Computational analysis of deep brain stimulation

Computational models advance deep brain stimulation for Parkinson's disease

Deep brain stimulation(DBS) has become an effective intervention for advanced Parkinson's disease(PD), but the exact mechanism of DBS is still unclear. In this review, we discuss the history of DBS, the anatomy and internal architecture of the basal ganglia (BG), the abnormal pathological changes of the BG in PD, and how computational models can help understand and advance DBS. We also describe two types of models: mathematical theoretical models and clinical predictive models. Mathematical theoretical models simulate neurons or neural networks of BG to shed light on the mechanistic principle underlying DBS, while clinical predictive models focus more on patients' outcomes, helping to adapt treatment plans for each patient and advance novel electrode designs. Finally, we provide insights and an outlook on future technologies.

Biophysical Principles and Computational Modeling of Deep Brain Stimulation

BACKGROUND

Deep brain stimulation (DBS) has revolutionized the treatment of neurological disorders, yet the mechanisms of DBS are still under investigation. Computational models are important in silico tools for elucidating these underlying principles and potentially for personalizing DBS therapy to individual patients. The basic principles underlying neurostimulation computational models, however, are not well known in the clinical neuromodulation community.

OBJECTIVE

In this study, we present a tutorial on the derivation of computational models of DBS and outline the biophysical contributions of electrodes, stimulation parameters, and tissue substrates to the effects of DBS.

RESULTS

Given that many aspects of DBS are difficult to characterize experimentally, computational models have played an important role in understanding how material, size, shape, and contact segmentation influence device biocompatibility, energy efficiency, the spatial spread of the electric field, and the specificity of neural activation. Neural activation is dictated by stimulation parameters including frequency, current vs voltage control, amplitude, pulse width, polarity configurations, and waveform. These parameters also affect the potential for tissue damage, energy efficiency, the spatial spread of the electric field, and the specificity of neural activation. Activation of the neural substrate also is influenced by the encapsulation layer surrounding the electrode, the conductivity of the surrounding tissue, and the size and orientation of white matter fibers. These properties modulate the effects of the electric field and determine the ultimate therapeutic response.

CONCLUSION

This article describes biophysical principles that are useful for understanding the mechanisms of neurostimulation.

Model-based quantitative optimization of deep brain stimulation and prediction of parkinson's states

Although the exact etiology of Parkinson's disease (PD) is still unknown, there are a variety of treatments available to alleviate its symptoms according to the development stage of PD. Deep brain stimulation (DBS), the most common surgical treatment for advanced PD, accurately locates and implants stimulating electrodes at specific targets in the brain to deliver high-frequency electrical stimulation that alters the excitability of the corresponding nuclei. However, for different patients and stages of PD development, there exists a choice of the optimal DBS protocol. In this paper, we propose a quantitative method (multi-dimensional feature indexes) to determine the stimulation pattern, stimulation parameters, and target of DBS from the perspective of the network model. On the other hand, based on this method, the development of PD can be predicted so that timely treatment can be given to patients. Simulation results show that, first, different network states can be distinguished by extracting features of the firing activity of neuronal populations within the basal ganglia network system. Secondly, the optimal DBS treatment can be selected by comparing the feature indexes vectors of the pre- and post-state of the network after the action of different modes of DBS. Lastly, the evolution of the network state from normal to pathological is simulated. The critical point of network state transitions is determined. These results provide a quantitative and qualitative method for determining the optimal regimen for DBS for PD, which is helpful for clinical practice.

Comparison of methodologies for modeling directional deep brain stimulation electrodes

Deep brain stimulation (DBS) is an established clinical therapy, and directional DBS electrode designs are now commonly used in clinical practice. Directional DBS leads have the ability to increase the therapeutic window of stimulation, but they also increase the complexity of clinical programming. Therefore, computational models of DBS have become available in clinical software tools that are designed to assist in the identification of therapeutic settings. However, the details of how the DBS model is implemented can influence the predictions of the software. The goal of this study was to compare different methods for representing directional DBS electrodes within finite element volume conductor (VC) models. We evaluated 15 different DBS VC model variants and quantified how their differences influenced estimates on the spatial extent of axonal activation from DBS. Each DBS VC model included the same representation of the brain and head, but the details of the current source and electrode contact were different for each model variant. The more complex VC models explicitly represented the DBS electrode contacts, while the more simple VC models used boundary condition approximations. The more complex VC models required 2-3 times longer to mesh, build, and solve for the DBS voltage distribution than the more simple VC models. Differences in individual axonal activation thresholds across the VC model variants were substantial (-24% to +47%). However, when comparing total activation of an axon population, or estimates of an activation volume, the differences between model variants decreased (-7% to +8%). Nonetheless, the technical details of how the electrode contact and current source are represented in the DBS VC model can directly affect estimates of the voltage distribution and electric field in the brain tissue.

Deep brain stimulation of terminating axons

BACKGROUND

Deep brain stimulation (DBS) of the subthalamic region is an established treatment for the motor symptoms of Parkinson's disease. Several types of neural elements reside in the subthalamic region, including subthalamic nucleus (STN) neurons, fibers of passage, and terminating afferents. Recent studies suggest that direct activation of a specific population of subthalamic afferents, known as the hyperdirect pathway, may be responsible for some of the therapeutic effects of subthalamic DBS.

OBJECTIVE

The goal of this study was to quantify how axon termination affects neural excitability from DBS. We evaluated how adjusting different stimulation parameters influenced the relative excitability of terminating axons (TAs) compared to fibers of passage (FOPs).

