Date of Award
Doctor of Philosophy (PhD)
Mohamed H. Al-Marzouqi
Dementia is a condition in which higher mental functions are disrupted. It currently affects an estimated 57 million people throughout the world. Dementia diagnosis is difficult since neither anatomical indicator nor functional testing are currently sufficiently sensitive or specific. There remains a long list of outstanding issues that must be addressed. First, multimodal diagnosis has yet to be introduced into the early stages of dementia screening. Second, there is no accurate instrument for predicting the progression of pre-dementia. Third, non-invasive testing cannot be used to provide differential diagnoses. By creating ML models of normal and accelerated brain aging, we intend to better understand brain development. The combined analysis of distinct imaging and functional modalities will improve diagnostics of accelerated decline with advanced data science techniques, which is the main objective of our study. Hypothetically, an association between brain structural changes and cognitive performance differs between normal and accelerated aging. We propose using brain MRI scans to estimate the cognitive status of the cognitively preserved examinee and develop a structure-function model with machine learning (ML). Accelerated aging is suspected when a scanned individual’s findings do not align with the usual paradigm. We calculate the deviation from the model of normal aging (DMNA) as the error of cognitive score prediction. Then the obtained data may be compared with the results of conducted cognitive tests. The greater the difference between the expected and observed values, the greater the risk of dementia. DMNA can discern between cognitively normal and mild cognitive impairment (MCI) patients. The model was proven to perform well in the MCI-versus-Alzheimer’s disease (AD) categorization. DMNA is a potential diagnostic marker of dementia and its types.
Svinterikos, Efstratios, "ACTIVATED CARBON NANOFIBERS FROM RENEWABLE (LIGNIN) AND WASTE RESOURCES (RECYCLED PET) AND THEIR ADSORPTION CAPACITY OF REFRACTORY SULFUR COMPOUNDS FROM FOSSIL FUELS" (2020). Dissertations. 159.