The human cochlea is mature at birth; however, axonal, dendritic, and synaptic maturation and myelination continue to develop in the brainstem into early childhood and in the cerebral cortex into late childhood [
In case of moderate HL in childhood, hearing aids (HAs) can improve speech audibility and facilitate language development, assuming that the auditory cortical areas are functional [
The successive peaks of CAEPs correspond to the spatiotemporal involvement of the cortical auditory generators and are therefore influenced by cortical maturation [
The aim of this study was to use CAEPs to investigate cortical auditory processing in regard to pediatric HA users with different levels of language ability. We hypothesized that temporal auditory responses and/or their sensitivity to stimulus rate would reflect different levels of language ability in these patients.
In this cortical electrophysiological study, we included children with symmetrical bilateral sensorineural HL fitted with HAs and aged between 8 and 12 years. Children in this age range were chosen for this pilot study as they are likely to understand and follow instructions during testing (i.e., CAEP, audiometry, and language tests). Participants were recruited from the Pediatric Unit of the Otolaryngology Department during clinical follow-up visits.
We reviewed patient charts of pediatric HA users. Thirty children had symmetrical bilateral sensorineural HL and were between 8 and 12 years old. Eleven children (8 males, 3 females), aged between 8.3 and 12.8 years (mean: 10.9 yrs), fitted with bilateral HAs for a bilateral moderate sensorineural HL were accepted into the study. A control group of 11 age- and gender-matched NH children with normal language development, as evaluated using the battery of oral language (evaluated with BILO battery including receptive and expressive language skills; see below for explanation), was also recruited for the study. For all participants, French was the main language spoken at home. All participants were right-handed.
Aided and unaided pure tone average (PTA) thresholds (averaged across audiometric frequencies 0.5, 1, 2, and 4 kHz) were <20 dB for pediatric HA and NH participants, respectively. Demographic data for pediatric HA users are shown in Table
Demographic information for the pediatric HA users with bilateral moderate sensorineural HL.
Subject | Gender | Age |
Duration of auditory |
Age at which child received |
Experience xith HA |
Aetiology | Right |
Left |
BILO |
Group |
---|---|---|---|---|---|---|---|---|---|---|
1 | M | 8; 3 | 6 | 3; 5 | 4; 10 | Unknown | 63 | 69 |
|
HL− |
2 | M | 8; 5 | 3 | 2; 1 | 6; 4 | Familial | 64 | 65 |
|
HL− |
3 | M | 8; 11 | 4 | 2 | 6; 11 | Familial | 66 | 66 |
|
HL− |
4 | F | 9; 7 | 2 | 6; 1 | 3; 6 | Familial | 56 | 56 |
|
HL− |
5 | M | 10; 5 | 2 | 5; 11 | 4; 6 | Unknown | 45 | 41 |
|
HL+ |
6 | M | 11; 5 | 2 | 5; 11 | 5; 6 | Unknown | 41 | 43 |
|
HL+ |
7 | M | 11; 8 | 1 | 3; 5 | 8; 3 | Familial | 55 | 66 |
|
HL+ |
8 | F | 12; 1 | 18 | 8; 11 | 3; 2 | Unknown | 40 | 41 |
|
HL+ |
9 | F | 12; 2 | 6 | 4; 10 | 7; 4 | Familial | 59 | 45 |
|
HL+ |
10 | M | 12; 6 | 5 | 8; 5 | 4; 1 | Familial | 41 | 41 |
|
HL− |
11 | M | 12; 10 | 6 | 5; 2 | 7; 8 | Unknown | 41 | 42 |
|
HL+ |
Note: M: male; F: female; yrs: years; HA: hearing aid; PTA: aided pure tone audiometric threshold averaged over 0.5, 1, 2, and 4 kHz; BILO: Batterie Informatisée du Langage Oral; HL+: good language ability; HL−: fair language ability.
The Ethics Committee of the University Hospital of Tours approved the protocol, and written informed consent was obtained from the parents and assent from the children.
Spoken language and literacy skills were assessed using BILO, a set of standardized, computerized French language tests [
As described in Delage and Tuller [
Participants were tested while sitting on an armchair in a dimly lit, sound-insulated room; pediatric HA participants were tested while wearing their HAs. Participants’ mother or father accompanied them in the room during testing. The stimuli were comprised of 50-ms tone bursts (1100 Hz) delivered through two loudspeakers placed symmetrically on each side of the computer screen. The tone stimuli were presented via Neuroscan Stim2 software. The stimuli were presented at four different interstimulus intervals: 700 (i1), 1100 (i2), 1500 (i3), and 3000 (i4) ms. The sound intensity was 70 dBA measured at the head of the participant.
