Computational Discovery
Machine learning, multimodal integration, disease subtyping, longitudinal prediction, and biomarker discovery for complex biomedical datasets.
Biomedical engineering, data science, and responsible AI
Raha Dastgheyb
Biomedical Engineer
Translational Data Scientist
Computational Neuroscientist
Scientific Software Builder
Reproducible Systems Architect
Assistant Professor of Neurology
Division of Neuroimmunology
Johns Hopkins School of Medicine
Building computational tools that transform complex biomedical data into scientific insight through reproducible analytics, open-source scientific software, and human-in-the-loop AI-assisted discovery.
Overall Positioning
The bottleneck is no longer data generation. The bottleneck is discovery.
Modern biomedical research can generate thousands of clinical, imaging, molecular, behavioral, and omics measurements from a single study. This research program builds computational infrastructure that helps teams explore complex datasets, interpret findings, and generate biologically meaningful hypotheses while maintaining scientific rigor.
Biomedical engineering and computer science training anchor a translational program that turns complex biomedical data into actionable insight for collaborators, cohorts, and grant-funded scientific teams.
Scientific Software Infrastructure
Open-source software that translates published methods into reproducible analytical workflows, reports, and reusable research systems.
Responsible AI-Assisted Science
Human-in-the-loop AI, conversational data exploration, scientific reasoning, and transparent workflows that help scientists think more effectively.
Methods and Toolchain
Scientific capabilities for finding hidden structure in biomedical data.
Statistical workflows, reports, visualization, and open-source biomedical data science.
MathWorks MATLAB app development for signal processing, GUI tools, and quantitative neuroscience.
Clustering, dimensionality reduction, prediction, and interpretable biomarker discovery.
Interactive figures and reports that help scientists inspect multidimensional evidence.
Apps and dashboards that make complex workflows usable by interdisciplinary teams.
Integrated molecular, imaging, clinical, cognitive, and behavioral analysis workflows.
Wearables, tablet-based neuropsychology, actigraphy, and longitudinal signal streams.
Open-source research infrastructure that makes methods reusable and auditable.
Featured Systems
Open-source software as scientific infrastructure.
These systems are positioned as research outputs: reusable, inspectable, and designed to turn complex biomedical data into reproducible scientific insight, from quality control to biomarker analysis to AI-assisted discovery.
Published, citable MATLAB app for multi-electrode array analysis.
released
MEAnalyzer
A downloadable point-and-click tool for spike-train statistics, burst detection, functional connectivity, periodicity analysis, visualization, and provenance-aware exports.
Open-source biomedical data analysis and reporting infrastructure.
released
SciDataReportR
Supports metadata-aware QC, statistical workflows, biomarker analysis, visualization, and reproducible scientific reasoning.
Reveals latent cognitive and biomarker patterns across heterogeneous cohorts.
capability area
Computational Phenotyping Pipelines
Moves beyond average effects toward clinically and biologically interpretable hidden structure.
Turns analysis workflows into publication-ready scientific reports.
capability area
Scientific Reporting Frameworks
Makes reproducible research outputs traceable, interpretable, shareable, and ready for collaborators.
In-development proteomics reporting workflows for multi-omics discovery.
in development
ProteomicsReportR
Designed to turn high-dimensional proteomics analyses into transparent QC, biomarker discovery, clustering, and interpretation reports.
In-development metabolomics reporting for biomedical discovery.
in development
MetabolomicsReportR
Designed for reproducible metabolomics reports that connect multi-omics measurements with statistical workflows and biological interpretation.
In-development multiplex biomarker reporting infrastructure.
in development
MultiPlexReportR
Designed for transparent QC, visualization, and biomarker discovery in multiplex translational studies.
Conceptual in-development human-in-the-loop AI reasoning layer.
concept
SciDataAgent
Frames conversational data exploration, transparent assumptions, and reproducible analysis planning as future scientific infrastructure.
Research Architecture
An interconnected research ecosystem for translational discovery.
The research architecture connects hidden structure in complex biomedical data to scientific infrastructure, translational science, biomarkers, multi-omics, visualization, reproducibility, people, and impact.
discovers subgroups
connects evidence
reveals mechanisms
builds reusable tools
integrates cohorts
enables reproducibility
Central system
Reproducible Infrastructure
Transparent workflows, reusable methods, and scientific software that make discovery auditable.
Hidden structure
Computational Phenotyping
Finds interpretable subgroups, trajectories, and cognitive profiles.
Evidence layer
Scientific Visualization
Makes multidimensional findings inspectable by collaborators.
Mechanisms
Biomarker Systems
Links molecular signals, multi-omics, and clinical brain health questions.
Reusable tools
Software Systems
Turns analytical methods into maintained scientific infrastructure.
Cohort translation
Translational Neuroinformatics
Connects cohorts, cognition, biomarkers, and brain health translation.
Research Threads
Papers, software, talks, and programs connected by scientific questions.
Research threads make the publication ecosystem navigable by showing how individual outputs become coherent programs of translational data science.
Hidden structure
Computational Phenotyping
Unsupervised learning, cognitive profiles, and interpretable models for heterogeneous brain health data.
Molecular signals
Biomarker Systems
Computational pipelines for linking multi-omics, neuroimaging, inflammation, and cognition.
Reusable systems
Scientific Software & Infrastructure
Open-source tools, reporting frameworks, dashboards, and analysis interfaces that make science reproducible.
Translational cohorts
Long COVID & Brain Health
Cognitive outcomes, biomarker signals, and clinical data streams for understanding post-acute brain health.
Emerging field
Organoid Intelligence
Field-building work around organoid intelligence, neural systems, and responsible scientific infrastructure.
Selected Publications
Highlighted papers as research portals.
BRACE-ing for the future: Establishing iPad-based norms for cognitive function in the MACS/WIHS Combined Cohort Study (Preprint)
JMIR mental health
Establishes regression-based iPad cognitive norms for BRACE in the MACS/WIHS Combined Cohort Study.
Identifying and distinguishing cognitive profiles among virally suppressed people with HIV.
Neuropsychology
Identifies six cognitive profiles among virally suppressed people with HIV and the factors that distinguish them.
Machine learning approaches to understand cognitive phenotypes in people with HIV
The Journal of infectious diseases
Frames machine learning as a tool for discovering cognitive biotypes in people with HIV.
MEAnalyzer-a spike train analysis tool for multi electrode arrays
Neuroinformatics
Introduces MEAnalyzer as reproducible software for spike train analysis in multi-electrode array experiments.
Blood-Brain barrier disruption in long COVID and cognitive correlates: A cross-sectional MRI study
Brain, Behavior, and Immunity
Uses non-contrast MRI to study blood-brain barrier permeability and cognition in Long COVID.
The Baltimore declaration toward the exploration of organoid intelligence
Frontiers in Science
Defines organoid intelligence as an emerging discipline requiring technical, ethical, and community infrastructure.
Patterns and predictors of cognitive function among virally suppressed women with HIV
Frontiers in neurology
Uses self-organizing maps and random forests to characterize cognitive profiles in virally suppressed women with HIV.
Metabolomic levels mediate the link between socioeconomic factors and changes in declarative memory in women with and without HIV
Brain, Behavior, & Immunity-Health
Links socioeconomic conditions, metabolomic profiles, and longitudinal declarative memory change in women with and without HIV.
Longitudinal effects of polypharmacy on cognitive function in people with HIV
AIDS
Examines how polypharmacy relates to cognitive function over time in people with HIV.
Tryptophan-Kynurenine Pathway Activation and Cognition in Virally Suppressed Women With HIV
JAIDS Journal of Acquired Immune Deficiency Syndromes
Studies whether tryptophan-kynurenine pathway activation is associated with cognition in virally suppressed women with HIV.
Talks and Scientific Communication
Field-building through talks, workshops, and technical storytelling.
Community Infrastructure
Scientific communities help people learn, connect, and build.
SynaptiCoders
Community infrastructure for scientific coding, mentoring, and reproducible practice.
R-Ladies Baltimore
Local data science community building, training, and inclusive technical leadership.
Neuromatch
Open computational neuroscience education and global scientific community.
