About

Steven E. Arnold, M.D., is an Emeritus Professor of Psychiatry at the University of Pennsylvania. He serves as Associate Director of the Institute on Aging and the Alzheimer Disease Core Center at the same institution. Dr. Arnold is also the Director of the Geriatric Psychiatry Section and the Penn Memory Center/Alzheimer's Disease Center Clinical Core at the University of Pennsylvania. His educational background includes an M.D. from Boston University, where he also earned a B.A. in Medical Science and Philosophy, graduating Summa cum laude in 1983. His research expertise encompasses schizophrenia, neuropsychiatry, cognition in aging and neurodegenerative dementias, clinicopathological correlation studies, postmortem research, cellular and molecular neuropathology, and quantitative microscopy. Dr. Arnold's clinical expertise involves evaluation, diagnosis, and medical management of complex neuropsychiatric disorders, including Alzheimer's disease, other neurodegenerative dementias, traumatic brain injury, and the cognitive, mood, and behavioral complications of neurological illness and injury. His contributions include advancing understanding in neurodegenerative diseases and improving diagnostic and treatment approaches for neuropsychiatric conditions associated with aging and neurological injury.

Research topics

• Medicine
• Biology
• Internal medicine
• Psychology
• Genetics
• Neuroscience
• Pathology
• Computer Science
• Bioinformatics
• Artificial Intelligence
• Psychiatry
• Machine Learning
• Oncology
• Virology
• Database
• Mathematics
• Pharmacology
• Immunology
• Computational biology
• Chemistry
• Data science
• Nuclear medicine
• Econometrics
• Audiology

Selected publications

• End-to-End Cortical Surface Reconstruction from Clinical Magnetic Resonance Images

  Lecture notes in computer science · 2026-01-01

  book-chapter
  PublisherDOI

• The NIH Toolbox Emotion Battery and Cerebrospinal Fluid Biomarkers in Alzheimer’s Disease: Findings from the Multisite ARMADA Study

  medRxiv · 2025-05-22

  preprintOpen access

  Objectives: Investigating the relationships between socioemotional functioning and Alzheimer's Disease (AD) pathology can contribute to screening and early detection of AD. This study explored the associations between socioemotional functioning and cerebrospinal fluid (CSF) AD biomarkers in older adults. Methods: We used baseline data from the Advancing Reliable Measurement in Alzheimer's Disease and Cognitive Aging (ARMADA) study. ARMADA is a multisite study with independent protocols for CSF assays at each site. The available sample size with comparable CSF assays had 31 participants with normal cognition (NC) and 28 with amnestic mild cognitive impairment (aMCI) or early-stage AD dementia. CSF-derived AD biomarkers included were: phosphorylated-tau 181 (p-Tau181), total tau (t-Tau), Aβ42, Aβ42/40 ratio, and p-Tau181/Aβ42 ratio. Socioemotional functioning (negative affect, psychological wellbeing, and social satisfaction) was measured with the self-reported NIH Toolbox Emotion Battery (NIHTB-EB). We ran linear regressions by cognitive subgroups (NC and aMCI/early-stage AD). Results: Among participants with NC, lower social satisfaction was associated with higher p-Tau181 and t-Tau; higher t-Tau was additionally associated with more negative affect. None of the CSF AD biomarkers were associated with the NIHTB-EB outcomes among participants with aMCI or early-stage AD. Discussion: These findings suggest that socioemotional functioning may be associated with tau pathology. Markers for amyloid were not related to socioemotional functioning regardless of disease stages. Future studies with larger, more diverse samples and harmonized CSF assay protocols are needed to further examine the role of early socioemotional change in early detection and prevention of dementia.

