The influence of sentence novelty and figurativeness on brain activity

Abstract

The predominance of the left hemisphere in language comprehension and production is well established. More recently, the right hemisphere's contribution to language has been examined. Clinical, behavioral, and neuroimaging research support the right hemisphere's involvement in metaphor processing. But, there is disagreement about whether metaphors, in and of themselves, engage the right hemisphere or if other factors that vary between metaphors and literal language elicit right hemisphere engagement. It is important to disambiguate these issues to improve our basic knowledge of figurative language processing, to more precisely define how the right hemisphere supports language, and to facilitate our ability to understand and treat language impairments. Here we investigated the role of the right hemisphere in language comprehension with functional magnetic resonance imaging (fMRI) by manipulating familiarity in both literal and metaphoric sentences. In an event-related design, participants viewed English sentences that appeared every 4.5-9s, and to which they made a pleasantness judgment. All sentences elicited activation in traditional language brain regions including left inferior frontal gyrus, left anterior inferior temporal and left posterior middle temporal gyri. Overall, metaphors and novel stimuli elicited activation in bilateral inferior frontal gyri and left temporal regions. Additionally, metaphors elicited greater activation than literal sentences in right temporal pole. Although our results are partially consistent with the graded salience hypothesis and the coarse coding hypothesis, the right hemisphere's sensitivity to familiar metaphors suggests that right hemisphere recruitment is most influenced by semantic integration demands.

Figure 1

Experimental Design. Participants viewed literal, metaphoric, and nonword sentences (duration=6s). Sentences occurred every 4.5 to 9 seconds, and a fixation cross was presented between sentences. Participants assessed each sentence for its valence.

Figure 2

Activation in Parahippocampal Gyrus and Inferior Temporal Gyrus. Regions in which novel literal sentences (blue), novel metaphors (green), and familiar metaphors (purple) elicited significant activation in left parahippocampal gyrus / medial inferior temporal regions (p<.01, cluster thresholded).

Figure 3

Activation in Inferior Frontal Gyri. Left: Regions in which metaphors elicited greater activation than literal sentences (blue) and in which literal sentences elicited greater activation than metaphors (red, p<.01, cluster thresholded). Metaphors elicited greater activation than literal sentences in bilateral inferior frontal gyri, and left medial superior frontal gyrus. Greater activation to literal sentences compared to metaphors can be seen in right medial superior frontal gyrus. Hemodynamic responses from the cluster in left inferior frontal gyrus (IFG) are shown below. All sentences elicited a significant hemodynamic response in left IFG; however, this response was significantly larger for metaphors. Right: Regions in which novel sentences elicited greater activation than familiar sentences (green) and in which familiar sentences elicited greater activation than novel sentences (purple). Novel sentences elicited activation in bilateral inferior frontal gyrus, and left medial frontal gyrus. There were no regions in which familiar sentences elicited greater activation than novel sentences. Hemodynamic responses from the cluster in left IFG are shown below. All sentences elicited a significant hemodynamic response in left IFG; however, this response was significantly larger for novel sentences compared to familiar sentences.

Figure 4

Activation in Lateral Temporal Gyri. Left: Regions in which metaphors elicited greater activation than literal sentences (blue) can be seen in anterior regions of middle temporal gyrus and left inferior frontal gyrus (p<.01, cluster thresholded). Hemodynamic responses (HDRs) from the cluster in left anterior middle temporal gyrus are shown. Metaphors elicited significantly larger hemodynamic responses in this region compared to literal sentences. Right: Regions in which novel sentences elicited greater activation than familiar sentences (green) can be seen in posterior middle temporal gyrus and left inferior frontal gyrus. Hemodynamic responses from the cluster in left posterior middle temporal gyrus are shown. Although both novel and familiar sentences elicited a hemodynamic response in this region; HDRs were larger for novel stimuli.