Deep learning for comprehensive ECG annotation.

BACKGROUND: Increasing utilization of long-term outpatient ambulatory electrocardiographic (ECG) monitoring continues to drive the need for improved ECG interpretation algorithms. OBJECTIVE: The purpose of this study was to describe the BeatLogic® platform for ECG interpretation and to validate the platform using electrophysiologist-adjudicated real-world data and publicly available validation data. METHODS: Deep learning models were trained to perform beat and rhythm detection/classification using ECGs collected with the Preventice BodyGuardian® Heart monitor. Training annotations were created by certified ECG technicians, and validation annotations were adjudicated by a team of board-certified electrophysiologists. Deep learning model classification results were used to generate contiguous annotation results, and performance was assessed in accordance with the EC57 standard. RESULTS: On the real-world validation dataset, BeatLogic beat detection sensitivity and positive predictive value were 99.84% and 99.78%, respectively. Ventricular ectopic beat classification sensitivity and positive predictive value were 89.4% and 97.8%, respectively. Episode and duration F1 scores (range 0-100) exceeded 70 for all 14 rhythms (including noise) that were evaluated. F1 scores for 11 rhythms exceeded 80, 7 exceeded 90, and 5 including atrial fibrillation/flutter, ventricular tachycardia, ventricular bigeminy, ventricular trigeminy, and third-degree heart block exceeded 95. CONCLUSION: The BeatLogic platform represents the next stage of advancement for algorithmic ECG interpretation. This comprehensive platform performs beat detection, beat classification, and rhythm detection/classification with greatly improved performance over the current state of the art, with comparable or improved performance over previously published algorithms that can accomplish only 1 of these 3 tasks.

Measuring In Vivo Changes in Extracellular Neurotransmitters During Naturally Rewarding Behaviors in Female Syrian Hamsters.

The ability to measure neurotransmitter release on a rapid time scale allows patterns of neurotransmission to be linked to specific behaviors or manipulations; a powerful tool in elucidating underlying mechanisms and circuitry. While the technique of microdialysis has been used for decades to measure nearly any analyte of interest in the brain, this technique is limited in temporal resolution. Alternatively, fast scan cyclic voltammetry is both temporally precise and extremely sensitive; however, because this technically difficult method relies on the electroactivity of the analyte of interest, the possibility to detect nonelectroactive substances (e.g., the neurotransmitter glutamate) is eliminated. This paper details the use of a turn-key system that combines fixed-potential amperometry and enzymatic biosensing to measure both electroactive and nonelectroactive neurotransmitters with temporal precision. The pairing of these two powerful techniques allows for the measurement of both tonic and phasic neurotransmission with relative ease, and permits recording of multiple neurotransmitters simultaneously. The aim of this manuscript is to demonstrate the process of measuring dopamine and glutamate neurotransmission in vivo using a naturally rewarding behavior (i.e., sexual behavior) in female hamsters, with the ultimate goal of displaying the technical feasibility of this assay for examining other behaviors and experimental paradigms.

Orientation selective deep brain stimulation.

OBJECTIVE: Target selectivity of deep brain stimulation (DBS) therapy is critical, as the precise locus and pattern of the stimulation dictates the degree to which desired treatment responses are achieved and adverse side effects are avoided. There is a clear clinical need to improve DBS technology beyond currently available stimulation steering and shaping approaches. We introduce orientation selective neural stimulation as a concept to increase the specificity of target selection in DBS. APPROACH: This concept, which involves orienting the electric field along an axonal pathway, was tested in the corpus callosum of the rat brain by freely controlling the direction of the electric field on a plane using a three-electrode bundle, and monitoring the response of the neurons using functional magnetic resonance imaging (fMRI). Computational models were developed to further analyze axonal excitability for varied electric field orientation. MAIN RESULTS: Our results demonstrated that the strongest fMRI response was observed when the electric field was oriented parallel to the axons, while almost no response was detected with the perpendicular orientation of the electric field relative to the primary fiber tract. These results were confirmed by computational models of the experimental paradigm quantifying the activation of radially distributed axons while varying the primary direction of the electric field. SIGNIFICANCE: The described strategies identify a new course for selective neuromodulation paradigms in DBS based on axonal fiber orientation.

Multimodal 7T Imaging of Thalamic Nuclei for Preclinical Deep Brain Stimulation Applications.

