Auditory Mismatch Response to Pitch and Duration Changes in Children with Developmental Language Disorder: A Longitudinal Approach

Auditory Mismatch Response to Pitch and Duration Changes in Children with Developmental Language Disorder: A Longitudinal Approach | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 28 July 2025 V1 Latest version Share on Auditory Mismatch Response to Pitch and Duration Changes in Children with Developmental Language Disorder: A Longitudinal Approach Authors : Liis Themas 0000-0002-7615-0895 [email protected] , Gesa Schaadt , Pärtel Lippus , Marika Padrik , Liis Kask , Ulvi Vaher , Mairi Männamaa , and Kairi Kreegipuu Authors Info & Affiliations https://doi.org/10.22541/au.175367725.54753416/v1 282 views 186 downloads Contents Abstract Liis Themas, Näituse 2, 50409, Tartu, Estonia; [email protected] 1.5 Goals, research questions and hypotheses Supplementary Material Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Developmental language disorder (DLD) is characterized by difficulties in auditory processing and prosody perception. Estonian word prosody features a unique three-way quantity distinction, characterized by changes in syllable duration and pitch contour, which differentiate word meaning and grammatical cases. In our longitudinal study, we investigated auditory discrimination in 50 children (25 with DLD), aged 4-7-years, using the event-related potential Mismatch Response (MMR) and behavioral tasks. In the duration change condition, we observed a shift from a positive at age 5;8 years to a negative MMR at age 6;8 years, which was less pronounced in the DLD group compared to their typically developing (TD) peers. In the duration and pitch change condition, no MMR was observed in either group at the first or second assessment. We did not find any associations between the MMR amplitudes and behavioral discrimination scores. However, there was a positive correlation between the MMR amplitude and participation in speech and language therapy. This study sheds light on the maturation of auditory change detection processes and deepens our understanding of the interplay between neurophysiological and behavioral indicators in Estonian children with and without language impairments. Auditory Mismatch Response to Pitch and Duration Changes in Children with Developmental Language Disorder: A Longitudinal Approach Running title Auditory Perception in DLD Authors * these authors contributed equally *Liis Themas, Institute of Estonian and General Linguistics, University of Tartu; Institute of Psychology, University of Tartu, Tartu, Estonia; https://orcid.org/0000-0002-7615-0895 *Gesa Schaadt, Department of Education and Psychology, Freie Universität Berlin; Department of Neuropsychology, Max Planck Insitute of Human Cognitive and Brain Sciences Leipzig, Germany; https://orcid.org/0000-0002-3192-3698 Pärtel Lippus, Institute of Estonian and General Linguistics, University of Tartu, Tartu, Estonia; https://orcid.org/0000-0003-4407-811X Marika Padrik, Institute of Educational Science, University of Tartu, Tartu, Estonia ; https://orcid.org/0009-0003-5851-2084 Liis Kask, Institute of Psychology, University of Tartu, Tartu, Estonia; https://orcid.org/0000-0003-1852-0058 Ulvi Vaher, Department of Radiology, Institute of Clinical Medicine, University of Tartu, Tartu, Estonia; Children’s Clinic, Tartu University Hospital, Tartu, Estonia; https://orcid.org/0000-0002-2546-1079 Mairi Männamaa, Department of Radiology, Institute of Clinical Medicine, University of Tartu, Tartu, Estonia; Children’s Clinic, Tartu University Hospital, Tartu, Estonia; https://orcid.org/0000-0001-5234-6474 Kairi Kreegipuu, Institute of Psychology, University of Tartu, Tartu, Estonia; https://orcid.org/0000-0002-0953-7264 Corresponding author information Liis Themas, Näituse 2, 50409, Tartu, Estonia; [email protected] Word count 9218 (including abstract, keywords, and main body: Introduction, Methods, Results, Discussion, and Conclusion) Figure count 8 (main text) + 6 (appendices) Acknowledgements This work was supported by the Estonian Research Council, Tallinn, Estonia [grant number PRG1151] and Estonian Center of Excellence for Well-Being Sciences (EstWell), funded by grant TK218 from the Estonian Ministry of Education and Research. List of abbreviations AIC – Akaike information criterion ANOVA – analyses of variances β - standardized beta regression coefficient CI – confidence intervals d’ – sensitivity index DD – developmental dyslexia DEV – deviant stimulus df – degrees of freedom DLD – developmental language disorder durQ1 – deviant stimulus in the first quantity degree, differing from the standard in syllable duration EEG – electroencephalography ERPs – event-related potentials group diff. – group differences hand. – handedness high. – higher education imp. – imperative inf. – infinitive form ISI – interstimulus interval IQ – intelligence quotient lang. test – the scaled overall score of the language test at T1 LDN – Late Discriminative Negativity lex. des. – lexical decision task LMM – linear mixed model M – mean MADE – the Maryland analysis of developmental EEG mat. edu. – maternal education level max – maximum value min – minimum value MMN – mismatch negativity MMR – mismatch response N - the number of observations NMAR – not missing at random nom. – nominative case p – probability value pitchdurQ3 – deviant stimulus in the third quantity degree, differing from the standard in syllable duration and pitch contour Q1 – first quantity degree Q2 – second quantity degree Q3 – third quantity degree quant. degree – quantity degree; eighter first, second or third quant. diff. – quantity discrimination task r – Pearson correlation coefficient RT – reaction time R² – R squared SD – standard deviation sec. – secondary education sg – singular sg2 – second person singular ST – standard stimulus t – a ratio of the estimated fixed effect to its standard error T1 – first assessment T2 – second assessment after one year TD – typically developing TW1 – first time window TW2 – second time window V – Wilcoxon test statistic WPPSI-IV UK – Wechsler Preschool and Primary Scale of Intelligence, fourth edition, United Kingdom version χ² - chi-squared statistic µV – microvolts Abstract Developmental language disorder is characterized by difficulties in auditory processing and prosody perception. Estonian word prosody features a unique three-way quantity distinction, characterized by changes in syllable duration and pitch contour, which differentiate word meaning and grammatical cases. In our longitudinal study, we investigated auditory discrimination in 50 children (25 with developmental language disorder), aged 4-7-years, using the event-related potential Mismatch Response and behavioral tasks. In the duration change condition, we observed a shift from a positive at age 5;8 years to a negative Mismatch Response at age 6;8 years, which was less pronounced in the developmental language disorder group compared to their typically developing peers. In the duration and pitch change condition, no Mismatch Response was observed in either group at the first or second assessment. We did not find any associations between the Mismatch Response amplitudes and behavioral discrimination scores. However, there was a positive correlation between the Mismatch Response amplitude and participation in speech and language therapy. This study sheds light on the maturation of auditory change detection processes and deepens our understanding of the interplay between neurophysiological and behavioral indicators in Estonian children with and without language impairments. Keywords EEG, MMN, auditory perception, developmental language impairment, longitudinal study, child development 1 Introduction 1.1 Developmental language disorder Developmental language disorder (DLD) emerges during development and is characterized by persistent difficulties in acquiring and using language across different domains, such as phonology, vocabulary, and morphology (Bishop et al., 2017; Gillam et al., 2021), independent of intellectual disabilities, hearing impairments, or primary physical or neurological disorders (Bishop et al., 2017). One of these children’s core deficits is an atypical auditory processing (Kujala et al., 2017; Näätänen et al., 2014), including prosody perception (Calet et al., 2021; Cumming et al., 2015; Leonard & Kueser, 2019; Richards & Goswami, 2019; Sundström et al., 2019). 1.2 The Estonian three-way quantity distinction Estonian is one of several quantity languages where phoneme length, as a prosodic feature, plays a crucial role in differentiating between word meanings and grammatical forms. Both vowels and consonants in the stressed syllable exhibit a three-way length distinction: they can be short (quantity 1 – Q1; e.g., [sɑtɑ], [kɑlɑ]), long (quantity 2 – Q2; e.g [sɑːtɑ], [kɑllɑ]) or overlong (quantity 3 – Q3; e.g., [sɑːːtɑ], [kɑlːlɑ]; Lippus et al., 2013). Quantity distinction is marked primarily by the duration of the stressed initial syllable, which is longer in the case of higher degrees of quantity, while the following unstressed syllable is compensatorily shortened. In addition, the quantity is marked with the fundamental frequency contour: in Q1 and Q2, the pitch fall is aligned with the end of the stressed syllable, whereas in Q3, the fall takes place in the first half of the stressed syllable (Lippus et al., 2013). To date, only one study (Themas et al., 2025a) has been conducted on auditory quantity perception with three-way quantity distinction in Estonian children. 