The behavioural patterns and neural correlates of concrete and abstract verb processing in aphasia: A novel verb semantic battery

Abstract

Typically, processing is more accurate and efficient for concrete than abstract concepts in both healthy adults and individuals with aphasia. While, concreteness effects have been thoroughly documented with respect to noun processing, other words classes have received little attention despite tending to be less concrete than nouns. The aim of the current study was to explore concrete-abstract differences in verbs and identify their neural correlates in post-stroke aphasia. Given the dearth of comprehension tests for verbs, a battery of neuropsychological tests was developed in this study to assess the comprehension of concrete and abstract verbs. Specifically, a sensitive verb synonym judgment test was generated that varied both the items' imageability and frequency, and a picture-to-word matching test with numerous concrete verbs. Normative data were then collected and the tests were administered to a cohort of 48 individuals with chronic post-stroke aphasia to explore the behavioural patterns and neural correlates of verb processing. The results revealed significantly better comprehension of concrete than abstract verbs, aligning with the existing aphasiological literature on noun processing. In addition, the patients performed better during verb comprehension than verb production. Lesion-symptom correlational analyses revealed common areas that support processing of concrete and abstract verbs, including the left anterior temporal lobe, posterior supramarginal gyrus and superior lateral occipital cortex. A direct contrast between them revealed additional regions with graded differences. Specifically, the left frontal regions were associated with processing abstract verbs; whereas, the left posterior temporal and occipital regions were associated with processing concrete verbs. Moreover, overlapping and distinct neural correlates were identified in association with the comprehension and production of concrete verbs. These patient findings align with data from functional neuroimaging and neuro-stimulation, and existing models of language organisation.

Keywords: Aphasia; Concreteness; Imageability; Lesion-symptom mapping; Verbs.

Fig. 1

An illustration of the stimuli presented to the patients. A: example stimulus from the verb synonym judgment test: probe ‘To suppose’, target ‘To assume’, and distractors ‘To happen’ and ‘To resolve’. B: example stimulus from the verb picture-to-word matching test: this is an illustration picture, the actual picture used was item number 97 from the action pictures (OANB: Druks and Masterson, 2000), target response ‘weaving’, semantic distractors ‘sewing’ and ‘knitting’, and unrelated distractors ‘swimming’ and ‘laughing’.

Fig. 2

Lesion overlap map across 48 post-stroke aphasia patients illustrating the distribution of lesions. Colour scale indicates number of patients with a lesion at that location. Image threshold = 1–40. The maximum number of participants who had a lesion in one voxel was 40 (MNI coordinate: − 38, − 9, 24; central opercular cortex). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Performance on the verb synonym judgment test. 2 × 2 repeated measure ANOVA showing an effect of imageability, with an advantage of concrete over abstract verbs, with no frequency or interaction effects on verb processing. Error bars show the standard error of mean correct responses.

Fig. 4

Lesion-symptom mapping showing neural correlates associated with single-word semantic verb comprehension. Results threshold at p < 0.0005 voxel-level and FWE cluster corrected p < 0.05, unless stated otherwise: (A) Verb synonym judgment test. (B) Abstract (blue) and concrete (red) verbs, and overlapping regions (violet). (C) Direct contrast between processing abstract (blue) and concrete verbs (red), thresholded at p < 0.001 voxel-level and FWE cluster-level corrected p < 0.05. (D) High (green) and low (red) frequency verbs, and overlapping regions (yellow). (E) Direct contrast between processing high (green) and concrete verbs, thresholded at p < 0.001 voxel-level and FWE cluster-level corrected p < 0.05. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5

Lesion-symptom mapping analysis showing: (A) Neural correlates associated with concrete verb naming (blue) and comprehension (red), and their overlapping regions (purple), thresholded at p < 0.0005 voxel-level, and FWE cluster-level correction at p < 0.05. (B) A direct contrast showing the neural correlates associated with concrete verb naming (red) and comprehension (blue), thresholded at p < 0.01 voxel-level, and FWE cluster-level correction of p < 0.05. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)