Current Projects
"We use MRI to catch the earliest problems in the brain that lead to Alzheimer’s, so we can start thinking about preventing it before it ever shows up."
Most people think Alzheimer’s starts with memory loss, but the brain actually shows changes decades earlier. We’re using MRI to see how blood vessels and the brain’s waste-clearance system start breaking down—and that could be the key to stopping the disease before it even begins. If we can spot those changes early, it means one day we could help people keep their brains healthy for decades longer—maybe even stop Alzheimer’s before it ever affects their lives.
Cerebrovascular disease in African Americans, Alzheimer’s Association and American Society for Neuroradiology, PI, 2025-2026
Cerebrovascular disease (CVD) contributes significantly to the development and progression of Alzheimer's disease (AD) by reducing cerebral blood flow, causing microhemorrhages, and leading to white matter lesions. While CVD risk factors such as hypertension, diabetes, and hyperlipidemia affect both African Americans and Non-Hispanic Whites, their impact on cognition appears to differ by race, likely due to social determinants of health (SDOH). Prior studies show that African Americans have a 64% higher rate of dementia progression and display unique CVD patterns, including stronger associations of hypertension with intracranial atherosclerosis and differing stroke presentations compared to Whites. Despite a higher stroke incidence, African Americans often experience more gradual symptom onset and better long-term outcomes, suggesting distinct underlying mechanisms. Current biomarker studies comparing white matter hyperintensity burden and hippocampal volume have not explained these racial disparities. No physiological biomarker yet exists to capture these mechanistic differences. This proposal addresses the gap by leveraging advanced MRI techniques to measure vascular reactivity, MRI-visible perivascular spaces, blood-brain barrier permeability, and Circle of Willis configuration. Importantly, it will employ a novel, contrast-free arterial spin labeling approach to assess blood-brain barrier permeability. This will be the first study to compare these comprehensive vascular measures between African Americans and Non-Hispanic Whites, aiming to clarify mechanisms linking CVD, SDOH, and AD risk.
Glymphatics imaging to probe sleep in Parkinson’s Disease, DoD, Site PI (Co-I, Main PI: Lim), 2024-2028
Rationale: Many U.S. Veterans suffer from neurotrauma, such as TBI and PTSD, which increases their risk of developing Parkinson's disease (PD). An early prodromal marker of PD is REM Sleep Behavior Disorder (RBD), which appears in a very high proportion of Veterans with TBI+PTSD compared to the general population.The glymphatic system, which clears proteins like amyloid, tau, and α-synuclein during sleep, becomes impaired with aging, neurotrauma, and sleep loss, and this dysfunction may link neurotrauma to PD pathology. Preclinical data from rodent models show that impaired glymphatic clearance accelerates α-synuclein aggregation, but the role of this system in Veterans with neurotrauma, RBD, and PD remains unknown. We hypothesize that neurotrauma- and sleep-related glymphatic impairment drives the development of PD-related pathology. To test this, we will conduct three synergistic studies: preclinical experiments in mouse models, advanced MRI imaging in Veterans, and post-mortem analyses of human brains to identify mechanisms and early predictors of PD. Our group at KUMC is involved in the processing of the advanced MRI data in Veterans and deriving measures of glymphatic clearance.
Glimpse TBI: Defining the Role of Post-TBI Sleep Disruption in the Development of CTE and Alzheimer's Disease-Related Neuropathology Rationale, NIH, Site PI (Co-I, Main PI Peskind, SIBCR, Puget Sound VA), 2022-2027
Rationale: Mild TBI (mTBI) is a risk factor for Alzheimer's disease (AD) and related tau pathologies, but the mechanisms remain unclear. Sleep-wake disruption (SWD), common after mTBI and in AD, may drive pathology, as shown by our finding that Veterans with mTBI and poor sleep have AD-like CSF biomarker profiles. We propose that post-TBI sleep disruption, glymphatic dysfunction, and altered noradrenergic signaling each contribute to AD pathology. Using a 10-year VA cohort of Veterans with blast mTBI, we will test: Aim 1) whether sleep disruption promotes AD pathology, Aim 2) whether glymphatic impairment predicts pathology, and Aim 3) whether noradrenergic tone changes are mechanistically linked to AD pathology. Our team will focus on acquisition and analyses of MR measurements of glymphatic impairment.
