Robert Eklund, SBI/KI, Voice Provider, Stockholm och ICSI, Berkeley, talar om "The neural correlates of filled pauses: A combined fMRI and EEG study of disfluency perception".

Seminariet arrangeras i samarbete med Institutionen för tal, musik och hörsel på KTH.

Tid och plats: Tisdagen den 10 november 2009, kl 15:15-17:00 i Fantum, adress Lindstedtsv. 24, 5 trappor.


Disfluency makes up around 6% of spontaneous speech, and after having been regarded as a mere performance error in the Chomskyan competence versus performance dichotomy, more and more research has focused on its potential beneficial role in conversation. Most such studies have been either purely observational (within general linguistics) or behavioral/psycholinguistic (measuring phenomena like reaction times), but a few neurocognitive studies have been carried out. Most of these neurocognitive studies, however, have in common that they use scripted disfluencies (written and read by actors) -- and thus do not constitute authentic disfluencies -- and that they have preponderatingly used electroencephalographic (EEG) paradigms, thus focusing on timing aspects of speech perception. Moreover, previous studies have focused not on the effect of disfluencies proper, but instead on ensuing words or phases in the speech stream, covering phenomena like word recognition reaction times or syntactic parsing of ambiguous sentences. This talk will present results from two experiments that use fMRI (functional Magnetic Resonance Imaging) and EEG -- thus obtaining results with both high spatial (fMRI) and high temporal (EEG) resolution -- to study the neurocognitive effect of filled pause perception, using authentic and spontaneous filled pauses.