Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Lesions to the left hemisphere language network alter word reading activity in the undamaged visual word form system
Poster Session B, Friday, October 25, 10:00 - 11:30 am, Great Hall 3 and 4
Sara M. Dyslin1, Andrew T. DeMarco1,2, Ryan Staples1,3, Peter E. Turkeltaub1,2,4; 1Center for Brain Plasticity and Recovery, Georgetown University, 2Center for Aphasia Research and Rehabilitation, Georgetown University, 3Georgetown-Howard Universities Center for Clinical and Translational Science, Georgetown University, 4MedStar National Rehabilitation Hospital
Reading relies on an area of the brain’s object recognition system in the ventral occipito-temporal cortex (vOTC) that becomes specialized for word recognition as we learn to read, called the visual word form area (VWFA). Prior fMRI studies in neurotypical controls have proposed a posterior-to-anterior gradient of increasing sensitivity to larger fragments of words in the VWFA, and also that the VWFA receives top-down information from language regions. Lesions to perisylvian language regions often cause alexia, which could be partly explained by abnormal visual word form processing due to the loss of top-down connections. This hypothesis has never been tested. Further, previous studies have shown that right hemisphere (RH) homotopes may partially support language after left hemisphere (LH) stroke, but these findings have not been extended to the RH homotope of the VWFA (rVWFA) in alexia. Here, we examine how lesions to the left hemisphere language network affect activation of the VWFA within the left and right vOTC during reading. Participants included 12 chronic left hemisphere stroke survivors (LHSS) (Age=58.3(16.1); Gender (7F/5M); Months-since-stroke=42.9(32.4); Oral reading accuracy: Words=84.0%(12.0%), Pseudowords=57.4%(38.2%)) and 17 demographically-matched controls. All participants completed three runs of a novel fMRI task consisting of 24 15-second blocks of rapidly presented words, pseudowords, false font strings, or fixation. Stimuli were presented for 150ms with a 150ms inter-stimulus interval and a total run time of 6 minutes. First-level contrasts of words>false fonts were combined in a mixed-effects model across three runs to provide a wholebrain t-statistic activation map for each participant. Group analyses determined the averaged magnitudes of activity within a Neurosynth-derived LH VWFA (lVWFA) mask and a corresponding rVWFA ROI. These ROIs were also divided into anterior and posterior subregions to assess spatial organization within the VWFA. Three 2x2 ANOVAs and post-hoc analyses assessed the main effects and interaction of group by hemisphere on magnitude of activation within 1) the entire lVWFA vs rVWFA, 2) the anterior subregions of the ROIs, and 3) the posterior subregions of the ROIs. There was a significant interaction of group and hemisphere in the whole VWFA (F(1,1)=4.46, p=0.039) and the anterior VWFA (F(1,1)=4.23, p=0.044), and a trending interaction in the posterior VWFA (F(1,1)=3.66, p=0.06). The nature of the interactions differed by subregion. The interaction in the anterior VWFA was driven mainly by greater right hemisphere activity for LHSS than controls, with five individuals who had especially high activation exceeding every control participant. In the posterior VWFA, the interaction was mainly driven by reduced activity for the LHSS in the left hemisphere. Eleven LHSS participants activated posterior lVWFA less than the control average. These results provide key lesion evidence that VWFA processing partly relies on top-down connections from the language network. Further characterizing the lesion effects on VWFA processing, including the factors that drive RH recruitment, will help reveal the nature of interactions between the vOTC and the language network. Additional work will also determine if the aberrant lVWFA activation underlies reading deficits in alexia, and if post-stroke reorganization in the rVWFA supports recovery.
Topic Areas: Reading, Disorders: Acquired