  PublisherOA PDFDOI

• Plasma brain-derived tau and phosphorylated tau at T181 are elevated in amyotrophic lateral sclerosis

  medRxiv · 2025-06-27 · 2 citations

  preprintOpen access

  Abstract There is an unmet need for reliable biomarkers for amyotrophic lateral sclerosis (ALS). Recent studies demonstrated that the levels of the microtubule-associated protein tau are altered in plasma and cerebrospinal fluid (CSF) from people living with ALS. Our previous findings revealed that while the ratio between tau and phosphorylated tau at T181 (pTau-T181) is lower in ALS CSF compared to healthy controls (HC), higher CSF tau levels correlated with faster disease progression. Here, we measured total tau and pTau-T181 in plasma samples from two cohorts of participants with ALS and HC using two platforms: Quanterix Simoa and Meso Scale Discovery (MSD). Both assays demonstrated significantly higher pTau-T181 and pTau-T181:tau ratio levels in ALS compared to HC. Longitudinal analysis revealed that higher pTau-T181 levels at baseline correlated with faster decline on the revised ALS functional rating scale (ALSFRS-R). Although the MSD analysis revealed higher plasma tau levels in ALS, the Quanterix Simoa assay demonstrated lower total tau levels in ALS plasma than in HC. To investigate the source of this discrepancy, we measured brain-derived tau (BD-tau) using a new Quanterix assay that specifically detects tau originating from the central nervous system (CNS). This analysis demonstrated significantly elevated plasma BD-tau levels in ALS, mirroring the results obtained with the MSD assay, suggesting that the elevated plasma tau observed in ALS is primarily CNS-derived. Collectively, our data indicate that plasma BD-tau and pTau-T181 levels are elevated in ALS, and future studies will aim to define their potential utility as diagnostic or prognostic biomarkers.

  PublisherOA PDFDOI

• Multi-ancestry genome-wide meta-analysis of 56,241 individuals identifies known and novel cross-population and ancestry-specific associations as novel risk loci for Alzheimer’s disease

  Genome biology · 2025-07-17 · 12 citations

  reviewOpen access

  BACKGROUND: Limited ancestral diversity has impaired our ability to detect risk variants more prevalent in ancestry groups of predominantly non-European ancestral background in genome-wide association studies (GWAS). We construct and analyze a multi-ancestry GWAS dataset in the Alzheimer's Disease Genetics Consortium (ADGC) to test for novel shared and population-specific late-onset Alzheimer's disease (LOAD) susceptibility loci and evaluate underlying genetic architecture in 37,382 non-Hispanic White (NHW), 6728 African American, 8899 Hispanic (HIS), and 3232 East Asian individuals, performing within ancestry fixed-effects meta-analysis followed by a cross-ancestry random-effects meta-analysis. RESULTS: We identify 13 loci with cross-population associations including known loci at/near CR1, BIN1, TREM2, CD2AP, PTK2B, CLU, SHARPIN, MS4A6A, PICALM, ABCA7, APOE, and two novel loci not previously reported at 11p12 (LRRC4C) and 12q24.13 (LHX5-AS1). We additionally identify three population-specific loci with genome-wide significance at/near PTPRK and GRB14 in HIS and KIAA0825 in NHW. Pathway analysis implicates multiple amyloid regulation pathways and the classical complement pathway. Genes at/near our novel loci have known roles in neuronal development (LRRC4C, LHX5-AS1, and PTPRK) and insulin receptor activity regulation (GRB14). CONCLUSIONS: Using cross-population GWAS meta-analyses, we identify novel LOAD susceptibility loci in/near LRRC4C and LHX5-AS1, both with known roles in neuronal development, as well as several novel population-unique loci. Reflecting the power of diverse ancestry in GWAS, we detect the SHARPIN locus with only 13.7% of the sample size of the NHW GWAS study (n = 409,589) in which this locus was first observed. Continued expansion into larger multi-ancestry studies will provide even more power for further elucidating the genomics of late-onset Alzheimer's disease.

