PUPIL DIAMETER AND MACHINE LEARNING FOR DEPRESSION DETECTION - A COMPARATIVE STUDY WITH DEEP LEARNING MODELS - Abstract

Aug 3, 2026

Upload by Levi

The study addresses depression detection challenges highlighted by the World Health Organization and the American Psychiatric Association, noting large-scale prevalence during COVID-19 and limited access to appropriate treatment. It proposes using machine learning with pupil diameter (PD) features, derived from Hilbert–Huang Transform statistics, to classify depressed versus healthy participants and benchmark against a prior deep learning model. With 58 participants, accuracy reaches 77.72%, and results compare ML Bagging versus deep learning (AlexNet) across left/right and combined eyes, while discussing runtime trade-offs.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages23

Reading Time58 min

AccessFree to download

Introduction

Depression detection challenges

Physiological markers and objective indicators

Eye movement and pupil diameter rationale

Related work on PD, EEG, fMRI, and response time

Submitted: 2024-09-26 | Revised: 2024-10-18 | Accepted: 2024-10-21 Keywords: Pupil Diameter (PD), Major Depressive Disorder (MDD), Machine Learning (ML), Hilbert–Huang Transform (HHT), Cross-Validation (CV) Islam MOHAMED [0009-0001-4408-7190]* , Mohamed EL-WAKAD [0000-0003-2637-1048]** , Khaled ABBAS [0009-0002-0913-4163]*** , Mohamed ABOAMER [0000-0002-4433-776X]**** , NaderA. Rahman MOHAMED [0000-0001-7680-306X]***** PUPIL DIAMETER AND MACHINE LEARNING FOR DEPRESSION DETECTION: A COMPARATIVE STUDY WITH DEEP LEARNING MODELS Abstract According to the World Health Organization, the Global Mental Health Report estimated that between 251 and 310 million individuals worldwide experienced depression during the first year of the COVID-19 pandemic. Most methods for detecting depression rely on clinical diagnoses and surveys. However, the American Psychiatric Association reports that over 50% of patients do not receive appropriate treatment. This study aims to utilize machine learning and pupil diameter features to identify depression and evaluate the accuracy of these classifiers in comparison to our previous deep learning model. While limited research has explored the use of pupillary diameter as a classification tool for distinguishing between individuals with and without depression, several studies have focused on EEG signals for this purpose. The study involved 58 participants, with 29 classified as depressed and 29 as healthy. The classification was based on statistical features extracted from the Hilbert-Huang Transform. Results showed a significant improvement in average accuracy compared to the authors’ prior work, with the current study achieving 77. 72% accuracy, compared to 64. 78% in their previous research. Machine learning methods, particularly Bagging, outperformed deep learning models such as AlexNet when classifying data from the left and right eyes individually (90.91% vs. 78.57% for the left eye; 90.91% vs. 71.43% for the right eye). However, when combining data from both eyes, deep learning using AlexNet demonstrated superior performance (98.28% accuracy compared to 93. 75% using Bagging with statistical features from both eyes). Despite the higher accuracy of deep learning, machine learning is recommended for its faster execution times. * Helwan University, Faculty of Engineering, Biomedical Engineering Department, Cairo, Egypt; Higher Technological Institute, Biomedical Engineering Department, 10th of Ramadan City, Egypt ** Future University, Faculty of Engineering and Technology, Biomedical Engineering Department, New Cairo, Egypt *** Higher Technological Institute, Electronics and Communication Department, 10th Ramadan City, Egypt **** Majmaah University, College of Applied Medical Sciences, Medical Equipment Technology Department, Majmaah 11952, Saudi Arabia ***** Misr University for Science and Technology, Faculty of Engineering, Biomedical Engineering Department, Giza, Egypt, [nader.shaaban@must.edu.eg](nader.shaaban@must.edu.eg) 1. INTRODUCTION The Global Mental Health Report indicated that between 251 and 310 million persons worldwide experienced depression during the first year of the COVID-19 pandemic ( World Health Organization, 2022) . Most depression detection methods use clinical diagnoses and subjective structured scales, which are subjective, time-consuming, and resource-intensive. As a result, traditional methods may delay diagnosis and treatment in many cases, even in severe cases. The American Psychiatric Association reports that over 50% of patients do not get appropriate therapy (Skowron et al., 2022) . There has been a growing focus in recent years on investigating objective physiological markers for the diagnosis of depression. Out of these markers, pupil diameter (PD) has shown as a very promising measure for differentiating between those with depression and those without. The path of this research, which directed author’s attention towards PD, represents

FAQs

What problem does the study target in depression detection?

The study targets the gap between clinical/survey-based depression detection and the need for more objective, efficient methods that can reduce delays in diagnosis and treatment.

How is pupil diameter (PD) used in the proposed approach?

PD is extracted as measurable eye-movement information and converted into statistical features using the Hilbert–Huang Transform for classification of depressed versus healthy individuals.

How do machine learning and deep learning compare in classification performance?

Machine learning with Bagging performs better for left-eye and right-eye data individually, while deep learning with AlexNet achieves higher accuracy when combining both eyes, reaching 98.28% versus 93.75%.

Aug 3, 2026

Upload by Levi