MRI Predictors of Cognitive Deficits in Multiple Sclerosis
Summary
Multiple Sclerosis (MS) is a chronic demyelinating autoimmune disease of the brain and spinal cord. MS is among the most prevalent causes of non-traumatic disability among young and middle-aged adults.1 While MS is primarily associated with motor disability, 40-65% of patients with MS also experience deficits of higher order cognition such as learning, attention, and planning.2 MS-related cognitive deficits frequently arise during the first year of diagnosis and may even precede motor deficits.3
For decades, neuroimaging has sought to relate MS lesion burden to clinical and cognitive disability. Cross-sectional studies have implicated both lesion number and lesion location (e.g. whether lesions exist focally in gray matter or interrupt white matter pathways) as predictors of clinical disability severity.4 However, lesion-based analyses are poor predictors of cognitive impairment, explaining <15% of variance in cognitive ability.5,6 We hypothesize that more sophisticated magnetic resonance imaging (MRI) approaches, such as functional MRI, will explain greater variance in cognitive disability than lesion analyses alone.
Our hypotheses are borne form the Cognitive Connectome Project, a KL2-funded initiative by Dr. Andrew James to map the neural basis of normative variance in cognition. We found that functional brain connectivity (patterns of co-occurring brain activity during wakeful rest) predicted individual differences in higher order cognitions (learning, attention, and planning) but not domain-specific cognitions (visual, motor, and language).7 This work, conducted within a healthy normative sample, was later replicated in stroke survivors; Siegel and colleagues found that stroke lesion location predicted deficits in domain-specific cognitions, but functional connectivity predicted deficits in higher order cognitions.8
The proposed project will build upon this past work to empirically evaluate gray matter lesions, white matter lesions, and functional connectivity as independent and cumulative predictors of cognitive deficits in MS. Specifically, ridge regression will relate gray matter lesion location, lesion load among white matter pathways, and bivariate functional connectivity to cognitive deficit severity. Neurobiology PhD candidate Ashley Pike will lead this project under the joint mentorship of Dr. James (Psychiatry; neuroimaging statistical analyses), Dr. Jenn Kleiner (Psychiatry; administration and interpretation of cognitive assessments), and Dr. Lee Archer (Neurology; MS neuropathology and clinical outcomes). Dr. Paul Drew (Neurobiology), as Ms. Pike’s graduate co-mentor, will provide additional training in MS neuropathology. The mentoring team will provide complementary training to help Ms. Pike achieve her career goals of becoming an independent MS neuroscience researcher. Of note, Drs. James and Kleiner previously collaborated in training UAMS PhD candidate Tonisha Kearney-Ramos PhD (now a K01-funded assistant professor at Columbia University) and in developing the Cognitive Connectome Project, but have not formally collaborated since 2016.
This project is enabled by UAMS’s commitment to upgrade the Brain Imaging Research Center MRI to a research-dedicated 3T Prisma MRI scanner, which will enable additional multimodal imaging functions including myelin imaging (via magnetization transfer imaging) and diffusion spectrum imaging (comprehensive mapping of white matter tracts). Although beyond the scope of Ms. Pike’s dissertation, these additional modalities will support future longitudinal collaborative projects for the investigative team.
Literature Cited: 1PMC6442006, 2PMC6715181, 3PMC6107340, 4PMID25662900, 5PMID23459568, 6PMID23468546, 7PMC5233432, 8PMC4968743
Keywords:
- multiple sclerosis
- neuroimaging
- cognition
- fMRI
- machine learning
- neurology