Predicting Receptivity to Health Messaging with fMRI

Differentiating the neural representations of self-relevance and valence is important because perceptions of message relevance and message valence may distinctively predict message effectiveness. In this study, I used a data-driven method to identify separate patterns of neural co-activation indexing judgments of self-relevance and valence and then explored how these patterns can be used to predict receptivity to health messaging.

Using fMRI data from a localizer task that asks participants to judge the self-relevance and valence of trait adjectives, I conducted two separate partial least squares (PLS) analyses. Each PLS analysis contrasted either self-relevance or valence judgments with a letter case judgment control task. Despite high overall similarity between the resulting maps, r=.95, I identified a cortical midline pattern (VMPFC/PCC) that showed maximal divergence between the self-relevance and valence judgment patterns (see Figure). For each whole-brain and midline-masked map, I computed trial-wise dot product similarity to betaseries images reflecting neural responses to messages in a separate messaging task promoting physical activity in 158 sedentary adults. I then used these trial-wise similarity scores to predict message ratings and subsequently measured changes in everyday physical activity.

Similarity to either whole-brain PLS map (i.e., for self-relevance or valence) was associated with: (1) lower in-scan ratings of relevance of messages, and (2) less post-scan message agreement. In sum, similarity to whole-brain neural patterns tracking deliberation about the self-relevance or valence of information was associated with reduced message receptivity, suggesting possible counterarguing. Within the midline-masked images, I found that similarity to the masked valence judgment map was generally associated with lower message receptivity, as in the whole-brain similarity analyses. However, trial-wise similarity to the masked self-relevance judgment map was associated with increased message receptivity.

These findings suggest that increased coactivation between the VMPFC and PCC that arises during deliberate judgments of self-relevance may be part of a neural signature underlying message receptivity. Moving beyond a focus on mean activity in singular brain regions, these findings highlight potential predictive utility of commonalities and differences in the neural patterns of coactivation underlying judgments about self-relevance and valence. Specifically, coordinated neural activity within the VMPFC/PCC resembling patterns that arise during judgments about self-relevance (and not valence) are associated with increased receptivity to message-based interventions promoting real-world physical activity.