Edited by: Gavin M. Bidelman, University of Memphis, USA
Reviewed by: Justin Aronoff, University of Illinois at Urbana-Champaign, USA; Saradha Ananthakrishnan, Towson University, USA; Samuel R. Atcherson, University of Arkansas at Little Rock, USA
*Correspondence: Niki K. Vavatzanidis
This article was submitted to Auditory Cognitive Neuroscience, a section of the journal Frontiers in Neuroscience
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Children with sensorineural hearing loss may (re)gain hearing with a cochlear implant—a device that transforms sounds into electric pulses and bypasses the dysfunctioning inner ear by stimulating the auditory nerve directly with an electrode array. Many implanted children master the acquisition of spoken language successfully, yet we still have little knowledge of the actual input they receive with the implant and specifically which language sensitive cues they hear. This would be important however, both for understanding the flexibility of the auditory system when presented with stimuli after a (life-) long phase of deprivation and for planning therapeutic intervention. In rhythmic languages the general stress pattern conveys important information about word boundaries. Infant language acquisition relies on such cues and can be severely hampered when this information is missing, as seen for dyslexic children and children with specific language impairment. Here we ask whether children with a cochlear implant perceive differences in stress patterns during their language acquisition phase and if they do, whether it is present directly following implant stimulation or if and how much time is needed for the auditory system to adapt to the new sensory modality. We performed a longitudinal ERP study, testing in bimonthly intervals the stress pattern perception of 17 young hearing impaired children (age range: 9–50 months; mean: 22 months) during their first 6 months of implant use. An additional session before the implantation served as control baseline. During a session they passively listened to an oddball paradigm featuring the disyllable “baba,” which was stressed either on the first or second syllable (trochaic vs. iambic stress pattern). A group of age-matched normal hearing children participated as controls. Our results show, that within the first 6 months of implant use the implanted children develop a negative mismatch response for iambic but not for trochaic deviants, thus showing the same result as the normal hearing controls. Even congenitally deaf children show the same developing pattern. We therefore conclude (a) that young implanted children have early access to stress pattern information and (b) that they develop ERP responses similar to those of normal hearing children.
Language acquisition is a marvelous thing: An infant arrives quite naïve into our world of sound and language and successfully accomplishes the task of attaching meaning to the streams of auditory information that is speech. In this process of turning sound into meaningful utterances, one of the most important steps the infant has to master is finding the word boundaries so as to have units to which meaning can be attached (Newman et al.,
Fortunately, infants are brilliant in grasping regularities (Mueller et al.,
Other cues for word boundaries may come from prosody. Stress pattern, for example, can mark the beginning or end of a word in rhythmic languages like German or English. Both German and English are characterized by a predominant trochaic meter (Cutler and Carter,
The perception of the overall stress pattern of the native language seems to be present from very early on, as newborns were shown to follow the native stress pattern already in their cries: Whereas German infants tend to stress the beginning of their cries, French infants stress the latter part (Mampe et al.,
When learning a language with a strong metric bias, relying on stress pattern for word segmentation is apparently a useful starting rule for identifying words in fluent speech. A series of behavioral experiments by Jusczyk et al. (
Electrophysiological studies confirm the early sensitivity of infants toward stress pattern. At 4–5 months of age the electrophysiological response of German infants already differentiate between iambic and trochaic stress patterns in a language specific manner (Weber et al.,
The importance of such cues for language acquisition becomes evident in children that are less sensitive to them. Several studies suggest that children with language impairments, like language impairment (SLI) or dyslexia, are not as sensitive to auditory features like duration, rise time, frequency, etc., as children with normal language development (Benasich et al.,
The cochlear implant (CI) has given children with severe to profound sensory hearing loss access to the auditory world with great success by directly stimulating the auditory nerve and thus bypassing the dysfunctioning inner ear. Yet, stimulating the auditory nerve with up to 22 active electrodes cannot be compared to the stimulation by thousands of inner ear hair cells. Consequently, frequency discrimination with the CI is lower (Zeng et al.,
A large number of studies describes language outcomes of children with a cochlear implant after several years of implant use (e.g., Svirsky et al.,
A stressed syllable is characterized by longer duration, higher amplitude (loudness) and a change of the fundamental frequency (pitch; Fry,
The ability of children's stress perception has been the focus of speech perception studies since the early days of the cochlear implant. Thielemeir et al. (
First of all, the studies vary greatly with respect to the age of their participants (infants—adolescents), their age at implantation (infancy—late adolescence), their amount of experience with the CI (months to several years) and the onset of deafness (congenital—prelingual—perilingual—postlingual). Also, most studies stem from the nineties (Thielemeir et al.,
Tasks more suitable for younger children are the visual habituation procedure and the preferential looking paradigm used in many stress pattern perception studies of normal hearing children (e.g., in the seminal studies by Peter Jusczyk) and employed by the few recent studies on stress pattern with implanted children. In the visual habituation procedure, the child hears one item repeatedly while looking at a screen. When the looking time toward the screen decreases, a novel item is presented and the change in looking time is evaluated. This has the benefit of being not only suitable for infants who are still in the early phases of language acquisition, but also suitable for testing a wide range of acoustic changes.
