Poster D34, Friday, August 17, 4:45 – 6:30 pm, Room 2000AB

A lesion study of semantic hubs: the anterior temporal lobe and temporo-parietal junction

Nicholas Riccardi1, Chris Rorden1, Julius Fridriksson1, Rutvik Desai1;1University of South Carolina

Introduction: Both the anterior temporal lobe (ATL) and temporo-parietal junction (TPJ) have been implicated as cross-modal hubs that are vital for conceptual processing. While some evidence indicates that damage to the left ATL is sufficient for cross-modal semantic impairment, other evidence suggests the left ATL is involved in lexical access and not conceptual representation per se. In a group of left hemisphere damaged stroke patients, we examined how damage to the ATL and TPJ affects performance on three tasks with varying lexical and semantic demands: lexical decision (LD), word-based semantic similarity judgment (SSJ), and picture-based Pyramids and Palm Trees (PPT). We hypothesized that damage to the ATL should impair accuracy on tasks with a high lexical demand (LD, SSJ), demonstrating its role in lexical access. Damage to the TPJ should impair tasks with high semantic demand (SSJ, PPT), demonstrating its role as a cross-modal hub. Methods: Lesion data of 63 left unilateral stroke patients were considered. The LD task required patients to make a binary choice as to whether the presented word (80 nouns, 80 verbs, 160 pronounceable nonwords) was real or not. For SSJ, 80 noun and 80 verb triplets were presented, and the patients had to determine which of the bottom two words most closely matched the meaning to the word on the top. PPT is similar to SSJ but uses 54 picture triplets instead of words. Region-based lesion-symptom mapping (RLSM), which determines the relationship between percent of voxels damaged within a region and a behavioral score, was then used in two masks made from the Johns Hopkins University atlas. The ATL mask consisted of the middle and superior temporal poles (MTGpole, STGpole), inferior temporal gyrus (ITG), and fusiform gyrus (FG). The TPJ mask consisted of the angular and supramarginal gyri (AG, SMG) and the posterior superior, middle, and inferior temporal gyri (pSTG, pMTG, pITG). RLSM was conducted in a pairwise manner for each task such that patient’s accuracies in the other tasks were used as a nuisance covariate (e.g. LD regressing out SSJ, etc.). In this way, each patient also served as their own control for lesion volume. Significance and correction for multiple comparisons was accomplished via permutation testing (5000 permutations). Results: In the ATL mask, damage to the MTGpole significantly predicted LD accuracy when regressing out PPT (p = .009). No other pairwise comparison was significant. For the TPJ mask, damage to the SMG, AG, pSTG, and pMTG predicted SSJ accuracy when regressing out LD (all p < .01). Damage to the AG predicted PPT performance when regressing out SSJ (p < .01). Conclusions: Damage to the left ATL causes lexical impairments, even after accounting for lesion size and performance on a picture-based semantic task, highlighting its role in lexical access. TPJ damage was associated with impairment to semantic tasks regardless of input modality and when regressing out lexical performance, substantiating its role in cross-modal conceptual processing. This suggests a stronger role of ATL in lexical processing, and TPJ as a semantic hub.

Topic Area: Meaning: Lexical Semantics