Close
Abstract
Introduction
We use language to communicate information, but a propositional or linguistic analysis does not recognize all the information available in spoken language. Emotional content is often conveyed in discourse, whether explicitly or implicitly. Virtually all social interactions involve some degree of emotional information, and understanding speech must involve processing and evaluating its emotional content. Previous biological studies of language and emotion have assumed that the emotional content of speech can be assessed at the single-word level. By presenting participants with stories with measurable emotional content, the present study aimed to investigate human language understanding in a broader context. We hypothesized that attending to stories high in positive or negative emotional ratings would differentially recruit brain regions important for both language and emotional processing compared to stories with lower ratings.
Methods
Eleven right-handed native English speakers (6 female) listened to stories presented in an audiovisual recording of a storyteller during spiral BOLD fMRI at 1.5T. Stimuli were separate stories of approximately 24 seconds that naturally varied along several emotional dimensions. To provide ratings on the emotional content of the stories, a separate group of 17 participants rated the degree to which each story produced different emotions (amused, happy, relaxed, upset, sad, bored). Stories were then assigned into two groups reflecting high and low levels for comparison within each emotion. A deconvolution analysis was used to reveal voxel-wise signal change for each condition, and the individual data were transformed into stereotactic coordinates, for analysis using a group ANOVA (p < .005, corrected).
Results
The results described here reflect statistical comparisons between hemodynamic responses to stories rated high and those rated low for each particular emotion. First, stories rated high in the “happy” condition demonstrated large clusters of activation in Broca’s area of the left inferior frontal gyrus, as well as left insula and left parahippocampal gyrus, possibly extending to the amygdala. Stories rated low for “sad” showed greater activation in the right middle frontal gyrus, left superior temporal gyrus, bilateral insula, and right putamen. In addition, stories rated low in the “upset” emotion revealed clusters in the right middle frontal gyrus and left insula. No clusters were found in comparing the “amused”, “bored”, or “relaxed” comparisons at this level of significance.
Conclusions
Areas previously implicated to be important to language processing demonstrated greater activity in response to stories that were rated higher in positive emotion and lower in negative emotion although linguistic demands on the listener did not change. These stories also activated regions involved in emotional processing of stimuli, providing information regarding how the brain processes emotional speech. The insula projects to the amygdala and striatal structures, including putamen, providing a means for evaluating the stimuli for emotional information. The active recruitment of Broca’s area suggests that this region is sensitive to the emotional content of language. The greater activation in response to positive stories provides evidence that listeners were more actively processing these stories, both in emotion and language regions.