EEG recordings were obtained using 28 Ag-AgCl cup electrodes (Ffz, Fz, Cz, Pz, O1, F3, FC1, FT3, C3, T3, CP1, TP3, P3, T5, and F7) and their counterparts on the right hemiscalp. Electrodes were placed according to the 10-20 system, as well as the left and right mastoids (M1 and M2), and referenced to the nose. In addition, to detect ocular artifacts, vertical electrooculogram (EOG) data were recorded from two electrodes above and below the right eye (vertical bipolar).
The EEG and EOG were digitized (Neuroscan Synamps amplifier, Scan 4.3, Compumedics Corp., El Paso, TX) at a sampling rate of 500 Hz. The EEG was amplified and bandpass-filtered (0.3–100 Hz). Electrode impedances were kept below 10 kΩ. Eye movement artifacts were eliminated using a spatial filter transform developed by Neuroscan, and EEG periods with movement artifacts were rejected manually. A digital zero-phase-shift low-pass filter (30 Hz) was then applied to the EEG.
CAEPs were analyzed with the ELAN software [
The influence of interstimulus interval on each peak of the CAEPs was analyzed using Friedman nonparametric analyses of variance. Amplitudes and latency peaks measured in HL+ and HL− children and the control group were compared using nonparametric Mann-Whitney rank tests.
HL+ and HL− children did not differ, using Wilcoxon Mann-Whitney test, on age at testing (HL+: 11.7 years old ± 0.8; HL−: 9.5 years old ± 1.7;
Frontocentral CAEPs displayed similar successive N1b-P2-N250 peaks across the four groups. No significant difference was observed for the N1b peak amplitudes and latencies between HL+ and HL+ controls, or between HL− and HL− controls (Figure
Cortical AEPs. Midline responses (Cz) at different interstimulus intervals (i1, i2, i3, and i4) for the HL+ (a), HL+ controls (b), HL− (c), and HL− controls (d). The asterisks indicate significant effect of interstimulus interval (
Because no significant effect of interstimulus interval was found for the amplitude and latency of the successive peaks recorded at temporal sites N1a, P1t, N1c, and P2t, the CAEPs were averaged across the 4 interstimulus intervals to increase the signal to noise ratio (SNR).
Although the grand average N1a peak amplitude was greater in the HL+ and HL− groups than in their respective controls (mainly on the left temporal site), the difference was not statistically significant. The subsequent P1t, N1c, and P2t waves also were not significantly different between the HL+ and HL+ control groups. However, the N1c and P2t amplitudes were significantly smaller (only on the right temporal site) in the HL− group than in the HL− controls (N1c: HL− = −0.8
CAEP T3 (a and c) and T4 temporal responses (b and d) at i1, i2, i3, and i4 for the HL+ and HL+ control groups (a and b) and for the HL− and HL− control groups (c and d). The asterisks indicate significant differences (
The present pilot study provided interesting preliminary findings regarding the relationship between CAEP characteristics and language ability in 8- to 12-year-old pediatric HA users. Atypical CAEPs were observed at temporal recording sites for pediatric HA users with some degree of language impairment. In normal development, temporal CAEPs typically display a stable morphology through childhood and particularly for the age range of the present study, with successive N1a and N1c peaks [
In our study, despite a normal (or greater than normal) amplitude of the early temporal peaks (N1a, P1t) for all pediatric HA patients, the later temporal responses (N1c and P2t) were reduced or absent in the HL− group. This does not appear to be due to an absence of cortical auditory input because waves N1a and P1t were present and normal in all the pediatric HA patients. N1c wave was absent or smaller in the HL− group. This relationship between N1c abnormalities (N1c being reduced or absent) and language impairment has previously been shown in other clinical populations with language impairment such as in children with autism [
Relatively few studies have investigated CAEP temporal responses in pediatric HA users. Most of these studies have focused on frontocentral responses (especially the latency of the P1 peak recorded at the vertex) in deaf children who use cochlear implants [
Unlike the temporal responses, the frontocentral responses of normally developing children are greatly influenced by age. The smaller N1b peak amplitude in the HL− and HL− control groups, compared to the well-defined peaks and greater amplitude of N1b in the HL+ and HL+ control groups, might be due to age differences between groups. This result is in agreement with the literature indicating the emergence of N1b at approximately 8–10 years of age and greater N1b amplitude at 10–12 years of age. The greater P2 peak amplitude observed in HL+ and HL− groups than in the controls may be related to HA amplification, as described in previous studies [
In this study, children with HA and good language ability (HL+) were older than those with language impairment (HL−). Language abilities were evaluated with BILO, which is calibrated in order to allow comparisons in language ability across age groups. Moreover, CAEPs were compared with age- and gender-matched control children.
Although further longitudinal studies are needed with larger sample of children, these preliminary results suggest that abnormal CAEP responses recorded at temporal sites might underlie language impairment in pediatric HA users who have a moderate sensorineural HL.
The authors declare that there is no conflict of interests regarding the publication of this paper.
This study was supported by the French ENT Society.