Working Memory Supports Listening in Noise for Persons with Hearing Impairment DOI: 10.3766/jaaa.22.3.4 Mary Rudner*† Jerker Ro ¨nnberg*† Thomas Lunner*†‡§ Abstract Background: Previous studies have demonstrated a relation between cognitive capacity, in particular working memory, and the ability to understand speech in noise with different types of hearing aid signal processing. Purpose: The present study investigates the relation between working memory capacity and the speech recognition performance of persons with hearing impairment under both aided and unaided conditions, following a period of familiarization to both fast- and slow-acting compression settings in the participants’ own hearing aids. Research Design: Speech recognition was tested in modulated and steady state noise with fast and slow compression release settings (for aided conditions) with each of two materials. Working memory capacity was also measured. Study Sample: Thirty experienced hearing aid users with a mean age of 70 yr (SD 5 7.8) and pure-tone average hearing threshold across the frequencies 0.25, 0.5, 1, 2, 3, 4, and 6 kHz (PTA 7 ) and for both ears of 45.8 dB HL (SD 5 6.6). Intervention: 9 wk experience with each of fast-acting and slow-acting compression. Data Collection and Analysis: Speech recognition data were analyzed using repeated measures anal- ysis of variance with the within-subjects factors of material (high constraint, low constraint), noise type (steady state, modulated), and compression (fast, slow), and the between-subjects factor working mem- ory capacity (high, low). Results: With high constraint material, there were three-way interactions including noise type and work- ing memory as well as compression, in aided conditions, and performance level, in unaided conditions, but no effects of either working memory or compression with low constraint material. Investigation of simple main effects showed a significant effect of working memory during speech recognition under con- ditions of both “high degradation” (modulated noise, fast-acting compression, low signal-to-noise ratio [SNR]) and “low degradation” (steady state noise, slow-acting compression, high SNR). The finding of supe- rior performance of persons with high working memory capacity in modulated noise with fast-acting com- pression agrees with findings of previous studies including a familiarization period of at least 9 wk, in contrast to studies with familiarization of 4 wk or less that have shown that persons with lower cognitive capacity may benefit from slow-acting compression. Conclusions: Working memory is a crucial factor in speech understanding in noise for persons with hearing impairment, irrespective of whether hearing is aided or unaided. Working memory supports speech understanding in noise under conditions of both “high degradation” and “low degradation.” A sub- component view of working memory may contribute to our understanding of these phenomena. The effect *Linnaeus Centre HEAD, Swedish Institute for Disability Research, Linko ¨ping University, Sweden; †Department of Behavioural Sciences and Learning, Linko ¨ping University, Sweden; ‡Department of Clinical and Experimental Medicine, Linko ¨ping University, Sweden; §Oticon A/S, Research Centre Eriksholm, Snekkersten, Denmark Mary Rudner, Ph.D., Department of Behavioural Sciences and Learning, Linko ¨ping University, SE-581 83 Linko ¨ping, Sweden; E-mail: mary. rudner@liu.se Parts of this paper were presented orally at the Workshop on Psycholinguistic Approaches to Speech Recognition in Adverse Conditions, March 8–10, 2010, University of Bristol. J Am Acad Audiol 22:156–167 (2011) 156