Gesture discrimination in primary progressive aphasia: the intersection between gesture and language processing pathways

Abstract

The issue of the relationship between language and gesture processing and the partial overlap of their neural representations is of fundamental importance to neurology, psychology, and social sciences. Patients suffering from primary progressive aphasia, a clinical syndrome characterized by comparatively isolated language deficits, may provide direct evidence for anatomical and functional association between specific language deficits and gesture discrimination deficits. A consecutive series of 16 patients with primary progressive aphasia and 16 matched control subjects participated. Our nonverbal gesture discrimination task consisted of 19 trials. In each trial, participants observed three video clips showing the same gesture performed correctly in one clip and incorrectly in the other two. Subjects had to indicate which of the three versions was correct. Language and gesture production were evaluated by means of conventional tasks. All participants underwent high-resolution structural and diffusion tensor magnetic resonance imaging. Ten of the primary progressive aphasia patients showed a significant deficit on the nonverbal gesture discrimination task. A factor analysis revealed that this deficit clustered with gesture imitation, word and pseudoword repetition, and writing-to-dictation. Individual scores on this cluster correlated with volume in the left anterior inferior parietal cortex extending into the posterior superior temporal gyrus. Probabilistic tractography indicated this region comprised the cortical relay station of the indirect pathway connecting the inferior frontal gyrus and the superior temporal cortex. Thus, the left perisylvian temporoparietal area may underpin verbal imitative behavior, gesture imitation, and gesture discrimination indicative of a partly shared neural substrate for language and gesture resonance.

Figure 1.

Static images from a representative trial. A–C, Intransitive gestures. D–F, Transitive gestures. For each trial, a correct version of a gesture is shown together with two distractors. Each trial consisted of one correctly executed gesture (A, D) and two distractors, shown in random order. For intransitive gestures, distractors consisted of an incorrect spatial configuration of either the proximal (B) or distal parts (C) of the involved limb. As distractors for transitive gestures, the correct object was replaced with an inappropriate object that had either a high (E) or low (F) degree of semantic similarity with the correct object.

Figure 2.

Correlation of gray matter volume with factor 3. A, Voxel-based partial correlation of GM volume with scores on factor 3 (voxel-level uncorrected, p < 0.001; cluster-level corr., p < 0.05). B, Correlation of mean gray matter volume in this cluster with gesture discrimination scores. C, Correlation of mean gray matter volume in this cluster with gesture imitation scores. D, Correlation of mean gray matter volume in this cluster with word repetition scores. Because case 13 was detected as an outlier statistically, this case was removed from the linear regression analysis. Case numbers refer to Table 1. Green, Control subjects; upward red triangle, semantic dementia; downward red triangle, progressive nonfluent aphasia; pink arrow, logopenic aphasia.

Figure 3.

Correlation of GM volume with the lexical-semantic factor. A, Voxel-based correlation of GM volume with scores on factor 1 (voxel-level uncorrected, p < 0.001; cluster-level corrected, p < 0.05). B, Correlation of mean gray matter volume in this cluster with gesture discrimination scores. C, Correlation of mean GM volume in this cluster with Boston Naming Test scores. D, Correlation of mean gray matter volume in this cluster with word–picture matching scores. Case numbers refer to Table 1. See Figure 2 for other information.

Figure 4.

Probabilistic fiber tracking with the temporoparietal cluster as seed region (shown in light blue). A, A random control subject (control 1, voxels are shown only when at least 500 streamlines from the seed region pass through the voxel). B, Composite maps showing voxels present in at least 75% (yellow/blue) of control subjects.