Particle.news

Dogs' Brains Segment Speech Using Humanlike Consonant Bias

Researchers say regular exposure to human speech may be sufficient to produce this pattern without requiring human-specific language wiring.

Overview

  • A peer-reviewed paper published on Thursday, September 24 found that companion dogs, like adult humans, preferentially use consonant patterns to break continuous nonsense speech into word-sized units.
  • The team recorded noninvasive EEG from 20 companion dogs and 20 adults as subjects listened to three types of continuous three-syllable streams—consonant-structured, vowel-structured, and random—designed so vowels were acoustically louder to rule out a simple volume effect.
  • Both species showed stronger neural entrainment to consonant-structured streams and a distinct event-related change about 400 milliseconds after word onset, a timing previously linked to speech segmentation in humans.
  • The consonant bias appeared in dogs regardless of early-life exposure to human speech, while humans showed finer-grained processing by also detecting vowel-defined patterns and synchronizing to speech-sound rhythm.
  • Authors and press coverage note key limitations and next steps: the sample was modest and limited to companion dogs, so further replication and broader species sampling are needed to test whether this reflects domestication, general mammalian auditory mechanisms, or experience-driven neural plasticity.