Verbal working memory and syntactic comprehension segregate into the dorsal and ventral streams, respectively

Abstract

Syntactic processing and verbal working memory are both essential components to sentence comprehension. Nonetheless, the separability of these systems in the brain remains unclear. To address this issue, we performed causal-inference analyses based on lesion and connectome network mapping using MRI and behavioural testing in two groups of individuals with chronic post-stroke aphasia. We employed a rhyme judgement task with heavy working memory load without articulatory confounds, controlling for the overall ability to match auditory words to pictures and to perform a metalinguistic rhyme judgement, isolating the effect of working memory load (103 individuals). We assessed non-canonical sentence comprehension, isolating syntactic processing by incorporating residual rhyme judgement performance as a covariate for working memory load (78 individuals). Voxel-based lesion analyses and structural connectome-based lesion symptom mapping controlling for total lesion volume were performed, with permutation testing to correct for multiple comparisons (4000 permutations). We observed that effects of working memory load localized to dorsal stream damage: posterior temporal-parietal lesions and frontal-parietal white matter disconnections. These effects were differentiated from syntactic comprehension deficits, which were primarily associated with ventral stream damage: lesions to temporal lobe and temporal-parietal white matter disconnections, particularly when incorporating the residual measure of working memory load as a covariate. Our results support the conclusion that working memory and syntactic processing are associated with distinct brain networks, largely loading onto dorsal and ventral streams, respectively.

Keywords: aphasia; lesion-symptom mapping; syntax; tractography; working memory.

Graphical Abstract

Figure 1

Lesion overlap maps and connectome graphs for the two groups of participants. (A) Lesion overlap map for the full set of 103 participants who were administered the WAB-R as well as the TALSA and PALPA rhyme judgement tasks. (B) Lesion overlap map for the sub-set of 78 participants who also were administered the non-canonical sentence comprehension task. (C) Average connectome graph for the set of 101 participants with DTI data who were administered the WAB-R as well as the TALSA and PALPA rhyme judgement tasks. (D) Average connectome graph for the sub-set of 77 participants with DTI data who also were administered the non-canonical sentence comprehension task. In panels (C) and (D), ROIs are plotted in order on the X and Y axes, separated into LH and RH.

Figure 2

Voxel-wise univariate lesion-symptom mapping regression analysis results, including total lesion volume as a covariate. (A) Performance on the TALSA triplets task, with AudWordRec and RhymeJudge as covariates. (B) Performance on the non-canonical sentence comprehension task, with residual triplets performance as a covariate. Colour bar indicates P-value for each voxel. Red colour indicates voxels that survived a permutation correction for multiple comparisons (4000 permutations). Note that the permutation-corrected threshold for the two analyses is different, indicated by the different solid red regions of the colour bar.

Figure 3

Connectome-based lesion-symptom mapping results and associated tractography. (A) and (C) Significant CLSM univariate results for (A) TALSA triplets with AudWordRec and RhymeJudge covariates, as reported in section ‘Connectome-based lesion-symptom mapping results’ and listed in Table 4, shown here as lines between the following pairs of regions: ‘angular gyrus’ < -> ‘anterior insula; angular gyrus’ < -> ‘lateral fronto-orbital gyrus’; ‘superior occipital gyrus’ < -> ‘anterior insula’ and (C) non-canonical sentence comprehension with covariate for WM load (residual TALSA triplets performance, as reported in section ‘Connectome-based lesion-symptom mapping results’ and listed in Table 4, shown here as lines between the following pairs of regions: ‘superior occipital gyrus’ < -> ‘central superior temporal gyrus’; ‘cuneus’ < -> ‘anterior inferior temporal gyrus’; ‘lingual gyrus’ < -> ‘pole of middle temporal gyrus’; ‘lingual gyrus’ < -> ‘anterior inferior temporal gyrus’. CLSM results are permutation-corrected for multiple comparisons (4000 permutations). (B) and (D) ROI-based tractography for (B) TALSA triplets with AudWordRec and RhymeJudge covariates, and (D) Non-canonical sentence comprehension with covariate for WM load (residual TALSA triplets performance). Fibre-tracking images show all fibres that intersected both pairs of regions implicated in each significant disconnection from CLSM analyses. The specific tract to which each fibre was associated are shown in color-coding, with only tracts containing 20 or more fibres shown. SLF, superior longitudinal fasciculus; AF, arcuate fasciculus; EC, extreme capsule; IFOF, inferior frontal occipital fasciculus; MLF, middle longitudinal fasciculus; ILF, inferior longitudinal fasciculus; AC, anterior commissure; PCST, posterior cortico-striatal tract; Tap., tapetum.

Figure 4

Supplementary multivariate LSM and CLSM regression analyses. Left: ROI-based multivariate LSM results, without including total lesion volume as a covariate for (A) performance on the TALSA triplets task, with AudWordRec and RhymeJudge as covariates and (C) performance on the non-canonical sentence comprehension task, with residual triplets performance as a covariate. Colour bar indicates P-value for each voxel. Red colour indicates voxels that survived a permutation correction for multiple comparisons (4000 permutations). Right: multivariate CLSM results, top 5 most statistically robust results (non-significant after correction for multiple comparisons) for (B) TALSA triplets with AudWordRec and RhymeJudge covariates, as reported in section ‘Connectome-based lesion-symptom mapping results’, shown here as red lines between the following pairs of regions: ‘postcentral gyrus < -> inferior frontal gyrus, pars triangularis; precentral gyrus < -> inferior frontal gyrus pars triangularis; angular gyrus < -> inferior frontal gyrus pars opercularis; angular gyrus < -> inferior frontal gyrus pars triangularis; middle occipital gyrus < -> inferior frontal gyrus pars triangularis’ and (D) non-canonical sentence comprehension with covariate for WM load (residual TALSA triplets performance), as reported in section ‘Connectome-based lesion-symptom mapping results’, shown here as red lines between the following pairs of regions: ‘pole of middle temporal gyrus < -> anterior superior frontal gyrus; pole of middle temporal gyrus < -> precuneus; lingual gyrus < -> pole of middle temporal gyrus; globus pallidus < -> middle occipital gyrus; posterior insula < -> inferior frontal gyrus pars orbitals’.