Augmented Neurophysiology of Sleep and Performance Readiness, MTEC, Site PI, 2021-2027
Sleep is essential for brain health, supporting learning, memory, repair, and waste clearance, but both acute and chronic sleep restriction, common in military operations, significantly impair cognitive and physical performance. The glymphatic system, which clears brain waste during deep slow-wave sleep, is thought to underlie the restorative effects of sleep, and its dysfunction may explain the detrimental impact of sleep loss on cognition. The Augmented Neural Oscillation Driver (AugNOD), a wearable EEG/TES headband, is designed to both monitor and enhance slow-wave sleep while directly improving glymphatic clearance through two complementary protocols. This project will test whether TES-enhanced slow-wave sleep improves glymphatic function and cognitive performance under sleep restriction, develop and refine the AugNOD device for operational usability, and validate its effects in both lab and at-home settings. In parallel, the effort will advance computational modeling of brain molecular transport and conduct contrast-enhanced MRI studies to define the dynamics of glymphatic clearance during sleep. Together, these studies aim to establish a novel device-based countermeasure to mitigate the cognitive impacts of sleep loss in military and civilian populations.
Pharmacological Modulation and Remote Measurement of Sleep-Driven Glymphatic Clearance, MTEC, Site PI (Co-I, PI: Carneiro K,UNC, Chapel Hill): 2023 - 2027
Acute and chronic sleep restriction impair near-term performance and readiness in military personnel while also increasing long-term risks for chronic disease, traumatic brain injury, and posttraumatic stress disorder. The glymphatic system, which clears metabolic waste during sleep, is thought to underlie the restorative effects of sleep on cognition, and its dysfunction may explain the cognitive decline associated with sleep loss. Detecting glymphatic impairment in real-world settings could help identify service members at risk and guide interventions to preserve safety, performance, and health. This study will test whether the pharmacological agent prazosin (PZN), which modulates central noradrenergic tone, can enhance glymphatic clearance in non-human primates (NHPs) and humans while also validating the AC Band, a wearable device for measuring glymphatic function. NHP studies at ONPRC will define dose-response effects of PZN using MRI and the AC Band, while a translational human study at Washington State University will examine whether PZN improves glymphatic function and cognition under sleep restriction and circadian disruption. Data from both arms will inform a computational model of intracranial fluid dynamics, aiming to establish glymphatic function as a target for countermeasures against the cognitive effects of sleep loss.
Pathological mechanisms of white matter hyperintensities, R01, PI, 2021-2026
Rationale: This project applies advanced in vivo imaging to determine the pathological mechanisms of white matter hyperintensities (WMHs) and their contribution to cognitive decline in older adults at risk of Alzheimer's disease (AD). WMHs are bright patches on T2- MRI but do not inform the pathology underlying their appearance. Our work is innovative because it utilizes, for the first time, non-invasive methods to quantify axon and myelin rarefaction, fluid retention, WMHs-associated cortical atrophy, and blood-brain barrier dysfunction as in vivo drivers of WMHs formation. We will also test typical pathological markers of WMHs, i.e., hypoperfusion and compromised vascular reactivity. All imaging measurements will be validated against well-established CSF markers. Our scientific premise is that different locations of WMHs (deep vs. periventricular) correspond to distinct mechanisms (vascular vs. neurodegenerative) and cognitive profiles. Thus, our work will allow the use pathological drivers of WMHs to develop targeted strategies to stop their growth and ameliorate associated cognitive decline in the future. In Aim 1, we will test the hypothesis that vascular pathology predominates in deep WMHs by measuring cortical cerebral blood flow, vascular reactivity, and blood-brain barrier dysfunction. We will use specialized arterial spin labeling and a hypercapnic functional MRI approach for these measurements. We will validate measurements with CSF markers of vascular injury (e-selectin), inflammation (adhesion molecules, VCAM/ICAM), and blood-brain barrier permeability (albumin extravasation). In Aim 2, we will test the hypothesis that neurodegenerative pathology predominates in periventricular WMHs by measuring axon rarefaction, demyelination, and fluid retention using advanced diffusion imaging and myelin water fraction imaging. Validation markers will be CSF levels of tau and myelin basic protein. In independent ex vivo samples of donor brain tissue, we will quantify vascular pathology using smooth muscle actin and albumin immunohistochemical stains in the two WMHs. We will also quantify axon and myelin density in the WMHs, and perform neuron counting in gray matter using an array of stains. Postmortem MRI will be used to identify WMHs on donor brains. The quantitative neuropathology will serve as an independent validation of our hypotheses regarding WMHs location and mechanisms. In Aim 3, we will test the hypothesis that deep WMHs interrupt discrete short-range association fibers and striatal fibers, causing specific cognitive deficits especially those related to processing speed. Periventricular WMHs, on the other hand, interrupt long-range association and inter-hemispheric commissural tracts causing global cognitive impairment. We will use mediation analysis to explain whether the two types of WMHs influence the relationship between the different imaging markers and distinct cognitive symptoms. The proposal takes advantage of recruitment, lumbar punctures, CSF analyses, postmortem MRI matched to ex vivo tissue sections, and systematic anatomic sampling of the ex vivo tissue sections, from the Cores of the University of Washington's Alzheimer's Disease Research Center, ensuring success of the PI's first independent R01.
