A study conducted by researchers at the Neuroethology of Communication Lab at ELTE Department of Ethology in Budapest shows that dog brains segment speech into words similarly to humans. This finding suggests that efficient speech processing may not be exclusive to humans' unique language abilities but can occur through regular exposure to speech, even among species evolutionarily distant from humans. The research was published in the journal Science.
Read More
Historically, the question has been whether humans process speech efficiently due to inherent brain structures or due to constant exposure to speech. "Words consist of two main types of speech sounds: vowels and consonants. Consonants typically form the framework of words, and studies indicate that the human brain focuses on consonants from infancy," stated Attila Andics, the cognitive neuroscientist leading the study.
The researchers sought to determine if a preference for consonants requires uniquely human language capabilities or can arise from simpler pattern learning, akin to processes observed in non-verbal species. They compared how 20 humans and 20 companion dogs responded to uninterrupted speech streams containing alternating consonants and vowels.
In this experiment, participants listened to three types of speech streams: one with recurring consonant patterns forming word structures, another with recurring vowel patterns, and a third with no discernible patterns. To measure brain activity, non-invasive EEG was employed, recording electrical activity through scalp electrodes. The researchers expected species showing a preference for consonants to synchronize brain activity more readily with the rhythm of consonant-structured words.
Results demonstrated that both dog and human brains synchronized more effectively to the rhythm of consonant-structured words compared to vowel-structured or control streams. Kinga G. Tóth, a psychologist and co-first author of the study, noted this revealed the first evidence of a consonant bias in non-human species.
Additionally, around 400 milliseconds after the onset of consonant-structured words, distinct changes in brain activity were observed in both species, indicating segmentation of speech into words. Importantly, this consonant bias was evident in dogs raised in various environments, including those from shelters or the street, suggesting that it does not necessitate pre-existing language processing mechanisms developed early in life.
While humans could also identify words defined by vowel patterns, dogs did not show the same capability, indicating that humans might better handle non-preferred linguistic structures. Furthermore, both species detected syllable rhythms, but only humans synchronized with the rhythm of individual speech sounds. This implies that the human brain can process continuous speech at a finer level of detail than dogs.
Overall, the findings indicate that tuning to speech and the consonant bias do not rely on pre-wired brain functions or human-like language abilities, but may instead develop through simple patterns of exposure to speech.