Core et al. (
Taken together, there is only scarce information on a) whether implanted children perceive and acquire a sensitivity toward their native stress pattern that may aid them to segment fluent speech into single words and b) whether the perception is present from the first day on or whether it evolves over time. On that ground we performed the present study that compares the ERPs of implanted children and normal hearing peers elicited by the native trochaic rhythm vs. the non-native iambic rhythm. We specifically used ERPs, as they offer several advantages. By measuring electrophysiological data, we obtain a direct and objective measure of ongoing brain processes without depending on any overt response of the child (e.g., pointing at the correct picture). This is important, as we observe that some congenitally deaf children lack auditory attention in the first weeks of implant use. That is, they rarely react to salient or loud stimuli, though it is clear from some isolated reactions that they are able to perceive them. Whereas in behavioral studies this would pose a problem, the ERP components will show whether the stimuli have been processed despite the lack of an overt behavioral reaction. Also, both behavioral and EEG/ERP studies suffer when the child is overactive or fussy and will not engage in a particular task. In our study, the oddball paradigm was to our great advantage as it works even when the participant does not pay any attention to the stimuli
One of the greatest challenges in clinical research is to arrive at a homogenous group of participants. Studies of cochlear implants have to address a wide range of variables that have potential influence on the research outcome. For children, the effect of
Nineteen bilaterally hearing impaired children, who received a cochlear implant participated in the study. For all children, CI indication was confirmed by pediatric audiological assessment consisting of a brain stem electric response audiometry (BERA) and subjective audiometry. When a period of bilateral hearing aid use proved to be without benefit, cochlear implantation was performed on both ears (see details in Table
1 | m | Simultaneous bilateral | CI512 | CP810 | 32 |
2 | m | Simultaneous bilateral | Concerto | Opus2 | 24 |
*3 | m | Simultaneous bilateral | Concerto | Opus2 | 15 |
4 | m | Sequential bilateral | Concerto | Opus2 | 25/27 |
5 | m | Sequential bilateral | CI422 & CI512 | CP810 | 50/57 |
*6 | m | Sequential bilateral | CI422 | CP810 | 11/15 |
*7 | m | Simultaneous bilateral | CI512 | CP810 | 37 |
*8 | m | Simultaneous bilateral | CI422 | CP810 | 11 |
*9 | f | Simultaneous bilateral | Concerto | Opus2 | 21 |
10 | f | Simultaneous bilateral | HiRes90K Advantage | Harmony | 31 |
*11 | m | Sequential bilateral | Concerto | Opus2 | 12/15 |
12 | f | Simultaneous bilateral | CI422 | CP810 | 39 |
13 | f | Simultaneous bilateral | Concerto | Opus2 | 11 |
*14 | f | Simultaneous bilateral | Concerto | Opus2 | 14 |
15 | f | Simultaneous bilateral | Concerto | Opus2 | 9 |
16 | m | Simultaneous bilateral | HiRes90K | Naida | 11 |
*17 | m | Simultaneous bilateral | CI522 | CP910 | 11 |
After implantation, all children entered the rehabilitation program at the Saxonian Cochlear Implant Center, University Hospital Dresden, Germany. There they received a bimonthly fitting of the speech processor and multidisciplinary speech and language therapy for up to 3 years. The first activation occurred 1 month postsurgically during a 5-day rehabilitation stay. Age at first activation of the implant ranged from 9 to 50 months (
