Graduation Year


Date of Submission


Document Type

Open Access Senior Thesis

Degree Name

Bachelor of Arts



Reader 1

Michael Spezio

Reader 2

Sebastien Naze

Terms of Use & License Information

Terms of Use for work posted in Scholarship@Claremont.

Rights Information

2019 Isabel R Milano


Alzheimer’s Disease (AD) is the only condition in the top ten leading causes of death for which we do not have an effective treatment that prevents, slows, or stops its progression. Our ability to design useful interventions relies on (a) increasing our understanding of the pathological process of AD and (b) improving our ability for its early detection. These goals are impeded by our current reliance on the clinical symptoms of AD for its diagnosis. This characterizations of AD often falsely assumes a unified, underlying AD-specific pathology for similar presentations of dementia that leads to inconsistent diagnoses. It also hinges on postmortem verification, and so is not a helpful method for identifying patients and research subjects in the beginning phases of the pathophysiological process. Instead, a new biomarker-based approach provides a more biological understanding of the disease and can detect pathological changes up to 20 years before the clinical symptoms emerge. Subjects are assigned a profile according to their biomarker measures of amyloidosis (A), tauopathy (T) and neurodegeneration (N) that reflects their underlying pathology in vivo. AD is confirmed as the underlying pathology when subjects have abnormal values of both amyloid and tauopathy biomarkers, and so have a biomarker profile of A+T+(N)- or A+T+(N)+. This new biomarker based characterization of AD can be combined with machine learning techniques in multimodal classification studies to shed light on the elements of the AD pathological process and develop early detection paradigms. A guiding research framework is proposed for the development of reliable, biologically-valid and interpretable multimodal classification models.