International Journal of Pediatric Otorhinolaryngology 140 (2021) 110519
Available online 24 November 2020
0165-5876/© 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Auditory event-related potentials and mismatch negativity in children with
hearing loss using hearing aids or cochlear implants – A three-year
follow-up study
Elisabet Engstr¨om
a,b,*, Petter Kallioinen
c, Cecilia Nakeva von Mentzer
d, Magnus Lindgren
e,f,
Birgitta Sahl´en
e,g, Bj¨orn Lyxell
h,i, Marianne Ors
j, Inger Uhl´en
a,baDepartment of Clinical Science, Intervention and Technology (CLINTEC), Karolinska Institutet, 171 77, Stockholm, Sweden bDepartment of Otoneurology, Karolinska University Hospital, 141 86, Stockholm, Sweden
cDepartment of Linguistics, Stockholm University, 106 91, Stockholm, Sweden dSchool Health Sciences, ¨Orebro University, 701 82, ¨Orebro, Sweden eCognition, Communication & Learning, Lund University, 221 00, Lund, Sweden fDepartment of Psychology, Lund University, 221 00, Lund, Sweden
gLund University, Faculty of Medicine, Department of Clinical Sciences, Logopedics, Phoniatrics & Audiology, Lasarettsgatan 21, 22185, Lund, Sweden hDepartment of Behavioral Sciences and Learning, Swedish Institute for Disability Research, Link¨oping University, 581 83, Link¨oping, Sweden iDepartment of Special Needs Education, University of Oslo, Oslo, Norway
jDepartment of Clinical Neurophysiology, Skåne University Hospit, 221 85, Lund, Sweden
A R T I C L E I N F O Keywords:
ERP MMN
Sensorineural hearing loss Hearing aids
Cochlear implants Children Optimum-1
A B S T R A C T
Objectives: The primary aim was to examine how event-related potentials (ERPs) and mismatch negativity (MMN) change and develop over time among children with hearing loss (HL) using hearing aids (HAs) or cochlear implants (CIs). Children with normal hearing (NH) were tested as a reference group.
Methods: This three-year follow-up study included 13 children with sensorineural HL (SNHL); 7 children using bilateral HAs and 6 children using CIs; and 10 children with NH as a reference group. ERPs were recorded at baseline and after three years. At time for the original study the children were approximately 5–8 years old and at the follow-up study 8–11 years old. ERP recordings and data processing were identical in both sessions. A standard stimulus alternated with five different deviants (gap, intensity, pitch, location and duration), presented in a pseudorandom sequence, thus following the multi-feature paradigm, Optimum-1. MMN was calculated from the average ERP of each deviant minus the standard stimuli. Repeated measures ANOVA was used for the sta-tistical analyses and the results were based on samples within a specific time interval; 80–224 ms.
Results: There was a statistically significant difference in the obligatory responses between the NH and HA groups at baseline, but this difference disappeared after three years in our follow-up study. The children with HA also showed a significant difference in mean ERP at baseline compared to follow-up, and significant differences between the deviants at follow-up but not at baseline. This suggests an improvement over time among the children with HAs. On the other hand, the children with CIs did not differ from the NH children at baseline, but after three years their mean ERP was significantly lower compared to both the children with HA and NH, indicating a reduced development of the central auditory system in this age span among the children with CIs. Regarding MMN, there was an interaction between the duration deviant and time for the children with HA, also indicating a possible improvement over time among the HA children.
Conclusions: This three-year follow-up study shows neurophysiological differences between children with HL and children with NH. The results suggest a delay in the central auditory processing among the HA children compared to children with NH, but a possible catch-up, over time, and this potential may be worth to be utilized. Regarding the CI children, similar improvement in this age span is missing, meaning there are differences be-tween the subgroups of children with HL, i.e. the children with HAs vs. CIs. The results highlight the importance of distinguishing between subgroups of children with HL in further research.
* Corresponding author. Department of Otoneurology, M53 Karolinska University Hospital Huddinge SE, 141 86, Stockholm Sweden. E-mail address: elisabet.engstrom@sll.se (E. Engstr¨om).
Contents lists available at ScienceDirect
International Journal of Pediatric Otorhinolaryngology
journal homepage: www.elsevier.com/locate/ijporlhttps://doi.org/10.1016/j.ijporl.2020.110519
1. Introduction
The central auditory system undergoes maturational changes over time, and a background to the neurophysiological findings, mainly referring to the auditory event-related potentials (ERPs) and mismatch negativity (MMN), follows below. In this study, the ERPs consider the auditory sensory responses mainly elicited by external stimuli. For this purpose, the multi-feature paradigm, Optimum-1, is used. The time in-terval for the analyses is 80–224 ms, involving the large positivity of the standard stimulus and the period of a typically MMN peak [1]. Furthermore, focus will be on changes of ERPs and MMN over time in children with sensorineural hearing loss (SNHL) using hearing aids (HAs) or cochlear implants (CIs). Children with hearing loss (HL) show speech and language deficits [2,3], making MMN of special interest to study, since it reflects the discrimination ability [4].
The auditory ERP technique provides a well-established method for examining the auditory cortical responses [1,5]. The P1, N1 and P2 components, i.e. positive (P1, P2) and negative (N2) auditory cortical responses, or waves, are elicited passively with auditory stimuli. They undergo maturational and age-related changes, observable in latencies and amplitudes [6–9] making them appropriate to assess the auditory central processing development in children with hearing loss (HL) [10], especially children with a profound SNHL fitted with CIs. Previous works indicate delayed or deviant auditory cortical responses in children with HL [11–13]. Understanding the central auditory development process has contributed to early CI surgery in congenitally deaf children leading to better patient outcome long term. Sharma et al. [12] showed that children implanted after age of 7 years had delayed P1 latencies, suggesting a reduced plasticity of the central auditory system after this age. Without normal auditory stimulation, such as with congenitally deaf children, there is a sensitive period of maximal plasticity of approximately 3.5 years. On the contrary, the latencies of P1 for early-implanted children did not differ from controls with normal hearing (NH). The latencies of the P1 cortical evoked potentials have also been used to infer the maturational status of auditory pathways in congenitally deafened children, who regain hearing after CI surgery [12].
However, studies investigating the auditory cortical responses in children with a mild-to-moderate SNHL are still scarce and the results are ambiguous. The N2 amplitude is reduced in 9–12 years old children with peripheral HL compared to children with NH, suggesting the N2 component as a possible neurophysiological marker of the deficit in the central auditory system in children with HL [14,15]. However, ERPs from 3 to 9 years old children with a mild-to-profound SNHL were similar to those of NH children [16].
