Wednesday, May 27, 2009

SRL Talk: Nicholas Altieri, Andrew Kirk and David B. Pisoni

The Speech Research Lab's next meeting will feature a talk by Nicholas Altieri, Andrew Kirk and Prof. David B. Pisoni of IU Bloomington. All are invited and welcome to attend.

Friday, May 29, 2009
1:30 - 3:00 p.m.
Psychology 128

" Can Visual-Spatial Working Memory Training Improve Lip-reading Ability? A first report of negative findings" will be presented by Nicholas Altieri, Andrew Kirk, and David B. Pisoni.

Abstract:
Decades of research on audiovisual speech perception has led to a large body work examining accuracy on visual-only speech recognition (i.e., lip-reading). In contrast to accuracy scores for auditory-only stimuli in quiet for normal hearing listeners, visual-only accuracy is highly variable and the sources of this variability in lip-reading skills remain unknown. The first goal of our recent study was to investigate whether cognitive functions such as attention, the ability to inhibit irrelevant information, and visual-spatial working memory skills correlate with lip-reading performance on CUNY sentences. Secondly, we investigated whether lip reading skills could be improved after several days of training on a visual-spatial working memory task. In our experiment, subjects participated in three days of visual-spatial working memory training in which they received either probabilistic-adaptive sequences or random sequences. We also assessed attention and inhibition using the Stroop color-word naming test, verbal working memory using forward and backward digit span, and lip-reading skills using visual only sentence recognition on separate days before and after training. The results failed to yield significant correlations between any of the cognitive factors and lip-reading ability. We also failed to find evidence that training facilitated visual-only accuracy. We are currently considering different approaches and several new training paradigms that might improve visual-only speech recognition. Suggestions on these matters would be greatly appreciated from the IU research community.

No comments: