Overview
- The peer‑reviewed study, published in Science on Sept. 24, 2026, recorded noninvasive EEG from 20 companion dogs and 20 adult humans as they listened to continuous three‑syllable nonsense streams.
- Both species showed stronger neural entrainment to word rhythms when consonant patterns defined the words, and both exhibited a roughly 400‑millisecond brain response associated with segmenting continuous speech into word‑like units.
- The consonant bias appeared in dogs regardless of early‑life exposure to human speech, including animals that spent their first three months in shelters or on the street, which the authors say points to statistical learning from regular exposure rather than a narrowly timed developmental window.
- Key differences remain: humans synchronized to finer speech‑sound rhythms and detected vowel‑based patterns better, while dogs processed speech at a coarser, syllable level, and the study notes limits such as a 20‑dog sample, restricted stimulus sets, and unresolved questions about domestication versus older mammalian auditory mechanisms.
- If replicated across breeds and contexts, the finding would mean routine exposure can reshape auditory processing in nonhuman animals, it may help explain how pets pick out words from conversation, and it sets out clear next steps for testing generalizability and evolutionary origin.