Two areas responded significantly differently to mixed and blocked talkers these were middle/superior temporal areas, and the superior parietal lobule. (2004) identified brain areas that were differentially active in blocked-talker and mixed-talker conditions. The results showed that listeners were slower to detect the target word in “mixed talker” conditions, which was expected from prior research (e.g., Nusbaum & Magnuson, 1997). In the “mixed talker” condition, the words were spoken by four different talkers. Each participant listened to lists in two conditions: in the “blocked by talker” condition, the words were all spoken by the same talker. In this study, participants listened to lists of individual spoken words, and their task was to recognize a target word in the list. (2004) used fMRI to investigate contributions to talker normalization, both from traditional superior temporal language areas, and from a more distributed attention network. In order to examine the underlying neural mechanism for the effects of talker change, Wong et al. Further, changes in talker result in slowed recognition and/or categorization for CV syllables, vowels, and whole words ( Mullennix & Pisoni, 1990, Nusbaum & Morin, 1992 Kaganovich, Francis, & Melara, 2006 Wong, Nusbaum, & Small, 2004). These results strongly suggest that adjusting to a new talker draws on WM resources. For example, listeners recognize spoken target syllables more slowly when the talker changes when listeners must maintain a high WM load compared to a low load yet when the talker does not change, the high WM load does not have the same effect ( Nusbaum & Morin, 1992). Though listeners can adjust to a new talker quickly, and are not usually aware of the fact that they are sensitive to the vocal characteristics of the new talker, there is evidence that this process may use working memory (WM), possibly to selectively direct perceptual attention towards acoustic cues needed to calibrate the speech for recognition. One explanation of this performance reduction is talker “normalization”: the process by which listeners use talker vocal characteristics to resolve acoustic-phonetic ambiguities that are introduced when a talker changes demands attention and extra processing ( Nusbaum & Morin, 1992). Despite this acoustic-phonetic variability, listeners appear to quickly and easily understand utterances from different talkers, albeit with a small but reliable recognition performance reduction ( Heald & Nusbaum, 2014b Nusbaum & Magnuson, 1997). This variability in the relationship between acoustic patterns and phonetic categories introduces ambiguity into the recognition of a speaker’s intended phoneme given that any particular acoustic pattern might map onto different phonetic categories ( Nusbaum & Magnuson, 1997). Even more acoustic-phonetic variability is introduced when accents and speech patterns specific to non-native speakers are taken into account. However, even among native speakers of a language, there is wide variability in the acoustic characteristics of various phonemes (Peterson & Barney, 1952 Heald & Nusbaum, 2014b). People seem to understand speech from different talkers with little difficulty. Though there may be differences in cortical recruitment to processing demands for non-speech sounds versus a changing talker, the underlying mechanisms are similar, supporting the view that shared cognitive-general mechanisms assist both talker normalization and speech-to-nonspeech transitions. We extend these findings to acoustic source change in general by examining understanding environmental sounds in spoken sentence context. Here we use source analysis from high-density EEG to show that perceiving fluent speech in which the talker changes recruits early involvement of parietal and temporal cortical areas, suggesting functional involvement of WM and attention in talker normalization. These studies have not examined how talker variability affects perception and neural responses in fluent speech. Previous research has indicated that adjusting to talker differences is an active cognitive process that depends on attention and working memory (WM). Adjusting to the vocal characteristics of a new talker is important for speech recognition.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |