WALID YASSIN

WalidYassin

D.M.Sc., M.M.Sc. (Wa-lead Ya-seen)

Assistant Professor of Psychiatry at Harvard Medical School
Staff Scientist at Beth Israel Deaconess Medical Center

Using AI, ML, statistical approaches and mathematical theories to understand the brain, mind and behavior across the psychosis and autism spectra.

Ongoing signal

01Research
Conditions

Psychosis spectrum/Autism spectrum

Our work also extends to conditions with high comorbidity, including:

  • Bipolar disorder
  • ADHD
  • among others
Biomarkers

Neuroimaging/Cognition

Complemented by physiological, behavioral, genetic and environmental measures, including:

  • Electrophysiology
  • Behavior
  • Genetics
  • Childhood trauma
  • Substance use
  • Social determinants of health
  • among others
Tools & theories

Artificial intelligence/Applied mathematics

Spanning machine learning, large language models and mathematical modeling, including:

  • Classification
  • Clustering
  • RNNs
  • LLMs
  • Graph theory
  • Bayes' theorem
  • Dynamical systems
  • Information theory
  • among others
AIMLLLMMathBrainMindBehaviorBiomarkersNeuropsychiatry
02For students

Three paths into the work.

AAI · ML · LLM
Path A

Machine learning

Use clustering, classification, recurrent neural networks and LLMs to understand neuropsychiatric conditions.

  • Clustering
  • Classification
  • RNNs
  • LLMs
BApplied math
Path B

Mathematical models

Use mathematical theory to model neuropsychiatric conditions and to explain how and why they develop.

  • Graph theory
  • Bayes' theorem
  • Dynamical systems
  • Information theory
  • Game theory
CStatistics
Path C

Statistical approaches

Apply statistical methods to delineate differences between and within conditions, and relative to the typically developing population.

  • SEM
  • Path analysis
  • Mixed-effects models
  • Latent growth curves
  • ANCOVA

Interested in joining a project? Tell me which path appeals to you and what you'd like to learn.

Get in touch
03Selected work

Tools, models and papers.

2020 — present

Diagnostic aid algorithms for psychiatry

Machine-learning algorithms that classify schizophrenia, autism and early psychosis from brain MRI scans. They are designed to make psychiatric assessment more reliable.

2020 — present

A recurrent neural network of attention to faces

A computational model of how the brain reads emotion in faces, extended to simulate how people with different psychiatric conditions perceive them.

Selected publications

  1. 2025

    Exposotypes in psychotic disorders

    Yassin W, Kromenacker B, Green JB, … Keshavan MS — Scientific Reports

    Exposome
  2. 2025

    Cognitive subtypes in youth at clinical high risk for psychosis

    Yassin W, Green JB, Keshavan M, … Stone WS — Psychiatry and Clinical Neurosciences

    Cognition
  3. 2025

    The neurodiversity framework in medicine: on the spectrum

    Miranda-Ojeda R, Wickramasinghe A, Ntolkeras G, Castanho I, Yassin W — Developmental Neurobiology

    Neurodiversity
  4. 2024

    Biomarker discovery using machine learning in the psychosis spectrum

    Yassin W, Loedige KM, Wannan CMJ, … De Nadai AS — Biomarkers in Neuropsychiatry

    Machine learning
  5. 2024

    The schizophrenia syndrome, circa 2024: what we know and how that informs its nature

    Tandon R, Nasrallah H, … Yassin W, Keshavan M — Schizophrenia Research

    Review
  6. 2024Perspective
  7. 2020Neuroimaging
  8. 2019

    Paternal age contribution to brain white matter aberrations in autism spectrum disorder

    Yassin W, Kojima M, Owada K, … Yamasue H — Psychiatry and Clinical Neurosciences

    Neuroimaging
04Contact

Curious?
Let's talk.

If you're interested in a research path, please include your CV and the path you'd like to explore. I also welcome collaborators, speaking invitations and media requests.

Based in Boston, Massachusetts