Mehran Spitmaan

The paper develops a pre-trained transformer model to predict cancer tissue of origin from bulk gene expression data, achieving high accuracy across 32 tissue types.

This paper adapts a pre-trained transformer model for bulk tumor data to predict patient survival outcomes, achieving strong correlations and outperforming traditional methods.

This study examines how training affects timescales of neural dynamics in primate prefrontal cortex and their role in working memory.

The paper introduces entropy-based metrics to predict choice behavior in mice and monkeys, explaining 50% and 41% of matching variance, and develops a more accurate model based on these metrics.

The paper estimates multiple timescales of neural dynamics and their integration of task-relevant signals across cortex, finding increasing timescales from parietal to prefrontal to cingulate cortex and no correlation with signal selectivity.

The paper investigates how low-level neural processes and high-level cognitive processes interact to influence context-dependent choice in humans, finding that both contribute to distinct behavioral patterns.

The paper investigates how humans evaluate complex risky options by showing subjects use integrated value functions for individual outcomes and differential weighting based on reward magnitude or probability for overall value construction.