Rare stimuli, or deviants, cause stronger responses compared to standard stimuli, and the difference in the ERPs is known as MMN [17]. MMN reflects the discrimination ability. The MMN technique is suitable for testing children, since the technique is non-invasive, safe and does not require any sedatives or active participation [1,5]. Maturational changes during life are also observable in MMN, e.g. changes in the amplitude and latency. The normal development is well outlined in previous research [18–22]. The latency is observed to decrease with increasing age. A positive mismatch response (pMMR) has been observed under the age of 5.5 years [21], likely reflecting an immaturity of the discrimination process. By using different deviants in an oddball tone discrimination paradigm, MMN latency decreases from 4 to 10 years of age; however, the amplitude is more stable in this period [23]. The mismatch responses may be obtained differently by different stimuli [24]-25].
Regarding patients with HL, MMN can be used to assess changes in central auditory pathway post CI implantation [11,26,27]. However, most MMN studies use adult patients with CIs, and data from children with HL remain scarce. Abnormal MMNs in children with a mild-to-moderate SNHL from early childhood suggest changes in the neural processing of sounds also in adolescence [10]. The maturational
changes may interfere with the possible deficits due to the HL and must be taken into consideration when analyzing the results among children. Regarding children with HAs and CIs around 5–8 years old, the presence of both MMN and pMMR have been observed, with an individual and seemingly unpredictable mix, meaning a child could have both MMNs and pMMRs in different deviants at the same time [28,29].
The multi-feature paradigm, Optimum-1, enables fast recordings, measuring five different deviant types (gap, intensity, pitch, location and duration) at the same session [30]. Optimum-1 has been used to objectively observe how 5–8 years old children with HAs or CIs have difficulty perceiving small sound contrasts [31]. The method was also used in an intervention study regarding children with HAs and CIs to determine whether a computer-assisted training program with a phonics approach could affect ERPs and MMN [28,29]. All these children were diagnosed with a SNHL but the degree of HL and type of devices, i.e. HAs vs. CIs, differed between the groups. Significant differences were observed both in the obligatory responses and the MMNs between children with NH and children using HAs before the computer-assisted training, but these differences disappeared after the intervention [28]. Interestingly, similar effect could not be seen among the children with CIs [29]. MMN is thought to be stable after the age of 5 years [24], but it is reasonable to assume that the central auditory pathways in children with HL develop differently from children with NH. The children with HL may have a developmental delay, motivating research in how these differences develop over time. Since previous studies [28,29,31] suggest possible differences in ERPs and MMNs among the subgroups of children with HL, i.e. the children using HAs vs. CIs, it seems important to distinguish between these groups. All the children participating in the original studies [28,29,31] were invited to this three-year follow-up study. Thus, the aim of the study is to examine how ERPs and MMNs change and develop over time among children with HL using HAs or CIs. Children with NH are tested as a reference group.
2. Material and methods
2.1. Participants
2.1.1. Eligibility and recruitment
This is a three-year follow-up study, focusing on electrophysiological recordings in children with HL using HAs and CIs, including children with NH as a reference group. It started as a comprehensive intervention study, investigating ERP and MMN in children with HAs and CIs [31] and whether computer-assisted phonics training could affect ERP and MMN in these children [28,29]. Within the ERP recordings not only MMN but also N400 was examined [32]. Furthermore, behavioral ef-fects and phonological processing skills were tested [33–36].
In the original studies, 30 children with HL using HA or CI partici-pated; 15 children (9 girls) in each group. The CI patients included 9 children (7 girls) with bilateral CIs and 6 children (2 girls) with a CI combined with HA. Since the subgroup of children with combined CI and HA mainly relied on their CI, due to severe or profound HL in the remaining hearing ear, they were studied together with the children Table 1
Characteristics at baseline and follow-up. Cochlear
implants Hearing aids Controls
N 6 7 10
Male sex (n, %) 2 (33%) 2 (29%) 7 (70%)
Age at baseline, years
(median, IQR) 6.4 (5.8–6.9) 5.6 (5.1–7.6) 7.0 (6.3–7.9) Age at follow-up, years
(median, IQR) 9.1 (8.9–9.5) 9.0 (8.5–10.6) 9.7 (9.2–10.2) Follow-up time, years*
(median, IQR) 3.0 (2.5–3.1) 3.3 (2.8–3.4) 2.8 (2.3–2.9) IQR, inter-quartile range. * Time from baseline assessment to follow-up.
with bilateral CI, constituting the CI group. As a reference group, 15 (5 girls) children with NH participated. At the first ERP test session, the age of all children ranged between approximately 5 and 8 years. All children were healthy and native Swedish-speaking. No diseases or syndromes affecting speech and language development were allowed. Most of the children were recruited from the Department of Audiology and Neuro-tology, Karolinska University Hospital, Stockholm but a few more chil-dren were added from Uppsala and Lund. All these chilchil-dren but 1 girl with NH fulfilled the intervention, i.e. four weeks of short and repeat-edly (10 min daily) computer-assisted phonics training. ERP were tested before and immediately after one month of training.
Three years later, all children listed above, were invited to this follow-up study, see Table 1. Among the children with HL, 13 families from Stockholm accepted the invitation of participation: 7 children (5 girls) with bilateral HAs, 4 children with bilateral CIs (2 girls) and 2 children (2 girls) who combined HA and CI at baseline and time for the intervention part. However, one of these latter girls added one CI before the follow-up. As in the previous study [29], the child with combined HA and CI was studied together with the children with bilateral CIs. Thus, this constituted one HA group with 7 children and one CI group with 6 children. In the reference group with NH, 10 children (3 girls) accepted invitation. Of these, 2 children were siblings to the children with HL. NH was ascertained at the ordinary Swedish hearing screening program at the Child Welfare Centers and School Health Service.
Regarding the children also participating at the follow-up, the age span at baseline among the children with HAs was 5 years and 1 month–7 years and 8 months; the children with CIs 4 years and 11 months to 6 years and 10 months; and, the children with NH 5 years and 11 months to 8 years and 1 month. For the follow-up study, the age span among the children with HAs was 7 years and 6 months to 11 years; among the children with CIs 8 years and 4 months to 9 years and 10 months; and, the children with NH 8 years and 10 months to 11 years. All parents or guardians were given written and spoken information about the study. Parental informed consent was obtained from all par-ticipants. The children were individually informed prior to the testing sessions. The study was approved by the Regional Committee for Med-ical Research Ethics in Stockholm, Sweden.