Early detection of Parkinson's disease using MRI-based nigral atrophy - The Dolsen Family Fund, PI, 2020-2022
Rationale: The substantia nigra produces a neurotransmitter called dopamine. Research shows that at least 40% of the substantia nigra’s dopamine-producing cells must be lost before any symptoms of Parkinson’s appear — which is late in the disease process. We want to develop an affordable, innovative scan which could tell us, earlier on, of a person’s risk for Parkinson’s. It’s called a neuromelanin-sensitive MRI, a reference to neuromelanin, a byproduct of oxidized dopamine. Lower levels of neuromelanin could help us identify patients with lower levels of dopamine, allowing us to predict who might develop Parkinson’s.
Cerebrovascular imaging of mild cognitive impairment with suspected non-amyloid pathology - NIA, K01, PI, 2017-2022
Rationale: The overall goal of this proposal is to compare vascular markers between amnestic mild cognitive impairment (MCI) subjects with amyloid or Alzheimer’s disease (AD) pathology, and the newly identified MCI subjects with suspected non-amyloid pathology (SNAP-MCI). The innovation of this work is that we will explore vascular pathology and its role as a driver of cognitive deficits in this new cohort.
The premise of this study is as follows: Proposed mechanisms for disease progression in MCI+ (Panel A) and SNAP-MCI (Panel B) subjects. ‘*’ in the gray box indicate the unsolved questions targeted by this K01. MCI+ subjects experience a dual amyloid-toxic hit (direct, black arrow + indirect vascular, blue arrows) causing neurodegeneration and dementia. SNAP-MCI subjects are likely to show aggravated neurodegeneration due to neuronal injury and vascular insufficiency. Interestingly, both panel pathways result in amnestic cognitive impairments.
Next steps: Longitudinal follow-up of CBF and CVR changes in all normal subjects, amyloid positive MCI participants, and SNAP-MCI participants. Our methods are well established and have high reproducibility. Our preliminary data shows longitudinal follow-up of older adult controls after 6 months to determine stability of CBF measurements using arterial spin labeling. Dotted line indicates unit correlation. CBF measurements in the gray matter are highly reproducible with our protocol and in accordance with other studies.
Characterization of SNAP MCI using large harmonized datasets - Alzheimer's Association, GEENA-Q, PI, 2019-2020
Rationale: The overall goal of this proposal is to identify cortical correlates of suspected non-amyloid pathology (SNAP) in which individuals without amyloid accumulation exhibit cognitive impairment and progression similar to AD. This work leverages the GAAIN platform that homogenizes large datasets across the world allowing us to identify a large sample of individuals with SNAP. Outcome of this work can be leveraged to study converging physiological mechanisms underlying SNAP and MCI+ as well as distinct mechanisms in the two groups for better therapeutic outcomes.
Non-invasive Quantification of Glymphatic Flow in Alzheimer's Disease, Royalty Research Fund, PI, 2019-2020
Rationale: The goal of this project is to develop quantitative MRI biomarkers of impaired glymphatic flow in older adults for detecting early Alzheimers Disease (AD).
Preliminary work presented at ISMRM 2019.
This figure shows our first attempts to measure slow diffusion processes in the brain that may be indicative of glymphatic flow. These are result from an experiment performed in eight older adults before and after dose titration of Prazosin. Prazosin is a alpha-adrenergic antagonist an believe to increase interstitial fluid volume and glymphatic transport. The top row shows conventional diffusion among neurons and is less likely to be sensitive to lymphatic transport. The second row is the pseudo diffusion coefficient, which we believe captures a component of lymphatic transport. The bottom row shows the fraction of fluid in the brain (blood and interstitial) that exhibits pseudo diffusion. With prazosin, we see an increase in fluid fraction an dan increase in pseudodiffusion. Both measures however, are confounded by perfusion occurring in the tissue, making it harder to detect the exact contribution associated with glymphatc flow.
