Automatic subthalamic nucleus detection from microelectrode recordings based on noise level and neuronal activity

A Systematic Review of Neurophysiology-Based Localization Techniques Used in Deep Brain Stimulation Surgery of the Subthalamic Nucleus

OBJECTIVE

This systematic review is conducted to identify, compare, and analyze neurophysiological feature selection, extraction, and classification to provide a comprehensive reference on neurophysiology-based subthalamic nucleus (STN) localization.

MATERIALS AND METHODS

The review was carried out using the methods and guidelines of the Kitchenham systematic review and provides an in-depth analysis on methods proposed on STN localization discussed in the literature between 2000 and 2021. Three research questions were formulated, and 115 publications were identified to answer the questions.

RESULTS

The three research questions formulated are answered using the literature found on the respective topics. This review discussed the technologies used in past research, and the performance of the state-of-the-art techniques is also reviewed.

CONCLUSION

This systematic review provides a comprehensive reference on neurophysiology-based STN localization by reviewing the research questions other new researchers may also have.

Median Nerve Stimulation Facilitates the Identification of Somatotopy of the Subthalamic Nucleus in Parkinson’s Disease Patients under Inhalational Anesthesia

Deep brain stimulation (DBS) improves Parkinson’s disease (PD) symptoms by suppressing neuropathological oscillations. These oscillations are also modulated by inhalational anesthetics used during DBS surgery in some patients, influencing electrode placement accuracy. We sought to evaluate a method that could avoid these effects. We recorded subthalamic nucleus (STN) neuronal firings in 11 PD patients undergoing DBS under inhalational anesthesia. Microelectrode recording (MER) during DBS was collected under median nerve stimulation (MNS) delivered at 5, 20, and 90 Hz frequencies and without MNS. We analyzed the spike firing rate and neuronal activity with power spectral density (PSD), and assessed correlations between the neuronal oscillation parameters and clinical motor outcomes. No patient experienced adverse effects during or after DBS surgery. PSD analysis revealed that peripheral 20 Hz MNS produced significant differences in the dorsal and ventral subthalamic nucleus (STN) between the beta band oscillation (16.9 ± 7.0% versus 13.5 ± 4.8%, respectively) and gamma band oscillation (56.0 ± 13.7% versus 66.3 ± 9.4%, respectively) (p < 0.05). Moreover, 20-Hz MNS entrained neural oscillation over the dorsal STN, which correlated positively with motor disabilities. MNS allowed localization of the sensorimotor STN and identified neural characteristics under inhalational anesthesia. This paradigm may help identify an alternative method to facilitate STN identification and DBS surgery under inhalational anesthesia.

Deep convolutional neural network for the automated detection of Subthalamic nucleus using MER signals

BACKGROUND

Deep brain stimulation (DBS) surgery has been extensively conducted for treating advanced Parkinson's disease (PD) patient's symptoms. DBS hinges on the localization of the subthalamic nucleus (STN) in which a permanent electrode should be accurately placed to produce electrical current. Microelectrode recording (MER) signals are routinely recorded in the procedure of DBS surgery to validate the planned trajectories. However, manual MER signals interpretation with the goal of detecting STN borders requires expertise and prone to inter-observer variability. Therefore, a computerized aided system would be beneficial to automatic detection of the dorsal and ventral borders of the STN in MER.

NEW METHOD

In this study, a new deep learning model based on convolutional neural system for automatic delineation of the neurophysiological borders of the STN along the electrode trajectory was developed.

COMPARISON WITH EXISTING METHODS

The proposed model does not involve any conventional standardization, feature extraction or selection steps.

RESULTS

Promising results of 98.67% accuracy, 99.03% sensitivity, 98.11% specificity, 98.79% precision and 98.91% F1-score for subject based testing were achieved using the proposed convolutional neural network (CNN) model.

CONCLUSIONS

This is the first study on the analysis of MER signals to detect STN using deep CNN. Traditional machine learning (ML) algorithms are often cumbersome and suffer from subjective evaluation. Though, the developed 10-layered CNN model has the capability of extracting substantial features at the convolution stage. Hence, the proposed model has the potential to deliver high performance on STN region detection which shows perspective in aiding the neurosurgeon intraoperatively.

