Frontiers in Psychology (Apr 2015)

Damage to Broca’s area OR the anterior temporal lobe is implicated in stroke-induced agrammatic comprehension: it depends on the task

  • Corianne Rogalsky,
  • Kuan-Hua Chen,
  • Hanna Damasio,
  • Tracy Love,
  • Greg Hickok

DOI
https://doi.org/10.3389/conf.fpsyg.2015.65.00077
Journal volume & issue
Vol. 6

Abstract

Read online

The neurobiology of sentence comprehension remains unresolved. Previous large-scale studies of stroke patients have yielded conflicting results regarding sentence comprehension, implicating inferior frontal, anterior temporal and/or posterior temporal regions (Dronkers et al., 2004; Magnusdottir et al., 2013; Thothathiri et al. 2012). Furthermore, only one large-scale lesion study (Magnusdottir et al. 2013) has examined the neural underpinnings of agrammatic comprehension (i.e. substantially worse performance on sentences with noncanonical word orders compared to canonical word order sentences in English), a hallmark of Broca’s aphasia. This one previous study of noncanonical < canonical sentence performance on a sentence picture-matching task implicated damage to the left anterior temporal lobe (ATL) and to a lesser degree Broca’s area damage (i.e. < 10% of significant voxels) (Magnusdottir et al. 2013). The present study investigated the neurobiology of agrammatic comprehension with two sentence comprehension tasks in the MARC test battery: a sentence-picture matching task (the SOAP Test: a test of syntactic complexity; Love & Oster, 2002) and a sentence plausibility judgment task. Each task contained active, passive, subject-relative and object-relative sentences. Participants included 91 patients with chronic focal cerebral damage. First, we conducted voxel-based lesion symptom mapping (VLSM; Bates et al. 2003) for each sentence type in each task. Consistent with previous studies (Magnusdottir et al. 2013; Thothathiri et al. 2012), the VLSMs identified a significant association between sentence comprehension impairments and damage to a large left temporal-inferior parietal network for all sentences (peak t values were in posterior temporal and inferior parietal voxels); no areas of frontal lobe damage were significant for any sentence type/task. We then conducted VLSMs to identify areas of damage associated specifically with agrammatic comprehension in each task. Agrammatic comprehension was indexed as the difference in performance of each patient between subject relative and object relative sentences. VLSMs using these indices identified that agrammatic comprehension in each task was associated with damage to different brain regions: In the sentence-picture matching task, agrammatic comprehension was associated with damage in the left putamen, external capsule and adjacent white matter underlying the left inferior frontal gyrus and left precentral gyrus. Conversely in the plausibility judgment task, agrammatic comprehension was associated with damage in the left anterior superior temporal gyrus. Thus agrammatic comprehension indexed using different tasks localized to different lesion patterns. Our findings suggest that the neurobiology of agrammatic comprehension is task-dependent. These results also provide possible neural bases for the behavioral dissociations between sentence comprehension and grammaticality judgment impairments that have been reported in patients with aphasia (Linebarger et al. 1983; Wulfeck, 1988), although grammaticality and plausibility judgments are certainly different computational tasks.

Keywords