AIM: To compare speech recognition, in quiet and in noise, among elderly people with and without a dead region in the cochlea. CASE REPORT: The study was performed at the Laboratory of Hearing Aids at the Universidade Federal de Santa Maria. Thirty-eight elderly individuals, aged 61 to 84 years, with mild to moderate sensorineural hearing loss and a descending audiometric curve participated in the study. Each participant submitted to a basic audiologic evaluation and then performed a task with white noise masking. The results were used to differentiate a Group A (GA), 24 subjects without dead regions, and Group B (GB), 14 individuals with presumed dead regions. After fitting each of these elderly individuals with hearing aids based on the results of the Portuguese Sentences Lists Test (PSL), each was evaluated using the Sentences Recognition Indexes in Quiet (SRIQ).Average GA age was 72.25 years and average GB age was77.71 years(p=0.010541). Regarding the SRIQ, the GA average was 61.24% and the GB average, 55.27% (p=0.340076). SRIN differed significantly between groups (p=0.49384), with an average of 55.99% in GA and 67.12% in GB. CONCLUSION: Subjects with presumed dead regions tended to be older than those without dead regions. The presence of dead regions did not appear to influence speech recognition, and was in fact associated with improved hearing in noisy environments. This suggests the development of compensation mechanisms in unfavorable situations.