This work is all based on our Interlaboratory Proposa Patterns of Myelination in Mild Cognitive Impairment, initially funded in 2015 for 2016, with funding renewed for the following year, and our funded 2017 Interlaboratory proposal, entitled: Establishing the Relationship Between Myelination Patterns and Regional Cerebral Blood Flow in Normative Aging, Mild Cognitive Impairment, and Dementia Mild cognitive impairment (MCI) is characterized by a progressive decline in cognitive abilities, including memory, language, and judgment. MCI increases the risk of progression to frank dementia, including Alzheimer's disease (AD). Although the majority of cases of MCI may be due to underlying AD pathology, MCI remains a heterogeneous condition. The development of non-invasive markers for MCI due to AD and, in particular, early pre-symptomatic stages of AD would provide important prognostic information and an opportunity for developing and evaluating targeted therapies. Moreover, such markers would provide important mechanistic information on the brain changes underlying impairment in memory and other cognitive functions. Conventional magnetic resonance imaging (MRI) using transverse relaxation time (T2) and magnetization transfer (MT) have been extensively applied in the evaluation of lesions, including therapeutic response, in multiple sclerosis (MS). Increases in T2 have been interpreted as indicating localized or diffuse edema, inflammation, or demyelination, likely in combination. However, while changes of T2 and MT are often attributed to myelin alteration, they do not directly measure myelin and are influenced by other microstructural changes. More advanced methods of myelin assessment are based on the multicomponent decomposition of the transverse (T2) decay signal. MWF correlates strongly with myelin stain, validating MWF as a measure of myelin density. However, these MR MWF analyses cannot be performed on the whole brain within clinically reasonable acquisition times due to the ill-posed nature of the signal analysis required to estimate MWF. Nevertheless, it is clear that quantification of MWF using quantitative MRI has tremendous potential for in vivo monitoring of demyelination and potential remyelination in response to therapies in MCI and, by extension, AD. We have studied the problem of multi-component relaxation extensively in our laboratory and have made several advances in this analysis. Much of our work has centered on cartilage, but we have successfully transferred this work to brain. An important recent extension of this has been the application of sophisticated Bayesian methods for multi-parameter data analysis using multicomponent driven equilibrium single pulse observation of T1 and T2 (mcDESPOT). This method permits whole-brain analysis at the expense of requiring analysis of a high-dimensional signal model, that is, model with several unknown parameters. However, we have specifically developed and applied Bayesian analysis methods to the mcDESPOT experiment, permitting whole-brain analysis with high-quality parameter estimates with acquisition times of only 30 - 45 minutes. Further decreases in acquisition time are possible with more advanced hardware. We anticipate that the currently proposed study may lead to a much larger initiative in understanding the role of myelin trajectory in cognitive function, including dementia. As one important direction, these studies may help distinguish cognitively normal A positive individuals who will ultimately develop cognitive impairment from those who will maintain cognitive health, i.e. remain resilient. Indeed, MWF changes even in pre-symptomatic stages of the pathway to cognitive impairment may serve as important biomarkers for development of clinical symptoms and has the potential to greatly facilitate development of early therapeutic interventions. We have evaluated the performance of our analytic technique applied to mcDESPOT over wide ranges of SNR and underlying input parameter values, and experimentally from MR imaging data of human brain obtained through use of different protocols reflectingdifferent noise levels. The accuracy and precision in the estimation of MWF obtained with our Bayesian Monte Carlo analysis showed substantial improvements over the currently used nonlinear least-squares methods. All volunteers undergo the following MRI protocol on the 3T Philips MRI system at Harbor Hospital: - MWF mapping using BMC-mcDESPOT analysis: spoiled recalled gradient echo (SPGR) and balanced steady-state free precession (bSSFP) brain datasets obtained at eight different flip angles and a TR of 8 ms will be acquired. In order to correct for off-resonance artifacts, bSSFP datasets will be acquired with two different phase increments of 00 and 180o. Inversion-recovery SPGR (IR-SPGR) images will also be acquired to be combined with the SPGR dataset in order to correct for flip angle heterogeneity. The total acquisition time is estimated to be approximately 30 minutes for whole brain coverage. Following acquisition, 3D MWF maps will be generated using the BMC-mcDESPOT analysis as indicated above. These methods have all been developed in Dr. Spencer's group. - Multiexponential relaxation mapping: Multi-spin-echo imaging will be used to acquire T2-weighted images at 32 echo-times increasing linearly from 9 to 288 ms, with TR fixed to 2000 ms. Through our published Bayesian analysis and conventional nonnegative least squares, MWF and transverse relaxation times maps will be generated. Results will be compared to those derived through the BMC-mcDESPOT analysis. Given the lengthy acquisition time required for this scan, only localized slices will be obtained instead of whole brain coverage. Additional conventional, non-specific, measures of myelination will also be applied: - Magnetization transfer (MT) mapping: two spin echo images will be obtained, with and without on-resonance saturating pre-pulses. MT maps will be generated through the calculated ratio between those two images. Given the long acquisition time required for this scan, only localized slices will be obtained instead of whole brain coverage. - Apparent diffusion coefficient (ADC) mapping: a single-shot spin-echo echo planar imaging sequence will be used to acquire up to three diffusion-weighted images at diffusion-sensitizing b-values of 0, 500 and 1000 s/mm2, with TR fixed to 2500 ms. Given the long acquisition time required for this scan, only localized slices will be obtained instead of whole brain coverage. EXPECTED OUTCOMES 1) We expect to obtain high-quality myelin water fraction maps from all subjects. While we have demonstrated this methodology in published work, the current proposal will provide the opportunity to test the approach in individuals across a wide range of age and cognition status. 2) We expect to obtain results that will correspond qualitatively to the conventional methods currently employed for white matter assessment, although with greatly improved speed, accuracy, and whole-brain coverage. 3) With age some myelin sheaths exhibit degenerative changes. We expect to find differences in myelin content between healthy young subjects and healthy old subjects in distinct brain regions. 4) We expect to find differences in myelin content between old subjects without cognitive impairment and old subjects with MCI in distinct brain regions. This is our principal hypothesis as detailed above. 5) We expect to determine quantitative associations between MWF and degree of cognitive impairment. Overall, our interest is in the relationship of myelination to MCI and preclinical AD, and the relationship between local cerebral blood flow and myelin deficits.
