Masking release, processing speed and listening effort in adults with traumatic brain injury

Abstract

Purpose: This preliminary study explored differences between adults with and without traumatic brain injury (TBI) for speech processing accuracy, processing speed and effort in various conditions of interference.

Methods: Ten adults with TBI and six adults without TBI participated. Speech processing was studied using sentence repetition in six listening conditions with different types of interference, including noise and two simultaneous talkers. Participants repeated sentences and rated effort. Participants also completed standardized tests of cognition, including working memory and processing speed measures.

Results: Sentence repetition accuracy did not differ between groups. However, the TBI group demonstrated slower processing speed than the control group and also reported significantly greater effort in the two-talker condition. Faster processing speed was also correlated with higher accuracy in the two-talker condition.

Conclusions: Results of this study show group similarities in repetition accuracy across listening conditions, but group differences in speed and effort. This preliminary finding, as well as the relationship between processing speed and repetition accuracy, suggests that it is only in the most complex listening conditions that the effects of brain injury may be detectable.

• Head injury
• sentence repetition
• speech processing

Acknowledgements

Many thanks to Liz Crump, Yingjiu Nie, Pradeep Ramanathan and Krystle Baumgarten for their help and support.

Notes

1. Standardized assessment group comparisons were repeated using non-parametric Mann-Whitney tests and the pattern of significant group differences remained the same.

2. The ANOVA was repeated using RAU-transformed accuracy data for all conditions in order to adjust further for floor and ceiling effects. This analysis showed no change in the results, so the findings of the original analysis are reported here.

3. Greenhouse-Geisser corrected results are reported when Mauchly’s test of sphericity showed a violation of sphericity with episilon < 0.75 [66]. Levene’s tests of equality of error variance were significant for the steady noise (p = 0.02) and two-talker (p = 0.006) conditions. In the steady noise condition, variance in the control group was near zero, while the TBI group had slightly greater variability in responses. The variability of the two-talker condition is examined further in the discussion section. Levene’s tests were not significant for the 4 Hz and 16 Hz gated noise and gated speech conditions (p = 0.26, 0.70 and 0.64, respectively).

4. Greenhouse-Geisser corrected results are again used here to adjust for Mauchly’s test of sphericity episilon <0.75.