Machine Learning Based
Research Projects
#1: Enhancing the Reliability of Electrocardiogram Signals for Stress Management: Learning Transferable Models from High-Resolution Electroencephalogram Supervision
This research is to develop a novel biomarker to detect stress levels by combining Electrocardiogram (ECG) and Electroencephalogram (EEG) signals.
​
I used Google Colab to code and construct convolutional neural networks (CNNs). I used 1D structured CNNs because both datasets had specific time settings.
​
I submitted to the Journal of Student Research (JSR) and am awaiting a response on the official dates of publication. Along with publication, I was awarded the Design for Society Award at the regional ISEF fair called the Greater Capital Region Science and Engineering Fair.
​
The research abstract is as follows...
​
Stress, though often overlooked, is a significant driver behind the high suicide rates in South Korea, which leads among OECD countries with 24.1 suicides per 100,000 people. This issue is not unique to South Korea; in the United States, adolescent suicide rates have surged by 35% since 1999, now affecting 12 per 100,000 individuals. To address these alarming trends and prevent stress-related health issues, schools frequently rely on traditional checklist tests. However, these tests are often inaccurate, are time-consuming, and lack scientific rigor. Recent studies have started to explore the use of electroencephalogram (EEG) technology, which provides a scientific and quantitative measure of stress by closely monitoring brain activity. Despite its promise, electrocardiogram (ECG) signal, which also correlates with stress, has not been as thoroughly investigated. To address this problem, I propose an integrated approach that combines EEG and ECG assessments to develop a more reliable and cost-effective method for stress detection using machine learning. During the training phase, the proposed system takes both EEG and ECG signals as input and learns to map these signals into an emotion-related feature space. After training, the pre-trained network is then used to predict arousal and valence from ECG signals. Extensive experiments demonstrated that the proposed approach significantly improved performance, reducing RMSE by 8.24.
#2: Electroencephalography Representation Learning with Spatiotemporal Reconstruction for Accurate Attention Deficit Hyperactivity Disorder Detection
This research is to develop a novel biomarker to detect ADHD by using EEG's spatiotemporal features, creating a novel EEG Representation Learning, and making a Reconstruction Time-Step CNN.
​
I used Google Colab to code and construct convolutional neural networks (CNNs). I used 1D structured CNNs because both datasets had specific time settings.
​
I submitted to the Journal of Student Research (JSR) and am awaiting a response on the official dates of publication. Along with publication, I was selected as a Finalist in the Genius Olympiad Science Innovation Category in 2025.
​
The research abstract is as follows...
​
Attention-Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental condition characterized by persistent patterns of inattention, hyperactivity, and impulsivity that interfere with daily functioning or development. Traditional diagnosis of ADHD often relies on self-report checklists and behavioral assessments, which can be subjective and prone to bias. These methods may overlook subtle neural patterns that are indicative of ADHD, which leads to misdiagnosis or delayed intervention. To address these limitations, I propose an Electroencephalography (EEG)-based ADHD screening system utilizing machine learning algorithms. The proposed network takes a set of EEG signals as input and outputs the probability that an individual has ADHD. To enhance accuracy, I introduce a novel EEG representation learning technique to capture the spatiotemporal features of EEG data. The trained network learns to extract rich features from the EEG signals, which significantly improves the accuracy of ADHD detection. Through extensive experiments, the proposed system achieved state-of-the-art performance in screening ADHD, reaching an accuracy of 99%. Additionally, further experiments were conducted to identify which parts of the brain are highly correlated with the screening of ADHD. In conclusion, this research not only validates the effectiveness of the EEG-based system but also contributes to our understanding of ADHD's neurobiological basis.