About

Michael I. Miller is the Bessie Darling Massey Professor and Director of Biomedical Engineering at Johns Hopkins University, as well as co-director of the Kavli Neuroscience Discovery Institute. He specializes in data science, computational neuroscience, medical imaging, computational anatomy, and pattern theory. His research focuses on understanding and diagnosing neurodegenerative diseases by analyzing the functional and structural characteristics of the human brain in health and disease, including conditions such as Huntington’s disease, Alzheimer’s disease, dementia, bipolar disorder, schizophrenia, and epilepsy. Miller develops new tools to analyze patient brain scans derived from advanced medical imaging technologies, aiming to predict the risk of neurological disorders years before clinical symptoms appear. His lab is working on cloud-based methods to build and share libraries of brain images and algorithms related to neuropsychiatric illnesses. He has co-founded four start-up companies and has authored over 200 peer-reviewed publications and two highly cited textbooks on random point processes and computational anatomy. Miller earned his BS from the State University of New York at Stony Brook and his MS and PhD from Johns Hopkins University. He has held faculty positions at Washington University in St. Louis before joining Johns Hopkins in 1998, where he was named the Herschel and Ruth Seder Professor in Biomedical Engineering in 2003 and the director of biomedical engineering in 2017.

Research topics

• Neuroscience
• Computer Science
• Sociology
• Political Science
• Medicine
• Psychology
• Engineering
• Materials science
• Pathology
• Engineering ethics
• Internal medicine
• Biology
• Nanotechnology
• Pedagogy

Selected publications

• Alzheimer’s disease cerebrospinal fluid biomarker levels and APOE genetic status are associated with hippocampal-cerebellar functional connectivity

  Neurobiology of Aging · 2025-04-18 · 5 citations

  articleOpen access

  Recent research suggests that hippocampal-cerebellar (Hp-CB) functional connectivity may be altered early in the course of Alzheimer’s disease (AD), given the early accumulation of AD pathology in the hippocampi and emerging evidence of cerebellar changes in early AD. This study analyzed the role of AD genetic risk (via APOE ε4 carrier status) and cerebrospinal fluid (CSF) biomarkers of AD pathology (ratio of phosphorylated tau (p-tau 181 ) to amyloid beta (Aβ 42 /Aβ 40 )) on the relationship between age and functional Hp-CB resting state fMRI connectivity in 161 cognitively unimpaired older adults ( M age =67.3; SD =9.0; 37 % APOE ε4 +). In multiple regression analyses with Hp-CB connectivity as the outcome, there were significant interactions between age and APOE ε4 status, and between age and CSF AD biomarkers. Older age was associated with greater Hp-CB connectivity in APOE ε4 non-carriers and participants with less abnormal CSF AD biomarkers. In contrast, Hp-CB connectivity was marginally lower with older age in ε4 carriers and those with more abnormal AD biomarkers. Furthermore, greater Hp-CB connectivity was associated with better episodic memory performance across all groups. These findings suggest that age-related increases in Hp-CB connectivity among APOE ε4 non-carriers and those with low AD biomarker levels reflect age-related changes that are largely unrelated to AD, while age-related decreases in Hp-CB connectivity in APOE ε4 carriers may reflect AD-related alterations. These findings also highlight the importance of cerebellar contributions to cognitive performance among older adults and suggest that Hp-CB connectivity may be altered in preclinical AD. • Examined hippocampal-cerebellar (Hp-CB) functional connectivity and AD risk. • APOE ε4 and abnormal AD markers associated with lower age-related Hp-CB connectivity. • Greater capacity for age-related Hp-CB recruitment with lower AD risk. • Greater Hp-CB connectivity associated with better episodic memory performance. • Greater age-related Hp-CB connectivity may be compensatory.

  PublisherDOI

• Single cell multiomics reveals drivers of metabolic dysfunction-associated steatohepatitis

  medRxiv · 2025-05-11 · 2 citations

  preprintOpen access

  Metabolic dysfunction-associated steatotic liver disease (MASLD) has limited treatments, and cell type-specific regulatory networks driving MASLD represent therapeutic avenues. We assayed five transcriptomic and epigenomic modalities in 2.4M cells from 86 livers across MASLD stages. Integrating modalities increased annotation of the genome in liver cell types several-fold over previous catalogs. We identified cell type regulatory networks of MASLD progression, including distinct hepatocyte networks driving MASL and mild and severe fibrosis MASH. Our single cell atlas annotated 88% of MASH-associated loci, including a third affecting hepatocyte regulation which we linked to distal target genes. Finally, we characterized hepatocyte heterogeneity, including MASH-enriched populations with altered repression, localization, and signaling. Overall, our results provide high-resolution maps of liver cell types and revealed novel targets for anti-MASH therapy.

