Abstract
In human face-to-face communication, the content of speech is often illustrated by coverbal gestures. Behavioral evidence suggests that gestures provide advantages in the comprehension and memory of speech. Yet, how the human brain integrates abstract auditory and visual information into a common representation is not known. Our study investigates the neural basis of memory for bimodal speech and gesture representations. In this fMRI study, 12 participants were presented with video clips showing an actor performing meaningful metaphoric gestures (MG), unrelated, free gestures (FG), and no arm and hand movements (NG) accompanying sentences with an abstract content. After the fMRI session, the participants performed a recognition task. Behaviorally, the participants showed the highest hit rate for sentences accompanied by meaningful metaphoric gestures. Despite comparable old/new discrimination performances (d′) for the three conditions, we obtained distinct memory-related left-hemispheric activations in the inferior frontal gyrus (IFG), the premotor cortex (BA 6), and the middle temporal gyrus (MTG), as well as significant correlations between hippocampal activation and memory performance in the metaphoric gesture condition. In contrast, unrelated speech and gesture information (FG) was processed in areas of the left occipito-temporal and cerebellar region and the right IFG just like the no-gesture condition (NG). We propose that the specific left-lateralized activation pattern for the metaphoric speech–gesture sentences reflects semantic integration of speech and gestures. These results provide novel evidence about the neural integration of abstract speech and gestures as it contributes to subsequent memory performance.