Neural representation of contextual hedonic valance across the lifespan

Affective experiences are an integral part of human life and significantly contribute to giving our existence meaning and purpose. They are a healthy part of our cognitive world, providing essential sources of motivation and teaching us to avoid harmful situations. Affect dysregulation is linked with significant and potentially deadly consequences, such as depression leading to suicide, a major cause of mortality with around 46,000 deaths per year in the US.

Identifying the physiological mechanisms associated with affect encoding in the brain is critical. Much is already known about this process with most of this knowledge coming from carefully controlled laboratory experiments performed in narrowly defined samples. It is now becoming increasingly pressing to study affect in a diverse population and in more naturalistic contexts, such as listening to narratives.

This overarching objective has direct and long-lasting benefits for society, both for our general understanding of how the brain works and for clinical applications associated with emotional dysregulation. This project addresses this need by using a large neuroimaging database to investigate how emotions are coded in the brain in situations similar to daily life.

This project analyzes a dataset that includes functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) recordings collected during narrative listening, as well as emotion assessment for a large and demographically diverse sample of adults between 20 and 80-years of age. It seeks to investigate the representations of affect (valence and arousal) across brain regions and frequencies from fMRI and EEG as a function of age and sex and relates these representations to individual differences in emotions.

First, this project relies on valence and arousal ratings to characterize the affective states associated with the narratives in a matched sample and compares these ratings with predictions from GPT and linguistic databases. Second, it attempts to map fMRI activations associated with affect during narrative listening. Third, it leverages of the high temporal resolution of EEG to refine our understanding of the spectral property of affect encoding.

Finally, it will attempt to develop a model of affect to uncover the mechanisms and causal relationships in the affective neural networks. This project is novel in the combination of elements it brings together, such as investigating the effect of age and sex on the representation of affect with dynamically unfolding stimuli, providing a more naturalistic evaluation of affect than traditional event-related paradigms.

It also relies on GPT to automate valence evaluation in narrative and uses EEG, fMRI, and modeling of the brain to push the current boundaries in our understanding of emotions. By investigating how affect is encoded in neurotypical individuals and naturalistic situations, this project seeks to build a foundation for future projects to understand better affect dysregulation and devise more effective preventive and therapeutic strategies.

This project adheres to open science principles and offers a multidisciplinary environment for students from underrepresented groups in STEM.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.