The dorsal stream contribution to phonological retrieval in object naming

Abstract

Meaningful speech, as exemplified in object naming, calls on knowledge of the mappings between word meanings and phonological forms. Phonological errors in naming (e.g. GHOST named as 'goath') are commonly seen in persisting post-stroke aphasia and are thought to signal impairment in retrieval of phonological form information. We performed a voxel-based lesion-symptom mapping analysis of 1718 phonological naming errors collected from 106 individuals with diverse profiles of aphasia. Voxels in which lesion status correlated with phonological error rates localized to dorsal stream areas, in keeping with classical and contemporary brain-language models. Within the dorsal stream, the critical voxels were concentrated in premotor cortex, pre- and postcentral gyri and supramarginal gyrus with minimal extension into auditory-related posterior temporal and temporo-parietal cortices. This challenges the popular notion that error-free phonological retrieval requires guidance from sensory traces stored in posterior auditory regions and points instead to sensory-motor processes located further anterior in the dorsal stream. In a separate analysis, we compared the lesion maps for phonological and semantic errors and determined that there was no spatial overlap, demonstrating that the brain segregates phonological and semantic retrieval operations in word production.

Figure 1

Maps depicting lesion overlaps of the 106 participants in the left hemisphere. Maps are superimposed on the MNI space Colin27 template. (A–D) MNI x coordinates: x = −62, x = −54, x = −46 and x = −38, respectively. (E) A single coronal slice at MNI y coordinate = −16 with white lines to indicate the location of sagittal slices A–D when viewed in the coronal plane. Voxels lesioned in at least 10 patients (the minimum allowed) are rendered in scale from red (10 patients) to yellow (57 patients).

Figure 2

Statistical map (t-statistic) of phonological errors in picture naming between patients with and without lesion in each voxel. Map is thresholded with a FDR q = 0.01 (t = 3.50). Voxels rendered in red start at t = 3.50 and scale to the maximum t-value (t = 5.17) rendered in yellow. The light green area locates the Sylvian-parieto-temporal (Spt) area mask. Voxels in area Spt that exceeded the threshold are rendered in blue; in F, these are indicated by the arrow. (A–D) At MNI x coordinates: x = −62, x = −54, x = −46 and x = −38, respectively. (E) A coronal slice at MNI y coordinate −16, with white lines to indicate the location of sagittal slices A–D when viewed in the coronal plane. (F) A coronal slice at MNI y coordinate −28.

Figure 3

From the VLSM of phonological errors regressing out lesion volume. Map shows voxels thresholded at t = 2.50 and above displayed on the MNI-space Colin27 template. Voxels with the highest t-value (4.31) are rendered in yellow. (A–D) Sagittal slices at x = −62, x = −54, x = −46 and x = −38, respectively. (E) A single coronal slice at y = −16, with white lines indicating the location of slices A–D when viewed in the coronal plane. (F) A coronal slice at MNI y coordinate −28.

Figure 4

From the VLSM of phonological errors including only the 83 patients without apraxia of speech. Map shows voxels thresholded at t = 2.50 and above displayed on the MNI-space Colin27 template. Voxels with the highest t-value (4.12) are rendered in yellow. (A–D) Sagittal slices at x = −62, x = −54, x = −46 and x = −38, respectively. (E) A single coronal slice at y = −16, with white lines indicating the location of slices A–D when viewed in the coronal plane. (F) A coronal slice at MNI y coordinate −28.

Figure 5

Map of the phonological errors results with a FDR q = 0.05 (t = 2.41) and a map of the arcuate fasciculus derived from the Johns Hopkins University Probabilistic atlas map. The probabilistic arcuate map is rendered at 0.8 probability. (A–E) MNI y coordinates: y = −9, y = −17, y = −25, y = −33 and y = −41, respectively. (F) A single sagittal slice at MNI x coordinate = −40 with white lines indicating the location of coronal slices (A–D) when viewed in the sagittal plane. Blue represents phonological errors results, yellow the white matter tracks of the arcuate fasciculus and green the region of overlap between phonological errors and the arcuate fasciculus.

Figure 6

Lesions masks derived from the VLSM of auditory discrimination errors and the VLSM of phonological errors, thresholded with the same FDR correction (q = 0.05) are rendered together on the MNI-space Colin27 template. The critical t-value for auditory discrimination errors was 2.72; the critical t-value for phonological errors was 2.41. Overlapping voxels, i.e. those that surpassed the threshold in both analyses, are shown in green. Voxels that were significant in the auditory discrimination errors analysis only are shown in red; those that were significant in the phonological errors analysis only are shown in blue. (A–D) Sagittal slices at x = −62, x = −54, x = −46 and x = −38, respectively. (E) A single coronal slice at y = −16, with white lines indicating the location of slices A–D when viewed in the coronal plane. (F) A coronal slice at MNI y coordinate −28.

Figure 7

Lesions masks derived from the VLSM of semantic errors and the VLSM of phonological errors, thresholded with the same FDR correction (q = 0.05), are rendered together on the MNI-space Colin27 template. The critical t-value for semantic errors was 2.82; the critical t-value for phonological errors was 2.41. Voxels that were significant in the VLSM of semantic errors are shown in purple; those that were significant in the phonological errors analysis are shown in blue. No voxels were significant in both analyses. (A–D) Sagittal slices at x = −62, x = −54, x = −46 and x = −38, respectively. (E) A single coronal slice at y = −16, with white lines indicating the location of slices A–D when viewed in the coronal plane. (F) A coronal slice at MNI y coordinate −28.