2.1.2. Etiology and hearing
In Supplementary table A, the hearing characteristics among the children with HAs and CIs are shown. The severity of the HL was mainly stable through the period. However, 1 girl received another CI one year before the follow-up. At baseline the pure tone average (PTA) on her remaining hearing ear was >101 dB.
Among the children with HAs, the PTA was generally better at the follow-up, probably due to age and participation in testing. Only 1 boy was showing a tendency to progress of the HL. He was falling from PTA 53 dB to 64 (right ear) and PTA 55 dB–61 dB (left ear). Nothing in his medical history pointed out any differences compared to the other children. His HL was diagnosed due to delayed speech and language development and he did not pass the hearing screening at 4 years. The audiogram showed a mid-frequency HL and he had mild heredity (uncle and cousin). The cause of the HL was not further investigated.
When most of these children were born, testing the otoacoustic emissions (OAE) in newborns was not yet implemented in the national hearing screening program in Sweden. However, some of the children (3 children with HA and 2 children with CI) were tested shortly after birth, resulting in early diagnosis in 2 children with HA and in 1 child with CI. However, diagnosis was still hard to determine in one girl in the HA group and HAs were not fitted in the two other HA children until the age of 1 year and 3 months vs. 2 years and 7 months. Further, one girl in the CI group did pass the OAE test at birth, however, later developing a SNHL due to congenital cytomegalovirus (CMV) infection, see below. Otherwise, a delay of speech and language development was the main reason causing further investigation. This was probably more obvious among the children with severe or profound HL; all the CI children were
diagnosed before the age of 2 years and 6 months. Use of HAs preceded CI surgery and HAs were fitted within 3 months (already within 1 month among 4 children) after diagnosis among the CI children. HL was not diagnosed in the remaining HA children, i.e. the HA children not tested with OAE (4 children), before the age of 4 years and 3 months. Though, HAs were then quickly fitted in these children; all within 4 months after diagnosis. Detailed information about each child with HL is presented in Supplementary table B. Regarding the children with HAs, the hearing thresholds between 250 and 8000 Hz at baseline and follow-up are individually presented in Supplementary table C.
CMV was only tested among the children with CIs and was positive in 1 girl, causing a progressive HL. Gene testing (Connexin 26) was per-formed in 1 CI girl with heredity and was negative. Magnetic resonance imaging (MRI) in the CI children prior to surgery showed no cochlear malformations. The HA children underwent neither MRI nor computed tomography (CT).
2.2. Method 2.2.1. Hearing tests
All participating children with HL were patients at the Department of ENT, Audiology and Neurotology, Karolinska University Hospital, Stockholm. Medical records were studied, and audiograms were plotted for both the original and follow-up studies.
2.2.2. Electrophysiological recordings
ERP were recorded at baseline and at follow-up after approximately 3 years. All children were tested at the Phonetics Laboratory at the Department of Linguistics, Stockholm University, using the same equipment at both sessions, i.e. at the original study and the follow-up study. The children were seated in sound booth watching a silent movie during the experimental procedures. They were asked not to pay any attention to the test stimuli and the testing sessions were designed to be comfortable, to reduce artifacts due to movements.
2.2.2.1. Stimuli and the multi-feature MMN-paradigm. With minor ad-justments due to the hearing devices (the stimuli had to be presented via loudspeakers in front of the child, 45◦on each side, 70 dB HL, and not
through headphones) and current age group (recording time was reduced from 15 to 12 min) [28,29,31], the multi-feature paradigm, Optimum-1, by N¨a¨at¨anen [30] was used at baseline and follow-up. Harmonic tones composed of 500, 1000 and 1500 Hz sinusoidal par-tials with a duration of 75 ms, including 5 ms rise and fall times, constituted the standard stimulus. The intensity of the second partial was 3 dB lower and the third partial was 6 dB lower. The standard stimulus alternated with five different deviants in a pseudorandom sequence. The deviants differed from standard as follows: the gap deviant had a 7 ms silent gap inserted in the middle of the tone; the intensity deviant was 10 dB higher or 10 dB lower (half of each); the pitch deviant was 10% higher or 10% lower (half of each); the location deviant showed an interaural time difference of 800 μs to the right or left side (half of each); the duration deviant was 25 ms. The stimuli were presented at a stimulus-onset-asynchrony (SOA) of 500 ms in two 6-min sequences. Every second stimulus was a deviant and there were around 120 stimuli of each deviant (1244 stimuli in total).
2.2.2.2. Electrophysiological processing. The electrophysiological cortical responses were digitally recorded using HydroCel Geodesic Sensor Net (Net Amp 300, Electrical Geodesics Inc, Eugene, OR, USA) with 129 electrodes. In patients with small head size, nets with 125 electrodes were used. Cz was used as the reference channel. The net was applied according to the technical manual of the manufacturer, for a more detailed description se previous study [28]. The impedance was kept below 50 kΩ.
and follow-up and are also described in previous studies [28,29,31]. The recordings were sampled at 20 000 Hz, followed by low pass filtering online with a cut off at 4000 Hz and resampling at 250 Hz, 1–40 Hz bandpass FIR filtering offline. Continuous data files were resampled to 125 Hz, then divided into epochs 100 ms before through 500 ms after the event. Epochs were excluded if the amplitudes exceeded ±500 μV. Thereafter, data underwent automatic preprocessing procedures and artifact correction performed in EP toolkit by Ref. [37]. If absolute correlation with neighboring channels did fall below 0.4, the channels were excluded. Detection and deletion of artifacts included blink correction by using independent component analysis (ICA). Regarding data from the children with CIs, the data was treated with an additional independent component analysis (ICA) [38–40] after the blink removal. The largest CI artifact is seen within 50 ms after stimulus onset, due to the CI magnetic pulse. Principal component analysis (PCA) was used to remove movement artifacts. Principal components >200 μV in ampli-tude changes were removed as well. Epochs with ampliampli-tude differences >300 μV and ≥25% bad channels were rejected, and these bad channels were replaced via interpolation from the remaining good channels. ERPs
were re-referenced to linked mastoids and baseline adjusted to 100 ms before onset of each stimuli.
2.2.2.3. Analyses of ERP data. Subject average files were created from ERPs from standard stimulus and all deviants. MMN was calculated from each deviant average ERP minus standard average ERP. Seven fronto- central electrodes (number 5, 6, 7, 12, 13, 106 and 112) were selected for the statistical analyses, since we assumed the largest responses were obtained from there [30,37]. A MMN peak could not visually be demonstrated [31]. The analyses were therefore based on samples within 80–224 ms. This time interval includes the large positivity of the standard stimulus, the period of a typically MMN peak [1] and was also used in previous studies involving the children with HAs [28] and CIs [29].