The EEG recordings were performed longitudinally at the regular bimonthly rehabilitation stays: in the week of initial activation (M0), after two (M2), four (M4), and six (M6) months of implant use plus an additional pre-operative measurement serving as baseline (Mpre). Not all of the above recordings could be obtained for each child due to occasional illness or restlessness of the child. Size and age distributions of the groups are listed in Tables
N | 8 | 9 | 8 | 11 | 11 |
Range | 8–29 | 11–39 | 13–27 | 12–44 | 14–56 |
Median | 10 | 14 | 14 | 26 | 30 |
N | 5 | 5 | 5 | 5 | 4 |
Range | 8–19 | 11–21 | 13–23 | 16–40 | 18–42 |
Median | 10 | 12 | 14 | 16 | 24 |
Two control groups of normal hearing (NH) full-term children were measured at the Max Planck Institute for Cognitive and Brain Sciences in Leipzig, Germany. The first group NH1 (
No additional control group matched for hearing age was obtained as the original study of this paradigm (Weber et al.,
For all children (implanted and controls), informed consent was signed by a parent or a person having the custody for the child. The following procedures were approved by the local ethics committee (Medical Faculty Carl Gustav Carus of the Technische Universität Dresden).
Stimuli and paradigm originate from the study by Weber et al. (
Data were obtained continuously with Ag-AgCl− electrodes positioned according to the International 10-20 System in an elastic electrode cap (EasyCap, GmbH, Herrsching, Germany). Nine scalp sites (F3, Fz, F4/C3, Cz, C4/P3, Pz, P4) and the left and right mastoid were recorded. In some cases where the position of the speech processor hindered the correct placement of the mastoid electrodes the speech processor was removed from the ear and taped onto the cap as close as possible to the original position but ensuring that proper placement of the mastoid electrodes was possible.
An electrooculogram was obtained from two horizontal electrodes at the outer canthi of the left and right eye and from a vertical electrode above the right eye. An additional vertical electrode was recorded below the right eye whenever possible. It was omitted if otherwise the child would not have tolerated the EEG measurement. The signal was sampled at 500 Hz and amplified with a PORTI-32/MREFA (Twente Medical Systems, Oldenzaal, The Netherlands) with electrode Cz as online reference.
Data were downsampled offline to 256 Hz and rereferenced to the average of both mastoids. If one mastoid was too corrupted by artifacts, the other mastoid served as single reference. A band-pass filter of 1–15 Hz reduced slow drifts and muscle artifacts. Trials with the signal at the midline electrodes (Fz, Cz, Pz) or eye electrodes exceeding 80 μV within a 200 ms sliding window were rejected. A subsequent correction of eye blinks and eye movements was applied (EEP 3.2.1, developed by the CBS MPI, Leipzig, Germany and distributed by ANT Neuro, Enschede, Netherlands). The standard trial immediately following a deviant trial was removed from analysis. All sessions had a minimum of 50 deviants (50%) and 200 standards (50%) with the exception of one participant's dataset. Because the respective single subject average had a good signal-to-noise ratio, it was still included in the analysis. Averaging occurred from -100 to 1200 ms with reference to stimulus onset. The 200 ms baseline was set from 100 ms before stimulus onset to 100 ms after stimulus onset, thus including the first 100 ms where both stimuli were identical.
