Topographic maps
How maps of the sensory world and of the body organize processing throughout the brain.
Knapen Lab
Computational cognitive neuroscience at the Vrije Universiteit Amsterdam and the Spinoza Centre for Neuroimaging
We study how the brain's topographic maps organize perception, action and social cognition. We combine intensive, high-resolution neuroimaging with computational models that link brain activity to the sensory world and to behavior.
Research
All research
Precision neuroimaging
Intensive, high-quality measurements in individual brains, collected to be shared as open datasets.
Computational models
Encoding models that connect sensory input, brain responses and behavior.
Perception and awareness
How the brain resolves ambiguity, and how attention and arousal shape what we perceive.
Selected work
All publications
prfmodel: population receptive field modeling in Python
prfmodel fits population receptive field (pRF) models and related models, such as connective field and contrast sensitivity models, on GPU-accelerated backends (TensorFlow, PyTorch or JAX). It is developed by the Popylar team with research software engineers of the Netherlands eScience Center.
pip install "prfmodel[tensorflow]"Version 0.2.0 (beta) was released for our hands-on workshop in Coimbra, 5–6 November 2026. The practicals and the slides of several lectures are public. About the workshop.
Arrow of Time dataset
An intensive 7T fMRI dataset of brain responses to 2,160 brief naturalistic videos, with about 21 hours of scanning per participant at 1.7 mm isotropic resolution, plus retinotopy and motion-energy localizers for every participant.
The first release is planned around the turn of the year (2026/27).
News
All news-
New project on body-based meditation and psilocybin
The Tiny Blue Dot Foundation funds a new project in the lab on body-based meditation and psilocybin. Xaver Funk joins as a postdoc in 2027.
-
Popylar workshop on pRF modeling in Coimbra
A two-day hands-on workshop on population receptive field modeling with the new prfmodel Python package, at the University of Coimbra.
-
PLoS Computational Biology paper combining fMRI and MEG published
A forward model that combines pRF maps from fMRI with MEG estimates when each visual field map contributes to the measured response.
Software & data
All software and dataprfpy
AvailablePython package to simulate and fit population receptive field (pRF) models to time series data, including the divisive normalization pRF model. prfmodel, developed in the Popylar project, builds on it.
nideconv
MaintainedPython package to estimate the shape of responses in fMRI, pupil and other time series by deconvolution, including hierarchical Bayesian estimation.
Cerebellar retinotopy atlas
AvailableAtlas of the retinotopic maps in the human cerebellum (oculomotor vermis, lobule VIIb and lobule VIIIb), derived from Human Connectome Project retinotopy data, in 2 mm MNI space, with the code that produced it.