Moveable Feast
Board involvement supporting community health, care, and service.
Roles & Affiliations
Leadership roles, institute affiliations, and collaborative infrastructure.
Leadership role
Brain Health Program
Co-director, Data Science and Mathematical Modeling Core, with Yanxun XuA translational engine for integrating cognitive data, biomarkers, neuroimaging, and computational methods across HIV, Long COVID, aging, and related brain health questions.
Leadership role
Johns Hopkins Center for the Advancement of HIV Neurotherapeutics (JH-CAHN)
Co-director, Biomarker Core, with Rebecca VeenhuisA collaborative infrastructure role focused on biomarker discovery, validation, and interpretation for HIV-associated neurocognitive and neuroimmune conditions.
Data Science and AI Institute
AffiliateInterdisciplinary data science, AI, modeling, and translational analytics.
IDIES
Voted memberLarge-scale data systems, computational research, and reproducible scientific workflows.
MWCCS / MACS-WIHS
Collaborative cohort infrastructureMember of the BRACE/tablet-based cognitive data reading center, Sleep Working Group, and Neuropsych Working Group.
CHARTER
Collaborative cohort scienceCollaborative research infrastructure for studying neurocognitive, clinical, and biopsychosocial phenotypes in people with HIV.
Brain Health Program
Program infrastructure for integrating cognitive, biomarker, neuroimaging, and computational methods across translational brain health questions.
Johns Hopkins Center for the Advancement of HIV Neurotherapeutics (JH-CAHN)
Collaborative neuroHIV infrastructure focused on biomarkers, neurotherapeutics, and translational discovery.
Organoid Intelligence
A cross-disciplinary research direction connecting stem-cell systems, electrophysiology, computational analysis, ethics, and community standards.
Training Arc
From biomedical engineering and computer science to translational data science infrastructure.
UVAB.S. Biomedical Engineering, minor in Computer Science
University of VirginiaEngineering, computation, and biomedical problem solving.
DUPh.D. Biomedical Engineering
Drexel UniversityAxonal injury, mechanobiology, and quantitative experimentation.
DUPostdoctoral training
Drexel University and Johns Hopkins NeurologyElectrophysiology, neurons, organoids, and quantitative neuroscience.
JHUFaculty and translational data science
Johns Hopkins NeurologyCohort science, biomarkers, software, and responsible AI-assisted discovery.
Selected Honors
Recognition of collaborative scientific impact.
ICTR 2026 Team Science Award
Recognition of the Brain Health Program's interdisciplinary work connecting cognitive impairment, HIV, Long COVID, translational neuroscience, and data science.
People
Mentorship, trainees, and collaborators extend the infrastructure.
Current trainees, former trainees, and collaborators make the scientific software, cohort analytics, and translational discovery ecosystem possible.
PeopleFuture Directions
AI that helps scientists think more effectively.
Through the Brain Health Program's Data Science and Mathematical Modeling Core and the JH-CAHN Biomarker Core, this work is extending reproducible computational science toward conversational data exploration, cross-domain biomedical integration, and SciDataAgent, an in-development AI reasoning layer for human-in-the-loop scientific discovery.
Human-in-the-loop AI
Scientific judgment stays central.
AI-assisted discovery should help researchers inspect evidence, surface assumptions, and reason transparently without replacing scientific expertise.
Conversational data exploration
Complex data should be explorable.
Researchers need interfaces that let them ask better questions of clinical, imaging, molecular, and omics datasets while preserving provenance.
Reproducible computational science
Discovery needs infrastructure.
Reports, workflows, dashboards, and open-source software make biomedical AI and machine learning results auditable, reusable, and easier to extend.
Cross-domain integration
Signals become insight together.
Clinical, biomarker, neuroimaging, behavioral, and multi-omics data become more powerful when integrated through transparent computational systems.
Contact
Collaborations, speaking invitations, trainee opportunities, software inquiries, and interdisciplinary research partnerships.
raha [at] jhmi [dot] edu
Open to collaborations, speaking invitations, trainee opportunities, software inquiries, and interdisciplinary research partnerships.