JohnMark Taylor

This paper introduces framed RSA, a method that evaluates neural models by incorporating both geometric patterns and population-mean response preferences, improving model selection accuracy in brain-region and network identification.

The paper introduces an open-source Python toolbox for representational similarity analysis, providing tools to evaluate neural representations and models.

TorchLens extracts and visualizes hidden activations of PyTorch models, providing full computational graph records and validation for analysis.

This paper uses fMRI and multivariate pattern analysis to determine whether person-knowledge for faces is localized or distributed, finding that unique information for person-knowledge is in vATL and mid-fusiform but only vATL represents explicit recall.

This study uses fMRI to examine how color and form are coded in the human ventral visual pathway, finding biased distributed representation and relative independence between color and form coding.

This paper examines how convolutional neural networks code color and shape features, finding stronger interaction in trained networks compared to untrained ones.