Fahd Yazin

We found three specialized computations in the midline prefrontal cortex across its subregions while participants learned virtual worlds. These were specialized for probabilistic inference of states (vmPFC), organizing these states into orthogonal coordinate frames (amPFC), and transitioning between these (dmPFC). These computations generalized across domains (spatial, social, and sequential) and phases (learning, inference), offering a unifying account of midline PFC function across tasks and domains.

We found that the midline prefrontal cortex simultaneously predicts changes in context, beliefs of others, and future changes. These fragmented predictions are integrated with sensory information within the Default-Mode Network to form a unified subjective experience. The precuneus toggles between various prediction threads that best explain sensory input. People who predict and integrate similarly share more similar experiences. This work offers a framework for understanding how shared realities and individual differences in perception may arise in the human brain.

We found that continuous dynamics of emotions are best explained by uncertainty in one’s predictions. Using a naturalistic movie paradigm and computational modeling, we show that the lateral orbitofrontal cortex tracks the ongoing uncertainty in these predictions, and this tracking correlates with participants’ anxiety scores. This work provides a computational framework for understanding the orbitofrontal cortex in the context of anxiety disorders and its role in representing predictive uncertainty.

We found that neural variability in the orbitofrontal cortex during emotional experiences carried unique information about aging. This effect was not present in average neural responses or spontaneous variability and was specific to this prefrontal region. The latent structure of neural responses mirrored participants’ emotional experience and its alteration with age. This work provides a simple, naturalistic approach to linking neural variability to emotional processing, advancing the study of neural volatility in aging.