You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

 
Poster B56, Wednesday, November 8, 3:00 – 4:15 pm, Harborview and Loch Raven Ballrooms

Lesion Localization of a Shared Phonologic Representation Deficit on Reading, Rhyming, Repetition, and Short-Term Memory Tasks

Sara Pillay1, Peter Kraegel1, Colin Humphries1, Diane Book1, Jeffrey Binder1;1Medical College of Wisconsin

Many tasks have been used to assess phonologic processing, including speech output, silent rhyming, and verbal short-term memory tasks. Despite differences in input modality, memory load, and response procedures, it is sometimes assumed that these various tasks all depend on a shared phonologic representation, i.e., an abstract representation of an ordered sequence of phoneme codes. Prior lesion correlation studies have focused on single tasks. Here we test for the first time whether there is a common lesion localization for impairments on phonologic short-term memory, oral reading, repetition, and silent rhyming tasks. We hypothesized that lesions affecting phonologic processing, regardless of input modality, production demand, or phonologic memory load, would localize to a common region in the posterior perisylvian cortex. Voxel-based lesion-symptom mapping (VLSM) was used in 57 patients with chronic left hemisphere stroke. All patients were right-handed, native English speakers and at least 6 months from stroke onset. The patients completed four tasks used to assess phonologic processing in different modalities: (#1) a visual pseudoword rhyming task in which patients matched a sample pseudoword to one of two choices based on phonologic similarity, (#2) a spoken pseudoword repetition task, (#3) an auditory phonologic delayed match-to-sample (DMS) task in which patients indicated whether two auditory natural CV syllable trains containing 1-5 syllables, separated by 5 seconds, were the “same” or “different”, and (#4) an overt reading task, where the rate of phonologic paraphasias was the variable of interest. To control for non-specific attention and executive control processes in measures #1-3, patients completed a semantic judgment task where they matched pictures based on meaning. Common lesion overlap across the four measures was examined using VLSM. Accuracy on the DMS (81%), pseudoword rhyming (75%), and pseudoword repetition (62%) tasks were strongly correlated (all p < .001). These three measures were also strongly correlated with the rate of phonologic paraphasias (mean 19%) made on the overt reading task (all p < .005). VLSM localized the responsible lesions in each of the tasks to the posterior perisylvian region. Overlap of these maps revealed a common lesion location involving the posterior medial planum temporale, junction of posterior superior temporal gyrus and supramarginal gyrus, parietal branch of the superior temporal sulcus, and white matter underlying these structures. There was variable involvement anterior to this core region in anterior supramarginal gyrus, inferiorly in the posterior temporal STS, and posteriorly in the inferior angular gyrus. The results confirm a common lesion location impairing performance on a range of phonologic tasks differing in input modality, output modality, and memory load. This overlap is unlikely to be due to nonspecific executive control deficits, which were controlled for in the VLSM analysis. The results suggest that all of the measures depend on a common phonologic representation, and they provide further support for an "emergent" view of phonologic short-term memory, in which activation of long-term phonologic representations and the "phonologic store" used for short-term memory are one and the same.

Topic Area: Phonology and Phonological Working Memory

Back to Poster Schedule