Graduation Semester and Year
Spring 2026
Language
English
Document Type
Dissertation
Degree Name
Doctor of Philosophy in Biomedical Engineering
Department
Bioengineering
First Advisor
Dr Hanli Liu
Abstract
Alzheimer's disease (AD) is the leading cause of dementia, and existing diagnostic methods such as PET scans, cerebrospinal fluid sampling and biomarker quantification, and gene sequencing are all either invasive, costly, or not sensitive enough for early detection. This dissertation introduces three different studies that develop a novel multimodal, non-invasive approach to diagnosing AD at its early stages by combining broad band near infrared spectroscopy (bbNIRS) and electroencephalography (EEG) technologies.
The first study showed cerebrovascular-cerebrospinal fluid coupling (CBV-CSF), which is measured by using 2-channel bbNIRS as an indicator of brain aging and early AD. Linear correlations between total blood (Δ[HbT]) and cerebrospinal fluid also be called as free water content in brain (Δ[H2Ofree]) were quantified through measurements carried out for seven minutes on three different groups, including healthy young individuals, healthy older participants, and those suffering from AD in three infraslow oscillatory frequency bands: endogenic (E; 0.005-0.02 Hz), neurogenic (N; 0.02-0.04 Hz), and myogenic (M; 0.04-0.1 Hz). CBV-CSF coupling was seen to increase in AD patients, and this is attributed to cerebrovascular stiffness with increasing age and AD.
The second study applied this framework to the assessment of how transcranial photobiomdulation (tPBM) modulates cerebrovascular-CSF and metabolic-CSF interactions in 27 young adults subjected to tPBM at 800 nm and 850 nm. After tPBM, there was an increase in cerebrovascular-CSF coupling as well as a decrease in metabolic-CSF coupling. These findings demonstrate a hierarchical coupling shift following tPBM, whereby metabolic activation effectively drives cerebral perfusion, and perfusion-induced volume compensation promotes CSF redistribution, enhancing CSF outflow and circulation. Consequently, CSF dynamics become less connected with metabolic fluctuations while vascular-CSF integration is enhanced.
The third study proposed an interpretable machine learning model for AD classification based on EEG, using spectral and temporal features from 17 normal older subjects and 15 AD subjects over five different frequency bands. The Random Forest model in combination with Segment-First method achieved a classification accuracy of 97.08%, whereas the accuracy for Split-First method was 84.2%. Complementing the study, this was followed by another longitudinal study of one AD subject undergoing tpbm treatments for eight months via a combination of system involving photobiomodulation treatment at Massachusetts General Hospital (MGH) (800 nm; 35-105 J/cm²) in office and a commercial tpbm device called iMedisync for neurostimulation at home (1-40 Hz). Electrophysiological changes such as decreased frontal theta power and increased posterior alpha power reflected the therapeutic course in AD.
All of these findings confirm that bbNIRS and EEG are feasible methods that can capture the vascular, metabolic, and neural biomarkers of AD, while tPBM may have the potential to induce neurophysiological changes in individuals with AD.
Keywords
Alzheimer's Disease, EEG, bbNIRS, CSF, Transcranial photobiomodulation, Machine learning, PSD, ANN, RF, Power ratios
Disciplines
Biomedical Engineering and Bioengineering
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Recommended Citation
Saeed, Fiza, "Development of Multimodal Measurements and Analysis for Early Detection of Alzheimer’s Disease" (2026). Bioengineering Dissertations. 1.
https://mavmatrix.uta.edu/bioengineering_dissertations2/1