“Statistical” Learning in Auditory Categorization and Speech Perception"
Mon, March 25, 2013 • 12:00 PM - 1:00 PM • SEA 4.244
ABSTRACT: The study of perceptual learning and categorization in audition has lagged behind comparable research in vision. Whereas there is plenty of overlap in the questions involved in the study of the two modalities, the areas of emphasis differ because of the different demands of the phenomena that are the major motivators of the research (object recognition in vision versus speech perception in audition). I will present some of our work on categorization of speech sounds and the formation of novel non‐ speech categories. In particular, I will discuss factors that influence the perceptual “weighting” applied to different auditory dimensions in making category judgments. These “weighting” strategies develop as a function of the structure of the training distributions and are adaptable to changes in the “informativeness” of the cues for particular situations. Inefficient dimension weighting is an important factor in the difficulty people have distinguishing the speech sounds of a foreign language (and learning to produce these sounds). Whereas this work will point to some of the successes in our understanding of auditory category learning, I will also point out where the major roadblocks for progress lie and how they relate to some of the work in vision.