Feature selection for KNN classifier to improve accurate detection of subthalamic nucleus during deep brain stimulation surgery in Parkinson’s patients

Machine learning for the localization of Subthalamic Nucleus during deep brain stimulation surgery: a systematic review and Meta-analysis

INTRODUCTION

Delineating subthalamic nucleus (STN) boundaries using microelectrode recordings (MER) and trajectory history is a valuable resource for neurosurgeons, aiding in the accurate and efficient positioning of deep brain stimulation (DBS) electrodes within the STN. Here, we aimed to assess the application of artificial intelligence, specifically Hidden Markov Models (HMM), in the context of STN localization.

METHODS

A comprehensive search strategy was employed, encompassing electronic databases, including PubMed, EuroPMC, and MEDLINE. This search strategy entailed a combination of controlled vocabulary (e.g., MeSH terms) and free-text keywords pertaining to "artificial intelligence," "machine learning," "deep learning," and "deep brain stimulation." Inclusion criteria were applied to studies reporting the utilization of HMM for predicting outcomes in DBS, based on structured patient-level health data, and published in the English language.

RESULTS

This systematic review incorporated a total of 14 studies. Various machine learning compared wavelet feature to proposed features in diagnosing the STN, with the HMM yielding a diagnostic odds ratio (DOR) of 838.677 (95% CI: 203.309-3459.645). Similarly, the K-Nearest Neighbors (KNN) model produced parameter estimates, including a diagnostic odds ratio of 25.151 (95% CI: 12.270-51.555). Meanwhile, the support vector machine (SVM) model exhibited parameter estimates, with a DOR of 13.959 (95% CI: 10.436-18.671).

CONCLUSIONS

MER data demonstrates significant variability in neural activity, with studies employing a wide range of methodologies. Machine learning plays a crucial role in aiding STN diagnosis, though its accuracy varies across different approaches.

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.