Exploring how AI is transforming the diagnosis and treatment of brain diseases
This paper shows that explainable machine-learning methods, combined with digital biomarkers such as eye tracking and cognitive tests, can help identify early signs of Parkinson's and Alzheimer's disease before clear clinical symptoms appear.
This paper applies interpretable machine-learning techniques to structural MRI data to distinguish between control, prodromal, and Parkinson's disease groups, supporting research on early diagnosis.
This book chapter describes explainable machine-learning workflows using DTI imaging, eye tracking, and cognitive testing to study Parkinson's disease progression and explore early screening approaches.
Artur Chudzik is an AI Engineer and PhD in Computer Science specializing in medical AI. He develops explainable machine learning methods to detect neurodegenerative diseases such as Parkinson's and Alzheimer's.
With over a decade of software engineering experience, he combines academic research to build intelligent systems that are interpretable, reliable, and clinically useful.
I work on building machine-learning systems for biomedical data that are not only accurate, but also understandable. I am especially interested in explainable AI — models that clinicians can trust and interpret. When AI is transparent, it can reveal meaningful patterns in data and make it easier to bring these tools into real healthcare settings.
My research explores how computational models can help us better understand the brain, particularly in neurodegenerative diseases. By connecting data analysis with biological processes, I aim to identify early changes in the brain and contribute to earlier detection and intervention strategies.
I am fascinated by complex systems — from high-dimensional data to mathematical models that describe how patterns emerge. Using tools from applied mathematics and data science, I study how structure and behavior arise in complicated systems, with the goal of improving both prediction and understanding.