1.3 Auditory discrimination and its development Auditory perception across development can be objectively and non-invasively investigated with electroencephalography (EEG) using event-related potentials (ERPs). An ERP reflecting the brain’s ability to discriminate different characteristics of speech, such as different quantity degrees, is the Mismatch Negativity (MMN), a well-studied fronto-central negativity elicited 100-200 ms after the onset of a deviating stimulus within a stream of standard stimuli (Näätänen et al., 1978). The calculation of the MMN requires the subtraction of the average ERP to the standard stimulus from the ERP to the deviant stimulus, resulting in a difference wave (Näätänen et al., 2007). The MMN is typically tested with an oddball paradigm (Squires et al., 1975), but can also be elicited with the multi-feature paradigm (Optimum-1; Näätänen et al., 2004), in which more than one deviant stimulus can be tested, such as different quantity degrees in the Estonian language. The MMN has been suggested to be associated with a person’s speech perception sensitivity, as its amplitude was shown to be linked to the ability to detect acoustic differences (Kujala, 2007; Näätänen, 2001; Näätänen et al., 2007; Näätänen et al., 2012). In studies investigating children, the MMN is often referred to as the Mismatch Response (MMR) because it can be of positive or negative polarity and its latency is delayed compared to the adult’s MMR latency (Campos et al., 2023; Cheng et al., 2015; Leppänen et al., 2004; Maurer et al., 2003; Mueller et al., 2012). EEG studies conducted on languages other than Estonian have found that children with DLD have brain-level differences in pitch and duration perception, manifested in the distinct morphology of the MMR compared to typically developing (TD) children (e.g. Bishop et al., 2007; Cantiani et al., 2016; Cheng et al., 2021; Friedrich et al., 2004; Kujala & Leminen, 2017). Fewer studies have investigated the maturation of brain-level differences in auditory perception, as reflected in the MMR, either longitudinally or by age group comparisons in the DLD population (Bishop et al., 2010; Chen et al., 2016; Choudhury & Benasich, 2011). To better understand DLD and its manifestations, it is, however, crucial not only to measure auditory processing at a single time point, but to also observe its maturation, as DLD itself evolves over time. Longitudinal studies offer valuable insights into several key aspects of DLD. The MMR measured at infancy and toddlerhood has been shown to be predictive of future language abilities (Cantiani et al., 2016; Chen et al., 2016; Choudhury & Benasich, 2011; Friedrich et al., 2009; Guttorm et al., 2005; Schaadt et al., 2015; Schaadt et al., 2023; Weber et al., 2005) and therefore could also be suitable for predicting future outcomes of DLD at later ages. Further investigation into how the MMR of children with DLD evolves in relation to participation in speech and language therapy could provide valuable insights. This is especially relevant as specific interventions targeting auditory processing or phonological skills have been shown to modulate the morphology of the MMR/MMN or other ERPs (Dacewicz et al., 2018; Heim et al., 2016; Pihko et al., 2006), informing how effective treatment should be conceptualized. 1.4 Relationship between neurophysiological and behavioral findings of auditory discrimination Currently, however, there is an ongoing discussion regarding the use of the MMR as a diagnostic tool and as a measure for evaluating treatment efficacy. For the MMR to be regarded as suitable for diagnostic purposes as well as for evaluating treatment efficacy, evidence should prove the MMR to be correlated with behavioral discrimination abilities, which, in turn, must be reliably reflected in overall language competences. However, findings on the relationship between auditory discrimination denoted by ERPs and behavioral auditory discrimination are contradictory so far. Some studies have found significant associations between some neurophysiological and behavioral measures of auditory discrimination (McArthur & Bishop, 2005; Nan et al., 2018; Pihko et al., 2006), while others have not (Ahmmed et al., 2008; Bishop et al., 2010; Halliday et al., 2014; McArthur & Bishop, 2005; Pihko et al., 2006; Uwer et al., 2002). Similarly, contradictory findings are observed concerning the relationship between behavioral auditory processing skills and overall language abilities (relationship: Corriveau et al., 2007; Datta et al., 2010 – with receptive language score; Davids et al., 2011; Eichhorn & Juur, 2024; Vandewalle et al., 2012 – associations between speech perception and oral language skills; no relationship: Beattie & Manis 2013; Coughler et al., 2021; Datta et al., 2010 - with expressive language score; Vandewalle et al., 2012 – no association with auditory processing tasks). Evidently, clarification is needed, as in clinical settings, understanding this link helps to verify if behavioral auditory discrimination tasks objectively measure these abilities. Finally, examining ERPs alongside behavioral tasks helps to validate each measure and would ensure that they accurately assess auditory discrimination. 