An intra-operative feature-based classification of microelectrode recordings to support the subthalamic nucleus functional identification during deep brain stimulation surgery

OBJECTIVE

The Subthalamic Nucleus (STN) is the most selected target for the placement of the Deep Brain Stimulation (DBS) electrode to treat Parkinson's disease. Its identification is a delicate and challenging task which is based on the interpretation of the STN functional activity acquired through microelectrode recordings (MER). Aim of this work is to explore the potentiality of a set of twenty-five features to build a classification model for the discrimination of MER signals belonging to the STN.

APPROACH

We explored the use of different sets of spike-dependent and spike-independent features in combination with an Ensemble Trees classification (ET) algorithm on a dataset composed of thirteen patients receiving bilateral DBS. We compared results from six subsets of features and two dataset conditions (with and without standardization) using performance metrics on a leave-one-patient-out validation schema.

MAIN RESULTS

We obtained statistically better results (i.e., higher accuracy p-value = 0.003) on the raw dataset than on the standardized one, where the selection of seven features using a minimum redundancy maximum relevance (MRMR) algorithm provided a mean accuracy of 94.1%, comparable with the use of the full set of features. In the same conditions, the spike-dependent features provided the lowest accuracy (86.8%), while a power density-based index was shown to be a good indicator of STN activity (92.3%).

SIGNIFICANCE

Results suggest that a small and simple set of features can be used for an efficient classification of microelectrode recordings to implement an intraoperative support for clinical decision during deep brain stimulation surgery.

Desflurane and sevoflurane differentially affect activity of the subthalamic nucleus in Parkinson's disease

BACKGROUND

Desflurane and sevoflurane are commonly used during inhalational anaesthesia, but few studies have investigated their effects on deep cerebral neuronal activity. In addition, the association between subthalamic nucleus (STN) neurophysiology and general anaesthesia induced by volatile anaesthetics are not yet identified. This study aimed to identify differences in neurophysiological characteristics of the STN during comparable minimal alveolar concentration (MAC) desflurane and sevoflurane anaesthesia for deep brain stimulation (DBS) in patients with Parkinson's disease.

METHODS

Twelve patients with similar Parkinson's disease severity received desflurane (n=6) or sevoflurane (n=6) during DBS surgery. We obtained STN spike firing using microelectrode recording at 0.5-0.6 MAC and compared firing rate, power spectral density, and coherence.

RESULTS

Neuronal firing rate was lower with desflurane (47.4 [26.7] Hz) than with sevoflurane (63.9 [36.5] Hz) anaesthesia (P<0.001). Sevoflurane entrained greater gamma oscillation power than desflurane (62.9% [0.9%] vs 57.0% [1.5%], respectively; P=0.002). There was greater coherence in the theta band of the desflurane group compared with the sevoflurane group (13% vs 6%, respectively). Anaesthetic choice did not differentially influence STN mapping accuracy or the clinical outcome of DBS electrode implantation.

CONCLUSIONS

Desflurane and sevoflurane produced distinct neurophysiological profiles in humans that may be associated with their analgesic and hypnotic actions.

SepaConvNet for Localizing the Subthalamic Nucleus Using One Second Micro-electrode Recordings

Micro-electrode recording (MER) is a powerful way of localizing target structures during neurosurgical procedures such as the implantation of deep brain stimulation electrodes, which is a common treatment for Parkinson's disease and other neurological disorders. While Micro-electrode Recording (MER) provides adjunctive information to guidance assisted by pre-operative imaging, it is not unanimously used in the operating room. The lack of standard use of MER may be in part due to its long duration, which can lead to complications during the operation, or due to high degree of expertise required for their interpretation. Over the past decade, various approaches addressing automating MER analysis for target localization have been proposed, which have mainly focused on feature engineering. While the accuracies obtained are acceptable in certain configurations, one issue with handcrafted MER features is that they do not necessarily capture more subtle differences in MER that could be detected auditorily by an expert neurophysiologist. In this paper, we propose and validate a deep learning-based pipeline for subthalamic nucleus (STN) localization with micro-electrode recordings motivated by the human auditory system. Our proposed Convolutional Neural Network (CNN), referred as SepaConvNet, shows improved accuracy over two comparative networks for locating the STN from one second MER samples.