We compared physically identical stimuli, resulting in two comparisons: iambic deviant–iambic standard (ID-IS) and trochaic deviant - trochaic standard (TD-TS). The difference wave was calculated by subtracting the standard stimulus from the deviant stimulus. This comparison ensures also that any artifact that could possibly arise from the implant in response to the stimulus would be eliminated by the subtraction (see also Lonka et al. (
The longitudinal data of the implanted children were statistically analyzed by applying a linear mixed effect model with R and the lme4 package (Bates et al.,
A separate test was run to determine whether the differentiation seen descriptively between the iambic and the trochaic stimuli was also statistically significant. Upon visual inspection, a 100 ms-window was determined for each group, and, analogously to the methods above, a repeated measure ANOVA (for the controls) or a mixed effect model (for the implanted children with
Both the implanted and the normal hearing group show a clear morphological separation of the trochaic vs. the iambic stimulus around 400–550 ms (Figure
For the implanted children, the negative peak of the difference wave at 496 ms reveals an interaction of
When considering only the congenitally deaf children, the difference wave peaks at 488 ms. The significant interaction of
The two control groups do not differ statistically and are therefore reported together. The control group shows a negative peak in the difference curve at 404 ms with a main effect for
For the implanted children, the difference wave of the trochaic stimuli displays a small negativity at 420 ms with significant main effect of
The subgroup of congenitally deaf children displays a negative peak at 504 ms. The negativity reveals a main effect for
The results of the normal hearing peers vary according to group. Whereas NH1 has a small negativity peaking at 480 ms but no significant effect, NH2 has a positive peak at 472 ms with a significant main effect of
Several auditory discriminative abilities are crucial for successful language acquisition. For rhythmic languages like English or German, one of these is the ability to differentiate between categories of stress patterns as the pattern is conveying information about word boundaries. The importance of this ability is highlighted by studies that link a diminished sensitivity to stress pattern deviations with impairments in language acquisition (Friedrich et al.,
Most encouragingly, the ERP curves of the implanted children strongly resemble that of normal hearing peers, and this is the case even for the congenitally deaf children. All children show a clear differentiation between the iambic and the trochaic stimuli between 400 and 550 ms. Moreover, between 400 and 500 ms the implanted children respond to a deviant iambic (non-native) stimulus with a significant negative mismatch response (MMR) as do the normal hearing controls. No comparable effect was seen for the trochaic stimulus in either the implanted or the normal hearing children.
The existence of a mismatch response for the foreign iambic but not the native trochaic stimulus parallels the findings of Friederici et al. (
A further difference to the results obtained by Segal et al. is that while the behavioral study did not find an effect of time with the CI, the ERP data show a development over the first 6 months. The negative MMR for the iambic stimulus is seen descriptively after 2 months of implant use and reaches significance after 4 months. In the subgroup of congenitally deaf children it reaches significance after 6 months of implant use. The latter finding has to be treated with care with regard to the small number of participants in the subgroup, though the small delay of 2 months for the congenitally deaf group is plausible. Considering that those children who had some residual hearing prior to implantation probably have been familiarized to some degree with the native stress pattern prior to implantation, the lag of only 2 months until a significant negative MMR is seen in the congenitally deaf group is remarkably small.
Also remarkable is the fact that even the congenitally deaf children respond with a negative MMR, which is considered to be the more mature mismatch response as opposed to a positive MMR (Kushnerenko et al.,
An important next step would be to increase the group size of implanted children. Recruiting sufficient participants from a clinical group is always a challenge and even more so, when the participants are infants and young children. It would be most valuable, however, to see whether the results can be replicated with a larger group of children. Of particular interest would be a larger group of congenitally deaf children, as they provide a unique insight into (a) the auditory system's performance, when the very first sensory input is delivered with a considerable delay, and (b) how this delay affects further language acquisition.
This is the first ERP study on stress pattern recognition in implanted children. We demonstrate that the cochlear implant allows the differentiation between native and foreign stress patterns and thus transmits necessary cues for language acquisition. Furthermore we show that even under the condition that there has been no auditory input prior to implantation, young implanted children manage to differentiate between stress patterns within the first 6 months of implant use.
NV, DM, AF, and AH created the design of the study. AF co-conceptualized a previous form of the paradigm. DM and AF enabled the collection of the clinical and the control data. NV analyzed the data and wrote the manuscript to which all co-authors contributed with critical revisions. AH was also involved to a great extent in data discussion and interpretation. All authors approved the final version and agree to be accountable for this work.
Partial funding of this work by the “Marga and Walter Boll Stiftung” (220-02-12).
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Special thanks go to Christina Rügen for collecting the control data and to all children and parents for their patience and willingness to participate. We would also like to thank the “Marga und Walter Boll Stiftung” for partially funding this work. Our acknowledgements go also to the German Research Foundations and the Open Access Publication Funds of the TU Dresden.