Matteo Visconti di Oleggio Castello, Ph.D.

Cognitive Computational Neuroscientist investigating individual differences in brain function

I'm an Assistant Project Scientist in the Department of Neuroscience at the University of California, Berkeley, working in Jack Gallant's lab. I'm interested in how our brains build and represent meaning from the world around us. My current research focuses on how these representations differ between people. To study these questions, I build computational models to predict brain activity when we watch movies, listen to stories, or interact with each other. Then I break these models apart, and see if I can learn something interesting about the brain. Before coming to Berkeley, I received a Ph.D. in Cognitive Neuroscience at Dartmouth, working with Ida Gobbini and Jim Haxby. At Dartmouth, I studied how our brains represent the identity of our friends and colleagues. I used psychophysics and fMRI to study familiar face perception.

Latest News

2026-03-13 — New preprint: Hyperface: a naturalistic fMRI dataset for investigating human face processing is available on bioRxiv. We released a comprehensive fMRI dataset of 21 participants watching 707 unique face video clips, along with code and three OpenNeuro datasets: raw data, fMRIPrep derivatives, and FreeSurfer derivatives.
2025-12-16 — I released autoflatten, a Python pipeline for automatically creating cortical flatmaps from FreeSurfer surfaces. It features automatic cut mapping using surface-based registration (no manual cut placement required) and a new JAX-accelerated Python implementation of FreeSurfer's mris_flatten algorithm that's much faster than the original.
2025-11-18 — I presented our work leveraging voxelwise encoding models to understand altered functional representations in frontotemporal dementia at SFN 2025 in San Diego.
2025-09-17 — My latest paper on the Voxelwise Encoding Model (VEM) framework is available as a preprint on PsyArXiv. This paper provides the first comprehensive guide for creating encoding models with fMRI data, and complements our VEM tutorials.
2025-08-23 — My latest work on individual differences in conceptual representation is available as a preprint on bioRxiv. This paper introduces a new statistical framework to measure and interpret person-specific functional brain representations, establishing a new paradigm for precision neuroscience.

Selected Publications

Hyperface: a naturalistic fMRI dataset for investigating human face processing. bioRxiv, 2026.
The Voxelwise Encoding Model framework: a tutorial introduction to fitting encoding models to fMRI data. Imaging Neuroscience, 2025.
Individual differences shape conceptual representation in the brain. bioRxiv, 2025.
Encoding models in functional magnetic resonance imaging: the Voxelwise Encoding Model framework. PsyArXiv, 2025.
Model connectivity: leveraging the power of encoding models to overcome the limitations of functional connectivity. bioRxiv, 2023.
Shared neural codes for visual and semantic information about familiar faces in a common representational space. Proceedings of the National Academy of Sciences, 2021.
The neural representation of personally familiar and unfamiliar faces in the distributed system for face perception. Scientific Reports, 2017.

GitHub Stats

Pinned Loading

automatically create cortical flatmaps from FreeSurfer surfaces

Python 29 2
Pycortex is a python-based toolkit for surface visualization of fMRI data

JavaScript 664 149
Voxelwise Encoding Model tutorials from the Gallant lab.

Jupyter Notebook 101 18
Flexible DICOM conversion into structured directory layouts

Python 283 138
A setup for automatic generation of shareable, version-controlled BIDS datasets from MR scanners

Shell 56 16
Quality assurance analyses of fMRI data collected while participants watched The Grand Budapest Hotel by Wes Anderson.

Jupyter Notebook 30 3