Characterization of Microelectrode Recordings for the Subthalamic Nucleus identification in Parkinson's disease

Deep brain stimulation (DBS) of the subthalamic nucleus (STN) is an effective treatment for Parkinson's disease, when the pharmacological approach has no more effect. DBS efficacy strongly depends on the accurate localization of the STN and the adequate positioning of the stimulation electrode during DBS stereotactic surgery. During this procedure, the analysis of microelectrode recordings (MER) is fundamental to assess the correct localization. Therefore, in this work, we explore different signal feature types for the characterization of the MER signals associated to STN from NON-STN structures. We extracted a set of spike-dependent (action potential domain) and spike-independent features in the time and frequency domain to evaluate their usefulness in distinguishing the STN from other structures. We discuss the results from a physiological and methodological point of view, showing the superiority of features having a direct electrophysiological interpretation.Clinical Relevance- The identification of a simple, clinically interpretable, and powerful set of features for the STN localization would support the clinical positioning of the DBS electrode, improving the treatment outcome.

Slim electrodes for improved targeting in deep brain stimulation

OBJECTIVE

The efficacy of deep brain stimulation can be limited by factors including poor selectivity of stimulation, targeting error, and complications related to implant reliability and stability. We aimed to improve surgical outcomes by evaluating electrode leads with smaller diameter electrode and microelectrodes incorporated which can be used for assisting targeting.

APPROACH

Electrode arrays were constructed with two different diameters of 0.65 mm and the standard 1.3 mm. Micro-electrodes were incorporated into the slim electrode arrays for recording spiking neural activity. Arrays were bilaterally implanted into the medial geniculate body (MGB) in nine anaesthetised cats for 24-40 h using stereotactic techniques. Recordings of auditory evoked field potentials and multi-unit activity were obtained at 1 mm intervals along the electrode insertion track. Insertion trauma was evaluated histologically.

MAIN RESULTS

Evoked auditory field potentials were recorded from ring and micro-electrodes in the vicinity of the medial geniculate body. Spiking activity was recorded from 81% of the microelectrodes approaching the MGB. Histological examination showed localized surgical trauma along the implant. The extent of haemorrhage surrounding the track was measured and found to be significantly reduced with the slim electrodes (541 ± 455 µm vs. 827 ± 647 µm; P < 0.001). Scoring of the trauma, focusing on tissue disruption, haemorrhage, oedema of glial parenchyma and pyknosis, revealed a significantly lower trauma score for the slim electrodes (P < 0.0001).

SIGNIFICANCE

The slim electrodes reduced the extent of acute trauma, while still providing adequate electrode impedance for both stimulating and recording, and providing the option to target stimulate smaller volumes of tissue. The incorporation of microelectrodes into the electrode array may allow for a simplified, single-step surgical approach where confirmatory micro-targeting is done with the same lead used for permanent implantation.

Relative Contribution of Magnetic Resonance Imaging, Microelectrode Recordings, and Awake Test Stimulation in Final Lead Placement during Deep Brain Stimulation Surgery of the Subthalamic Nucleus in Parkinson's Disease

INTRODUCTION

For deep brain stimulation (DBS) surgery of the subthalamic nucleus (STN) in Parkinson's disease (PD), many centers employ visualization of the nucleus on magnetic resonance imaging (MRI), intraoperative microelectrode recordings (MER), and test stimulation in awake patients. The value of these steps is a subject for ongoing debate. In the current study, we determined the relative contribution of MRI targeting, multitrack MER, and awake test stimulation in final lead placement during STN DBS surgery for PD.