METHODS

We used finite element electric field models of DBS, coupled to multi-compartment cable models of axons, to calculate activation thresholds for populations of TAs and FOPs. These generalized models were used to evaluate the response to anodic vs. cathodic stimulation, with short vs. long stimulus pulses.

RESULTS

Terminating axons generally exhibited lower thresholds than fibers of passage across all tested parameters. Short pulse widths accentuated the relative excitability of TAs over FOPs.

CONCLUSION(S)

Our computational results demonstrate a hyperexcitability of terminating axons to DBS that is robust to variation in the stimulation parameters, as well as the axon model parameters.

Quantifying volume conducted potential using stimulation artefact in cortico-cortical evoked potentials

BACKGROUND

Cortico-cortical evoked potentials (CCEP) are a technique using low frequency stimulation to infer regions of cortical connectivity in patients undergoing Stereo-electroencephalographic (SEEG) monitoring for refractory epilepsy. Little attention has been given to volume conducted components of CCEP responses, and how they may inflate CCEP connectivity.

NEW METHOD

Using data from 37 SEEG-CCEPs patients, a novel method was developed to quantify stimulation artefact by measuring the peak-to-peak voltage difference in the first 10 ms after CCEP stimulation. Early responses to CCEP stimulation were also quantified by calculating the root mean square of the 10-100 ms period after each stimulation pulse. Both the early CCEP responses and amplitude of stimulation artefact were regressed by physical distance, stimulation waveform, stimulation intensity and tissue type to identify conduction related properties.

RESULTS

Both stimulation artefact and early responses were correlated strongly with the inverse square of the distance from the stimulating electrode. Once corrected for the inverse square distance from the electrode, stimulation artefact and CCEP responses showed a linear relationship, indicating a volume conducted component.

COMPARISON WITH EXISTING METHODS

This is the first study to use stimulation artefact to quantify volume conducted potentials, and is the first to quantify volume conducted potentials in SEEG. A single prior study utilizing electrocorticography has shown that parts of early CCEP responses are due to volume conduction.

CONCLUSIONS

The linear relationship between stimulation artefact amplitude and CCEP early responses, once corrected for distance, suggests that stimulation artefact can be used as a measure to quantify the volume conducted components.

Consensus Paper: Experimental Neurostimulation of the Cerebellum

The cerebellum is best known for its role in controlling motor behaviors. However, recent work supports the view that it also influences non-motor behaviors. The contribution of the cerebellum towards different brain functions is underscored by its involvement in a diverse and increasing number of neurological and neuropsychiatric conditions including ataxia, dystonia, essential tremor, Parkinson's disease (PD), epilepsy, stroke, multiple sclerosis, autism spectrum disorders, dyslexia, attention deficit hyperactivity disorder (ADHD), and schizophrenia. Although there are no cures for these conditions, cerebellar stimulation is quickly gaining attention for symptomatic alleviation, as cerebellar circuitry has arisen as a promising target for invasive and non-invasive neuromodulation. This consensus paper brings together experts from the fields of neurophysiology, neurology, and neurosurgery to discuss recent efforts in using the cerebellum as a therapeutic intervention. We report on the most advanced techniques for manipulating cerebellar circuits in humans and animal models and define key hurdles and questions for moving forward.

A Driving-Force Predictor for Estimating Pathway Activation in Patient-Specific Models of Deep Brain Stimulation

OBJECTIVE

Detailed biophysical modeling of deep brain stimulation (DBS) provides a theoretical approach to quantify the cellular response to the applied electric field. However, the most accurate models for performing such analyses, patient-specific field-cable (FC) pathway-activation models (PAMs), are so technically demanding to implement that their use in clinical research is greatly limited. Predictive algorithms can simplify PAM calculations, but they generally fail to reproduce the output of FC models when evaluated over a wide range of clinically relevant stimulation parameters. Therefore, we set out to develop a novel driving-force (DF) predictive algorithm (DF-Howell), customized to the study of DBS, which can better match FC results.

METHODS

We developed the DF-Howell algorithm and compared its predictions to FC PAM results, as well as to the DF-Peterson algorithm, which is currently the most accurate and generalizable DF-based method. Comparison of the various methods was quantified within the context of subthalamic DBS using activation thresholds of axons representing the internal capsule, hyperdirect pathway, and cerebellothalamic tract for various combinations of fiber diameters, stimulus pulse widths, and electrode configurations.

RESULTS

The DF-Howell predictor estimated activation of the three axonal pathways with less than a 6.2% mean error with respect to the FC PAM for all 21 cases tested. In 15 of the 21 cases, DF-Howell outperformed DF-Peterson in estimating pathway activation, reducing mean-errors up to 22.5%.

CONCLUSIONS

DF-Howell represents an accurate predictor for estimating axonal pathway activation in patient-specific DBS models, but errors still exist relative to FC PAM calculations. Nonetheless, the tractability of DF algorithms helps to reduce the technical barriers for performing accurate biophysical modeling in clinical DBS research studies.

