Computational intelligence in subthalamic nucleus deep brain stimulation: A case study in Parkinson`s disease using machine learning supervised techniques

---

Original Article

Author Details : Venkateshwarla Rama Raju*

Volume : 7, Issue : 2, Year : 2021

Article Page : 156-163

https://doi.org/10.18231/j.ijn.2021.026

Suggest article by email

Get Permission

---

Abstract

Deep brain stimulation (DBS) is a complex procedure for subjects experiencing with Parkinson disease (PD) medically resistant neurologic neurodegenerative features (the signs and symptoms). Its impediments are singular; detecting predictors involve several minimal invasive neurosurgical operations. Artificial intelligence (AI) machine learning techniques (MLT) can be employed to well predict these outcomes. The goal of this study is to investigate pre operative quantifiable risk factors experimentally, and to build ML models to predict unfavorable outcomes. Based on the UPDRS stage III+ scale, the subjects were selected. We have gathered clinical - demographic characteristics of PDs undergoing DBS and tabulated occurrence of hurdles. Logistic Regression (LR) is employed to compute risk factors and supervised learning techniques (SLT) were imparted training plus corroborated on 70% and 30% of oversampled and novel registry data. The performance was authenticated exploiting vicinity in the receiver working characteristic curve (A U C), sensitivity, specificity, and accuracy. LR proved that the peril of snag was linked to the working institute wherein the brain-operation done. Odds-ratio(OR): 0.44, confidence-intervals(CI) 0.25e0.78, body-mass-index: BMI OR- 0.94, CI: 0.89e0.99, and diabetics: OR- 2.33, CI:1.18e4.60. PD subjects in diabetics were nearly~33 more accountable to return to the working room OR: 2.78, CI:1.31e5.88. PD subjects by a record of smoking were 43 more probable to practice post operative (post op) infection: OR- 4.20, CI:1.21e14.61. AI-SLTs verified high bias recital while predicting some snag (AUC: 0.86), a snag within dozen months (AUC: 0.91), return to the operating/working room (AUC: 0.88), and bug (AUC: 0.97). Age, BMI, procedure-side, gender, and a diagnosis of Parkinson disease were influential features. Many snag peril factors were recognized, and SLT successfully predicted critical outcomes in D B neurosurgery.
 

Keywords : Data, Data base, Data set, Deep brain Stimulator (DBS), Machine Learning (ML), Machine Learning Algorithms (MLA), Machine Learning Techniques (MLT), Gradient Boosting Machines (GBM), Neuro­Surgical­Operation (NSO), Supervised Learning (SL), Confidence interval (CI), Magnetic resonance imaging (MRI), Computed Axial Tomography (CAT), microelectrode recording (MER), Parkinson Disease (PD), Subthalamic nucleus (STN), Gradient Boosting Machine (GBM)

How to cite : Raju V R, Computational intelligence in subthalamic nucleus deep brain stimulation: A case study in Parkinson`s disease using machine learning supervised techniques. IP Indian J Neurosci 2021;7(2):156-163

This is an Open Access (OA) journal, and articles are distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 License, which allows others to remix, tweak, and build upon the work non-commercially, as long as appropriate credit is given and the new creations are licensed under the identical terms.