Tom Dupré la Tour

This paper develops a statistical framework to measure individual differences in conceptual brain representations using fMRI and finds that these differences are systematic and highest in social information processing regions.

The paper develops a hierarchical convolutional energy model to predict V4 neuron responses to natural videos and finds it captures complex spatiotemporal selectivity in V4.

The paper examines whether language timescale representations in the brain are shared between reading and listening using fMRI data.

This paper introduces model connectivity as a method to recover functional brain networks by analyzing encoding model weights, showing it outperforms functional connectivity in noise robustness and interpretability.

This study shows that context significantly affects the quality of neuroimaging data and the representation of semantic information in the brain during language comprehension.

This paper introduces banded ridge regression for feature-space selection in brain recordings, improving prediction accuracy and interpretability by adapting regularization per feature space.

The paper investigates how phase-amplitude coupling in the alpha-beta band affects the precision of self-generated timing tasks.

This paper introduces non-linear auto-regressive models to detect and quantify cross-frequency coupling in neural time series, showing improved performance over traditional methods.