A Study on the Feasibility of the Deep Brain Stimulation (DBS) Electrode Localization Based on Scalp Electric Potential Recordings

Deep Brain Stimulation (DBS) is an effective therapy for patients disabling motor symptoms from Parkinson's disease, essential tremor, and other motor disorders. Precise, individualized placement of DBS electrodes is a key contributor to clinical outcomes following surgery. Electroencephalography (EEG) is widely used to identify the sources of intracerebral signals from the potential on the scalp. EEG is portable, non-invasive, low-cost, and it could be easily integrated into the intraoperative or ambulatory environment for localization of either the DBS electrode or evoked potentials triggered by stimulation itself. In this work, we studied with numerical simulations the principle of extracting the DBS electrical pulse from the patient's EEG - which normally constitutes an artifact - and localizing the source of the artifact (i.e., the DBS electrodes) using EEG localization methods. A high-resolution electromagnetic head model was used to simulate the EEG potential at the scalp generated by the DBS pulse artifact. The potential distribution on the scalp was then sampled at the 256 electrode locations of a high-density EEG Net. The electric potential was modeled by a dipole source created by a given pair of active DBS electrodes. The dynamic Statistical Parametric Maps (dSPM) algorithm was used to solve the EEG inverse problem, and it allowed localization of the position of the stimulus dipole in three DBS electrode bipolar configurations with a maximum error of 1.5 cm. To assess the accuracy of the computational model, the results of the simulation were compared with the electric artifact amplitudes over 16 EEG electrodes measured in five patients. EEG artifacts measured in patients confirmed that simulated data are commensurate to patients' data (0 ± 6.6 μV). While we acknowledge that further work is necessary to achieve a higher accuracy needed for surgical navigation, the results presented in this study are proposed as the first step toward a validated computational framework that could be used for non-invasive localization not only of the DBS system but also brain rhythms triggered by stimulation at both proximal and distal sites in the human central nervous system.

ESM-CT: a precise method for localization of DBS electrodes in CT images

BACKGROUND

Deep brain stimulation (DBS) of the subthalamic nucleus produces variable effects in Parkinson disease. Variation may result from different electrode positions relative to target. Thus, precise electrode localization is crucial when investigating DBS effects.

NEW METHOD

We developed a semi-automated method, Electrode Shaft Modeling in CT images (ESM-CT) to reconstruct DBS lead trajectories and contact locations. We evaluated methodological sensitivity to operator-dependent steps, robustness to image resampling, and test-retest replicability. ESM-CT was applied in 56 patients to study electrode position change (and relation to time between scans, postoperative subdural air volume, and head tilt during acquisition) between images acquired immediately post-implantation (DBS-CT) and months later (DEL-CT).

RESULTS

Electrode tip localization was robust to image resampling and replicable to within ∼ 0.2 mm on test-retest comparisons. Systematic electrode displacement occurred rostral-ventral-lateral between DBS-CT and DEL-CT scans. Head angle was a major explanatory factor (p < 0.001,Pearson's r = 0.46, both sides) and volume of subdural air weakly predicted electrode displacement (p = 0.02,r = 0.29:p = 0.1,r = 0.25 for left:right). Modeled shaft curvature was slightly greater in DEL-CT. Magnitude of displacement and degree of curvature were independent of elapsed time between scans.

COMPARISON WITH EXISTING METHODS

Comparison of ESM-CT against two existing methods revealed systematic differences in one coordinate (1 ± 0.3 mm,p < 0.001) for one method and in three coordinates for another method (x:0.1 ± 0.1 mm, y:0.4 ± 0.2 mm, z:0.4 ± 0.2 mm, p < 10-10). Within-method coordinate variability across participants is similar.

CONCLUSION

We describe a robust and precise method for CT DBS contact localization. Application revealed that acquisition head angle significantly impacts electrode position. DBS localization schemes should account for head angle.

Quantifying axonal responses in patient-specific models of subthalamic deep brain stimulation

Medical imaging has played a major role in defining the general anatomical targets for deep brain stimulation (DBS) therapies. However, specifics on the underlying brain circuitry that is directly modulated by DBS electric fields remain relatively undefined. Detailed biophysical modeling of DBS provides an approach to quantify the theoretical responses to stimulation at the cellular level, and has established a key role for axonal activation in the therapeutic mechanisms of DBS. Estimates of DBS-induced axonal activation can then be coupled with advances in defining the structural connectome of the human brain to provide insight into the modulated brain circuitry and possible correlations with clinical outcomes. These pathway-activation models (PAMs) represent powerful tools for DBS research, but the theoretical predictions are highly dependent upon the underlying assumptions of the particular modeling strategy used to create the PAM. In general, three types of PAMs are used to estimate activation: 1) field-cable (FC) models, 2) driving force (DF) models, and 3) volume of tissue activated (VTA) models. FC models represent the "gold standard" for analysis but at the cost of extreme technical demands and computational resources. Consequently, DF and VTA PAMs, derived from simplified FC models, are typically used in clinical research studies, but the relative accuracy of these implementations is unknown. Therefore, we performed a head-to-head comparison of the different PAMs, specifically evaluating DBS of three different axonal pathways in the subthalamic region. The DF PAM was markedly more accurate than the VTA PAMs, but none of these simplified models were able to match the results of the patient-specific FC PAM across all pathways and combinations of stimulus parameters. These results highlight the limitations of using simplified predictors to estimate axonal stimulation and emphasize the need for novel algorithms that are both biophysically realistic and computationally simple.

