Our results show that the head-tilt behavior observed in dogs in response to human cues is right-lateralized, consistent with left-hemisphere lexical processing. Specifically, we found that dogs tilted their heads more to the right in response to familiar speech, and male dogs tilted more frequently than female dogs in a right-lateralized manner.
The only known prior work on the head-tilt behavior of dogs reported a relationship between head-tilt and the processing of meaningful stimuli10. In that study, individual dogs’ head-tilts were consistent across experiments, but did not show consistent evidence of population-level lateralization. However, data on the direction of the head-tilt were limited to six dogs of one breed. In contrast, the present study includes over 40 dogs that performed the head-tilt behavior across all seven breed groups, plus mixed-breed dogs.
Work utilizing the head-turn behavior has suggested lateralization of processing human vocalizations in the canine brain, but produced contradictory results regarding hemispheric bias for processing familiar words8,15. The majority of head-tilts in the present study occurred in Condition 4, where owners spoke in a neutral to happy tone using words their dog was familiar with. The recorded number of head-tilts in all other conditions were not significantly different from zero (including Condition 3, in which the owner was speaking for the same duration of time, but utilizing unfamiliar words). Our results reveal that differences in owner communication between conditions had an effect on the frequency of the head-tilting behavior, consistent with the hypothesis that the head-tilt behavior is related to the processing of meaningful speech-stimuli, specifically familiar words and phrases.
Because only Condition 4 contained enough head-tilts for further statistical comparisons, directionality and sex differences were explored in this condition only. There were significant sex differences in head-tilts between neutered males and spayed females, with neutered males-tilting their heads significantly more often. In humans, the processing of semantic information occurs primarily in the left hemisphere for both sexes. However, females have been reported to show more bilateral activation during these tasks16,17. The results of this study demonstrate that dogs follow this pattern as well - if males show a more obvious pattern of left-hemispheric lateralization in auditory processing, it is likely they more prevalently display an asymmetrical behavior related to it. Why unaltered males and females did not display significant differences (c.f. Supplementary Materials), however, remains unclear.
Orienting asymmetries have evidenced hemispheric biases in primates, pigs, cats, and goats; largely suggesting a left-hemispheric bias for processing species-specific, emotionally relevant auditory stimuli18,19,20,21,22,23. Assuming, as we do, that the head-tilt behavior in dogs indexes lateralized brain processing, the results of the present study provide evidence that dogs process familiar human speech in the left hemisphere of the brain. These results concur with findings from Barton et al.24, which discovered leftward asymmetry of canine brain regions that are activated during human speech, and Boros et al.25, who concluded dogs’ left basal ganglia was activated for word segmentation in a word learning task. Left-brain lateralization in dogs in conjunction with the hemispheric lateralization patterns seen in the human literature suggest that dogs process familiar words not just as affective cues, but similarly to how humans process non-emotive language.
Limitations and alternative explanations
Prior work has suggested lateralization of brain patterns in response to non-verbal vocalizations of different human emotions, with the right hemisphere being associated with fear and sadness, and the left associated with happiness26. Because the familiar words and phrases used in the present study were chosen by the owners and almost all had a positive connotation, we cannot rule out that the left-hemispheric bias recorded here was due to asymmetric emotional modulation of the canine brain. Future research should explore if verbal vocalizations elicit the same hemispheric biases as these non-verbal vocalizations.
Data collection was done throughout the COVID-19 pandemic utilizing community science, meaning no in-person collection could be performed. See Supplementary Materials and Sexton et al.13.
In our analysis of breed groups differences of head-tilt occurrence, there were multiple breed groups made up of fewer than ten individuals. These results should be interpreted with caution as statistical power to detect breed difference was low.
We also cannot rule out the possibility that condition sequence, in which speaking familiar words always occurred last, contributed to our results. While owners were instructed to allow at least 30 minutes between the filming of conditions, it is possible that the longer dogs are engaged with, the more likely they are to tilt their heads. We suggest future research reverse or randomize the order of the conditions or manipulate the duration of engagement using dog-directed speech.
Alternative potential explanations for observed results regarding sex differences include potential differences in lateral bias between sexes. One study of lateral biases in paw preference observed significantly more male dogs being classified as right-pawed, while more females were more likely to be classified as ambilateral than left or right-pawed27. Our data is consistent with the idea that male dogs are right side dominant, as we saw male dogs tilting their head more frequently to the right than to the left. If female dogs are truly more ambidextrous, recording their head-tilting data as binomial would have underrepresented the frequency of the behavior.