FN400 and LPC’s Sensitivity to Phonological Processes: An ERP Study

Abstract

In the literature, there are few studies addressing phonological processing within the context of vowel harmony (Aaltonen et al., 2008; Scharinger et al., 2011; Ylinen et al., 2016). These studies adhere to the oddball paradigm (e.g., 85% harmonic, 15% disharmonic items) and report MMN differences between harmonic and disharmonic conditions. In this study, we examined phonological processing through Turkish backness harmony using N400 paradigm (50% harmonic, 50% disharmonic items). The first experiment included harmonic (e.g., aca, ehe) and disharmonic (e.g., ace, eha) V(owel)-C(onsonant)-V(owel)pseudowords. The second experiment involved the same pseudowords inflected with harmonic (e.g., acalar, afeler) and disharmonic (e.g., acaler, afelar) forms of the Turkish plural suffix “-lAr” to yield harmonic and disharmonic stems. We implemented cluster-based permutation tests via the FieldTrip Toolbox. We observed FN400 in the first experiment, but there was no significant cluster between the harmonic and disharmonic conditions. In the second experiment, we observed FN400 and LPC components and there were two significant negative clusters (p<.05) and one positive cluster (p<.05). The disharmonic condition showed greater amplitude than the harmonic condition in the negative (FN400) and positive (LPC) clusters. Although FN400 has been linked with semantic processing and familiarity-based memory, our study involved phonological processes and no recognition task (e.g. studied/test phases) was present. Therefore, our findings indicate that i) FN400 and LPC are sensitive to phonological processes and ii) phonological rules can trigger familiarity-based memory.