Presentation

Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams

Spatiotemporal dynamics of concreteness

Poster A54 in Poster Session A, Thursday, October 6, 10:15 am - 12:00 pm EDT, Millennium Hall

Elliot Murphy¹, Oscar Woolnough¹, Cale Morse¹, Xavier Scherschligt¹, Nitin Tandon¹; ¹University of Texas Health Science Center at Houston

Language comprehension relies on rapidly navigating the boundary between concrete words (perceptual: ‘table’) and abstract words (non-perceptual: ‘justice’). Consistent behavioral differences have been found between these categories, such that concrete words are faster to make lexical decisions about and offer processing advantages. Any neurobiological model of human semantics will ultimately need to make reference to the concrete/abstract formatting of conceptual representations. While behavioral effects between concrete/abstract words are consistent, in stark contrast the neurobiological effects differ substantially. Controversy exists over regions of activation corresponding to concrete/abstract words, often due to the range of methodologies and differences in task complexity. There remains no consensus concerning the contribution of nodes in the semantics network to concreteness. Many previous studies have not separated single-word from sentence-level paradigms, confounding potential effects. We present an intracranial investigation in a cohort of epilepsy patients (n = 15) of the spatiotemporal dynamics of concreteness using an orthographic single-word concreteness judgment paradigm. We additionally conducted cortical stimulation mapping in a group of these patients (n = 6) to attribute causality to nodes in the semantic network. Patients were presented with a single word on a screen and asked to judge whether it referred to something you could touch, taste, smell, see or hear. A fixation cross was presented in the center of the screen for 700 ms, after which orthographic single-word exposures lasting 1000 ms were presented. 1500 ms of blank screen followed during which patients had to press either the left arrow (Concrete) or right arrow (Abstract). We recruited a large list of concrete, abstract and ‘midscale’ words (that occupy the middle of the concreteness scale; e.g. ‘magic’, ‘translation’, and for which there is no incorrect answer) from a comprehensive review of databases for concreteness ratings. Data were acquired from either subdural grid electrodes or stereotactically placed depth electrodes. Typical coverage was frontotemporal, dictated by location of the epilepsy in the antero-mesial temporal lobe in the majority of patients. Behavioral performance was high (>80% for all) and only correct trials were analyzed. Analyzing broadband high gamma activity (BGA; 70–150 Hz), we discovered a frontotemporal network for concreteness, implicating mid-fusiform gyrus (mFus), parahippocampal cortex (PHC), inferior frontal gyrus (IFG), orbitofrontal cortex (OFC) and frontal operculum (FO). The earliest effects were found first in OFC, followed by mFus and PHC. Analyzing directionality via amplitude-envelope correlations in BGA, information flowed from OFC to FO, PHC, temporoparietal junction (TPJ) and both aIFG and pIFG (~350 ms for pIFG), and slightly later (~500 ms) from pIFG to OFC and FO. Only in FO at approximately 800–1000 ms did patient subjective responses for midscale words impact BGA, showing greater activity for concrete-rated midscale items. Lastly, we used direct cortical stimulation to attribute causal involvement of certain nodes in this network (ventral temporal, inferior frontal) to concreteness, successfully disrupting judgments for concrete words. Our work unveils the fine spatiotemporal dynamics of the semantic network and affords causal power to our documented concreteness network.

Topic Areas: Methods, Speech Motor Control