Bruno Direito

This paper proposes a structurally constrained message-passing framework for brain decoding on small datasets, achieving competitive performance while maintaining interpretability.

The paper examines how signal quality affects inter-individual variability in fNIRS data and introduces an open dataset with multi-task recordings and peripheral physiology measures.

This paper develops a music-based real-time fMRI neurofeedback interface that modulates interhemispheric connectivity and improves mood in healthy adults.

This paper shows that neurofeedback therapy for autism improves facial expression recognition by increasing connectivity in higher-level visual regions and reducing connectivity in lower-level regions, correlating with clinical improvements.

This paper introduces a shallow neural network with structural connectivity message-passing for brain decoding, achieving 83.0% classification accuracy using fMRI data from 30 subjects.

The paper uses multivariate pattern analysis to decode emotion categories from brain activity during self-selected music listening, revealing consistent neural correlates across auditory, motor, and interoceptive regions.

This paper develops and tests a multichannel electrocutaneous device for non-invasive somatosensory stimulation during fMRI, confirming its safety and ability to elicit expected brain activity.

The paper uses fMRI to decode music-evoked valence and arousal by analyzing neural correlates and musical features, finding expressive features modulate reward and emotion networks.

This paper evaluates State-Space Granger Causality with Time Reversed Testing on simulated BOLD data, finding it outperforms classical methods in detecting causal relationships.