Imaging physiological mechanisms in healthy aging and dementia (PI: Tom Grabowski, The ACT Study, UW ADRC)
Rationale: To determine alterations in neuronal function in older adults with and without cognitive decline using functional imaging approaches.
Arterial Spin labeling (ASL): One way to quantify baseline brain function is by measuring perfusion or cerebral blood flow (CBF), which serves as a surrogate for cerebral metabolism. This image shows averaged quantitative CBF map obtained using a pCASL sequence in older adult subjects.
Breath-hold BOLD MRI: We use breath-holds to quantify cerebrovascular reserve (CVR = dilatatory ability of microvessels) in older adults. Typically, CVR is decreased in neurodegenerative diseases. CVR changes during breathhold are much less than those observed with gas challenges but are more feasible in older adults. To overcome the low signal, we use BOLD instead of ASL. Adjacent is a CVR map of 28 older adults. (2<Z<7).
Multi-modal noninvasive vascular imaging of early Alzheimer's disease, Charleston Conferences on Alzheimer's Disease (Role: PI)
Rationale: To compare relationships among markers of microvascular health (cerebral blood flow; CBF), cerebral blood volume; CBV) and macrovascular health (vessel wall thickening) in older adults with and without mild cognitive impairment (MCI), and subsequently assess their utility as novel imaging markers of risk in Alzheimer’s disease (AD).
Outcome: We developed a whole brain inflow Vascular Space Occupancy (iVASO) sequence to measure CBV without using any exogenous contrast. Below is the video of the presentation showcasing our approach. We then assessed the relationship between CBF and CBV in early AD i.e., MCI individuals. We found that for the same increase in CBV, the increase in CBF in MCI individuals was significantly lower in the hippocampus as well as the cortex. Furthermore, precuneal hypoperfusion was associated with poorer cognitive performance. Although rare in incidence, the number vascular wall abnormalities were also associated with reduced cortical perfusion and lower scores on delayed recall tests.
Biomarkers Across Neurodegenerative Diseases (PI) - Michael J Fox Foundation, Alzheimer's Association, and W. Garfield Weston Foundation. (Role: PI)
Rationale: The overall goal of this work is to integrate MRI data from ADNI and PPMI cohorts to test fundamental hypotheses about Alzheimer’s disease (AD) and Parkinson’s disease (PD) pathology. To accomplish this, we will apply advanced computational methods developed at the Vanderbilt University Institute of Imaging Science (VUIIS) to compare tissue atrophy and functional deficits with alpha synuclein (α-SYN) concentration and APOE status in patients with early-stage AD and PD.
Outcome: Our study found that T1 MRI datasets from ADNI and PPMI that are acquired on the same scanner manufacturer can be combined to increase statistical power even if they have slightly different acquisition protocols. The advantage of this result is that statistical power for future studies for understanding aging, as well as disease mechanisms in Alzheimer’s (AD) and Parkinson’s disease (PD) can be increased by including data from the two repositories. Additional data gathering may not be necessary. Leveraging this finding, we found that both AD and PD patients with poor memory function showed significant cortical thinning in the parahippocampal regions of the brain. These regions also showed significant hippocampal and amygdala atrophy with age, when compared to cognitively normal older adults (Figure below). Furthermore, the inferior parietal region was markedly thinner in APOE- ε4 carriers (a genetic risk factor implicated in both diseases as well as in memory function) in both AD and PD. Thus, we believe that we identified brain regions that are commonly affected in AD and PD populations with memory impairment. In addition, the amnestic AD patients showed additional thinning in the superior frontal and the precuneual regions of the brain. They also had significantly smaller hippocampal and amygdala volumes compared to the normal older adults and the amnestic PD patients. No such PD-specific brain regions were detected in this study.
We also evaluate the concordance between subcortical measurements by different software packages. In 52 individuals, we evaluated volumes of the hippocampus, amygdala, thalamus, caudate, putamen, pallidum, accumbens using FreeSurfer v5.0, FIRST (FSL 5.0), and Multi-Atlas segmentation. While large structures (< 3 cc) with clear white-gray matter boundaries, there was a strong correlation between measurements by the three software packages, measurements in smaller structures (< 3 cc) were quite disparate (Adjacent figure).