2.2.3. Statistical analyses
We completed a three-way mixed ANOVA to understand the effects of type of deviant, time (change from baseline to follow-up) and groups (children with HAs or CIs and the controls with NH) on the obligatory
Fig. 1. The average responses from a fronto-central electrode (number 6, near Fz) to standard (thin line), the deviants (thick line) and MMN (dotted line), all deviants and groups (NH, HA and CI), at baseline and at follow-up. Durd, duration deviant; Gapd, gap deviant; Intd, intensity deviant; Locd, location deviant; Pitd, pitch deviant; NH, normal hearing; HA, hearing aid; CI, cochlear implant.
responses of ERPs and the corresponding MMN. The CI, HA and NH groups were entered as the between-subjects factor. Type of deviant (1 standard and 5 deviants) and time (baseline and follow-up) were entered as the within-subject factors. The obligatory responses in ERP and MMN were analyzed separately. Continuous variables with a normal distri-bution were reported as mean with standard deviation (SD) and 95% confidence interval (95% CI) for the mean. The obligatory responses and MMNs were normally distributed, as assessed by Shapiro-Wilk’s test (p >.05), with the exception for the location response in NH subjects at baseline (ERP: p = .040; MMN: p = .005), caused by an unexpected high value. There were 8 (ERP; the obligatory responses) vs. 7 (MMN) out-liers, as assessed by inspection of a boxplot. The outliers were kept in the analysis because they did not materially affect the results as assessed by a comparison of the results with and without the outliers. There was homogeneity of variances, as assessed by Levene’s test for equality of variances (p > .05) except for the obligatory responses of intensity at baseline (p = .012). For the three-way interaction effect, Mauchly’s test of sphericity indicated that the assumption of sphericity was met, χ2(2) =11.432, p = .655 (ERP; the obligatory responses) vs. χ2(9) = 4.900, p =.844 (MMN). Therefore, we did not use any epsilon (ε) estimate to adjust the df.
We completed a two-way mixed ANOVA to compare the mean dif-ferences between groups, i.e. the CI, HA and NH groups were entered as the between-subject factor and time (baseline and follow-up) as the within-subject factor. There were no outliers, as assessed by examina-tion of studentized residuals for values greater than ±3. The mean of the obligatory responses in ERP and the mean of MMN at baseline and follow-up was normally distributed, as assessed by Shapiro-Wilk’s test (p > .05). The mean of responses was normally distributed, as assessed by Normal Q-Q Plot. There was homogeneity of variances, as assessed by Levene’s test for equality of variances (p > .05). (ERP: p = .057 for median responsepre and .329 median responsef-u; MMN: p = .350 for
mean responsepre and .485 mean responsef-u). Regarding the obligatory
responses in ERP, there was close to homogeneity of covariances, as assessed by Box’s test of equality of covariance matrices (p = .049) for model with mean responses at baseline and follow-up as dependent variables. Regarding MMN, there was homogeneity of covariances, as assessed by Box’s test of equality of covariance matrices (p = .794) for model with mean responses at baseline and follow-up as dependent variables.
All analyses were carried out using IBM SPSS® Statistics version 23.0.0.3. P-values < .05 were considered significant.
3. Results
The results of the obligatory responses in ERP and the corresponding MMN, both regarding the mean amplitudes in the time interval between 80 and 224 ms, are presented separately below. The results are analyzed in all 3 groups (the children with CIs; the children with HAs; the children with NH as a reference group), time (at baseline, i.e. time for the original studies; and after three years, i.e. at the follow-up study), and all stimuli (standard and 5 deviants; gap, intensity, pitch, location and duration). The results at baseline refer to the children also participating after three years, thus, not including all children participating in the previous studies [28,29,31]. No significant correlations were observed in oblig-atory responses in ERP or MMN regarding age, age of hearing, age of implantation or hearing thresholds. There were no significant relations to sex. Fig. 1 offers an overall display of the obligatory responses in ERP and the corresponding MMN, in all groups at baseline vs. follow-up. 3.1. Obligatory responses
The obligatory responses to all stimuli are presented in Table 2; ERP data of the standard stimuli and each deviant are specified in all 3 groups (CI, HA and NH), at baseline and at follow-up, and, also the difference, i. e. the mean amplitudes at follow-up minus the mean amplitudes at
Table 2
Event-related potentials at baseline and long-term follow-up in children with cochlear implants, hearing aids, and normal hearing.
Deviant Group Baseline Long-term follow-up* Difference* p
Mean SD (95% CI) Mean SD (95% CI) Mean SD (95% CI)
Gap (μV) CI 4.38 1.62 (2.06–6.70) 2.09 1.62 (0.39–3.80) −2.29 2.83 (− 5.26–0.68) .104 HA 1.67 1.44 (0.34–3.01) 4.34 2.18 (2.32–6.36) 2.66 1.68 (1.11–4.22) .006 NH 4.89 1.33 (3.94–5.84) 4.27 2.01 (2.84–5.71) −0.62 2.10 (− 2.12–0.88) .377 Intensity (μV) CI 4.58 2.86 (1.58–7.58) 1.75 1.53 (0.15–3.35) −2.83 3.23 (− 6.22–0.56) .085 HA 2.48 0.55 (1.97–2.98) 5.07 1.93 (3.28–6.85) 2.59 2.24 (0.52–4.67) .022 NH 4.70 2.16 (3.16–6.35) 4.19 2.04 (2.74–5.65) −0.51 0.60 (− 1.88–0.85) .419 Pitch (μV) CI 4.80 2.19 (2.50–7.10) 2.61 1.29 (0.81–3.51) −2.64 3.05 (− 5.84–0.55) .087 HA 3.13 1.67 (1.59–4.67) 4.92 1.74 (3.31–6.53) 1.79 0.92 (0.95–2.64) .002 NH 4.89 2.35 (3.21–6.58) 4.62 2.18 (3.06–6.18) −0.27 1.80 (− 1.56–1.02) .419 Location (μV) CI 4.09 1.80 (2.19–5.98) 1.16 0.73 (0.39–1.92) −2.93 1.90 (− 4.93–− 0.93) .013 HA 2.55 1.61 (1.06–4.03) 4.61 1.92 (2.83–6.39) 2.06 1.99 (0.22–3.90) .034 NH 5.03 2.47 (3.26–6.80) 4.70 1.97 (3.29–6.11) −0.33 1.64 (− 1.50–0.85) .963 Duration (μV) CI 1.25 1.27 (− 0.09–2.59) 0.54 1.14 (− 0.66–1.74) −0.71 2.01 (− 2.82–1.41) .429 HA 2.58 1.46 (1.22–3.93) 2.82 1.82 (1.14–4.50) 0.24 2.28 (− 1.87–2.35) .789 NH 2.53 1.77 (1.26–3.80) 2.27 1.96 (0.87–3.67) −0.26 1.68 (− 1.46–0.94) .630 Standard (μV) CI 3.88 2.60 (1.15–6.60) 2.41 1.42 (0.92–3.89) −1.47 3.46 (− 5.11–2.16) .345 HA 2.57 0.67 (1.95–3.18) 5.23 0.56 (4.71–5.74) 2.66 0.90 (1.83–3.49) .0002 NH 4.67 1.45 (3.63–5.71) 4.67 1.62 (3.52–5.83) 0.00 1.49 (− 1.06–1.07) .992
95% CI, 95% confidence interval for mean. p, 2-tailed significance (paired t-test). All values are mean amplitudes (μV) in time window 80–224 ms. *Based on 6 subjects in CI group, 7 in HA group, and 10 in NH groups, respectively. ERP, event-related potential; CI, cochlear implant; HA, hearing aid; NH normal hearing; μV, microvolt. Fig. 2. Mean ERPs (mean amplitude of the obligatory responses, all deviants and standard; time interval 80–224 ms) in the NH, HA and CI groups at baseline and after three years. ERP, event-related potentials; NH, normal hearing; HA, hearing aid; CI, cochlear implant.
