Poster D17, Friday, August 17, 4:45 – 6:30 pm, Room 2000AB

Perceptual processing of pre-boundary lengthening during phrase segmentation in English: Preliminary ERP evidence.

Annie Gilbert1,2, Jasmine Lee3, Max Wolpert2,4, Shari Baum1,2;1School of Communication Sciences and Disorders, McGill University, Canada., 2Centre for Research on Brain, Language and Music, Canada., 3Honours in Cognitive Science, McGill University, Canada., 4Integrated Program in Neuroscience, McGill University, Canada.

The study of phrase processing has greatly benefited from the identification of a specific Event-Related Potential (ERP) component associated with the perception of phrase boundaries, namely the Closure Positive Shift (CPS). Most of the work on the CPS has looked at higher-level phrase processing where the phrasing informs the syntactic-semantic parsing of the sentence (see for example the difference between “When the man was parking the truck # ...” and “When the man was parking # the truck …”). Interestingly, a few studies have also found CPSs (or CPS-like positivities, hereafter included in CPS) associated with the acoustic processing of the prosodic markers of phrase boundaries, even if such boundaries did not have a significant impact on the interpretation of the stimuli (in nonsense stimuli, for example). Among these studies, Gilbert et al. (2015) demonstrated that phrase-final lengthening (without F0 modulation) is sufficient to trigger a CPS-like positivity in both simple French sentences and nonsense series of syllables. They interpreted their results as reflecting a domain-general perceptual process related to phrase segmentation and independent of the content of the utterance. It should be noted that there are topographical differences between the classic CPS and the ‘perceptual’ CPS, which might indicate that they index slightly different processes. Moreover, French has a very simple prosodic system with no lexically-coded prosody; thus, it is unclear if a similar perceptual CPS would emerge in languages from different prosodic typologies. To answer this question, we designed an ERP experiment to determine if phrase-final lengthening would trigger a perceptual CPS in simple English utterances, as was shown in French. Stimuli consisted of fifty sentence pairs allowing for the comparison of the same target word in phrase-final (e.g.: Last year’s flu / caused…) and non-phrase-final position (e.g.: Last year’s flu scare / caused…). Stimuli were recorded by a native speaker of English with no salient phrase-final F0 rise and phrase-final words being on average 1.4 times longer than their non-phrase-final counterpart. Sentences were presented in pseudo-random order among fillers to ten monolingual English speakers so far. Their ERPs were time-locked to the target-word offset in both conditions and averaged from -500ms before to 1000 ms after the time-locking point. Average ERP amplitudes at electrode Cz were compared across conditions in 15 consecutive 100ms time-windows. Wilcoxon Signed Rank tests revealed no significant amplitude differences between conditions prior and up to 100ms after the time-locking point, but yielded significant differences across conditions from 100ms to 700ms after the time-locking point, with the phrase-final condition triggering a positive deflection in the ERP compared to the non-phrase final condition. Visual inspection of voltage maps revealed that the distribution of the present CPS is similar to the one found for French by Gilbert et al., with both CPSs being maximal over a left fronto-central region. These preliminary results demonstrate that phrase-final lengthening triggers a perceptual CPS in English, which supports the domain-general interpretation of phrase-final lengthening acting as a lower-level phrase segmentation cue.

Topic Area: Perception: Speech Perception and Audiovisual Integration