Computational Modeling of Neurotransmitter Release Evoked by Electrical Stimulation: Nonlinear Approaches to Predicting Stimulation-Evoked Dopamine Release

Neurochemical changes evoked by electrical stimulation of the nervous system have been linked to both therapeutic and undesired effects of neuromodulation therapies used to treat obsessive-compulsive disorder, depression, epilepsy, Parkinson's disease, stroke, hypertension, tinnitus, and many other indications. In fact, interest in better understanding the role of neurochemical signaling in neuromodulation therapies has been a focus of recent government- and industry-sponsored programs whose ultimate goal is to usher in an era of personalized medicine by creating neuromodulation therapies that respond to real-time changes in patient status. A key element to achieving these precision therapeutic interventions is the development of mathematical modeling approaches capable of describing the nonlinear transfer function between neuromodulation parameters and evoked neurochemical changes. Here, we propose two computational modeling frameworks, based on artificial neural networks (ANNs) and Volterra kernels, that can characterize the input/output transfer functions of stimulation-evoked neurochemical release. We evaluate the ability of these modeling frameworks to characterize subject-specific neurochemical kinetics by accurately describing stimulation-evoked dopamine release across rodent (R² = 0.83 Volterra kernel, R² = 0.86 ANN), swine (R² = 0.90 Volterra kernel, R² = 0.93 ANN), and non-human primate (R² = 0.98 Volterra kernel, R² = 0.96 ANN) models of brain stimulation. Ultimately, these models will not only improve understanding of neurochemical signaling in healthy and diseased brains but also facilitate the development of neuromodulation strategies capable of controlling neurochemical release via closed-loop strategies.

Computer-Guided Deep Brain Stimulation Programming for Parkinson's Disease

OBJECTIVE

Pilot study to evaluate computer-guided deep brain stimulation (DBS) programming designed to optimize stimulation settings using objective motion sensor-based motor assessments.

MATERIALS AND METHODS

Seven subjects (five males; 54-71 years) with Parkinson's disease (PD) and recently implanted DBS systems participated in this pilot study. Within two months of lead implantation, the subject returned to the clinic to undergo computer-guided programming and parameter selection. A motion sensor was placed on the index finger of the more affected hand. Software guided a monopolar survey during which monopolar stimulation on each contact was iteratively increased followed by an automated assessment of tremor and bradykinesia. After completing assessments at each setting, a software algorithm determined stimulation settings designed to minimize symptom severities, side effects, and battery usage.

RESULTS

Optimal DBS settings were chosen based on average severity of motor symptoms measured by the motion sensor. Settings chosen by the software algorithm identified a therapeutic window and improved tremor and bradykinesia by an average of 35.7% compared with baseline in the "off" state (p < 0.01).

CONCLUSIONS

Motion sensor-based computer-guided DBS programming identified stimulation parameters that significantly improved tremor and bradykinesia with minimal clinician involvement. Automated motion sensor-based mapping is worthy of further investigation and may one day serve to extend programming to populations without access to specialized DBS centers.

Mapping the "Depression Switch" During Intraoperative Testing of Subcallosal Cingulate Deep Brain Stimulation

IMPORTANCE

The clinical utility of monitoring behavioral changes during intraoperative testing of subcallosal cingulate deep brain stimulation is unknown.

OBJECTIVE

To characterize the structural connectivity correlates of deep brain stimulation-evoked behavioral effects using probabilistic tractography in depression.

DESIGN, SETTING, AND PARTICIPANTS

Categorization of acute behavioral effects was conducted in 9 adults undergoing deep brain stimulation implantation surgery for chronic treatment-resistant depression in a randomized and blinded testing session at Emory University. Patients were studied from September 1, 2011, through June 30, 2013. Post hoc analyses of the structural tractography patterns mediating distinct categories of evoked behavioral effects were defined, including the best response overall. Data analyses were performed from May 1 through July 1, 2015.

MAIN OUTCOMES AND MEASURES

Categorization of stimulation-induced transient behavioral effects and delineation of the shared white matter tracts mediating response subtypes.

RESULTS

Among the 9 patients, 72 active and 36 sham trials were recorded. The following stereotypical behavior patterns were identified: changes in interoceptive (noted changes in body state in 30 of 72 active and 4 of 36 sham trials) and in exteroceptive (shift in attention from patient to others in 9 of 72 active and 0 sham trials) awareness. The best response was a combination of exteroceptive and interoceptive changes at a single left contact for all 9 patients. Structural connectivity showed that the best response contacts had a pattern of connections to the bilateral ventromedial frontal cortex (via forceps minor and left uncinate fasciculus) and to the cingulate cortex (via left cingulum bundle), whereas behaviorally salient but nonbest contacts had only cingulate involvement. The involvement of the 3 white matter bundles during stimulation of the best contacts suggests a mechanism for the observed transient "depression switch."

CONCLUSIONS AND RELEVANCE

This analysis of transient behavior changes during intraoperative deep brain stimulation of the subcallosal cingulate and the subsequent identification of unique connectivity patterns may provide a biomarker of a rapid-onset depression switch to guide surgical implantation and to refine and optimize algorithms for the selection of contacts in long-term stimulation for treatment-resistant depression.

Modeling sequence and quasi-uniform assumption in computational neurostimulation

Computational neurostimulation aims to develop mathematical constructs that link the application of neuromodulation with changes in behavior and cognition. This process is critical but daunting for technical challenges and scientific unknowns. The overarching goal of this review is to address how this complex task can be made tractable. We describe a framework of sequential modeling steps to achieve this: (1) current flow models, (2) cell polarization models, (3) network and information processing models, and (4) models of the neuroscientific correlates of behavior. Each step is explained with a specific emphasis on the assumptions underpinning underlying sequential implementation. We explain the further implementation of the quasi-uniform assumption to overcome technical limitations and unknowns. We specifically focus on examples in electrical stimulation, such as transcranial direct current stimulation. Our approach and conclusions are broadly applied to immediate and ongoing efforts to deploy computational neurostimulation.

