How Computers Use Brain Maps to Diagnose Alzheimer's Disease

A method for classifying stages of Alzheimer's disease by turning brain scan data into mathematical graphs and using machine learning to identify patterns.

What it covers

The patent describes a computational pipeline that turns raw brain scan data—like EEG or MRI—into a mathematical representation of how different brain regions communicate. First, it extracts data from a scan and builds a 'brain network graph,' where brain regions are nodes and their connections are edges. It then uses a technique called node2vec to turn this complex graph into a simplified list of numbers, known as a feature vector. Finally, it uses machine learning algorithms like Support Vector Machines to analyze these vectors and classify whether a patient is healthy, has mild cognitive impairment, or has Alzheimer's disease.

What it doesn't cover

• —Does not cover the physical hardware or medical devices used to capture the initial EEG, fMRI, or PET scans.
• —Does not cover diagnostic methods that rely solely on clinical interviews or memory tests without the graph-based computational analysis.
• —Does not cover the specific medical treatment or drug intervention prescribed after a diagnosis is made.
• —Does not cover general-purpose machine learning algorithms that are not specifically applied to the described brain network graph methodology.

The clever bit

The innovation lies in applying node2vec—a technique usually used to map social networks or web links—to the structural and functional connections of the human brain to detect disease progression.

Why it matters

As the global population ages, early and accurate diagnosis of neurodegenerative diseases is critical. By standardizing how we turn brain imaging into data that computers can 'read,' this approach helps move Alzheimer's diagnosis from subjective clinical observation toward more objective, data-driven analysis.

Real-world examples

• 1.Computer-aided diagnostic software for neurologists
• 2.Automated analysis tools for clinical research studies
• 3.Digital health platforms processing EEG data