Investigation of Cerebrovascular Reactivity Using Hypercapnia and Optical Brain Imaging
Wednesday, June 7, 2023
10:00 AM-12:00 PM
BIOMED PhD Thesis Defense
Title:
Investigation of Cerebrovascular Reactivity Using Hypercapnia and Optical Brain Imaging
Speaker:
Pratusha Reddy, PhD Candidate
School of Biomedical Engineering, Science and Health Systems
Drexel University
Advisors:
Kurtulus Izzetoglu, PhD
Associate Professor
School of Biomedical Engineering, Science and Health Systems
Drexel University
Ramon R. Diaz-Arrastia, MD, PhD
Director of Traumatic Brain Injury (TBI)
Clinical Research Center
Associate Director for Clinical Research
Center for Neurodegeneration and Repair
John McCrea Dickson MD Presidential Professor
Department of Neurology
University of Pennsylvania Perelman School of Medicine
Details:
Damage to the cerebral microvasculature network that regulates cerebral blood flow (CBF) is a universal feature of aging and many neurological disorders including Alzheimer’s disease, small vessel disease, stroke, traumatic brain injury (TBI) and others. One way to quantify this damage is to assess cerebrovascular reactivity (CVR), which is the ability of cerebral microvasculature to increase or decrease CBF in response to a vasoactive stimulus, such as hypercapnia, i.e., increasing partial arterial carbon dioxide (PaCO2) levels. In comparison to CBF, CVR offers increased sensitivity to disease progression and greater precision in quantifying effects of therapeutic intervention aimed at improving cerebrovascular function. Despite these advantages, CVR has not been implemented into the routine clinical settings yet due to the prohibitive cost of existing techniques and the lack of practical methods and sensors.
Functional near infrared spectroscopy (fNIRS) is a noninvasive and wearable neuroimaging modality that offers the potential to assess CVR in routine clinical settings. Few studies have explored the utility of fNIRS to measure hemodynamic changes during hypercapnic stimulus and assess CVR in healthy and TBI patients. However, large variability in fNIRS-derived CVR measures was observed across these studies. This variability has been proposed to be due to inadequate detection and removal of fNIRS signal components arising from extracerebral tissue layers and those associated with systemic factors such as Mayer waves, which share similar time and frequency characteristics as the signal of interest, PaCO2. Therefore, the primary goal of this thesis is to investigate signal components confounding fNIRS measures and evaluate fNIRS measure’s ability to assess CVR in healthy adults and TBI patients.
To achieve this goal, my first specific aim focused on investigating time, frequency, and spatial characteristics of fNIRS measures from healthy adults during resting, cognitive and hypercapnic conditions. Successful completion of this aim improved our understanding of how signal components confounding fNIRS measures vary with stimulus type and provided quantitative evidence of the source of variability observed in CVR measures. Additionally, the results from this aim identified that the frequency band of 0.009 Hz to 0.06 Hz was associated with PaCO2 related activity and indicated that de-oxygenated hemoglobin, as assessed by fNIRS measures, was the most sensitive to PaCO2 related changes in the cerebral layer. The second specific aim focused on investigating differences in time, frequency, and spatial characteristics of fNIRS measures between healthy controls and TBI patients. This aim verified that the frequency band (0.005 to 0.1 Hz), commonly used for CVR quantification, is affected by PaCO2 and Mayer wave related activity. More importantly, it demonstrated that these physiological mechanisms are differently affected in TBI patients compared to healthy controls. Hence, these results posit that proper assessment of CVR using fNIRS requires separation of effects arising from PaCO2, and Mayer wave related activity. The final specific aim of this thesis was to develop a novel signal-processing methodology to extract PaCO2-related effects from cerebral layers and compare the resulting fNIRS-derived CVR measures against those derived from commonly used signal-processing methodologies. The results confirm that the current analysis methodologies result in highly variable CVR measures and verifies that the proposed signal-processing methodology led to more valid and reliable fNIRS-derived CVR measures.
In summary, this thesis demonstrated that fNIRS is an effective and viable modality to measure CVR during hypercapnic stimulus in healthy controls and TBI patients. The approaches, i.e., combination of features, methods, and techniques, developed in this thesis can provide increased monitoring capability of microvasculature dysfunction in various clinical settings and conditions, including TBI and other neurological disorders such as Alzheimer’s, small vessel disease, and stroke.
Contact Information
Natalia Broz
njb33@drexel.edu