You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

 
Poster C73, Thursday, November 9, 10:00 – 11:15 am, Harborview and Loch Raven Ballrooms

Neural entrainment to acoustic edges in speech

Maria Oana Cucu1, Nina Kazanina1, Conor Houghton1;1University of Bristol

Introduction The theta rhythm of neural oscillations was shown to reliably track speech at the level of syllables (Luo & Poeppel, 2007). Syllable onsets may aid neural entrainment by acting as acoustic edges for phase resetting (Doelling et al., 2014). Furthermore, the level of entrainment depends on the intelligibility of the stimulus, with native speech eliciting stronger responses than foreign speech (Perez et al., 2015). Our aim was to investigate whether the acoustic properties in the syllable onsets of natural speech modulate neural entrainment. The effects of syllable onsets on neural entrainment was explored both in the native (English) and foreign (Russian) language. Methods The quality of the acoustic edge provided by the syllable onsets was varied in each language condition by using either ‘strong’ (‘b’,’d’,’g’,’k’,’p’,’t’) or weak (‘f’,’l’,’r’,’s’,’v’,’z’) consonants. For each language, we used 20 sets of stimuli comprising of a strong, weak and filler sentence. Sentences were between 8-11 syllables long and were all matched for length (in syllables), lexical stress and syntactic structure. Each sentence was repeated four times. Nineteen native speakers of English listened to the stimuli and were required to press ‘Space’ as soon as they heard a cough (that always occurred within a filler sentence). EEG was recorded using a 32-channel BrainProducts system at 1KHz sampling rate. Analysis For each subject, we computed the cerebro-acoustic coherence between the EEG signal of individual channels in each trial, and the sound envelope of the stimulus. Sharpness values for each stimulus were obtained as the positive derivative of the sound envelope, normalised by the integral of the envelope. Results (preliminary) Sharpness values were significantly greater in the strong versus weak condition across both language, but there was no difference between English and Russian. Coherence values clearly follow different trends in each condition. The present analysis focused on the theta (4-8 Hz) and gamma (25-40 Hz) ranges only. In line with previous findings, coherence was significantly greater in the theta than in the gamma range (p<.05). In the theta range, we obtained greater coherence to weak than strong consonants, and this was only significant in the native language condition. In the gamma range, strong English stimuli elicited higher coherence than the other conditions, although not significantly. While we expected more entrainment to strong versus weak stimuli in both the theta and gamma ranges, it is possible that acoustic edges as defined by their phonetic characteristics may only aid neural entrainment in their corresponding frequency domain. Instead, neural entrainment in the theta range may reflect a process of tracking syllable onsets, and therefore temporal properties of speech, rather than acoustic ones. We are currently conducting further analyses to shed light onto this phenomenon.

Topic Area: Perception: Speech Perception and Audiovisual Integration

Back to Poster Schedule