No Learning Effect Observed for Reception Thresholds for Sentences in Noise

Abstract

This study examined reception thresholds for sentences (RTSs) as a function of test session (N = 5) and noise (continuous and interrupted) in normal-hearing adults. It was hypothesized that RTSs would be superior in interrupted noise and would be stable across repeated testing. Twenty-five normal-hearing adults participated. RTSs were determined with Hearing in Noise Test sentences in continuous and interrupted noise presented at 65 dBA. An adaptive technique was used where sentences varied in intensity to converge on a level of 50% of correct performance. Sentence lists were counterbalanced with 5 unique lists in both continuous and interrupted noise. RTS signal-to-noise ratios were significantly better in the interrupted noise (p < .0001). There was no effect of test session (p = .12) or a Test Session × Noise interaction (p = .13). Stable RTS signal-to-noise ratios across test sessions in both noises are consistent with the notion that a learning effect was not present in noise. Further, one may conclude that Hearing in Noise Test sentences provide stable measures of sentence recognition thresholds in normal-hearing adults over time so long as sentences are unique or are not repeated.