The effect of three practice conditions on the consistency of chronic dysarthric speech

This study investigated whether it is possible for people with chronic dysarthria to adjust their articulation in three practice conditions. A speaker dependent, speech recognition system was used to compare participants' practice attempts with a model of a word made from previous recordings to give a recognition score. This score was used to indicate changes in production of practice words with different conditions. The three conditions were reading of written target words, visual feedback, and an auditory model followed by visual feedback. For eight participants with dysarthria, the ability to alter speech production was shown, together with a differential effect of the three conditions. Copying an auditory target gave significantly better recognition scores than just repeating the word. Visual feedback was no more effective than repetition alone. For four control participants, visual feedback did produce significantly better recognition scores than just repetition of written words, and the presence of an auditory model was Significantly more effective than visual feedback. Possible reasons for differences between conditions are discussed