Takashi-Hudson Naito
Class of 2026Okayama, Okayama
About
To my sensei, Hello! By way of introduction, I'm going to write a little about where my topic ideas for the research come from. I'm afraid they are still fairly general at this point but they stem from my keen interest in neuroscience and technology. For the past few years I've been drawn to any studies I could find online connecting the human neurological system to computers. My interest started to develop around the time I lost my grandfather; he helped raise me and passed away right before I graduated from middle school. Up untill then death felt surreal, something I’d read about or heard of but was never really close to home. It was during this period of grieving and thinking philosophical about life and death, I stumbled upon a movie titled Chappie. In it, a robot achieves self-consciousness and gets to resurrect a human friend by transferring her brain into a digital format, essentially converting her consciousness into data. While such technology may be far beyond our reach today, besides some obvious ethical concerns of course, the silly movie inspired me. The idea that human consciousness could be copied fascinated me and from then on I knew this was the field I wanted to explore in a rigorously scientific manner. One of my primary goals in completing a Polygence project is to strengthen my college applications by highlighting my sincere interest in biology/neuroscience and technology. That said, I'm very much open to suggestions. By conducting a research project under your kind guidance, I would like to start building a robust research portfolio in that domain. I'm hopeful that in completing my project I could further develop lab-relevant coding proficiency, readiness for scientific research and gain more experience in working with complex datasets. Right now I would want nothing more than to hone my analytical and investigate abilities as well as hard skills including but not limited to Python, Matlab and data visualization.gProjects
- "How can deep learning models, particularly CNN-based architectures, be used to classify emotional states from EEG-derived power spectral density (PSD) topomaps?" with mentor David (June 14, 2025)
Project Portfolio
How can deep learning models, particularly CNN-based architectures, be used to classify emotional states from EEG-derived power spectral density (PSD) topomaps?
Started Dec. 17, 2024
Abstract or project description
The project focuses on developing a deep learning model to classify emotional states based on EEG data and map them onto Russell’s Circumplex Model of emotions. Using power spectral density (PSD) topomaps extracted from EEG recordings, the model will leverage convolutional neural networks (CNNs) to recognize patterns in brain activity associated with different emotions. The classified emotions will then be projected onto a two-dimensional space (valence-arousal) to provide a structured representation of emotional states. The study also aims to refine EEG data preprocessing techniques, evaluate different neural network architectures, and explore the effectiveness of spatial EEG representations for emotion classification. The ultimate goal is to create an AI system capable of making accurate emotion predictions from EEG signals, contributing to advancements in brain-computer interfaces and affective computing.