METHODS

Data on PD patients undergoing MRI-targeted STN DBS surgery with three-channel MER and awake test stimulation between February 2010 and January 2014 were analyzed to determine in which MER trajectory final leads were implanted and why this tract was chosen.

RESULTS

Seventy-six patients underwent implantation of 146 DBS leads. In 92% of the STN, the final leads were implanted in one of the three planned channels. In 6%, additional channels were needed. In 2%, surgery was aborted before final lead implantation due to anxiety or fatigue. The final leads were implanted in the channels with the longest STN MER signal trajectory in 60% of the STN (38% of the bilaterally implanted patients). This was the central channel containing the MRI target in 39% of the STN (18% bilaterally). The most frequently noted reasons why another channel than the central channel was chosen for final lead placement were (1) a lower threshold for side effects (54%) and (2) no or a too short trajectory of the STN MER signal (40%) in the central channel. The latter reason correlated with larger 2D (x and y) errors in our stereotactic method.

CONCLUSIONS

STN DBS leads were often not implanted in the MRI-planned trajectory or in the trajectory with the longest STN MER signal. Thresholds for side effects during awake test stimulation were decisive for final target selection in the majority of patients.

Optimization of the KNN Supervised Classification Algorithm as a Support Tool for the Implantation of Deep Brain Stimulators in Patients with Parkinson's Disease

Deep Brain Stimulation (DBS) of the Subthalamic Nuclei (STN) is the most used surgical treatment to improve motor skills in patients with Parkinson’s Disease (PD) who do not adequately respond to pharmacological treatment, or have related side effects. During surgery for the implantation of a DBS system, signals are obtained through microelectrodes recordings (MER) at different depths of the brain. These signals are analyzed by neurophysiologists to detect the entry and exit of the STN region, as well as the optimal depth for electrode implantation. In the present work, a classification model is developed and supervised by the K-nearest neighbour algorithm (KNN), which is automatically trained from the 18 temporal features of MER registers of 14 patients with PD in order to provide a clinical support tool during DBS surgery. We investigate the effect of different standardizations of the generated database, the optimal definition of KNN configuration parameters, and the selection of features that maximize KNN performance. The results indicated that KNN trained with data that was standardized per cerebral hemisphere and per patient presented the best performance, achieving an accuracy of 94.35% (p < 0.001). By using feature selection algorithms, it was possible to achieve 93.5% in accuracy in selecting a subset of six features, improving computation time while processing in real time.

Electrode Location in a Microelectrode Recording-Based Model of the Subthalamic Nucleus Can Predict Motor Improvement After Deep Brain Stimulation for Parkinson's Disease

Motor improvement after deep brain stimulation (DBS) in the subthalamic nucleus (STN) may vary substantially between Parkinson’s disease (PD) patients. Research into the relation between improvement and active contact location requires a correction for anatomical variation. We studied the relation between active contact location relative to the neurophysiological STN, estimated by the intraoperative microelectrode recordings (MER-based STN), and contralateral motor improvement after one year. A generic STN shape was transformed to fit onto the stereotactically defined MER sites. The location of 43 electrodes (26 patients), derived from MRI-fused CT images, was expressed relative to this patient-specific MER-based STN. Using regression analyses, the relation between contact location and motor improvement was studied. The regression model that predicts motor improvement based on levodopa effect alone was significantly improved by adding the one-year active contact coordinates (R² change = 0.176, p = 0.014). In the combined prediction model (adjusted R² = 0.389, p < 0.001), the largest contribution was made by the mediolateral location of the active contact (standardized beta = 0.490, p = 0.002). With the MER-based STN as a reference, we were able to find a significant relation between active contact location and motor improvement. MER-based STN modeling can be used to complement imaging-based STN models in the application of DBS.

