Brain Scans Predict Which Wildlife Photos Drive Conservation Donations

When a person glances at a wildlife photo on social media and decides whether to tap the heart icon or click a donation link, the decision feels instantaneous and personal. But two overlapping brain regions appear to govern that choice - and their activity can forecast, with measurable accuracy, how thousands of strangers will respond to the same image.

That is the central finding from a brain imaging study led by researchers at Stanford University, published in PNAS Nexus in February 2026. The work scanned 34 adults while they viewed 56 wildlife images drawn from National Geographic's Instagram feed, asking each participant to decide in real time whether to "like" a post or donate money to protect the depicted species. The results were then compared against actual engagement data from the same social media account.

Two Brain Regions, Two Distinct Signals

The nucleus accumbens, a structure associated with reward anticipation and motivation, and the medial prefrontal cortex, which plays a central role in social cognition and mentalizing - inferring the mental states of others - both predicted individual choices to like images and donate. That dual involvement points to something specific about how wildlife imagery works on viewers: it triggers both a reward-seeking impulse and a social reasoning process simultaneously.

The medial prefrontal cortex result proved particularly useful for predicting large-scale behavior. When the researchers averaged activity in that region across all 34 participants for each image, they could forecast how those same images would perform on Instagram, measured by likes relative to follower counts. The signal from inside individual skulls, aggregated across a small laboratory sample, translated to actual public behavior at scale.

A further analysis found that medial prefrontal cortex activity correlated with brain regions involved in face processing. This led the researchers to code all 56 images for whether an animal face was visible and for the phylogenetic distance of the depicted species from humans. Both variables predicted engagement. Mammals - closer to humans on the evolutionary tree - and animals photographed with their faces clearly visible consistently drove stronger responses.

Testing the Model on New Images

The researchers then applied a neural-inspired model to 276 additional wildlife images from the same National Geographic feed that participants had never seen. The model successfully forecasted engagement for this broader set, suggesting the predictive relationship is not specific to the original 56 images but reflects something more general about how socioemotional visual features activate neural pathways tied to generosity and social engagement.

"If you want to encourage people to protect an animal, you might depict it in a way that evokes a social or emotional connection," said Brian Knutson, a professor of psychology at Stanford and co-author of the study. "For instance, emphasizing facelike features or attention to the viewer."

Lead researcher Tara Srirangarajan noted that the findings could inform how environmental organizations design visual content: prioritizing images of mammalian species and choosing photographs that show an animal's face rather than its body may produce measurably stronger responses in terms of both engagement and donations.

Limitations Worth Noting

The sample of 34 adults is small, and the participants were likely drawn from a university population - relatively young, educated, and already engaged enough with wildlife content to participate in such a study. Whether the same brain patterns would appear in demographically different populations is unknown.

The engagement metric used - likes relative to follower counts - captures one narrow dimension of what conservation organizations actually want, which includes sustained donations, volunteer recruitment, and policy support. A photograph that performs well on Instagram may not translate to a meaningful increase in conservation funding.

The study also creates a tension it does not fully resolve. The species most threatened by habitat loss are not necessarily the ones that most resemble humans or have the most expressive faces. Charismatic megafauna already dominate conservation imagery - the neural data confirms this is not irrational, but it also means that the species generating the strongest neural responses may not be the ones most in need of conservation attention.

The researchers noted that future work using generative AI to modify image features could allow systematic testing of whether neural-predicted variables can be deliberately optimized to increase engagement and charitable giving.