Xinyuan Yan

The paper investigates how bilinguals represent semantic concepts in two languages, finding shared neural geometries that support translation despite language-specific neuron responses.

This study uses large language models to analyze neural responses in autistic individuals, finding reduced contextualization signatures compared to controls.

This study identifies distinct functional roles of human amygdala subregions in fearful-expression recognition through intracranial EEG and stimulation, revealing bidirectional interactions and a temporal hierarchy.

The paper maps semantic organization in the human hippocampus, finding semantotopic patterns and multiscale semantic similarity along the mediolateral axis.

The paper shows hippocampal neurons encode meanings of spoken and heard words using shared geometric embeddings, enabling abstract cross-personal representations while preserving speaker identity binding.

The paper shows that human exploration behavior follows a U-shaped pattern with the ratio of volatility to stochasticity in uncertain environments, with optimal performance at intermediate uncertainty ratios.

The paper identifies a dominant nonlinear axis of individual variability in decision-making tasks, showing that exploration tendency, rather than personality or learning parameters, drives differences in performance.

This study identifies distinct computational mechanisms of uncertainty processing in anxiety and apathy, showing how they influence exploratory behaviors through different perceptions of volatility and stochasticity.

The paper identifies a low-dimensional behavioral manifold that captures exploration-exploitation decision-making and links it to symptoms of apathy and anxiety.