  PublisherOA PDFDOI

• Plasma NfL and cognitive functioning in older adults: The moderating role of HDL cholesterol

  Alzheimer s & Dementia Diagnosis Assessment & Disease Monitoring · 2025-10-01 · 1 citations

  articleOpen access

  BACKGROUND AND OBJECTIVES: Plasma neurofilament light chain (NfL) is a marker of neuroaxonal injury associated with cognitive decline. High-density lipoprotein (HDL) cholesterol has neuroprotective properties, but its interaction with neurodegeneration remains unclear. This study examined whether HDL moderates the association between NfL and cognitive performance. METHODS: Baseline data from 417 participants in the Aging Adult Brain Connectome study were analyzed. Plasma NfL and HDL were measured via Simoa and enzymatic assays; cognition was assessed using Montreal Cognitive Assessment (MoCA) and Preclinical Alzheimer Cognitive Composite (PACC). Generalized linear models were used to evaluate NfL and HDL interactions, adjusting for demographics. Sensitivity analyses included apolipoprotein E ε4, body mass index, total cholesterol, LDL, and triglycerides. RESULTS: = 0.004), indicating HDL moderates the negative association between NfL and cognition. DISCUSSION: These findings suggest that HDL modifies the cognitive impact of neurodegeneration, highlighting the importance of metabolic-neurological interactions. Highlights: High-density lipoprotein (HDL) cholesterol moderates the negative association between plasma neurofilament light chain (NfL) and cognition.Higher HDL levels intensify the negative effect of NfL on cognitive performance.Findings challenge the assumption of HDL's uniformly protective role.Results support the integrated use of metabolic and neurodegenerative biomarkers.

  PublisherOA PDFDOI

• Efflux Pumps

  RCSB Protein Data Bank · 2025-03-26

  article1st authorCorresponding
  PublisherDOI

• Association among cardiorespiratory fitness, plasma biomarkers of pathology and astrocyte reactivity, and cognition in autosomal-dominant Alzheimer's disease

  Journal of Alzheimer s Disease · 2025-05-27 · 2 citations

  article

  Background High cardiorespiratory fitness has been associated with greater neuroplasticity, and slower neurodegeneration and cognitive decline in healthy adults. Yet, less is known about whether low-to-intermediate cardiorespiratory fitness is associated with lower markers of disease progression in the preclinical stage of Alzheimer's disease (AD). Objective We investigated whether cardiorespiratory fitness was associated with plasma biomarkers for AD-related pathology, neural injury and astrocyte reactivity, and episodic memory in Presenilin-1 E280A carriers without dementia from the world's largest kindred with autosomal-dominant AD. Methods Twenty-seven mutation carriers (25 cognitively unimpaired, 2 with mild cognitive impairment; ages: μ=30.22 years, SD = 5.24; 74% female) participated in the study. Participants underwent a graded aerobic fitness test to assess cardiorespiratory fitness, measured in maximum metabolic equivalent of task (MET). Plasma biomarkers included amyloid 42/40, phosphorylated tau-181, neurofilament light chain, and glial fibrillary acidic protein. Participants completed the Consortium to Establish a Registry for AD word list learning and delayed recall. We conducted multiple linear regressions controlling for age, sex, and years of education. Results Fourteen participants’ MET values were indicative of low cardiorespiratory fitness (< 9 MET), and 13 participants’ MET values of intermediate cardiorespiratory fitness (9–14 MET). METs were not associated with age, biomarkers, or episodic memory. Conclusions Our findings suggest that low-to-intermediate cardiorespiratory fitness may not be associated with biomarkers for AD-related pathology, neural injury and astrocyte reactivity, or memory in people at genetic risk for dementia. Longitudinal studies and randomized-controlled trials are needed to better understand the relationships among cardiorespiratory fitness and AD progression.

  PublisherDOI

• Kinematic Correlates of Early Speech Motor Changes in Cognitively Intact APOE-ε4 Carriers: A Preliminary Study Using a Color-Word Interference Task

