In the framework of the “Fruits of Ethology” guest talk series we will have the pleasure to welcome our next guest speakers, Laura V. Cuaya and Raúl Hernández-Pérez (Instituto de Neurobiología, Universidad Nacional Autónoma de México).
Date: 11th July, Monday, 10.15 am
Location: ELTE, Department of Ethology – South Building, Pázmány Péter sétány 1/c, 6th floor
Title: Processing of faces in dogs through fMRI
Laura V. Cuaya
Dogs are a unique model in the study of face recognition, because they not only have the capacity to discriminate between dog faces but they have also developed a remarkable capacity to extract valuable information from human faces. We have explored the cerebral correlates of face processing in seven dogs through fMRI. The stimuli were images of objects, human faces and dog faces. We found a posterior to anterior pattern of cerebral activity depending of the kind of visual stimuli: occipital cortex responds to all categories, while temporal cortex responds to all facial stimuli, regardless of species, and the frontal cortex responds preferentially to human faces. Our findings are consistent with the importance of temporal cortex in face processing in humans, non-human primates and sheep. And the activity in the frontal cortex related to human faces could be the brain correlate of the differential behavioral patterns that dogs show towards human faces.
Also because of the importance of human emotions for dogs we explored the cerebral correlates of one emotion in eight dogs: happiness in faces. We found brain activity related to happy human faces in several regions, including temporal cortex, frontal cortex and caudate. All the activity was found in the right hemisphere. We think that these studies can contribute to an integral understanding of the foundations of social cognition in dogs.
Title: Predicting dog’s perception: Multivariate pattern analysis as tool to study the process of emotions in faces
Dogs use emotional cues from humans to guide their behavior. There is behavioral evidence that shows that dogs are capable of discriminate expression of emotions in human faces. We explored the brain correlates of perception of emotions in human faces in dogs using fMRI.
We used a block design with four emotions in humans: happiness, sadness, fear and anger. Four dogs participated in this experiment; for each dog we acquired ten runs. To analyze the data, we employed a new analysis technique: multivariate pattern analysis (MVPA). The traditional analysis method in fMRI matches a stimulus with a change in the activity of an individual voxel, so if the voxel increments its activity is then considered to be related to the stimulus being presented, this approach has pointed out regions that are related to specific cognitive states, however, the sensitivity of this method relies on the intensity of the response of a single voxel to a cognitive state. However, we know that a cognitive state is not only the sum of activity of single brain pieces, but in fact, emerges from the collective activity of several brain regions. MVPA takes this into account, this analysis uses the pattern of activity of several voxels at the same time. It considers not only the increase of activity, but small changes in multiple voxels that can inform about a cognitive state. The analysis uses machine learning techniques not only to assess the relationship between a stimulus and a brain region, but to predict the stimulus being presented to the participant.
We were capable to predicted above of chance (p < 0.05) from the pattern of activity of all brain the emotion that dogs was observed. Our findings show that at the cerebral level dogs can discriminate basic emotions in human faces, this maybe reflect an adaptations of dogs to anthropogenic niche. Beyond the importance of these results to understanding of emotional processing in dogs, we introduce the MVPA as a useful tool to explore the cerebral correlates of dog’s cognition.