baseline. Based on paired t-test, there was a significant difference for all deviants except duration in the HA group. The difference in the location deviant was also significant among the CI children, but none of the deviants in the NH group. The analyses proceeded with repeated mea-sures ANOVAs as follows.
Two-way mixed ANOVA showed statistically significant interactions between the groups (HA vs. CI vs. NH) and time (baseline vs. follow-up) on mean response, F(2, 20) = 12.445, p = .0003, partial η2 =0.554, see Fig. 2. There was a statistically significant difference in mean ERP response between groups at baseline, F(2, 20) = 4.078, p = .033, partial η2 =0.290. Data are mean ± standard error, unless otherwise stated. Mean ERP was statistically significantly greater in the NH group (1.9580 ± 0.68883 μV, p = .026) compared to the HA group. Mean ERP in the CI group was not statistically significantly lower than the NH group (− 0.6227 ± 0.72181 μV, p = .669) or greater than the HA group, (1.3353 ± 0.77765 μV, p = .224). There was a statistically significant difference in mean ERP between groups at follow-up, F(2, 20) = 6.620, p = .006, partial η2 =0.398. Data are mean ± standard error, unless otherwise stated. Mean ERP at follow-up was statistically significantly lower in the CI group (− 2.4376 ± 0.78223 μV, p = .014) compared to the NH group and the HA group (− 2.8131 ± 0.84275 μV, p = .009). Mean ERP in the HA group was not statistically significantly higher than the NH group (0.3755 ± 0.74650 μV, p = .871).
Further, there was a significant difference in mean ERP at baseline compared to follow-up F(1, 6) = 17.534, p = .006, partial η2 =0.745 in the HA group. There was a marginally significant difference in mean ERP at baseline compared to follow-up F(1, 5) = 5.152, p = .072, partial η2 =0.507 in the CI group. There was no significant difference in mean ERP at baseline compared to follow-up F(1, 9) = 1.095, p = .323, partial η2 =0.109 in the NH group.
Three-way mixed ANOVA showed a statistically significant three- way interaction between deviant (gap, intensity, pitch, location and duration), time (baseline and follow-up) and group (HA, CI and NH), F (10, 100) = 1.936, p = .049, partial η2 =0.162.
When running a two-way mixed ANOVA for separate analyses of each deviant, there was a statistically significant interaction between the groups (HA vs. CI vs. NH) and time (baseline vs. follow-up) regarding ERP of the location deviant, F(2, 20) = 4.395, p = .026, partial η2 = 0.305. There was a simple main effect of time on mean amplitude of location ERP only in subjects with CIs, who had a lower ERP at follow-up compared to baseline (4.088 ± 1.805 (SD) vs. 1.278 ± 2.764 (SD) μV); F (1,5) 10.474, p = .023. In HA and NH subjects, time had no effect on mean amplitude of location ERP; F(1,6) 3.194, p = .124 and F(1,9) 0.027, p = .873, respectively. There was no significant differences in mean amplitude of location ERPs between the subject groups at baseline,
F(2,20) 2.935, p = .076, or at follow-up F(2,20) 2.488, p = .108 (uni-variate ANOVA with Bonferroni adjustment for multiple comparisons). The main effect of group showed that there was a statistically significant difference in mean ERP of the gap deviant between groups F(2, 20) = 4.674, p = .022, partial η2 =0.319. The mean amplitude for gap ERP at baseline was significantly lower in subjects with HAs compared to both subjects with NH 4.894 μV vs. 1.675 (HA), p = .002, and CI users 4.384 μV vs. 1.675 (HA), p = .021 (from univariate ANOVA, F(2,20) 8.643, p = .002) with pairwise comparisons adjusting for multiple comparisons according to Bonferroni. This difference disappeared at follow-up, ANOVA F(2,20) 0.618, p = .549.
3.2. MMN
Terminology in this part includes all mismatch responses, no matter MMN or pMMR. Paired t-test preceded repeated measures ANOVA an-alyses, and the difference, i.e. the mean amplitudes at follow-up minus the mean amplitudes at baseline, was significant for the duration deviant in the HA group (p = .016), and the location deviant (p = .030) in the CI group. MMNs of each deviant at baseline and follow-up among all groups are shown in Fig. 3.
Two-way mixed ANOVA showed no statistically significant in-teractions between the groups (HA vs. CI vs. NH) and time (baseline vs. follow-up) on mean response, F(2, 20) = 0.115, p = .892, partial η2 = 0.011. The main effect of time showed no statistically significant dif-ference in mean MMN at the different time points, F(1, 20) = 0.091, p = .766, partial η2 =0.005. The main effect of group showed that there was no statistically significant difference in mean MMN between groups F(2, 20) = 0.088, p = .917, partial η2 =0.009.
Three-way mixed ANOVA showed a statistically significant three- way interaction between deviant (gap, intensity, pitch, location and duration), time (baseline vs. follow-up) and group (whether NH, CI or HA), F(8, 80) = 2.446, p = .020, partial η2 =0.197. We considered a Bonferroni-adjusted alpha level of 0.025 as statistically significant for the simple two-way interaction. The assumption of sphericity was met for both simple two-way interactions effects, as assessed by Mauchly’s test of sphericity (p > .05). There was a statistically significant simple two-way interaction between deviant and time for subjects with HA subjects, F(4, 24) = 3.415, p = .024, ε =0.363, due to the significant difference in duration between baseline and follow-up (p = .016). There was no statistically significant simple two-way interaction between deviant and time for subjects with NH, F(4, 36) = 0.102, p = .981, or for subjects with CI, F(4, 20) = 1.732, p = .182.