  medRxiv · 2025-06-17

  preprintOpen access

  Alzheimer's disease (AD) is the most prevalent form of dementia and a major public health challenge. In the absence of a cure, accurate and innovative early diagnostic methods are essential for proactive life and healthcare planning. Speech metrics have shown promising potential for identifying individuals with mild cognitive impairment (MCI) and AD, prompting investigation into whether speech motor features can detect elevated risk even prior to cognitive decline. This study examined whether speech kinematic features measured during a color-word interference task could distinguish cognitively normal APOE-ε4 carriers (E4+) from non-carriers (E4-). Lip movement properties were extracted across pre-, during-, and post-interference sentence segments. Descriptive statistics and independent t-tests were conducted to examine group-level trends in lip movement duration, average speed, and range. Although no group differences reached statistical significance, several features showed moderate effect sizes, suggesting potential neuromotor differences. A support vector machine model with a degree-2 polynomial kernel achieved 87.5% accuracy in classifying APOE-ε4 status using three features: lip movement duration prior to interference, average lip speed during interference, and the change in lip movement range from pre- to during-interference segments. These features reflect subtle differences between the two groups in baseline motor planning, susceptibility to cognitive-motor interference, and articulatory adaptability. The model's high precision (88.90%), sensitivity (88.90%), and specificity (85.70%) underscore the potential clinical utility of speech kinematics for early risk identification. These findings support the use of non-invasive, low-burden speech analysis as a promising digital biomarker for AD risk in cognitively intact individuals and highlight its potential for scalable, remote screening and longitudinal monitoring in diverse populations.

  PublisherOA PDFDOI

• Digital twins techniques in the design of Alzheimer’s disease clinical trials: an application to the Internet‐Based Conversational Engagement Clinical Trial (I‐CONECT)

  Alzheimer s & Dementia · 2025-12-01

  articleOpen access

  BACKGROUND: With U.S. Food and Drug Administration (FDA)-approved anti-amyloid, partially disease-modifying treatments now available, the ethical justification for randomly assigning patients to placebo has become controversial. The idea of twin-controls (or virtual-controls) has gained significant attention in recent years, with the aim of creating twin patient cohorts that can be used as a surrogate to evaluate the effects of treatment on a personalized level. While promising, the feasibility of digital twins techniques remains largely untested. METHOD: I-CONECT is a multi-site, single-blind, randomized controlled trial (RCT) examining the effects of conversational interactions on cognition among socially isolated subjects aged ≥ 75 years (normal cognition; mild cognitive impairment). 186 participants were randomized into experimental or control groups. The experimental group engaged in video chats with study staff 4 times/week for 6 months, while the control groups received weekly 10-minute phone calls. The current analysis focused on the efficacy-shown Montreal Cognitive Assessment (MoCA; global cognition) and category fluency animals (CFA; language-based executive function) at 6-month follow-up. Data from the National Alzheimer's Coordinating Center-Uniform Data Set (NACC-UDS) were used to create digital twins for treatment participants through two methods. Method 1 involved twin mapping, matching participants with 1 to 20 twins who had similar demographic, biological, and social factors, and comparing change scores between each participant and their twins. Method 2 used direct modeling by building random forest models to predict change scores as if participants were assigned to a "usual care" control group. Effect sizes were compared between original and twin-controls trials, as well as between the two methods. RESULT: Approximately 10% of NACC-UDS participants (5,332 out of 50,259) were eligible for I-CONECT. For parallel-group designs, treatment effect sizes on MoCA closely aligned between original (β=1.67) and twin-control (β=1.46-1.97) trials when the Euclidean distance mapping was applied. Similar findings were found in CFA (original trial β=2.56; twin-control trial β=2.31-3.34). For single-case, n-of-1 designs, methods 1 and 2 showed substantial agreement in identifying treatment responders (Cohen's Kappa=1 for MoCA; 0.68 for CFA). CONCLUSION: Digital twins from publicly available datasets enhance the rigor of RCTs by providing mapped twins as controls for early-phase dementia trials.

  PublisherOA PDFDOI

• Effects of nicotinamide riboside on NAD+ levels, cognition, and symptom recovery in long-COVID: a randomized controlled trial

