Bilateral auditory and left-dominant premotor activity in prosodic focus processing

Résumé En

Introduction: Contrastive focus is used to emphasize a constituent in an utterance as opposed to another. In French, it can be conveyed by prosody using a specific intonational contour on the constituent pointed at (THOMASF a mangé la pomme. 'THOMASF ate the apple.'). The studies of the neural correlates of the perception of prosody have led to different conclusions (see Baum et al., 1199 for a review). Early works claimed that prosody is processed in the right hemisphere (e.g. Brådvik et al., 1991). Other studies have shown that prosodic processing cannot be restricted to the right hemisphere, however (e.g. Astésano et al.,2004). In this framework, this fMRI study deals with the perception of prosodic contrastive focus in French, according to the hemispheric specialization. Methods: An fMRI experiment was performed on 22 French adults (all were right-handed according to Oldfield, 1971). Two conditions were compared: 'sentences with narrow contrastive focus' (Focus condition, F) and 'sentences with broad focus' (Neutral condition, N). All sentences had the same syntactic and syllabic structure (e.g., 'Thomas cassait le vélo'. 'Thomas broke the bike'). The subjects were asked to judge whether the audio stimuli contained contrastive focus. The responses were recorded and the performance of task execution was evaluated. Functional MR imaging was performed on a Bruker 3T MR imager. Data analyses were performed by using SPM5 (www.fil.ion.ucl.ac.uk/spm). A random-effect group analysis on the F vs. N contrast images from individual analyses by using a one-sample t-test has been subsequently performed. We identified clusters of activated voxels (p < 0.001, uncorrected, height threshold: T = 3.53, K > 15). In addition, we defined seven Regions of Interest based on the whole-brain activation obtained by contrasting F vs. N in all subjects. ROIs were delineated in the left and right hemispheres symmetrically. For each ROI, the parameter estimates values were analyzed by means of a repeated ANOVA with hemisphere as a within-subject factor. Results: Behavioral responses obtained during the fMRI experiment were correct on most trials: Focus (M = 92.99%, SD = 6.73%) and Neutral (N = 97.72 %, SD = 3.85%). Our fMRI results showed significant left hemisphere predominance of the inferior frontal (BA 47) and supramarginal (BA 40) gyri, as well as for left anterior insula (BA 13) in the F vs. N contrast (c.f. Figure 1). Conclusions: Contrary to many brain imaging studies, these results suggest that both hemispheres participate in the auditory perception of prosody, but with a left-dominant contribution for morpho-syntactic processes and thematic role monitoring.