Fig. 3. MMNs (mean amplitudes; time interval 80–224 ms) of all deviants at baseline and 3-year follow-up in (a) the NH group (n = 10), in (b) the HA group (n = 7), and in (c) the CI group (n = 6). Error bars are 95% confidence intervals. MMN, mismatch negativity; HA, hearing aid; CI, cochlear implant; NH, normal hearing; Durd, duration deviant; Gapd, gap deviant; Intd, intensity deviant; Locd, location deviant; Pitd, pitch deviant.
4. Discussion
Our results in this follow-up-study are based on recordings of ERPs and MMNs using the multi-feature paradigm Optimum-1 [30] and show neurophysiological differences between children with HL and children with NH. There was no statistically significant difference in the mean amplitude of the obligatory responses regarding the children with NH after three years. However, the mean responses at baseline was lower in the children with HAs compared to the children with NH, but this dif-ference disappeared after three years during this follow-up study. The mean ERPs of the children with CIs did not significantly differ from the NH group at baseline. Though, after three years their mean ERPs was significantly lower compared to both the children with NH and HAs. The children with HAs also showed a significant difference in mean ERP at baseline compared to follow-up and there were significant differences between the deviants at follow-up but not at baseline. Together, this suggests a possible improvement, or catch-up, over time among the children with HAs compared to their peers with CIs. The possible reduced plasticity among the children with CIs in the ages around 7 years, is consistent with results in other CI studies [12].
Regarding MMN, there was no statistically significant difference in mean MMN at follow-up compared to baseline, and there were no sig-nificant differences in mean MMN between groups. There was a statis-tically significant interaction in duration between baseline and follow- up for subjects with HA subjects. The MMN of the duration deviant was more negative after three years, suggesting the children with HA might improve their ability to distinguish differences in duration over time in this agespan. In comparison with other ERP components, MMN is considered quite stable during childhood [24]. However, the partly diverging results in different MMN studies challenge MMN as a neuro-physiological marker of deficit in the central auditory system in children with HL. Sharma et al. [41] showed a fundamentally different pattern of development of P1 cortical response latency for early- and late-implanted children. The P1 latencies have been used to infer the maturational status of auditory pathways in congenitally deafened children fitted with CIs [12] and P1 latencies seem to provide a clinically useful biomarker of central auditory system development in children after cochlear implantation [42]. However, regarding children with HAs, another study could not reveal that P1 significantly differed from children with NH. Instead, the N2 component seemed to have potential to better reflect the deficit among these children [15]. However, the different results may not be contradictory, rather elucidating the complexity and suggest possible differences between the subgroups of children with HL, i.e. the children with either HAs or CIs.
Different stimuli can be used in the ERP recordings and may vary in stability. The duration deviant seems to elicit stable MMN in NH chil-dren and NH adults [43–45]. Thus, it is important to notice the choice of stimuli, when evaluate, or maybe compare, different studies. However, it seems reasonable to use MMN as a tool for comparison between groups within each study. To account the uncertainty in the method, we considered the obligatory responses as well as the MMNs. Our results indicate that the obligatory responses might be a more sensitive tool for showing differences in the auditory processing. Further, we used the same stimuli as in previous ones [28,29,31], and we have also been following an already established method, the multi-feature paradigm Optimum-1, offering five deviants at the same time [30]. Based on this study, it is difficult to draw any conclusions regarding the different stimuli. The maturational and age-related changes [6,7], must also be taken into consideration. The amplitudes and latencies differ not only with age but may also be affected by HL. Latency appears to decrease with increasing age, while the amplitude is more stable in children [20, 46], meaning possible amplitude changes may reflect effects of the HL only. In the present study, our analyses are based on mean amplitudes in a specific time interval and we could not significantly prove any time effects regarding the MMNs, however, ERPs did reveal some significant results, see above.
We expected only small individually changes in the obligatory re-sponses and MMNs, which, given the small sample size, reduce the probability of demonstrating significant results. Previous studies have shown presence of both MMNs and pMMR among these children [28,29, 31], which probably also hampers the possibilities to statistically prove changes over time. At the follow-up study, the participation was limited because of drop out, and the groups were too small to allow subgroup analysis of MMNs vs. pMMRs, not least since previous studies showed an unpredictable mix of MMNs and pMMRs regarding all deviants in each child.
Children with HL often show phonological impairment [3] and less academic success than hearing peers [47,48]. Thus, since MMN reflects the discrimination ability, it is of interest when testing children with HL. However, MMN itself implies some difficulties in interpretations and analyses. Our study helps to illuminate the complexity in the neuro-physiological field and the need of research in this field. The changes over time in ERPs suggest possible improvement among the children with HAs, but not with CIs, both groups in the age span around 5–11 years. The differences between the children with HL using HAs vs. CIs may have different explanations, the degree of HL seems to be the most obvious, even if the hearing level following the implantation was acceptable. None of the CI children had received any CI under the age of 1.5 years (4 of the children were implanted under the age of 2 years; 1 child at 3 years and 2 months; 1 child with unilateral CI at 4 years and 7 months), which, despite HA fitting preceding CI surgery, might have caused too poor stimulation and reduced plasticity of their central auditory system.
Further knowledge regarding the central auditory processing among subgroups of children with HL may be of importance for individually designed support at audiological, pedagogical, and speech and language pathology services. In future research, larger samples would improve the statistical analyses. It would also be of interest to follow the ERPs and MMNs in children after another few years and studying possible corre-lations with behavioral findings might be helpful. Examination of P300 and other high-level cognitive components, as well as language-related ERP’s [1], would also contribute to more comprehensive information about the children with HL. Enhanced support may be of importance in different time periods in these subgroups of children with HL. Maybe an intensified support under the age of 7 years, i.e. within the maximum of the plasticity of the central auditory system [12], would be beneficial for both groups. However, our study suggests a delay in the central auditory processing among the children with HAs compared to children with NH, but a possible catch-up after time, and this potential must not be over-looked. The lack of similar results regarding the CI children, suggests need of further efforts from the habilitation centers.
