
Research Topic (Updated)
Improving Anomaly Detection in EEG Data Using Machine Learning for Monitoring and Aiding Diagnosis Planning of Diseases Related to Brain Health
Research Process Overview

Abstract
Context: Paediatric patients with brain diseases including white matter diseases (WMD) face diagnostic challenges due to subtle changes in symptoms. Electroencephalography (EEG) offers a non-invasive method to monitor brain activity. Manual interpretation is time-consuming and prone to errors, but AI and ML methods can solve the above problems.
Objective: This research aims to develop advanced machine learning (ML) models to improve EEG anomaly detection in paediatric patients with WMD, enhancing early diagnosis and continuous monitoring.
Method: This study adopts an applied research approach using inductive logic to develop ML models from EEG data for anomaly detection in brain health including paediatric White Matter Disease (WMD). Both explanatory and exploratory aims will be addressed through a design science methodology, combining qualitative with quantitative analysis. Data will be gathered from archival research, surveys (optional), and observations, and analysed through statistical and thematic methods to ensure rigorous and ethically rich results. The study will utilize convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to analyse publicly available EEG data and real-world clinical data (optional), while ensuring strict ethical standards for patient privacy and data security.
Expected Results: The research aims to develop an accurate ML process framework that autonomously detects subtle EEG anomalies, improving monitoring, diagnosis, and treatment planning for paediatric neurological conditions.
Expected Contribution: This research will contribute to AI-driven diagnostic tools, enabling non-invasive, real-time monitoring while prioritizing ethical AI practices to enhance patient outcomes and quality of life.
Research Questions
- What are the most effective ML techniques for detecting subtle anomalies in EEG data related to pediatric brain health, especially WMD, for assessing and monitoring QoL in patients?
- How can ML models be optimized to autonomously analyze EEG data, enhancing accuracy and reducing false positives and negatives in anomaly detection?
- What ethical safeguards are necessary to ensure secure, privacy-compliant, and clinically validated AI-driven EEG anomaly detection for pediatric patients with WMD?
Research Objectives
- Develop and Optimize Machine Learning Algorithms for Anomaly Detection of EEG data.
- Reduce Human Intervention and Enhance Diagnostic Accuracy.
- Ensure Ethical AI Implementation and Clinical Validation.
About Me

I am a PhD researcher at the University of Oulu, specializing in artificial intelligence (AI) and machine learning (ML) applications in mental workload classification. With a background in electrical engineering and a master’s degree in computer science, I have extensive experience in developing intelligent systems that integrate physiological data, such as EEG, to assess cognitive states. My research interests include mobile applications for real-time workload management, AI-driven systems for workforce optimization, IoT solutions, and Robotics for industry and healthcare. My project portfolio includes humanoid robot design, smart agriculture systems, and AI-based mobile applications. Notably, I secured second place in the Hack4Health hackathon conducted by the Oulu Incubator Program and the University of Oulu, demonstrating my ability to create innovative solutions with real-world impact.