Predictive phenotyping of Alzheimer’s and Parkinson’s disease from multifactor biomarker and neuroimaging data

Individuals diagnosed with either Alzheimer’s disease (AD) or Parkinson’s disease (PD) typically exhibit a distinctive profile of progressive impairments. AD primarily affects cognitive functions (e.g., dementia), whereas PD affects motor functions (e.g., tremor). However, in many cases, this symptomatic distinction is not apparent, because some AD and PD patients experience a mixture of cognitive and motor functions of varying severities. These observations have called into question our current definitions of AD and PD as being two separate diseases. An alternative framework proposes that AD and PD may represent extreme phenotypes of a continuum where mixed AD/PD subtypes lie in the middle. To test this, we will obtain longitudinal neuroimaging, protein pathology and clinical test data of AD and PD patients from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and the Parkinson’s Progression Markers Initiative (PPMI), respectively. Then, we will use a novel data integration technique called similarity network fusion to combine all of the different data types, producing networks of patients who share common patterns in their data. We expect to see unique clusters of patients corresponding to the distinct disease subtypes.

Faculty Supervisor:

Taylor Schmitz;Marieke Mur

Student:

Partner:

Parkinson Society Southwestern Ontario

Discipline:

Life Sciences

Sector:

Other services (except public administration)

University:

The University of Western Ontario

Program: