Overview
- The study published in iScience in early August used awake fMRI on trained pet dogs to test responses to human faces and found happy expressions activated the dogs' temporal cortex and caudate nucleus, areas tied to reward processing.
- A second experiment using whole‑brain machine‑learning pattern analysis on a larger set of dogs produced separable neural signatures for fear versus anger and sadness rather than a single good/bad split.
- Sample sizes were small — eight dogs in the first task and twelve in the second — and most animals were companion dogs, which limits how far the results can be generalized to all breeds or free‑ranging dogs.
- Authors and reporters stressed the study used static photos and noted dogs also rely on body language, voice and scent, so the team plans follow‑ups with moving, multisensory stimuli and cross‑species comparisons to test real‑world relevance.
- The research builds on prior work showing dogs' social attunement to humans and offers a proof‑of‑concept that neural discrimination of human expressions is possible while cautioning that neural signals do not prove identical subjective emotions across species.