Machine Learning Approach to Optimizing Combined Stimulation and Medication Therapies for Parkinson's Disease

BACKGROUND

Deep brain stimulation (DBS) of the subthalamic region is an established therapy for advanced Parkinson's disease (PD). However, patients often require time-intensive post-operative management to balance their coupled stimulation and medication treatments. Given the large and complex parameter space associated with this task, we propose that clinical decision support systems (CDSS) based on machine learning algorithms could assist in treatment optimization.

OBJECTIVE

Develop a proof-of-concept implementation of a CDSS that incorporates patient-specific details on both stimulation and medication.

METHODS

Clinical data from 10 patients, and 89 post-DBS surgery visits, were used to create a prototype CDSS. The system was designed to provide three key functions: (1) information retrieval; (2) visualization of treatment, and; (3) recommendation on expected effective stimulation and drug dosages, based on three machine learning methods that included support vector machines, Naïve Bayes, and random forest.

RESULTS

Measures of medication dosages, time factors, and symptom-specific pre-operative response to levodopa were significantly correlated with post-operative outcomes (P < 0.05) and their effect on outcomes was of similar magnitude to that of DBS. Using those results, the combined machine learning algorithms were able to accurately predict 86% (12/14) of the motor improvement scores at one year after surgery.

CONCLUSIONS

Using patient-specific details, an appropriately parameterized CDSS could help select theoretically optimal DBS parameter settings and medication dosages that have potential to improve the clinical management of PD patients.

Defining critical white matter pathways mediating successful subcallosal cingulate deep brain stimulation for treatment-resistant depression

BACKGROUND

Subcallosal cingulate white matter (SCC) deep brain stimulation (DBS) is an evolving investigational treatment for depression. Mechanisms of action are hypothesized to involve modulation of activity within a structurally defined network of brain regions involved in mood regulation. Diffusion tensor imaging was used to model white matter connections within this network to identify those critical for successful antidepressant response.

METHODS

Preoperative high-resolution magnetic resonance imaging data, including diffusion tensor imaging, were acquired in 16 patients with treatment-resistant depression, who then received SCC DBS. Computerized tomography was used postoperatively to locate DBS contacts. The activation volume around the contacts used for chronic stimulation was modeled for each patient retrospectively. Probabilistic tractography was used to delineate the white matter tracts traveling through each activation volume. Patient-specific tract maps were calculated using whole-brain analysis. Clinical evaluations of therapeutic outcome from SCC DBS were defined at 6 months and 2 years.

RESULTS

Whole-brain activation volume tractography demonstrated that all DBS responders at 6 months (n = 6) and 2 years (n = 12) shared bilateral pathways from their activation volumes to 1) medial frontal cortex via forceps minor and uncinate fasciculus; 2) rostral and dorsal cingulate cortex via the cingulum bundle; and 3) subcortical nuclei. Nonresponders did not consistently show these connections. Specific anatomical coordinates of the active contacts did not discriminate responders from nonresponders.

CONCLUSIONS

Patient-specific activation volume tractography modeling may identify critical tracts that mediate SCC DBS antidepressant response. This suggests a novel method for patient-specific target and stimulation parameter selection.

Deep brain stimulation in the treatment of depression

Major depressive disorder is a worldwide disease with debilitating effects on a patient's life. Common treatments include pharmacotherapy, psychotherapy, and electroconvulsive therapy. Many patients do not respond to these treatments; this has led to the investigation of alternative therapeutic modalities. Deep brain stimulation (DBS) is one of these modalities. It was first used with success for treating movement disorders and has since been extended to the treatment of psychiatric disorders. Although DBS is still an emerging treatment, promising efficacy and safety have been demonstrated in preliminary trials in patients with treatment-resistant depression (TRD). Further, neuroimaging has played a pivotal role in identifying some DBS targets and remains an important tool for evaluating the mechanism of action of this novel intervention. Preclinical animal studies have broadened knowledge about the possible mechanisms of action of DBS for TRD, Given that DBS involves neurosurgery in patients with severe psychiatric impairment, ethical questions concerning capacity to consent arise; these issues must continue to be carefully considered.

Functional anatomy of subthalamic nucleus stimulation in Parkinson disease

OBJECTIVE

We developed a novel method to map behavioral effects of deep brain stimulation (DBS) across a 3-dimensional brain region and to assign statistical significance after stringent type I error correction. This method was applied to behavioral changes in Parkinson disease (PD) induced by subthalamic nucleus (STN) DBS to determine whether these responses depended on anatomical location of DBS.

METHODS

Fifty-one PD participants with STN DBS were evaluated off medication, with DBS off and during unilateral STN DBS with clinically optimized settings. Dependent variables included DBS-induced changes in Unified Parkinson Disease Rating Scale (UPDRS) subscores, kinematic measures of bradykinesia and rigidity, working memory, response inhibition, mood, anxiety, and akathisia. Weighted t tests at each voxel produced p images showing where DBS most significantly affected each dependent variable based on outcomes of participants with nearby DBS. Finally, a permutation test computed the probability that this p image indicated significantly different responses based on stimulation site.

RESULTS

Most motor variables improved with DBS anywhere in the STN region, but several motor, cognitive, and affective responses significantly depended on precise location stimulated, with peak p values in superior STN/zona incerta (quantified bradykinesia), dorsal STN (mood, anxiety), and inferior STN/substantia nigra (UPDRS tremor, working memory).

