INTEGRATING fNIRS AND MACHINE LEARNING: SHEDDING LIGHT ON PARKINSON'S DISEASE DETECTION

Aug 2, 2026

Upload by Asher

This research presents an approach that supports Parkinson’s disease (PD) diagnosis by classifying functional near-infrared spectroscopy (fNIRS) studies as PD positive or negative. fNIRS provides a noninvasive optical measurement of cerebral hemodynamic responses and blood oxygenation changes, offering a safer and more cost-effective alternative to many neuroimaging modalities. Multiple implementations based on logistic regression are evaluated using 792 extracted temporal and spectral features per participant.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Abstract

Introduction

Parkinson’s disease and machine learning in neuroimaging

fNIRS as a non-invasive neuroimaging modality

fNIRS integration with machine learning for PD detection

Original article: INTEGRATING FNIRS AND MACHINE LEARNING: SHEDDING LIGHT ON PARKINSON'S DISEASE DETECTION Edgar Guevara 1, Gabriel Solana-Lavalle2, Roberto Rosas-Romero2* 1 CONAHCYT – Universidad Autónoma de San Luis Potosí 2 Electrical & Computer Engineering Department, Universidad de las Américas-Puebla * Corresponding author: Roberto Rosas-Romero, Electrical & Computer Engineering Department, Universidad de las Américas-Puebla. E-mail: [roberto.rosas@udlap.mx](roberto.rosas@udlap.mx) [https://dx.doi.org/10.17179/excli2024-7151](https://dx.doi.org/10.17179/excli2024-7151) This is an Open Access article distributed under the terms of the Creative Commons Attribution License ([http://creativecommons.org/licenses/by/4.0/](http://creativecommons.org/licenses/by/4.0/)) . ABSTRACT The purpose of this research is to introduce an approach to assist the diagnosis of Parkinson’s disease (PD) by classifying functional near-infrared spectroscopy (fNIRS) studies as PD positive or negative. fNIRS is a noninvasive optical signal modality that conveys the brain’s hemodynamic response, specifically changes in blood oxygenation in the cerebral cortex; and its potential as a tool to assist PD detection deserves to be explored since it is non-invasive and cost-effective as opposed to other neuroimaging modalities. Besides the integration offNIRSand machine learning, a contribution of this work is that various approaches were implemented and tested to find the implementation that achieves the highest performance. All the implementations used a logistic regression model for classification. A set of 792 temporal and spectral features were extracted from each participant’s fNIRS study. In the two best performing implementations, an ensemble of feature-ranking techniques was used to select a reduced feature subset, which was subsequently reduced with a genetic algorithm. Achieving optimal detection performance, our approach reached 100 % accuracy, precision, and recall, with an F1 score and area under the curve (AUC) of 1, using 14 features. This significantly advances PD diagnosis, highlighting the potential of integrating fNIRS and machine learning for non-invasive PD detection. Keywords: Parkinson’s disease, functional near-infrared spectroscopy, machine learning, feature subset selection, genetic algorithms INTRODUCTION Parkinson's disease (PD), a progressive neurodegenerative disorder, is typified by motor symptoms such as rigidity, tremors, and bradykinesia (Váradi, 2020) . The integration of neuroimaging techniques such as Positron Emission Tomography (PET) and functional Magnetic Resonance Imaging (fMRI) with machine learning has significantly advanced research in PD, offering profound insights into the disease's neural mechanisms. Functional imaging studies have provided powerful tools to study the functional anatomy and pathophysiology of PD, enabling the analysis of task-specific changes in regional cerebral blood flow and blood oxygenation level dependent (BOLD) effects (CeballosBaumann, 2003) . Resting State fMRI (RSfMRI) has been utilized to identify frequencyspecific changes in resting brain activity, offering a novel perspective in PD diagnosis through machine learning approaches (Tian et al., 2020) . Additionally, advancements in data analysis methods using Empirical Mode Decomposition (EMD) have been explored to investigate temporal changes in early PD, further emphasizing the role of neuroimaging in understanding the disease's progression (Cordes et al., 2018) . These studies under- score the potential of combining advanced neuroimaging with machine learning to enhance the diagnosis, understanding, and treatment of PD, paving the way for more effective interventions and improved patient outcomes. In summary, neuroimaging coupled with machine learning techniques is a promising approach for detecting PD potentially leading to earlier diagnosis and more personalized treatment strategies. Functional Near-

FAQs

How does the approach detect Parkinson’s disease using fNIRS?

It classifies fNIRS studies into PD positive or negative based on functional near-infrared spectroscopy signals that reflect cerebral hemodynamic responses and blood oxygenation changes.

What models and feature-selection strategy are used?

All implementations use logistic regression for classification. The best-performing systems apply an ensemble of feature-ranking methods to select a reduced feature subset, followed by reduction using a genetic algorithm.

What performance results does the proposed method achieve?

The reported optimal detection performance reaches 100% accuracy, precision, recall, with an F1 score and area under the curve (AUC) of 1, using 14 features.

Aug 2, 2026

Upload by Asher