A selected collection of datasets, research initiatives, and software that I have helped develop or that were developed in my lab. For a complete and more frequently updated overview of our code, please visit the Vision and Computational Cognition Group on GitHub.
Featured resources
A global research initiative built around a shared database of 1,854 object concepts and 26,107 naturalistic images, together with open behavioral and neural datasets.
A densely sampled 7T-fMRI dataset designed to cover a broad distribution of natural images.
A community-driven replication and generalization initiative for visual neuroscience, built on LAION-fMRI.
A Matlab toolbox for multivariate analyses of functional and structural MRI data that I developed with Kai Görgen and John-Dylan Haynes. TDT supports decoding, representational similarity analysis, feature selection, and SPM and AFNI workflows.
Software and computational methods
SRF — Similarity-based Representation Factorization recovers interpretable dimensions directly from complete or partially observed similarity matrices.
SPoSE — Sparse Positive Similarity Embedding learns interpretable object dimensions from large-scale behavioral similarity judgments.
VICE — Variational Interpretable Concept Embeddings extends this approach with uncertainty estimates and automatic selection of informative dimensions.
DimPred — A Python package for predicting perceived similarity of new images through interpretable dimensions.
THINGSvision — A Python package for extracting representations from a broad range of state-of-the-art computer-vision models.
Additional analysis code and project repositories are available through our GitHub organization. Earlier resources remain available in the archive.