Precise neurosurgical targeting of electrode arrays within the brain is essential to the successful treatment of a range of brain disorders with deep brain stimulation (DBS) therapy. Here, we describe a set of computational tools to generate in vivo, subject-specific atlases of individual thalamic nuclei thus improving the ability to visualize thalamic targets for preclinical DBS applications on a subject-specific basis. A sequential nonlinear atlas warping technique and a Bayesian estimation technique for probabilistic crossing fiber tractography were applied to high field (7T) susceptibility-weighted and diffusion-weighted imaging, respectively, in seven rhesus macaques. Image contrast, including contrast within thalamus from the susceptibility-weighted images, informed the atlas warping process and guided the seed point placement for fiber tractography. The susceptibility-weighted imaging resulted in relative hyperintensity of the intralaminar nuclei and relative hypointensity in the medial dorsal nucleus, pulvinar, and the medial/ventral border of the ventral posterior nuclei, providing context to demarcate borders of the ventral nuclei of thalamus, which are often targeted for DBS applications. Additionally, ascending fiber tractography of the medial lemniscus, superior cerebellar peduncle, and pallidofugal pathways into thalamus provided structural demarcation of the ventral nuclei of thalamus. The thalamic substructure boundaries were validated through in vivo electrophysiological recordings and post-mortem blockface tissue sectioning. Together, these imaging tools for visualizing and segmenting thalamus have the potential to improve the neurosurgical targeting of DBS implants and enhance the selection of stimulation settings through more accurate computational models of DBS.

Model-Based Comparison of Deep Brain Stimulation Array Functionality with Varying Number of Radial Electrodes and Machine Learning Feature Sets.

Deep brain stimulation (DBS) leads with radially distributed electrodes have potential to improve clinical outcomes through more selective targeting of pathways and networks within the brain. However, increasing the number of electrodes on clinical DBS leads by replacing conventional cylindrical shell electrodes with radially distributed electrodes raises practical design and stimulation programming challenges. We used computational modeling to investigate: (1) how the number of radial electrodes impact the ability to steer, shift, and sculpt a region of neural activation (RoA), and (2) which RoA features are best used in combination with machine learning classifiers to predict programming settings to target a particular area near the lead. Stimulation configurations were modeled using 27 lead designs with one to nine radially distributed electrodes. The computational modeling framework consisted of a three-dimensional finite element tissue conductance model in combination with a multi-compartment biophysical axon model. For each lead design, two-dimensional threshold-dependent RoAs were calculated from the computational modeling results. The models showed more radial electrodes enabled finer resolution RoA steering; however, stimulation amplitude, and therefore spatial extent of the RoA, was limited by charge injection and charge storage capacity constraints due to the small electrode surface area for leads with more than four radially distributed electrodes. RoA shifting resolution was improved by the addition of radial electrodes when using uniform multi-cathode stimulation, but non-uniform multi-cathode stimulation produced equivalent or better resolution shifting without increasing the number of radial electrodes. Robust machine learning classification of 15 monopolar stimulation configurations was achieved using as few as three geometric features describing a RoA. The results of this study indicate that, for a clinical-scale DBS lead, more than four radial electrodes minimally improved in the ability to steer, shift, and sculpt axonal activation around a DBS lead and a simple feature set consisting of the RoA center of mass and orientation enabled robust machine learning classification. These results provide important design constraints for future development of high-density DBS arrays.

A Novel Lead Design for Modulation and Sensing of Deep Brain Structures.

GOAL: Develop and characterize the functionality of a novel thin-film probe technology with a higher density of electrode contacts than are currently available with commercial deep brain stimulation (DBS) lead technology. Such technology has potential to enhance the spatial precision of DBS and enable a more robust approach to sensing local field potential activity in the context of adaptive DBS strategies. METHODS: Thin-film planar arrays were microfabricated and then assembled on a cylindrical carrier to achieve a lead with 3-D conformation. Using an integrated and removable stylet, the arrays were chronically implanted in the subthalamic nucleus and globus pallidus in two parkinsonian nonhuman primates. RESULTS: This study provides the first in vivo data from chronically implanted DBS arrays for translational nonhuman primate studies. Stimulation through the arrays induced a decrease in parkinsonian rigidity, and directing current around the lead showed an orientation dependence for eliciting motor capsule side effects. The array recordings also showed that oscillatory activity in the basal ganglia is heterogeneous at a smaller scale than detected by the current DBS lead technology. CONCLUSION: These 3-D DBS arrays provide an enabling tool for future studies that seek to monitor and modulate deep brain activity through chronically implanted leads. SIGNIFICANCE: DBS lead technology with a higher density of electrode contacts has potential to enable sculpting DBS current flow and sensing biomarkers of disease and therapy.

