Neural processing of mass, count and dual nouns: an ERP investigation

Abstract

Our aim was to examine differences in lexical processing of verbs and nouns by manipulating noun type: count (e.g. pen), mass (e.g. milk) and dual (e.g. lamb), the latter representing either mass or count qualities. A covert judgment task (verb/noun) was administered to 20 participants from Sussex University using a within participants design. Behavioral analyses revealed a main effect of noun type, i.e., significantly faster RTs for dual (794 ms) and count nouns (881 ms) compared to mass nouns (934 ms) and verbs (926 ms). EEG was recorded using a 128-channel Net connected to an AC-coupled 128-channel high-input impedance amplifier. Electrode impedance was kept below 50 kOhms. Voltage was digitized at a rate of 250 Hz and band-pass filtered between 1 and 30Hz. EEG was segmented into windows from100 ms before to 2000 ms after stimulus onset. Segments were baseline corrected to the average of the pre-stimulus epoch and re-referenced using an average across all channels. Noun type had a main effect on voltage change with dual nouns differentiated at posterior sites between 400 and 600 ms and count nouns eliciting greater negative activation in the N400 component over centroparietal sites. Verbs showed an increased positivity between 400 and 600 ms, particularly in frontal regions. The results distinguish the processing of dual and count nouns from mass nouns at a semantic level and nouns from verbs at the level of motor processing. We submit that differential responses for nouns reflect numerical quantification that is available for dual and count nouns and not mass nouns.