  PublisherOA PDFDOI

• Cerebellar White Matter Microstructure Is Associated With Age, Cerebrospinal Fluid Amyloid Beta Levels, and Cognition in Cognitively Unimpaired Older Adults

  Human Brain Mapping · 2025-11-01 · 1 citations

  articleOpen access

  ABSTRACT Structural changes in the cerebellum contribute to cognitive decline due to aging and Alzheimer's disease (AD). However, it is unclear whether age and AD pathology are associated with structural alterations in the cerebellum among cognitively unimpaired individuals and how these alterations relate to cognition. This study examined the association of age and cerebrospinal fluid (CSF) AD biomarkers (amyloid beta [Aβ 42 /Aβ 40 ], phosphorylated tau [p‐tau 181 ]) with cerebellar gray matter (GM) and white matter (WM) volumes and cerebellar WM microstructure, measured via magnetic resonance imaging (MRI) among 176 cognitively unimpaired middle‐aged and older adults (mean age = 66.70, range = 34–89). Cognition was measured with executive function and visuospatial composite scores. Older age was associated with lower cerebellar GM and WM volumes ( p s < 0.01) and greater mean diffusivity (MD) in the cerebellar peduncles ( p < 0.01). In contrast, more abnormal Aβ levels were associated with lower MD in three regions of interest, including the middle cerebellar peduncle (MCP, p < 0.01), a composite of superior, middle, and inferior peduncles ( p < 0.05), and within‐cerebellar WM ( p < 0.05). Patterns were similar when comparing biomarker positive versus negative groups, particularly for the MCP. Further, lower MD in the peduncles and cerebellar WM was associated with better executive function and visuospatial composite scores ( p s < 0.05), whereas cerebellar volumetric measures were not related to cognition. Results suggest that older age is associated with microstructural and volumetric cerebellar GM and WM alterations. In contrast, Aβ levels are associated with WM microstructural properties in cognitively unimpaired individuals. These findings highlight the importance of cerebellar WM microstructure to cognition and are consistent with, and expand on, previous reports that have linked more abnormal amyloid levels to WM microstructure in cerebral tracts. They also suggest that cerebellar WM alterations may be markers of preclinical AD.

  PublisherOA PDFDOI

• Quantification of perforant path fibers for early detection of Alzheimer's disease

  Alzheimer s & Dementia · 2025-04-01 · 11 citations

  articleOpen access

  INTRODUCTION: The entorhinal cortex (ERC) and perforant path (PP) fibers are critical structures in the pathology of Alzheimer's disease (AD). This study aims to explore these regions using high-field magnetic resonance imaging (MRI), with the goal of identifying reliable biomarkers based on histopathological observations. METHODS: Twenty post mortem brain specimens were scanned with 11.7T MRI, including diffusion tensor imaging and tractography, and were cut for subsequent histological examinations. The entorhinal cortical thickness and number of PP fibers derived from MRI were compared across neuropathological and premortem clinical diagnoses of AD. RESULTS: The entorhinal cortical thickness and number of PP fibers decreased along with severities of neurofibrillary tangles in the ERC. Meanwhile, a reduction in the number of PP fibers, but not the entorhinal cortical thickness, was observed during the preclinical stage of AD. CONCLUSIONS: Degeneration of PP fibers was observed in early AD and progressed along with neuropathological changes. HIGHLIGHTS: Twenty post mortem brain tissues were scanned with 11.7T MRI. Degeneration of PP fibers was observed at 250 µm isotropic resolution. PP fiber indices were linked with severities of NFTs. The number of PP fibers was decreased in preclinical AD.

  PublisherOA PDFDOI

• Solving the where problem and quantifying geometric variation in neuroanatomy using generative diffeomorphic mapping

  Nature Communications · 2025-11-24

  articleOpen access

  A current focus in neuroscience is to map neuronal cell types in whole vertebrate brains using different imaging modalities. Mapping modern molecular and anatomical datasets into a common atlas includes challenges that existing workflows do not adequately address: multimodal signals, missing data or non reference signals, and quantification of individual variation. Our solution implements a generative model describing the likelihood of data given a sequence of transforms of an atlas, and a maximum a posteriori estimation framework. Our approach allows composition of mappings across chains of datasets rather than only pairs, and computes metrics for geometric quantification. We study a range of datasets (in/ex-vivo MRI, STP and fMOST, 2D serial histology, snRNAseq prepared tissue), quantifying cell density and geometric fluctuations across covariates, and reveal that individual variation is often greater than differences due to tissue processing techniques. We provide open source code, dataset standards, and a web interface. This establishes a quantitative workflow for unifying multi-modal whole-brain images in an atlas framework, validated using mouse datasets, enabling large scale integration of datasets essential to modern neuroscience.

