Bálint Kincses

This study develops a brain-based biomarker, PARES, that predicts both pain increases and relief in individuals using fMRI data.

This paper introduces RLBF, a framework for real-time fMRI-based adaptive stimulus generation using reinforcement learning, validated in a study optimizing visual stimuli to enhance V1 responses.

The paper uses high-resolution fMRI to show that sensory and cognitive influences on pain are processed in distinct cortical layers, linking circuit organization to behavioral outcomes.

This paper evaluates the replicability of diffusion-weighted MRI-based brain-behavior models and finds moderate replicability with small sample sizes, but higher replicability for trait-like phenotypes.

This study compares neural mechanisms of placebo analgesia induced by verbal suggestions and conditioning, finding distinct brain activity patterns and enhanced analgesia when combined.

The paper proposes an adaptive splitting method for external validation of machine learning models to improve generalizability and replicability in predictive modeling studies.

This paper introduces fcANNs, a model of brain dynamics based on functional connectivity, showing it can reconstruct and predict brain activity in rest, tasks, and disease states.

This study develops a multi-center machine learning model using gray matter cortical thickness to predict individual pain sensitivity, achieving a moderate predictive performance.

The paper identifies a resting-state brain connectivity pattern predictive of individual pain sensitivity without painful stimulation.

The paper examines how motion artifacts may bias parameter choices in functional connectome-based predictive models, suggesting the need for careful consideration of sparsity and non-neural confounds.