  EClinicalMedicine · 2025-11-01 · 2 citations

  articleOpen access

  Background: Long-COVID often involves cognitive difficulties, immune dysregulation, and mitochondrial dysfunction. Studies suggest nicotinamide adenine dinucleotide (NAD+) precursors like nicotinamide riboside (NR) may reduce inflammation and support mitochondrial and neurological function. This double-blind, placebo (PBO)-controlled clinical trial with a placebo lead-in phase evaluated the effects of NR (2000 mg/day) on NAD+ and changes in cognitive and long-COVID symptoms. Methods: This was a 24-week, double-blind, placebo-controlled trial at a single center in Boston, USA, between August 2021 and September 2023. 58 community-dwelling participants with long-COVID were randomized 2:1 to the NR-NR group (NR for 20 weeks) or the PBO-NR group (PBO for 10 weeks, followed by NR for 10 weeks). The primary outcome was cognition, assessed using the Everyday Cognition scale (ECog), Repeatable Battery for the Assessment of Neuropsychological Status (RBANS), and Trail Making Test-B (TMT-B). Secondary outcomes included the Fatigue Severity Scale (FSS), Beck Depression Inventory (BDI), Beck Anxiety Inventory (BAI), and Pittsburgh Sleep Quality Index. We conducted a mixed model for repeated measures to compare groups, then post-hoc and unadjusted for multiplicity, combined both groups to explore changes from baseline after 10 weeks of NR. This trial was registered with ClinicalTrials.gov (NCT04809974) in 2021. Findings: 37 participants (64%) were assigned to NR-NR, and 21 participants (36%) to PBO-NR. There was a 32.4% and 51.4% dropout in the NR-NR group at 10 weeks and 20 weeks, respectively, vs. 14.3% dropout at each timepoint in the PBO-NR group. In the NR-NR group, NAD+ levels increased by 2.6- to 3.1-fold after 5-10 weeks of supplementation, respectively, and remained elevated at 20 weeks. In the PBO-NR group, NAD+ levels remained close to baseline (0.93- to 1.0-fold change, 95% CI: 0.5-1.4) during the initial 5 and 10 weeks of PBO. After switching to NR, levels rose to a 2.6-fold and 2.1-fold increase after 5 and 10 weeks of NR, respectively. No significant between-group differences were observed for cognitive outcomes (ECog, RBANS, TMT-B; p-values = 0.47-0.74). There were no significant differences in fatigue severity (p = 0.59), sleep quality p = 0.69), and symptoms of anxiety (p = 0.84) or depression (p = 0.20) between PBO and NR groups. In post-hoc exploratory analysis, examining within-group changes during 5 and 10 weeks of NR intake by grouping all participants during the first 10 weeks of the NR phase, there were significant differences from baseline after 10 weeks of NR in executive functioning, fatigue severity, sleep quality, and symptoms of depression (compared with no significant changes in TMT-B, FSS, PSQI, BAI, or BDI scores during the PBO phase). One serious adverse event was reported, deemed unrelated to the study drug or trial. Interpretation: In long-COVID, NR increased NAD+ within 5 weeks but did not significantly improve cognition, fatigue, sleep, or mood vs. PBO. Exploratory analyses suggested within-group benefits after 10 weeks of NR, supporting the need for larger trials. Funding: This work was supported by Niagen Bioscience, the MGH McCance Center for Brain Health, Lavine Brain Health Innovation Fund, MGH ECOR CDI Physician-Scientist Development Award, and the Alzheimer's Association (grant no. AARGD-23-114103).

  PublisherDOI

Recent grants

• NIH Grant K20MH000978

  NIH · $679k · 1997

• NIH Grant R01AG03947802

  NIH · $505k · 2016

• NIH Grant R29MH055199

  NIH · $550k · 2001

• NIH Grant R21TR003040

  NIH · $485k · 2021

• NIH Grant R01NS084965

  NIH · $3.2M · 2021

Frequent coauthors

• Keith A. Johnson

  Massachusetts General Hospital

  3052 shared

• D. Cheng

  2268 shared

• Joseph C. Wu

  1944 shared

• Monte S. Buchsbaum

  University of California, Irvine

  1620 shared

• Marcelo F. Di Carli

  Harvard University

  1620 shared

• Carl K. Hoh

  University of California, San Diego

  1620 shared

• Carolyn C. Meltzer

  University of Southern California

  1296 shared

• Darin E. Olson

  University of Alabama at Birmingham

  972 shared

Education

• MD

  Boston University School of Medicine

  1983