  PublisherOA PDFDOI

• Change points for dynamic biomarkers in the Alzheimer's disease pathological cascade: A 30‐year cohort study

  Alzheimer s & Dementia · 2025-07-31 · 8 citations

  articleOpen access

  INTRODUCTION: This study aimed to delineate trajectories of biomarkers-amyloid beta (Aβ), phosphorylated tau, neurodegeneration, and inflammation-and to identify change points along these trajectories. METHODS: Longitudinal data were collected over 30 years from 349 cognitively unimpaired individuals enrolled in the Biomarkers for Older Controls at Risk for Dementia study. Piecewise regression models were used to identify change points in cerebrospinal fluid biomarkers, brain magnetic resonance imaging volumes, and a composite measure of global cognition. RESULTS: Eighty-two participants progressed to mild cognitive impairment or dementia during the follow-up period. Change points were identified in years prior to clinical symptom onset: Aβ at -17.1 years, phosphorylated tau at -15.8 years, neurofilament light chain and whole-brain white matter at -11.6 years, and total ventricles at -9.7 years. CONCLUSION: These findings support the temporal sequence proposed by the dynamic biomarker model of Alzheimer's disease and underscore the significance of white matter degeneration as an early marker for disease progression in the pathological cascade. HIGHLIGHTS: Data were collected from 349 participants in Biomarkers for Older Controls at Risk for Dementia, a 30-year cohort study. Eighty-two participants progressed to mild cognitive impairment or dementia over an average of 11 to 12 years. Core cerebrospinal fluid biomarkers for Alzheimer's disease began to accelerate 15 to 20 years before clinical onset. Magnetic resonance imaging volumes accelerated with variations across brain structures. Volumetric changes in the white matter and ventricles preceded the hippocampus.

  PublisherOA PDFDOI

• The xIV-LDDMM toolkit of image-varifold based technologies for mapping 3D images and spatial-omics across scales

  Communications Biology · 2025-09-30

  articleOpen accessSenior author

  Advancements in imaging and molecular techniques enable the collection of subcellular-scale data. Diversity in measured features, resolution, and physical scope of capture across technologies and experimental protocols pose numerous challenges to integrating data with reference coordinate systems and across scales. This paper describes a collection of technologies that we have developed for mapping data across scales and modalities, such as genes to tissues, specifically in a 3D setting. Our collection of technologies include (i) an explicit censored data representation for the partial matching problem mapping whole brains to subsampled subvolumes, (ii) a multi, scale-space optimization technology for generating resampling grids optimized to represent spatial geometry at fixed complexities, and (iii) mutual-information based functional feature selection. We integrate these technologies with our cross-modality mapping algorithm through the use of image-varifold measure norms to represent universally data across scales and imaging modalities. Collectively, these methods afford efficient representations of peta-scale imagery providing the algorithms for mapping from the nano to millimeter scales, which we term cross-modality image-varifold LDDMM (xIV-LDDMM).

  PublisherOA PDFDOI

• Predicting future cognitive impairment in preclinical Alzheimer’s disease using amyloid PET and MRI: a multisite machine learning study

  medRxiv · 2025-10-17

  preprintOpen access

  Abstract Predicting the likelihood of developing Alzheimer’s disease (AD) dementia in at-risk individuals is important for the design of and optimal recruitment for clinical trials of disease-modifying therapies. Machine learning (ML) has been shown to excel in this task; however, there remains a lack of models developed specifically for the preclinical AD population, who display early signs of abnormal brain amyloidosis but remain cognitively unimpaired. Here, we trained and evaluated ML classifiers to predict whether individuals with preclinical AD will progress to mild cognitive impairment or dementia within multiple fixed time windows, ranging from one to five years. Models were trained on regional imaging features extracted from amyloid positron emission tomography and magnetic resonance imaging pooled across seven independent sites and from two amyloid radiotracers ([ 18 F]-florbetapir and [ 11 C]-Pittsburgh-compound-B). Out-of-sample generalizability was evaluated via a leave-one-site-out and leave-one-tracer-out cross-validation. Classifiers achieved an out-of-sample receiver operating characteristic area-under-the-curve of 0.66 or greater when applied to all except one hold-out sites and 0.72 or greater when applied to each hold-out radiotracer. Additionally, when applying our models in a retroactive cohort enrichment analysis on A4 clinical trial data, we observed increased statistical power of detecting differences in amyloid accumulation between placebo and treatment arms after enrichment by ML stratifications. As emerging investigations of new disease-modifying therapies for AD increasingly focus on asymptomatic, preclinical populations, our findings underscore the potential applicability of ML-based patient stratification for recruiting more homogeneous cohorts and improving statistical power for detecting treatment effects for future clinical trials. Highlights Machine learning can predict future cognitive impairment in preclinical Alzheimer’s Models achieved high out-of-sample ROC-AUC on external sites and PET tracers Models were able to distinguish cognitively stable from decliners in the A4 cohort ML cohort enrichment enhanced secondary treatment effect detection in the A4 cohort

