Hi 👋, I'm Hamza

HAM-za ab-del-HE-di he/him

I use M/EEG and computational models to study brain dynamics, cognition, and representations in brains and AI.

PhD student in Biomedical Engineering, Faculty of Medicine, Université de Montréal

Get in touch CV (PDF)

In my PhD with Prof. Karim Jerbi, I study how changing evidence shapes decisions. I also collaborate on pediatric EEG research at CHU Sainte-Justine.

I build and maintain open-source tools for analyzing brain signals, including CoCo-PiPe and MNE-Denoise, and contribute to packages across the MNE ecosystem. I share practical methods through teaching, mentoring, and community work, with a focus on open, accessible science.

At Abundant Intelligences, I support AI development for low-resource and Indigenous-language applications and deliver technical workshops.

Originally from Sfax 🇹🇳, now based in Tiohtià:ke / Montréal 🇨🇦. More about my path.

Task-optimized neural networks reveal distinct contributions of specialized and broader visual learning to neural representations of face familiarity

Hamza Abdelhedi, Shahab Bakhtiari, Karim Jerbi

bioRxiv · 2026

Five Minutes With Hamza Abdelhedi: Bringing New AI Tools to Neuroscience

Abundant Intelligences, Jan 2026

I discuss my path into Neuro-AI, clinical EEG research, and my work with Abundant Intelligences.

Foundation models for M/EEG

Brainhack Montréal · Oct 2026

I taught classical EEG decoding alongside frozen and fine-tuned foundation models.

MNE-Denoise

Co-Creator & Core Maintainer

Artifact suppression and signal denoising for EEG and MEG, with NumPy and MNE-Python integration.

Research directions

Explore my research

My work connects questions about cognition and learning with methods for studying neural activity.

Neural dynamics & cognition

I study the brain dynamics of decision-making and conscious experience.
I ask how they vary across brain states, disorders, and individuals.

Explore this direction

Representations in brains & AI

I investigate how learning shapes representations in brains and AI.
I compare what neural networks learn with patterns of human brain activity.

Explore this direction

Computational methods for M/EEG

I develop methods to extract and interpret information from M/EEG.
I test how they generalize across people, tasks, and recordings.

Explore this direction

Selected publications

View all publications

Selected work on visual representations, neural dynamics, and MEG methods.

Five-panel study-design diagram showing familiar, unfamiliar and scrambled face stimuli; MEG sensor and cortical source activity; CNNs trained for face identity, objects, or objects including faces; and brain and CNN dissimilarity matrices compared over time.

2026 · bioRxiv · Preprint

Task-optimized neural networks reveal distinct contributions of specialized and broader visual learning to neural representations of face familiarity

Networks trained on faces or broader visual tasks align with different stages of familiar-face processing.

Face familiarity
MEG
Visual recognition
CNNs

Hamza Abdelhedi, Shahab Bakhtiari, Karim Jerbi

Ten-step PCA workflow grouped into data preparation, dimensionality reduction, and analysis and interpretation, with neural representations, preprocessing, data matrices, principal components, scalp loadings, trajectories, behavioural measures and validation.

2026 · arXiv · Preprint

From Signals to Trajectories: A Primer on Low-Dimensional Dynamics in Human EEG and MEG

A practical guide to tracing M/EEG activity through low-dimensional trajectories, with an EEG tutorial.

EEG
MEG
PCA
Neural trajectories

Vanessa Hadid, Hamza Abdelhedi et al.

Diagram of routes toward a reusable MEG foundation model, including MEG-native pretraining, EEG model adaptation and continued pretraining, cross-modal training, and transfer from generic time-series models, with trade-offs in data, compute and MEG-specific expressiveness.

2026 · arXiv · Preprint

A Roadmap for MEG Foundation Models

Routes to reusable MEG models, from pretraining and EEG transfer to evaluation across people and tasks.

MEG
Foundation models
Neuroimaging
Pretraining

Philipp Thölke, Hamza Abdelhedi et al.

Literature-review flowchart from database search through screening and data extraction to 119 included studies: 70 classification, 16 modeling, and 33 other methodological studies, with application subcategories.

2025 · Journal of Neural Engineering

Artificial neural networks for magnetoencephalography: a review of an emerging field

A review of 119 studies using neural networks for MEG decoding, brain–model comparisons, and signal analysis.

MEG
ANNs
Machine Learning
Brain Decoding
Review
Neuroscience

Arthur Dehgan, Hamza Abdelhedi et al.

Teaching & mentoring

View courses & workshops

I develop courses, lead workshops, and mentor students in AI and neuroscience.

Hamza teaching beside representational similarity analysis slides in a classroom. — Teaching representational similarity analysis at MAIN Educational, Montréal, October 2024.
Tutorial materials

Brainhack Montréal · October 2, 2026 · Workshop instructor

Foundation models for M/EEG

A hands-on comparison of classical EEG decoding and foundation models, from pretrained representations to fine-tuning.

Mila · Summers 2024–2026 · Teaching assistant & instructor

Indigenous Pathfinders in AI

I teach programming and machine learning to First Nations, Inuit, and Métis participants through lectures and hands-on AI projects.

Open-source software

View all software

I build and contribute to tools that make brain-signal methods easier to inspect, reuse, and share.

CHU Sainte-Justine 2025–Present Research Collaborator

CHU Sainte-Justine

Jan 2025–Present

Research Collaborator

Le Centre hospitalier universitaire Sainte-Justine

Collaborating with Dr. Alexander G. Weil and Dr. Aristides Hadjinicolaou on clinical EEG research.
Abundant Intelligences 2024–Present Research contributor

Abundant Intelligences

2024–Present

AI Lab Member / Research Contributor

Abundant Intelligences

Abundant Intelligences; part-time since March 2026.
Université de Montréal 2024–Present PhD · Biomedical Engineering 2022–2024 MSc · Computer Science / AI

Université de Montréal

Sep 2024–Present

PhD in Biomedical Engineering

Université de Montréal

Supervisor: Prof. Karim Jerbi. Research affiliations: CoCo Lab and Mila.

Sep 2022–Aug 2024

MSc in Computer Science — Artificial Intelligence

Université de Montréal

Supervisor: Prof. Karim Jerbi; co-supervisor: Prof. Shahab Bakhtiari. Research affiliations: CoCo Lab and Mila.
CoCo Lab · Mila Graduate research assistant 2024–Present

CoCo Lab · Mila

Sep 2024–Present

Graduate Research Assistant

MILA - Quebec Artificial Intelligence institute

Cognitive & Computational Neuroscience Lab (CoCo Lab) and Mila; supervisor: Prof. Karim Jerbi.
SUP’COM, Tunisia 2018–2021 Telecommunications engineering diploma

SUP’COM, Tunisia

Sep 2018–Oct 2021

Engineering Diploma in Telecommunications — Smart Imaging Applications

Higher School of Communication of Tunis

University of Carthage — Higher School of Communication of Tunis (SUP'COM).