Spatio-spectral characterization of local field potentials in the subthalamic nucleus via multitrack microelectrode recordings

Deep brain stimulation of the subthalamic nucleus (STN) is a highly effective treatment for motor symptoms of Parkinson's disease. However, precise intraoperative localization of STN remains a procedural challenge. In the present study, local field potentials (LFPs) were recorded from three tracks during microelectrode recording-based (MER) targeting of STN, in five patients. The raw LFP data were preprocessed in original recording setup and then data quality was compared to data with common average derivation. The depth-frequency maps were generated according to preprocessing results for each patient and spectral characteristics of LFPs were explored at each depth across different tracks and different subjects. Spatio-spectral analysis of LFP was investigated to see whether LFP activity can be used for optimal track selection and STN border identification. Analysis show that monopolar derivation suffer from various artifacts and/or power line noise which makes the interpretation of target localization very difficult in most of the subjects. Unlikely, bipolar derivation helps to recover the neurological signals and investigation of signal characteristics. The frequency-vs-depth maps using a modified Welch periodogram with robust statistics, demonstrated that a median-based spectrum estimation approach eliminates outliers pretty well by preserving band-specific LFP activity. The results indicate that there is a clear oscillatory beta activity around 20 Hz in all subjects. 1/f normalization reveals the high frequency oscillations (HFOs) between 200-to-350 Hz in two subjects. It's noted that the optimal track selection is not consistent with the track having highest beta band oscillations in two out of five subjects. In conclusion, microelectrode-derived LFP recordings may provide an alternative approach to single unit activity (SUA)-based MER, for localizing the target STN borders during DBS surgery. Despite the small number of subjects, the present study adds to existing knowledge about LFP-based pathophysiology of PD and its target-based spectral activities.

The Effects of Dexmedetomidine on Microelectrode Recordings of the Subthalamic Nucleus during Deep Brain Stimulation Surgery: A Retrospective Analysis

BACKGROUND

The placement of subthalamic nucleus (STN) deep brain stimulation (DBS) electrodes can be facilitated by intraoperative microelectrode recording (MER) of the STN.

OBJECTIVES

Optimal anesthetic management during surgery remains unclear because of a lack of quantitative data of the effect of anesthetics on MER. Therefore, we measured the effects of dexmedetomidine (DEX) on MER measures of the STN commonly taken intraoperatively.

METHODS

MER from 45 patients was retrospectively compared between patients treated with remifentanil (REMI) alone or both REMI and DEX, which are the 2 main standards of care at our center. The measures examined were population activity, such as root mean square, STN length, and number of passes yielding STN, and the single-neuron measures of firing rate and variability.

RESULTS

The addition of DEX does not affect population measures (number of passes: DEX+REMI, n = 68, REMI only, n = 154), or neuronal firing rates (number of neurons: DEX+REMI, n = 64, REMI only, n = 72), but firing rate variability was reduced.

CONCLUSIONS

In this cohort, population-based measures routinely used for electrode placement in the STN were unaffected by DEX when added to REMI. Neuronal firing rates were also unaffected, but their variability was reduced, even beyond 20 min after cessation.

Stop! border ahead: Automatic detection of subthalamic exit during deep brain stimulation surgery

BACKGROUND

Microelectrode recordings along preplanned trajectories are often used for accurate definition of the subthalamic nucleus (STN) borders during deep brain stimulation (DBS) surgery for Parkinson's disease. Usually, the demarcation of the STN borders is performed manually by a neurophysiologist. The exact detection of the borders is difficult, especially detecting the transition between the STN and the substantia nigra pars reticulata. Consequently, demarcation may be inaccurate, leading to suboptimal location of the DBS lead and inadequate clinical outcomes.

METHODS

We present machine-learning classification procedures that use microelectrode recording power spectra and allow for real-time, high-accuracy discrimination between the STN and substantia nigra pars reticulata.

RESULTS

A support vector machine procedure was tested on microelectrode recordings from 58 trajectories that included both STN and substantia nigra pars reticulata that achieved a 97.6% consistency with human expert classification (evaluated by 10-fold cross-validation). We used the same data set as a training set to find the optimal parameters for a hidden Markov model using both microelectrode recording features and trajectory history to enable real-time classification of the ventral STN border (STN exit). Seventy-three additional trajectories were used to test the reliability of the learned statistical model in identifying the exit from the STN. The hidden Markov model procedure identified the STN exit with an error of 0.04 ± 0.18 mm and detection reliability (error < 1 mm) of 94%.