5. Conclusion
In this three-year follow-up study, we examine the obligatory re-sponses in ERPs and MMNs in a specific time interval among children with HL and children with NH in the total age range around 5–11 years. The main finding is a statistically significant difference in the obligatory responses in ERP between the NH and HA groups at baseline, which disappears after three years in our follow-up study. Further, there is a significant difference in mean ERP at baseline compared to follow-up in the HA group. This suggests a possible catch-up over time among the children with HAs. On the other hand, the children with CIs do not differ from the children with NH at baseline, but after three years their mean ERPs are significantly lower compared to both the children with NH and children with HAs, indicating a poorer development of the central auditory system among the children with CI within this age span. Thus, we show a difference between the subgroups of children with HL, i.e. the children with HAs vs. CIs. Regarding MMN, there is an interaction be-tween the duration deviant and time for the children with HA, but otherwise analyses of the MMNs do not statistically demonstrate any changes over time. The small sample, the normal development of MMN
in this age span, and the CI artifacts affect the certainty in our inter-pretation of results and highlight the need of more research in this field.
Declaration of competing interest
No conflict of interest is declared, including financial, personal, or other relationships with other people or organizations for any of the authors in this study. All authors have approved the final article.
Acknowledgements
The authors want to thank the Phonetics Laboratory at the Depart-ment of Linguistics, Stockholm University; the Humanities Laboratories at Lund University; Linnaeus Centre HEAD at Link¨oping University; and Cognition, Communication and Learning (CCL) at Lund University; and, Professor Risto N¨a¨at¨anen for giving access to Optimum-1.
Supplementary tables A - C
Supplementary data to this article can be found online at https://doi. org/10.1016/j.ijporl.2020.110519.
Funding
This work was supported by the Swedish Research Council for Working Life and Social Sciences (Forskningsrådet f¨or Arbetsliv och Social Vetenskap).
References
[1] S.J. Luck, An Introduction to the Event-Related Potential Technique, MIT press, USA, 2014.
[2] A. Kral, A. Sharma, Developmental neuroplasticity after cochlear implantation, Trends Neurosci. 35 (2) (2012) 111–122.
[3] J. Briscoe, D.V. Bishop, C.F. Norbury, Phonological processing, language, and literacy: a comparison of children with mild-to-moderate sensorineural hearing loss and those with specific language impairment, JCPP (J. Child Psychol. Psychiatry) 42 (3) (2001) 329–340.
[4] N. Kraus, M. Cheour, Speech sound representation in the brain, Audiol. Neuro. Otol. 5 (3–4) (2000) 140–150.
[5] S.J. Luck, E.S. Kappenman, The Oxford Handbook of Event-Related Potential Components, Oxford University Press, 2011.
[6] C.W. Ponton, J.J. Eggermont, B. Kwong, M. Don, Maturation of human central auditory system activity: evidence from multi-channel evoked potentials, Clin. Neurophysiol. 111 (2) (2000) 220–236.
[7] D.S. Goodin, K.C. Squires, B.H. Henderson, A. Starr, Age-related variations in evoked potentials to auditory stimuli in normal human subjects,
Electroencephalogr, Clin. Neurophysiol. 44 (4) (1978) 447–458.
[8] D.V. Bishop, M.J. Hardiman, J.G. Barry, Is auditory discrimination mature by middle childhood? A study using time-frequency analysis of mismatch responses from 7 years to adulthood, Dev. Sci. 14 (2) (2011) 402–416.
[9] S. Lippe, E. Martinez-Montes, C. Arcand, M. Lassonde, Electrophysiological study of auditory development, Neuroscience 164 (3) (2009) 1108–1118.
[10] A. Calcus, O. Tuomainen, A. Campos, S. Rosen, L.F. Halliday, Functional brain alterations following mild-to-moderate sensorineural hearing loss in children, Elife 8 (2019).
[11] C.W. Ponton, J.J. Eggermont, M. Don, M.D. Waring, B. Kwong, J. Cunningham, P. Trautwein, Maturation of the mismatch negativity: effects of profound deafness and cochlear implant use, Audiol. Neuro. Otol. 5 (3–4) (2000) 167–185. [12] A. Sharma, M.F. Dorman, A.J. Spahr, A sensitive period for the development of the
central auditory system in children with cochlear implants: implications for age of implantation, Ear Hear. 23 (6) (2002) 532–539.
[13] J.J. Eggermont, C.W. Ponton, M. Don, M.D. Waring, B. Kwong, Maturational delays in cortical evoked potentials in cochlear implant users, Acta Otolaryngol. 117 (2) (1997) 161–163.
[14] A. Koravand, B. Jutras, M. Lassonde, Cortical auditory evoked potentials in children with a hearing loss: a pilot study, Int. J. Pediatr. 2012 (2012) 250254. [15] A. Koravand, B. Jutras, M. Lassonde, Auditory event related potentials in children
with peripheral hearing loss, Clin. Neurophysiol. 124 (7) (2013) 1439–1447. [16] G. Rance, B. Cone-Wesson, J. Wunderlich, R. Dowell, Speech perception and
cortical event related potentials in children with auditory neuropathy, Ear Hear. 23 (3) (2002) 239–253.
[17] R. Naatanen, A.W. Gaillard, S. Mantysalo, Early selective-attention effect on evoked potential reinterpreted, Acta Psychol. 42 (4) (1978) 313–329. [18] R. Draganova, H. Eswaran, P. Murphy, M. Huotilainen, C. Lowery, H. Preissl,
Sound frequency change detection in fetuses and newborns, a magnetoencephalographic study, Neuroimage 28 (2) (2005) 354–361.
[19] M. Cheour, R. Ceponiene, A. Lehtokoski, A. Luuk, J. Allik, K. Alho, R. Naatanen, Development of language-specific phoneme representations in the infant brain, Nat. Neurosci. 1 (5) (1998) 351–353.
[20] M.L. Morr, V.L. Shafer, J.A. Kreuzer, D. Kurtzberg, Maturation of mismatch negativity in typically developing infants and preschool children, Ear Hear. 23 (2) (2002) 118–136.
[21] V.L. Shafer, Y.H. Yu, H. Datta, Maturation of speech discrimination in 4- to 7-yr-old children as indexed by event-related potential mismatch responses, Ear Hear. 31 (6) (2010) 735–745.
[22] C. Weber, A. Hahne, M. Friedrich, A.D. Friederici, Discrimination of word stress in early infant perception: electrophysiological evidence, Brain Res. Cogn. Brain Res. 18 (2) (2004) 149–161.
[23] V.L. Shafer, M.L. Morr, J.A. Kreuzer, D. Kurtzberg, Maturation of mismatch negativity in school-age children, Ear Hear. 21 (3) (2000) 242–251. [24] M. Cheour, P.H. Leppanen, N. Kraus, Mismatch negativity (MMN) as a tool for
investigating auditory discrimination and sensory memory in infants and children, Clin. Neurophysiol. 111 (1) (2000) 4–16.
