Neurophysiological conditions for hearing in children using hearing aids or cochlear implants : an intervention and follow-up study

Abstract

OBJECTIVES: The four studies in this thesis examine the central auditory pathways in children with hearing loss (HL) through recording event-related potentials (ERP) and mismatch negativity (MMN). The design of the project is partly experimental, and it also includes an intervention part and a follow-up-study after three years. The primary aims were the following: I. Explore whether a multi-feature paradigm (Optimum-1) for eliciting MMN could characterise difficulties in perceiving small sound contrasts in children with HL using their hearing aids (HAs) or cochlear implants (CIs) (Paper I). II. Investigate whether a computer-assisted reading intervention programme with a phonics approach (GraphoGame) could affect ERP and MMN in children with HL using HAs (Paper II) and CIs (Paper III). III. Examine the developmental changes of ERP and MMN over time (three years) among children with HL using HAs or CIs (Paper IV).

METHODS: In total, children with HL (n=30) using HAs (n=15) or CIs (n=15), as well as a reference group of children with normal hearing (NH) (n=16), participated. All children were approximately 5–8 years old at baseline (the first ERP recording session), and 8–11 years old at the follow-up after three years. Paper I includes all participants at baseline. Paper II (children with HAs) and Paper III (children with CIs) are the intervention part: one month of repeatedly training with a computer-assisted reading intervention programme with a phonics approach (GraphoGame) and involve two ERP recording sessions: before and after the training. Paper IV comprises the follow-up study after three years and includes all three groups: the children using HAs, the children with CIs, and the children with NH. All studies are based on ERP recordings that, including the data processing, were identical for each of the three sessions. The ERP-recordings followed the multi-feature paradigm, Optimum-1. Thus, a standard stimulus alternated among five different deviants (gap, intensity, pitch, location, and duration), presented in a pseudorandom sequence. MMN was calculated from the average ERP of each deviant minus the standard stimulus. Analysis of variance (ANOVA) was used for the statistical analyses. Paper I served as a model for analyses and interpretations of the results. The results in Paper II–IV were based on samples within a specific time interval: 80–224 ms. The method is non-invasive and safe for the participants. MMN is independent of attention, measured by using an electrode net. Optimum-1 enables short recording sessions. Together, it is a method appropriate for testing children.

RESULTS: Paper I demonstrated that the multi-feature paradigm, Optimum-1, could elicit responses in children with HL using their HAs or CIs. Four time windows (TW) were created to structure and facilitate further analyses. TW 2 of 80–220 ms was considered appropriate for the primary test of MMN effects. Overall, the response amplitudes were smaller in the HA group than the NH groups in TW 2. Otherwise, the results could not statistically prove any major differences in discrimination between groups. Paper II established significant differences in the obligatory responses in both ERP and MMN between the NH and HA groups before the computer-assisted training, which disappeared after the intervention. This suggests possible training effects among the children with HAs. This paper also provides a description of MMN and positive mismatch response (pMMR) in all deviants and groups (NH vs HA). Paper III could not statistically demonstrate any computer-assisted training effects detectable with ERPs and MMN among children with CIs. Thus, the results differ from the results regarding children with HAs in Paper II, suggesting that there are differences between these two subgroups of children with HL. The paper also offers a description of MMN and pMMR regarding all deviants in each group (NH vs CI). Paper IV found a significant difference in mean ERP at baseline compared to the time of follow-up three years later in the HA group. This suggests a possible catch-up over time among the children with HAs. On the contrary, the obligatory responses in ERP among the children with CI were significantly lower than both the children with NH and HAs after three years, indicating impaired development of the central auditory system among the children with CIs. There was a high degree of inter-individual variability of both MMNs and pMMRs due to maturational changes in the current age groups. Only small changes were to be expected after training (Paper II and III) and after follow-up (Paper IV), which, together with small sample sizes, may have diminished the ability to demonstrate significant results.