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

Poster A39, Wednesday, November 8, 10:30 – 11:45 am, Harborview and Loch Raven Ballrooms

Semantic similarity effect for written words in left perirhinal cortex: influence of type of property retrieved, visual versus nonvisual

Antonietta Gabriella Liuzzi1, Patrick Dupont1, Ronald Peeters2, Simon De Deyne3, Gerrit Storms3, Rik Vandenberghe1,4;1Laboratory for Cognitive Neurology, Department of Neurosciences, KU Leuven, Belgium, 2Radiology Department, University Hospitals Leuven, 3000 Leuven, Belgium., 3Laboratory of Experimental Psychology, Humanities and Social Sciences Group, KU Leuven, Belgium, 4Neurology Department, University Hospitals Leuven, 3000 Leuven, Belgium

INTRODUCTION: An increasing number of studies have demonstrated that left perirhinal activity patterns code for semantic similarity between concrete written words. Here we examine how these effects are influenced by the type of property retrieved, visual versus nonvisual, and by the input-modality, picture versus written word. METHODS: An event-related fMRI study was run on a Philips Achieva dstream 3T equipped with a 32-channel head coil in 18 subjects. Twelve animate (mammals, birds, insects) and 12 inanimate entities (kitchen tools, clothes, music instruments) (De Deyne et al.,2008) were presented as either a written word or a picture. From the concept-feature matrix 52 properties were evaluated online by 11 healthy volunteers who judged on a 1-7 rating scale the degree to which each property was visual or nonvisual. Four visual and 4 nonvisual properties for each subcategory were selected. From the word association matrix (De Deyne et al.,2016), the pairwise semantic cosine similarity was calculated for each pair of items (semantic matrix). During fMRI, subjects performed a property verification task. fMRI data were modelled using a General Linear Model (GLM). By calculating the pairwise cosine similarity between every pair of trials, 10 fMRI matrices in perirhinal cortex were generated: for written words, pictures, written words and pictures pooled, crossmodal effect, visual properties, nonvisual properties, visual properties for written words, nonvisual properties for written words, visual properties for pictures, nonvisual properties for pictures. A representational similarity analysis between the semantic matrix and each fMRI matrix was conducted by using one-tailed statistical threshold of P <0.05. RESULTS: A 3-way repeated measures ANOVA with reaction times as outcome showed a main effect of input-modality (written words: 1.7s pictures: 1.63s) (F(1,15)=31.9 p=.000) and property (non-visual: 1.69s; visual: 1.63s) (F(1,15)=17.60 p=.001). The accuracy of responses was 70.8%. The correlation between semantic similarity and the similarity of activity patterns in left perirhinal cortex was significant for written words (Pearson correlation (r) = 0.15 p=0.0086), written words and pictures pooled (r=0.20 p=0.001), for crossmodal word-picture pairs (r=0.20 p=0.0002) and for the retrieval of visual properties for written words (r=0.14 p=0.013) but not for nonvisual properties (r = -0.03 p = 0.7) and neither for visual or nonvisual properties of pictures (r= -0.0058 p =0.52, r = 0.07 p = 0.13). CONCLUSION: The effect of type of property and input-modality leads us to hypothesize that left perirhinal cortex mediates mnemonic retrieval of properties of referents of concrete written words.

Topic Area: Meaning: Lexical Semantics

Back to Poster Schedule