INTERPRETATION

Our method identified DBS-induced behavioral changes that depended significantly on DBS site. These results do not support complete functional segregation within STN, because movement improved with DBS throughout, and mood improved with dorsal STN DBS. Rather, findings support functional convergence of motor, cognitive, and limbic information in STN.

Engineering the next generation of clinical deep brain stimulation technology

Deep brain stimulation (DBS) has evolved into a powerful clinical therapy for a range of neurological disorders, but even with impressive clinical growth, DBS technology has been relatively stagnant over its history. However, enhanced collaborations between neural engineers, neuroscientists, physicists, neurologists, and neurosurgeons are beginning to address some of the limitations of current DBS technology. These interactions have helped to develop novel ideas for the next generation of clinical DBS systems. This review attempts collate some of that progress with two goals in mind. First, provide a general description of current clinical DBS practices, geared toward educating biomedical engineers and computer scientists on a field that needs their expertise and attention. Second, describe some of the technological developments that are currently underway in surgical targeting, stimulation parameter selection, stimulation protocols, and stimulation hardware that are being directly evaluated for near term clinical application.

Bimanual force coordination in Parkinson's disease patients with bilateral subthalamic deep brain stimulation

OBJECTIVE

Studies of bimanual actions similar to activities of daily living (ADLs) are currently lacking in evaluating fine motor control in Parkinson's disease patients implanted with bilateral subthalamic deep brain stimulators. We investigated basic time and force characteristics of a bimanual task that resembles performance of ADLs in a group of bilateral subthalamic deep brain stimulation (DBS) patients.

METHODS

Patients were evaluated in three different DBS parameter conditions off stimulation, on clinically derived stimulation parameters, and on settings derived from a patient-specific computational model. Model-based parameters were computed as a means to minimize spread of current to non-motor regions of the subthalamic nucleus via Cicerone Deep Brain Stimulation software. Patients were evaluated off parkinsonian medications in each stimulation condition.

RESULTS

The data indicate that DBS parameter state does not affect most aspects of fine motor control in ADL-like tasks; however, features such as increased grip force and grip symmetry varied with the stimulation state. In the absence of DBS parameters, patients exhibited significant grip force asymmetry. Overall UPDRS-III and UPDRS-III scores associated with hand function were lower while patients were experiencing clinically-derived or model-based parameters, as compared to the off-stimulation condition.

CONCLUSION

While bilateral subthalamic DBS has been shown to alleviate gross motor dysfunction, our results indicate that DBS may not provide the same magnitude of benefit to fine motor coordination.

Artificial neural network based characterization of the volume of tissue activated during deep brain stimulation

OBJECTIVE

Clinical deep brain stimulation (DBS) systems can be programmed with thousands of different stimulation parameter combinations (e.g. electrode contact(s), voltage, pulse width, frequency). Our goal was to develop novel computational tools to characterize the effects of stimulation parameter adjustment for DBS.

APPROACH

The volume of tissue activated (VTA) represents a metric used to estimate the spatial extent of DBS for a given parameter setting. Traditional methods for calculating the VTA rely on activation function (AF)-based approaches and tend to overestimate the neural response when stimulation is applied through multiple electrode contacts. Therefore, we created a new method for VTA calculation that relied on artificial neural networks (ANNs).

MAIN RESULTS

The ANN-based predictor provides more accurate descriptions of the spatial spread of activation compared to AF-based approaches for monopolar stimulation. In addition, the ANN was able to accurately estimate the VTA in response to multi-contact electrode configurations.

SIGNIFICANCE

The ANN-based approach may represent a useful method for fast computation of the VTA in situations with limited computational resources, such as a clinical DBS programming application on a tablet computer.

MRI-based multiscale model for electromagnetic analysis in the human head with implanted DBS

Deep brain stimulation (DBS) is an established procedure for the treatment of movement and affective disorders. Patients with DBS may benefit from magnetic resonance imaging (MRI) to evaluate injuries or comorbidities. However, the MRI radio-frequency (RF) energy may cause excessive tissue heating particularly near the electrode. This paper studies how the accuracy of numerical modeling of the RF field inside a DBS patient varies with spatial resolution and corresponding anatomical detail of the volume surrounding the electrodes. A multiscale model (MS) was created by an atlas-based segmentation using a 1 mm³ head model (mRes) refined in the basal ganglia by a 200 μm² ex-vivo dataset. Four DBS electrodes targeting the left globus pallidus internus were modeled. Electromagnetic simulations at 128 MHz showed that the peak of the electric field of the MS doubled (18.7 kV/m versus 9.33 kV/m) and shifted 6.4 mm compared to the mRes model. Additionally, the MS had a sixfold increase over the mRes model in peak-specific absorption rate (SAR of 43.9 kW/kg versus 7 kW/kg). The results suggest that submillimetric resolution and improved anatomical detail in the model may increase the accuracy of computed electric field and local SAR around the tip of the implant.

High-resolution modeling assisted design of customized and individualized transcranial direct current stimulation protocols

OBJECTIVES

Transcranial direct current stimulation (tDCS) is a neuromodulatory technique that delivers low-intensity currents facilitating or inhibiting spontaneous neuronal activity. tDCS is attractive since dose is readily adjustable by simply changing electrode number, position, size, shape, and current. In the recent past, computational models have been developed with increased precision with the goal to help customize tDCS dose. The aim of this review is to discuss the incorporation of high-resolution patient-specific computer modeling to guide and optimize tDCS.

