Spectro-Temporal Modulation Sensitivity Prospectively Predicts Speech-in-Noise Recognition in Cochlear Implant Users.
OBJECTIVES For cochlear implant (CI) patients, reliable predictors of speech outcome are desirable. In this longitudinal study, we examined the contributions of spectro-temporal sensitivity to successful speech recognition following cochlear implantation. DESIGN We assessed N = 46 recently implanted adult patients shortly after CI activation (T1) and 6 months later (T2) with an adaptive ripple discrimination paradigm where dynamic ripples varied around the temporal rate of 4 Hz and spectral scale of 0.5 cyc/oct. At time points T1 and T2, we evaluated speech-in-quiet recognition using the Freiburg number and monosyllabic word test. The Oldenburg speech-in-noise test was administered 1 year after implantation (T3) to a subset of N = 36 CI recipients. RESULTS Shortly after implantation (T1), temporal ripple discrimination thresholds predicted speech-in-noise recognition 1 year later (T3; Pearson's r = 0.51). In a linear model predicting 1-year speech-in-noise outcome, the predictors temporal or spectral thresholds at T1 and age performed better than speech measures such as Freiburg wordrecognition. CONCLUSIONS A simple spectro-temporal ripple discrimination test is a reliable predictor of speech-in-noise outcome 1 year after cochlear implantation, over and above established clinical speech tests. It offers an efficient method in clinical settings to improve the early prediction of speech outcome.