Influence of Internal Phonetic Category Structure in On-Line Speech Processing

Author(s):  
Sarah C. Wayland ◽  
Joanne L. Miller
1997 ◽  
Vol 40 (6) ◽  
pp. 1406-1424 ◽  
Author(s):  
Joan E. Sussman ◽  
Brian Gekas

The current investigation examined the structure of the phonetic category [I] for 13 listeners. Experiments reported are results from identification, "best exemplar," and discrimination tasks using 105 [I] stimuli. The tokens were synthesized along a mel-spaced vowel continuum that differed in first and second formants. All stimuli ended in a 30 ms [b] sound. Results showed that 10 of 13 listeners demonstrated differing choices of the best exemplars, although most were within 37.5 mels of the central best exemplar chosen in the first experiment. Seven of the participants demonstrated "circular" patterns in identification of the [I] category that appeared to be organized around a central "best exemplar." Six participants showed other identification patterns: "downward, " "upward," and "left-extending," with "best exemplars" on an edge or border of the phonetic categories. Graded category structure from a central "best exemplar" was apparent only in the averaged identification results, and not for individual participants. The size of the [I] category was significantly smaller than that surrounding the [i] best exemplar reported in a prior study by Sussman and Lauckner-Morano (1995). Finally, listeners had equivalent or better discrimination sensitivity with the best exemplar as the fixed standard compared to that for a "poor" exemplar token 45 mels away from the best exemplar. Results showed that phonetic category structure for the lax vowel [I] was different from the similar, but tense vowel [i]. The findings question whether prototype theory is generalizable to vowel categories other than [i].


2016 ◽  
Vol 140 (4) ◽  
pp. EL307-EL313 ◽  
Author(s):  
Julia R. Drouin ◽  
Rachel M. Theodore ◽  
Emily B. Myers

2018 ◽  
Author(s):  
Sevada Hovsepyan ◽  
Itsaso Olasagasti ◽  
Anne-Lise Giraud

ABSTRACTSpeech comprehension requires segmenting continuous speech to connect it on-line with discrete linguistic neural representations. This process relies on theta-gamma oscillation coupling, which tracks syllables and encodes them in decipherable neural activity. Speech comprehension also strongly depends on contextual cues predicting speech structure and content. To explore the effects of theta-gamma coupling on bottom-up/top-down dynamics during on-line speech perception, we designed a generative model that can recognize syllable sequences in continuous speech. The model uses theta oscillations to detect syllable onsets and align both gamma-rate encoding activity with syllable boundaries and predictions with speech input. We observed that the model performed best when theta oscillations were used to align gamma units with input syllables, i.e. when bidirectional information flows were coordinated, and internal timing knowledge was exploited. This work demonstrates that notions of predictive coding and neural oscillations can usefully be brought together to account for dynamic on-line sensory processing.


Sign in / Sign up

Export Citation Format

Share Document