CONCLUSIONS

The results indicate that robust, accurate, and automatic real-time electrophysiological detection of the ventral STN border is feasible. © 2016 International Parkinson and Movement Disorder Society.

Localization of subthalamic nucleus borders using macroelectrode local field potential recordings

Deep brain stimulation of the subthalamic nucleus (STN) is a highly effective treatment for motor symptoms of Parkinson's disease. However, precise intraoperative localization of STN remains a procedural challenge. In the present study, local field potentials (LFPs) were recorded from DBS macroelectrodes during trajectory to STN, in six patients. The frequency-vs-depth map of LFP activity was extracted and further analyzed within different sub-bands, to investigate whether LFP activity can be used for STN border identification. STN borders identified by LFPs were compared to border predictions by the neurosurgeon, based on microelectrode-derived, single-unit recordings (MER-SUA). The results demonstrate difference between MER-SUA and macroelectrode LFP recording with respect to the dorsal STN border of -1.00 ±0.84 mm and -0.42 ±1.07 mm in the beta and gamma frequency bands, respectively. For these sub-bands, RMS of these distances was found to be 1.26 mm and 1.06 mm, respectively. Analysis of other sub-bands did not allow for distinguishing the caudal border of STN. In conclusion, macroelectrode-derived LFP recordings may provide an alternative approach to MER-SUA, for localizing the target STN borders during DBS surgery.

NEUROZONE: on-line recognition of brain structures in stereotactic surgery--application to Parkinson's disease

The success of stereotactic surgery for Deep Brain Stimulation depends critically on the exact positioning of a microelectrode recording in a target area of the brain. This paper presents the software system NEUROZONE composed of two main applications: first, it allows online recognition of brain structures by the analysis of signals from microelectrode recordings (MER), and second, it processes and analyses off-line databases allowing the inclusion of new trained classifiers for automatic identification. The software serves as a support to the analysis done by a medical specialist during surgery, and seeks to reduce the adverse side effects that may occur because of inadequate identification of the target areas. The software also allows the specialists to label recordings obtained during surgery, in order to generate a new off-line database or increase the amount of records in an already existing off-line database. NEUROZONE has been tested for Deep Brain Stimulation performed at the Institute for Epilepsy and Parkinson of the Eje Cafetero (Colombia), achieving positive identifications of the Subthalamic Nucleus (STN) over to 85% using a naive Bayes classifier.

Microelectrode recording duration and spatial density constraints for automatic targeting of the subthalamic nucleus

BACKGROUND

Accurate detection of the boundaries of the subthalamic nucleus (STN) in deep brain stimulation (DBS) surgery using microelectrode recording (MER) is considered to refine localization and may therefore improve clinical outcome. However, MER tends to extend operation time and its cost-utility balance has been debated.

OBJECTIVES

To quantify the tradeoff between accuracy of STN localization and the spatial and temporal parameters of MER that effect the operation time using an automated detection method.

METHODS

We retrospectively estimated the accuracy of STN detection on data from 100 microelectrode trajectories. Our dense (average step = 0.12 mm) and long (average duration = 22.5 s) MER data was downsampled in the spatial and temporal domains. Then, the STN borders were detected automatically on both the downsampled and original data and compared to each other.

RESULTS

With a recording duration of 16 s, average accuracy for detecting STN entry ranged from 0.06 mm for a 0.1-mm step to 0.51 mm for a 1.0-mm step. Smaller effects were found along the temporal axis. For example, a 0.1-mm recording step yielded an STN entry average accuracy ranging from 0.06 mm for a 16-second recording duration to 0.16 mm for 0.1 s.

CONCLUSIONS

STN entry detection error was about half of the step size. Sampling duration of STN activity can be minimized to 1 s/record without compromising accuracy. We conclude that bilateral DBS surgery time utilizing MER may be significantly shortened without compromising targeting accuracy.