METHODS

In this review, we discuss the following topics: 1) The clinical motivation and rationale for models of transcranial stimulation is considered pivotal in order to leverage the flexibility of neuromodulation; 2) the protocols and the workflow for developing high-resolution models; 3) the technical challenges and limitations of interpreting modeling predictions; and 4) real cases merging modeling and clinical data illustrating the impact of computational models on the rational design of rehabilitative electrotherapy.

CONCLUSIONS

Though modeling for noninvasive brain stimulation is still in its development phase, it is predicted that with increased validation, dissemination, simplification, and democratization of modeling tools, computational forward models of neuromodulation will become useful tools to guide the optimization of clinical electrotherapy.

Current steering to activate targeted neural pathways during deep brain stimulation of the subthalamic region

Deep brain stimulation (DBS) has steadily evolved into an established surgical therapy for numerous neurological disorders, most notably Parkinson's disease (PD). Traditional DBS technology relies on voltage-controlled stimulation with a single source; however, recent engineering advances are providing current-controlled devices with multiple independent sources. These new stimulators deliver constant current to the brain tissue, irrespective of impedance changes that occur around the electrode, and enable more specific steering of current towards targeted regions of interest. In this study, we examined the impact of current steering between multiple electrode contacts to directly activate three distinct neural populations in the subthalamic region commonly stimulated for the treatment of PD: projection neurons of the subthalamic nucleus (STN), globus pallidus internus (GPi) fibers of the lenticular fasiculus, and internal capsule (IC) fibers of passage. We used three-dimensional finite element electric field models, along with detailed multicompartment cable models of the three neural populations to determine their activations using a wide range of stimulation parameter settings. Our results indicate that selective activation of neural populations largely depends on the location of the active electrode(s). Greater activation of the GPi and STN populations (without activating any side effect related IC fibers) was achieved by current steering with multiple independent sources, compared to a single current source. Despite this potential advantage, it remains to be seen if these theoretical predictions result in a measurable clinical effect that outweighs the added complexity of the expanded stimulation parameter search space generated by the more flexible technology.

Asymptotic model of electrical stimulation of nerve fibers

We present a novel theory and computational algorithm for modeling electrical stimulation of nerve fibers in three dimensions. Our approach uses singular perturbation to separate the full 3D boundary value problem into a set of 2D "transverse" problems coupled with a 1D "longitudinal" problem. The resulting asymptotic model contains not one but two activating functions (AF): the longitudinal AF that drives the slow development of the mean transmembrane potential and the transverse AF that drives the rapid polarization of the fiber in the transverse direction. The asymptotic model is implemented for a prototype 3D cylindrical fiber with a passive membrane in an isotropic extracellular region. The validity of this approach is tested by comparing the numerical solution of the asymptotic model to the analytical solutions. The results show that the asymptotic model predicts steady-state transmembrane potential directly under the electrodes with the root mean square error of 0.539 mV, i.e., 1.04% of the maximum transmembrane potential. Thus, this work has created a computationally efficient algorithm that facilitates studies of the complete spatiotemporal dynamics of nerve fibers in three dimensions.

Subthalamic nucleus stimulation affects incentive salience attribution in Parkinson's disease

Deep brain stimulation (DBS) of the subthalamic nucleus (STN) can induce nonmotor side effects such as behavioral and mood disturbances or body weight gain in Parkinson's disease (PD) patients. We hypothesized that some of these problems could be related to an altered attribution of incentive salience (ie, emotional relevance) to rewarding and aversive stimuli. Twenty PD patients (all men; mean age ± SD, 58.3 ± 6 years) in bilateral STN DBS switched ON and OFF conditions and 18 matched controls rated pictures selected from the International Affective Picture System according to emotional valence (unpleasantness/pleasantness) and arousal on 2 independent visual scales ranging from 1 to 9. Eighty-four pictures depicting primary rewarding (erotica and food) and aversive fearful (victims and threat) and neutral stimuli were selected for this study. In the STN DBS ON condition, the PD patients attributed lower valence scores to the aversive pictures compared with the OFF condition (P < .01) and compared with controls (P < .01). The difference between the OFF condition and controls was less pronounced (P < .05). Furthermore, postoperative weight gain correlated with arousal ratings from the food pictures in the STN DBS ON condition (P < .05 compensated for OFF condition). Our results suggest that STN DBS increases activation of the aversive motivational system so that more relevance is attributed to aversive fearful stimuli. In addition, STN DBS-related sensitivity to food reward stimuli cues might drive DBS-treated patients to higher food intake and subsequent weight gain.

Predicted effects of pulse width programming in spinal cord stimulation: a mathematical modeling study

To understand the theoretical effects of pulse width (PW) programming in spinal cord stimulation (SCS), we implemented a mathematical model of electrical fields and neural activation in SCS to gain insight into the effects of PW programming. The computational model was composed of a finite element model for structure and electrical properties, coupled with a nonlinear double-cable axon model to predict nerve excitation for different myelinated fiber sizes. Mathematical modeling suggested that mediolateral lead position may affect chronaxie and rheobase values, as well as predict greater activation of medial dorsal column fibers with increased PW. These modeling results were validated by a companion clinical study. Thus, variable PW programming in SCS appears to have theoretical value, demonstrated by the ability to increase and even ‘steer’ spatial selectivity of dorsal column fiber recruitment. It is concluded that the computational SCS model is a valuable tool to understand basic mechanisms of nerve fiber excitation modulated by stimulation parameters such as PW and electric fields.

Patient-specific models of deep brain stimulation: influence of field model complexity on neural activation predictions

Deep brain stimulation (DBS) of the subthalamic nucleus (STN) has become the surgical therapy of choice for medically intractable Parkinson's disease. However, quantitative understanding of the interaction between the electric field generated by DBS and the underlying neural tissue is limited. Recently, computational models of varying levels of complexity have been used to study the neural response to DBS. The goal of this study was to evaluate the quantitative impact of incrementally incorporating increasing levels of complexity into computer models of STN DBS. Our analysis focused on the direct activation of experimentally measureable fiber pathways within the internal capsule (IC). Our model system was customized to an STN DBS patient and stimulation thresholds for activation of IC axons were calculated with electric field models that ranged from an electrostatic, homogenous, isotropic model to one that explicitly incorporated the voltage-drop and capacitance of the electrode-electrolyte interface, tissue encapsulation of the electrode, and diffusion-tensor based 3D tissue anisotropy and inhomogeneity. The model predictions were compared to experimental IC activation defined from electromyographic (EMG) recordings from eight different muscle groups in the contralateral arm and leg of the STN DBS patient. Coupled evaluation of the model and experimental data showed that the most realistic predictions of axonal thresholds were achieved with the most detailed model. Furthermore, the more simplistic neurostimulation models substantially overestimated the spatial extent of neural activation.

Neuromodulation: advances in the next five years

Neuromodulation (deep brain stimulation; DBS) has become an established treatment for movement disorders (e.g., Parkinson's disease), and is in trials for refractory epilepsy, headache, and certain mood disorders. Two main themes will advance DBS significantly in the next five years: closed-loop DBS, that is, feedback from brain electrical activity to direct the stimulation; and computational analysis (CA)--electrophysiological modeling to enhance DBS. Closed-loop DBS is currently in clinical trials for refractory epilepsy. New imaging techniques offer preoperative modeling for DBS surgery, including nerve fiber tracts (diffusion tensor imaging), and imaging of volume of tissue activated by a specific electrode. CA techniques for DBS include mathematical models of the abnormally synchronized electrical activity which underlies epilepsy, movement disorders, and likely many mood disorders as well. By incorporating feedback loops and multiple recording and/or stimulating sites, the abnormally synchronized brain electrical activity can be desynchronized, then "unlearned" ("unkindling" in epilepsy). Characteristics of DBS utilizing CA include low frequency rather than high frequency stimulation; multiple stimulation and/or recording sites; likely 10-fold or more reduction in electrical current needs (much smaller "pulse generators"); more focused and less disruptive stimulation--fewer unwanted side effects; and potential to "cure" certain disorders by resetting abnormal firing patterns back to normal. These advantages of more sophisticated DBS techniques bring the following challenges, which may require a decade of research before reaching clinical practice because many brain disorders involve neurotransmitter abnormalities (e.g., dopamine in Parkinson's disease and certain mood disorders). Namely, how do we monitor and modulate neurotransmitters in addition to electrical activity? How do we get multiple microelectrodes into the brain in a minimally invasive manner? In the accompanying article, I address these two issues and offer some potential solutions.

Network perspectives on the mechanisms of deep brain stimulation

Deep brain stimulation (DBS) is an established medical therapy for the treatment of movement disorders and shows great promise for several other neurological disorders. However, after decades of clinical utility the underlying therapeutic mechanisms remain undefined. Early attempts to explain the mechanisms of DBS focused on hypotheses that mimicked an ablative lesion to the stimulated brain region. More recent scientific efforts have explored the wide-spread changes in neural activity generated throughout the stimulated brain network. In turn, new theories on the mechanisms of DBS have taken a systems-level approach to begin to decipher the network activity. This review provides an introduction to some of the network based theories on the function and pathophysiology of the cortico-basal-ganglia-thalamo-cortical loops commonly targeted by DBS. We then analyze some recent results on the effects of DBS on these networks, with a focus on subthalamic DBS for the treatment of Parkinson's disease. Finally we attempt to summarize how DBS could be achieving its therapeutic effects by overriding pathological network activity.

A tractography analysis of two deep brain stimulation white matter targets for depression

BACKGROUND

Deep brain stimulation (DBS) of the subcallosal cingulate white matter (SCCwm) or anterior limb of the internal capsule (ALIC) may be effective in treating depression. Connectivity patterns of these regions may inform on mechanisms of action for DBS of these targets.

METHODS

Diffusion tensor imaging (DTI) and probabilistic tractography were performed in 13 nondepressed subjects to determine connectivity patterns of SCCwm and ALIC. Tract maps were generated for each target in each subject, and tract voxels were coded as being unique to either target or shared. Group level tract maps were generated by including only those voxels common to at least 10 of 13 (>75%) subjects.

RESULTS

The two targets have distinct patterns of connectivity with regions of overlap. The SCCwm showed consistent ipsilateral connections to the medial frontal cortex, the full extent of the anterior and posterior cingulate, medial temporal lobe, dorsal medial thalamus, hypothalamus, nucleus accumbens, and the dorsal brainstem. The ALIC seed, in contrast, demonstrated widespread projections to frontal pole, medial temporal lobe, cerebellum, nucleus accumbens, thalamus, hypothalamus, and brainstem. Common to both targets, albeit through distinct white matter bundles, were connections to frontal pole, medial temporal lobe, nucleus accumbens, dorsal thalamus, and hypothalamus.

CONCLUSIONS

Connectivity patterns of these two DBS white matter targets suggest distinct neural networks with areas of overlap in regions implicated in depression and antidepressant response.