A frequency-based review

Deep brain stimulation (DBS) surgery offers a unique opportunity to record local field potentials (LFPs), the electrophysiological population activity of neurons surrounding the depth electrode in the target area. With direct access to the subcortical activity, LFP research has provided valuable insight into disease mechanisms and cognitive processes and inspired the advent of adaptive DBS for Parkinson's disease (PD). A frequency-based framework is usually employed to interpret the implications of LFP signatures in LFP studies on PD. This approach standardizes the methodology, simplifies the interpretation of LFP patterns, and makes the results comparable across studies. Importantly, previous works have found that activity patterns do not represent disease-specific activity but rather symptom-specific or task-specific neuronal signatures that relate to the current motor, cognitive or emotional state of the patient and the underlying disease. In the present review, we aim to highlight distinguishing features of frequency-specific activities, mainly within the motor domain, recorded from DBS electrodes in patients with PD. Associations of the commonly reported frequency bands (delta, theta, alpha, beta, gamma, and high-frequency oscillations) to motor signs are discussed with respect to band-related phenomena such as individual tremor and high/low beta frequency activity, as well as dynamic transients of beta bursts. We provide an overview on how electrophysiology research in DBS patients has revealed and will continuously reveal new information about pathophysiology, symptoms, and behavior, e.g., when combining deep LFP and surface electrocorticography recordings.

Is there a single beta oscillation band interfering with movement in Parkinson's disease?

Beta oscillations in basal ganglia are considered to contribute to motor dysfunction in Parkinson's disease (PD). However, there is a high variety in frequency borders for beta oscillations between studies, which complicates the comparison and interpretation of results. Here we aimed to study the homogeneity of oscillations in the broad "beta" range (8-30 Hz) and their implication to motor functioning in PD. For this purpose, we recorded local field potentials (LFP) in the subthalamic nucleus (STN) during 34 deep brain stimulation surgeries. We identified spectral features of LFP recordings in the range 8-30 Hz to search for candidate sub-regions of stable oscillations and assessed their association with clinical scores on the contralateral side of the body and sensitivity to motor tests. Lower frequency oscillations (8-16 Hz) had a significant positive association with bradykinesia score. During voluntary movements, we observed a significant increase in LFP power in the 12-16 Hz range and a decrease in the 18-26 Hz range. We may conclude that the 8-30 Hz oscillation range includes oscillations with different functional features-sensitivity and responsiveness to movement, and clinical symptoms, which should be taken into account in further studies of beta oscillations association with PD pathophysiology. These data assume the coexistence of several frequency domains within beta range that are modulated in different ways under dopaminergic regulation and motor processing in human STN.

Machine Learning's Application in Deep Brain Stimulation for Parkinson's Disease: A Review

Deep brain stimulation (DBS) is a surgical treatment for advanced Parkinson's disease (PD) that has undergone technological evolution that parallels an expansion in clinical phenotyping, neurophysiology, and neuroimaging of the disease state. Machine learning (ML) has been successfully used in a wide range of healthcare problems, including DBS. As computational power increases and more data become available, the application of ML in DBS is expected to grow. We review the literature of ML in DBS and discuss future opportunities for such applications. Specifically, we perform a comprehensive review of the literature from PubMed, the Institute for Scientific Information's Web of Science, Cochrane Database of Systematic Reviews, and Institute of Electrical and Electronics Engineers' (IEEE) Xplore Digital Library for ML applications in DBS. These studies are broadly placed in the following categories: (1) DBS candidate selection; (2) programming optimization; (3) surgical targeting; and (4) insights into DBS mechanisms. For each category, we provide and contextualize the current body of research and discuss potential future directions for the application of ML in DBS.

Prediction of mild parkinsonism revealed by neural oscillatory changes and machine learning

Neural oscillatory changes within and across different frequency bands are thought to underlie motor dysfunction in Parkinson's disease (PD) and may serve as biomarkers for closed-loop deep brain stimulation (DBS) approaches. Here, we used neural oscillatory signals derived from chronically implanted cortical and subcortical electrode arrays as features to train machine learning algorithms to discriminate between naive and mild PD states in a nonhuman primate model. Local field potential (LFP) data were collected over several months from a 12-channel subdural electrocorticography (ECoG) grid and a 6-channel custom array implanted in the subthalamic nucleus (STN). Relative to the naive state, the PD state showed elevated primary motor cortex (M1) and STN power in the beta, high gamma, and high-frequency oscillation (HFO) bands and decreased power in the delta band. Theta power was found to be decreased in STN but not M1. In the PD state there was elevated beta-HFO phase-amplitude coupling (PAC) in the STN. We applied machine learning with support vector machines with radial basis function (SVM-RBF) kernel and k-nearest neighbors (KNN) classifiers trained by features related to power and PAC changes to discriminate between the naive and mild states. Our results show that the most predictive feature of parkinsonism in the STN was high beta (∼86% accuracy), whereas it was HFO in M1 (∼98% accuracy). A feature fusion approach outperformed every individual feature, particularly in the M1, where ∼98% accuracy was achieved with both classifiers. Overall, our data demonstrate the ability to use various frequency band power to classify the clinical state and are also beneficial in developing closed-loop DBS therapeutic approaches.NEW & NOTEWORTHY Neurophysiological biomarkers that correlate with motor symptoms or disease severity are vital to improve our understanding of the pathophysiology in Parkinson's disease (PD) and for the development of more effective treatments, including deep brain stimulation (DBS). This work provides direct insight into the application of these biomarkers in training classifiers to discriminate between brain states, which is a first step toward developing closed-loop DBS systems.

The exploration of Parkinson's disease: a multi-modal data analysis of resting functional magnetic resonance imaging and gene data

Parkinson's disease (PD) is the most universal chronic degenerative neurological dyskinesia and an important threat to elderly health. At present, the researches of PD are mainly based on single-modal data analysis, while the fusion research of multi-modal data may provide more meaningful information in the aspect of comprehending the pathogenesis of PD. In this paper, 104 samples having resting functional magnetic resonance imaging (rfMRI) and gene data are from Parkinson's Progression Markers Initiative (PPMI) and Alzheimer's Disease Neuroimaging Initiative (ADNI) database to predict pathological brain areas and risk genes related to PD. In the experiment, Pearson correlation analysis is adopted to conduct fusion analysis from the data of genes and brain areas as multi-modal sample characteristics, and the clustering evolution random forest (CERF) method is applied to detect the discriminative genes and brain areas. The experimental results indicate that compared with several existing advanced methods, the CERF method can further improve the diagnosis of PD and healthy control, and can achieve a significant effect. More importantly, we find that there are some interesting associations between brain areas and genes in PD patients. Based on these associations, we notice that PD-related brain areas include angular gyrus, thalamus, posterior cingulate gyrus and paracentral lobule, and risk genes mainly include C6orf10, HLA-DPB1 and HLA-DOA. These discoveries have a significant contribution to the early prevention and clinical treatments of PD.