  PublisherOA PDFDOI

• Tau <scp>PET</scp> Burden Is Associated With Hippocampal Microstructure and Episodic Memory in Amyloid Positive Older Adults

  Hippocampus · 2025-09-29 · 1 citations

  article

  Growing evidence suggests that hippocampal gray matter microstructure, assessed through diffusion-weighted imaging (DWI), is a sensitive marker of neurodegeneration in Alzheimer's disease (AD). While hippocampal atrophy is a characteristic feature of AD, microstructural changes likely precede macrostructural changes such as volumetric loss, offering important insights into the early phases of the disease. This study assessed the relationships between hippocampal microstructure (assessed with mean diffusivity [MD] from DWI) and Braak-staged tau burden (measured by positron emission tomography [PET]) with performance on an episodic memory composite score, among individuals with and without amyloid burden, assessed by PET imaging. The study included 192 participants without dementia (14 with mild cognitive impairment [MCI]) from the BIOCARD cohort (mean age = 68), of which 52 (27%) were amyloid positive. In multiple linear regression analyses, increased hippocampal MD was associated with worse memory and greater tau PET burden in Braak stages II-IV, but only in individuals who were amyloid positive (e.g., significant amyloid × hippocampal MD interactions). Building on prior findings linking early Braak-staged tau to memory, we further assessed whether tau PET burden statistically mediated the relationship between elevated hippocampal MD and poorer memory performance. Tau PET burden in Braak stages II-IV was found to statistically mediate the relationship between elevated hippocampal MD and poorer memory performance, independent of hippocampal volume, but only in amyloid-positive participants. These associations were only significant when MCI participants were included in the analysis. These findings suggest hippocampal microstructure may be sensitive to AD-related pathological burden and associated neurodegeneration, particularly in the early symptomatic phase, and is associated with tau PET and cognitive decline, even after accounting for hippocampal volume.

  PublisherDOI

• Susceptibility MRI Helps Predict Mild Cognitive Impairment Onset and Cognitive Decline in Cognitively Unimpaired Older Adults

  Radiology · 2025-09-01 · 7 citations

  articleOpen access

  Tissue magnetic susceptibility elevations measured at MRI in the entorhinal cortex and putamen were significant predictors of onset of mild cognitive impairment and cognitive decline in cognitively unimpaired older adults, especially those with amyloid neuropathologic abnormalities.

  PublisherOA PDFDOI

Recent grants

• Tracing Spread of Pathology Within The HD Brain via Automated Neuroimaging

  NIH · $2.4M · 2018–2023

• NIH Grant U01NS082085

  NIH · $1.1M · 2015

• Continued Development and Maintenance of MriStudio

  NIH · $3.6M · 2013–2023

• NIH Grant R01EB020062

  NIH · $2.8M · 2021

• NIH Grant R01NS023007

  NIH · $344k · 1988

Frequent coauthors

• Susumu Mori

  Johns Hopkins University

  234 shared

• Laurent Younès

  Johns Hopkins University

  112 shared

• Ulf Grenander

  110 shared

• Daniel J. Tward

  University of California, Los Angeles

  101 shared

• J. Tilak Ratnanather

  Johns Hopkins University

  100 shared

• Andréia V. Faria

  88 shared

• Kenichi Oishi

  75 shared

• Marilyn Albert

  Johns Hopkins University

  74 shared

Education

• PhD, Biomedical Engineering

  The Johns Hopkins University

  1983

Awards & honors

• IEEE Biomedical Engineering Thesis Award (1982)
• Johns Hopkins Paul Ehrlich Graduate Student Thesis Award (19…
• NSF Presidential Young Investigator Award (1986)
• Inaugural Johns Hopkins University Gilman Scholar (2011)
• Fellow of the American Institute for Medical and Biological…