The effects of brain injury and talker characteristics on speech processing in a single-talker interference task.

Abstract

People who have experienced traumatic brain injury (TBI) often report increased difficulty processing speech than they experienced before their injuries; and this difficulty is not always accompanied by impaired performance on standardized tests of language, auditory processing, or other cognitive factors. Unfortunately, there is a paucity of research directly addressing this issue. Prior studies have mostly focused on measures of characteristics such as attention (e.g. Schmitter-Edgecombe & Nissley, 2000), dichotic listening (e.g. Meyers, Roberts, Bayless, Volkert, & Evitts, 2002), and masking release (e.g. Krause, Nelson, & Kennedy, 2009) in people with TBI. All of these factors may play their part in functional listening tasks, but separately they cannot capture the complex task of speech processing (e.g. Wilson 2003). The goal of this dissertation is to build on existing work to explore and compare the issue of speech processing after brain injury across several levels: sentence repetition with single-talker interference, using targets spoken by native- and non-native-accented speakers of English; standardized testing of cognition and auditory processing; and semi-structured interviews about participants’ subjective experiences with complex speech processing. This study comprised two experiments. The purpose of Experiment 1 was to demonstrate the effects of simulated peripheral hearing loss on performance on the sentence repetition task. Participants in Experiment 1 were 30 healthy young adults, 15 of whom completed the sentence repetition task with stimuli that were filtered with a 1400 Hz low-pass filter to simulate peripheral hearing loss. The other 15 participants completed the task with unfiltered stimuli. The purpose of Experiment 2 was to compare adults with and without TBI on the (unfiltered) sentence repetition task, and included 13 people in each participant group. Groups were matched for age, education, and estimated verbal IQ within each experiment. Dependent variables for the sentence repetition task were accuracy and subjective effort. Participants in Experiment 2 also completed a battery of standardized tests and a semi-structured interview about their subjective experiences with speech processing. Results of Experiment 1 showed that the filtered group was less accurate and reported greater effort than the unfiltered group, and the filtered group showed greater effects of speaker accent. Results of Experiment 2 demonstrated that the TBI group had poorer accuracy than controls, but not significantly higher effort. As predicted, the TBI group also reported more subjective difficulty with complex speech processing than healthy controls, but there was no direct correlation between interview reports and accuracy on the sentence repetition task. Comparisons between the TBI group of Experiment 2 and the filtered group of Experiment 1 suggest that, although the TBI group did show impaired performance on the speech processing task, their performance was not consistent with the peripheral auditory effects that were modeled in Experiment 1. This suggests that the speech processing difficulty experienced by people with TBI cannot be explained by bottom-up processing (e.g. energetic masking) alone. Further research is needed in order to better understand the nature of speech processing problems after TBI, the ultimate goal of which is to be able to develop therapies and strategies that will allow people with TBI to communicate successfully even in difficult circumstances.