Subject-specific computational modeling of DBS in the PPTg area.

Deep brain stimulation (DBS) in the pedunculopontine tegmental nucleus (PPTg) has been proposed to alleviate medically intractable gait difficulties associated with Parkinson's disease. Clinical trials have shown somewhat variable outcomes, stemming in part from surgical targeting variability, modulating fiber pathways implicated in side effects, and a general lack of mechanistic understanding of DBS in this brain region. Subject-specific computational models of DBS are a promising tool to investigate the underlying therapy and side effects. In this study, a parkinsonian rhesus macaque was implanted unilaterally with an 8-contact DBS lead in the PPTg region. Fiber tracts adjacent to PPTg, including the oculomotor nerve, central tegmental tract, and superior cerebellar peduncle, were reconstructed from a combination of pre-implant 7T MRI, post-implant CT, and post-mortem histology. These structures were populated with axon models and coupled with a finite element model simulating the voltage distribution in the surrounding neural tissue during stimulation. This study introduces two empirical approaches to evaluate model parameters. First, incremental monopolar cathodic stimulation (20 Hz, 90 µs pulse width) was evaluated for each electrode, during which a right eyelid flutter was observed at the proximal four contacts (-1.0 to -1.4 mA). These current amplitudes followed closely with model predicted activation of the oculomotor nerve when assuming an anisotropic conduction medium. Second, PET imaging was collected OFF-DBS and twice during DBS (two different contacts), which supported the model predicted activation of the central tegmental tract and superior cerebellar peduncle. Together, subject-specific models provide a framework to more precisely predict pathways modulated by DBS.

In Vivo 7T MRI of the Non-Human Primate Brainstem.

Structural brain imaging provides a critical framework for performing stereotactic and intraoperative MRI-guided surgical procedures, with procedural efficacy often dependent upon visualization of the target with which to operate. Here, we describe tools for in vivo, subject-specific visualization and demarcation of regions within the brainstem. High-field 7T susceptibility-weighted imaging and diffusion-weighted imaging of the brain were collected using a customized head coil from eight rhesus macaques. Fiber tracts including the superior cerebellar peduncle, medial lemniscus, and lateral lemniscus were identified using high-resolution probabilistic diffusion tractography, which resulted in three-dimensional fiber tract reconstructions that were comparable to those extracted from sequential application of a two-dimensional nonlinear brain atlas warping algorithm. In the susceptibility-weighted imaging, white matter tracts within the brainstem were also identified as hypointense regions, and the degree of hypointensity was age-dependent. This combination of imaging modalities also enabled identifying the location and extent of several brainstem nuclei, including the periaqueductal gray, pedunculopontine nucleus, and inferior colliculus. These clinically-relevant high-field imaging approaches have potential to enable more accurate and comprehensive subject-specific visualization of the brainstem and to ultimately improve patient-specific neurosurgical targeting procedures, including deep brain stimulation lead implantation.

Computational modeling of an endovascular approach to deep brain stimulation.

OBJECTIVE: Deep brain stimulation (DBS) therapy currently relies on a transcranial neurosurgical technique to implant one or more electrode leads into the brain parenchyma. In this study, we used computational modeling to investigate the feasibility of using an endovascular approach to target DBS therapy. APPROACH: Image-based anatomical reconstructions of the human brain and vasculature were used to identify 17 established and hypothesized anatomical targets of DBS, of which five were found adjacent to a vein or artery with intraluminal diameter ≥1 mm. Two of these targets, the fornix and subgenual cingulate white matter (SgCwm) tracts, were further investigated using a computational modeling framework that combined segmented volumes of the vascularized brain, finite element models of the tissue voltage during DBS, and multi-compartment axon models to predict the direct electrophysiological effects of endovascular DBS. MAIN RESULTS: The models showed that: (1) a ring-electrode conforming to the vessel wall was more efficient at neural activation than a guidewire design, (2) increasing the length of a ring-electrode had minimal effect on neural activation thresholds, (3) large variability in neural activation occurred with suboptimal placement of a ring-electrode along the targeted vessel, and (4) activation thresholds for the fornix and SgCwm tracts were comparable for endovascular and stereotactic DBS, though endovascular DBS was able to produce significantly larger contralateral activation for a unilateral implantation. SIGNIFICANCE: Together, these results suggest that endovascular DBS can serve as a complementary approach to stereotactic DBS in select cases.