The table at the end shows intra-class correlation coefficients showing agreement between different software packages.
| ICC [95%] |
Thalamus | Caudate | Putamen | Pallidum | Hippocampus | Amygdala | Accumbens |
|---|---|---|---|---|---|---|---|
| FIRST & Multi-Atlas | 0.75 [0.61,0.85] |
0.85 [0.76, 0.91] |
0.78 [0.65,0.87] |
0.64 [0.45,0.77] |
0.78 [0.66,0.87] |
0.06 [-0.21,0.32] |
0.83 [0.72,0.90] |
| FreeSurfer & Multi-Atlas | 0.69 [0.52,0.82] |
0.86 [0.78,0.92] |
0.77 [0.63,0.86] |
0.74 [0.59,0.84] |
0.79 [0.67,0.87] |
0.39 [0.14,0.60] |
0.45 [0.21,0.64] |
| FIRST & FreeSurfer | 0.83 [0.73,0.89] |
0.85 [0.76,0.91] |
0.71 [0.55, 0.82] |
0.54 [0.32,0.70] |
0.82 [0.70,0.89] |
-0.17 [0.14,0.10] |
0.44 [0.20,0.73] |
The role of dopamine in Parkinson's Disease - The Dolsen Family Fund, Role: PI, Pacific Udall Center (P2 PI: Tom Grabowski)
Rationale: To understand the role of dopamine and its interaction with cerebral perfusion in Parkinson's Disease
Outcome: We applied arterial spin labeling (ASL) MRI for this purpose. Uniquely, we assessed this pattern separately in PD individuals ON and OFF dopamine medications. We further compared the existence of these patterns and their strength in each individual with their Movement Disorder Society-Unified Parkinson's Disease Rating Scale motor (MDS-UPDRS) scores, cholinergic tone as indexed by short-term afferent inhibition (SAI), and other neuropsychiatric tests. Results. We observed a PD-related perfusion pattern that was similar to previous studies. The patterns were observed in both ON and OFF states but only the pattern in the OFF condition could significantly (AUC=0.72) differentiate between PD and healthy subjects.
In the ON condition, PD subjects were similar to controls from a CBF standpoint (AUC=0.45). The OFF pattern prominently included the posterior cingulate, precentral region, precuneus, and the subcallosal cortex. Individual principal components from the ON and OFF states were strongly associated with MDS-UPDRS scores, SAI amplitude and latency. Conclusion. Using ASL, our study identified patterns of abnormal perfusion in PD and were associated with disease symptoms. (Work performed in conjunction with other projects).
Non-invasive MR assessment in normal pressure hydrocephalus - UW-Philips Grant (PI: Jalal Andre), Co-I, 2018-2019
Rationale: The pathophysiology of normal pressure hydrocephalus (NPH) is poorly understood. Normally, CSF flows out of the cerebral tissue via the ventricles. However, in the setting of NPH, studies have reported a “reversal” of net flow through the aqueduct of Sylvius, resulting in retrograde flow into the ventricles. This reversal of flow may also be seen following ventriculoperitoneal (VP) shunt catheter placement. However, clinically significant hydrocephalus can be difficult to distinguish from general cerebral atrophy, yet the idiopathic variant has a reported prevalence of 5.9% in persons aged 80 and older. In the absence of specific imaging or clinical symptoms, the idiopathic variant of NPH can be suspected in the setting of the classical “clinical triad” (gait, urinary incontinence, and cognitive impairment), but the diagnosis often relies on the relatively subjective judgment of ventriculomegally out of proportion to cortical cerebral atrophy. Intracranial compliance (ICC) is typically reduced as hydrocephalus develops, such that traditional cine MR imaging measurements show only small increases in ventricular wall displacement during hydrocephalus. Because there is no robust and clinically proven non-invasive method to assess ICC and intracranial pressure (ICP), existing imaging does not discriminate whether patients would benefit from surgery such as VP shunt placement. Since up to 30% of patients do not benefit from VP shunt placement, the inability to determine benefit leads to a “trial and error” approach, requiring multiple surgeries and incurring additional patient risk as well as increased health care costs. The development of a non-invasive, clinically available MR method for assessment of ICC and ICP would be extremely beneficial.