[25] N. Kraus, T. McGee, A. Micco, A. Sharma, T. Carrell, T. Nicol, Mismatch negativity in school-age children to speech stimuli that are just perceptibly different, Electroencephalogr. Clin. Neurophysiol. 88 (2) (1993) 123–130.
[26] D.V. Bishop, Using mismatch negativity to study central auditory processing in developmental language and literacy impairments: where are we, and where should we be going? Psychol. Bull. 133 (4) (2007) 651–672.
[27] R. Naatanen, B. Petersen, R. Torppa, E. Lonka, P. Vuust, The MMN as a viable and objective marker of auditory development in CI users, Hear, Res. 353 (2017) 57–75.
[28] E. Engstrom, P. Kallioinen, C. Nakeva von Mentzer, M. Lindgren, M. Ors, B. Sahlen, B. Lyxell, I. Uhlen, Computer-assisted reading intervention for children with sensorineural hearing loss using hearing aids: effects on auditory event-related potentials and mismatch negativity, Int. J. Pediatr. Otorhinolaryngol. 117 (2019) 17–25.
[29] E. Engstrom, P. Kallioinen, M. Lindgren, C. Nakeva von Mentzer, B. Sahlen, B. Lyxell, I. Uhlen, Computer-assisted reading intervention for children with hearing impairment using cochlear implants: effects on auditory event-related potentials and mismatch negativity, Int. J. Pediatr. Otorhinolaryngol. 137 (2020) 110229.
[30] R. Naatanen, S. Pakarinen, T. Rinne, R. Takegata, The mismatch negativity (MMN): towards the optimal paradigm, Clin. Neurophysiol. 115 (1) (2004) 140–144. [31] I. Uhlen, E. Engstrom, P. Kallioinen, C. Nakeva von Mentzer, B. Lyxell, B. Sahlen,
M. Lindgren, M. Ors, Using a multi-feature paradigm to measure mismatch responses to minimal sound contrasts in children with cochlear implants and hearing aids, Scand. J. Psychol. 58 (5) (2017) 409–421.
[32] P. Kallioinen, J. Olofsson, C. Nakeva von Mentzer, M. Lindgren, M. Ors, B. S. Sahlen, B. Lyxell, E. Engstrom, I. Uhlen, Semantic processing in deaf and hard-of- hearing children: large N400 mismatch effects in brain responses, despite poor semantic ability, Front. Psychol. 7 (2016) 1146.
[33] C. Nakeva von Mentzer, B. Lyxell, B. Sahlen, M. Wass, M. Lindgren, M. Ors, P. Kallioinen, I. Uhlen, Computer-assisted training of phoneme-graphene correspondence for children who are deaf and hard of hearing: effects on phonological processing skills, Int. J. Pediatr. Otorhinolaryngol. 77 (12) (2013) 2049–2057.
[34] C. Nakeva von Mentzer, B. Lyxell, B. Sahlen, O. Dahlstrom, M. Lindgren, M. Ors, P. Kallioinen, I. Uhlen, Computer-assisted reading intervention with a phonics approach for children using cochlear implants or hearing aids, Scand. J. Psychol. 55 (5) (2014) 448–455.
[35] C.N. Von Mentzer, B. Lyxell, B. Sahl´en, ¨O. Dahlstr¨om, M. Lindgren, M. Ors, P. Kallioinen, I. Uhl´en, The phonics approach in Swedish children using cochlear implants or hearing aids: inspecting phonological gain, Commun Disord Deaf Stud Hearing Aids 2 (3) (2014) 1–11.
[36] C. Nakeva Von Mentzer, B. Lyxell, B. Sahlen, O. Dahlstrom, M. Lindgren, M. Ors, P. Kallioinen, E. Engstrom, I. Uhlen, Segmental and suprasegmental properties in nonword repetition–an explorative study of the associations with nonword decoding in children with normal hearing and children with bilateral cochlear implants, Clin. Linguist. Phon. 29 (3) (2015) 216–235.
[37] J. Dien, The ERP PCA Toolkit: an open source program for advanced statistical analysis of event-related potential data, J. Neurosci. Methods 187 (1) (2010) 138–145.
[38] T.P. Jung, S. Makeig, M. Westerfield, J. Townsend, E. Courchesne, T.J. Sejnowski, Removal of eye activity artifacts from visual event-related potentials in normal and clinical subjects, Clin. Neurophysiol. 111 (10) (2000) 1745–1758.
[39] A. Delorme, S. Makeig, EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis, J. Neurosci. Methods 134 (1) (2004) 9–21.
[40] P.M. Gilley, A. Sharma, M. Dorman, C.C. Finley, A.S. Panch, K. Martin, Minimization of cochlear implant stimulus artifact in cortical auditory evoked potentials, Clin. Neurophysiol. 117 (8) (2006) 1772–1782.
[41] A. Sharma, K. Martin, P. Roland, P. Bauer, M.H. Sweeney, P. Gilley, M. Dorman, P1 latency as a biomarker for central auditory development in children with hearing impairment, J. Am. Acad. Audiol. 16 (8) (2005) 564–573.
[42] P.W. Bauer, A. Sharma, K. Martin, M. Dorman, Central auditory development in children with bilateral cochlear implants, Arch. Otolaryngol. Head Neck Surg. 132 (10) (2006) 1133–1136.
[43] R. Uwer, W. von Suchodoletz, Stability of mismatch negativities in children, Clin. Neurophysiol. 111 (1) (2000) 45–52.
[44] N. Kathmann, T. Frodl-Bauch, U. Hegerl, Stability of the mismatch negativity under different stimulus and attention conditions, Clin. Neurophysiol. 110 (2) (1999) 317–323.
[45] S. Singh, A. Liasis, K. Rajput, A. Towell, L. Luxon, Event-related potentials in pediatric cochlear implant patients, Ear Hear. 25 (6) (2004) 598–610. [46] M. Gomot, M.H. Giard, S. Roux, C. Barthelemy, N. Bruneau, Maturation of frontal
and temporal components of mismatch negativity (MMN) in children, Neuroreport 11 (14) (2000) 3109–3112.
[47] O. Hendar, Elever Med Hørselshemning I Skolen: En Kartleggingsundersøkelse Om Læringsutbytte, Skådalen Kompetansesenter, 2012.
[48] Coordination, Responsibility and Communication – the Road to Increased Quality for Students with Disabilities (SOU 2016:46), Fritzes offentliga publikationer, Stockholm, 2016.