1.5 Goals, research questions and hypotheses As discrimination of quantities is crucial for understanding the Estonian language, we aimed at investigating the development of duration and pitch perception, the two main cues for distinguishing between quantity degrees, in Estonian children with DLD compared to their TD peers across one year. Further, we were interested in whether behavioral discrimination abilities are reflected in the MMR’s amplitude in both groups. Additionally, we wanted to investigate whether participation in speech and language therapy, as indicated by children’s parents (years of participation, frequency of sessions per months, duration of each session), influences the maturation of auditory duration and pitch perception in children with DLD over the course of one year. We investigated the children’s ability to distinguish three-way quantity degrees behaviorally as well as at the neurophysiological level, by measuring the MMR in the ERP when children were 4;6 to 6;5 years old (first assessment, T1) and one year later, when children were 5;6 to 7;5 years old (second assessment, T2). We used a two-deviant Optimum-1 paradigm, where the standard stimulus was a meaningful word in Q2. For the first deviant we changed the quantity degree of the standard to be in Q1 (syllable duration ratio change) and for the second deviant we changed the quantity degree of the standard to be in Q3 (both, syllable duration ratio and pitch contour change). Previous studies found duration perception to be quite stable at the age of 6-13 years (Engström et al., 2021; Putkinen et al., 2014). However, Linnavalli et al., (2018) found that duration perception denoted by the amplitude of the MMR is still ongoing at ages 5-6 years (Linnavalli et al., 2018). In our previous research (T1: Themas et al., 2025a), our statistical model showed changes in amplitude with age, which suggests the development to still be in progress at T2 and we found positive MMRs, which should in time change to a negative more adult-like reaction (i.e., MMN; e.g. Norton et al., 2021; Themas et al., 2023). Hypothesis 1: We therefore expected significant changes in the MMR’s amplitude in both groups when comparing the results of T1 and the results of T2 after one year, specifically in the duration ratio change condition (standard stimulus in Q2 and deviant in Q1). Hypothesis 2: Additionally, we expected that the direction of change in the MMRs would be towards negativity as decrease in amplitude with age has been demonstrated in several studies (Lee et al., 2012; Linnavalli et al., 2018; Themas et al., 2023). We found some unexpected results at T1 (Themas et al., 2025a), showing that the MMR decreased with age in the DLD group but increased in the TD group, which might suggest the MMR amplitude does not shift linearly from positivity to negativity (Werwach et al., 2022). However, since the overall trend is reported to move towards negativity, we expected to observe this shift in our findings at T2. Even though there is little research on differences between children with DLD and TD, showing no group differences on duration change detection (Uwer et al., 2002) in older children comparable to the present sample, research on children with developmental dyslexia (DD) can shed light on to be expected group differences. DLD and DD seem to stem from similar language processing difficulties (Leonard, 2014) and studies on duration change detection in school-aged children (older than the current sample) with DD showed reduced negative MMR amplitudes compared to TD children (Männel et al., 2017; Schaadt & Männel, 2019). In addition, behavioral evidence suggests duration perception difficulties in DLD children aged 7–10-years (Corriveau et al., 2007; Cummings et al., 2015). And finally, there is one Estonian study done specifically on the three-way quantity distinction that suggests that even at school-age, children with reading and writing difficulties make many mistakes in marking the quantities in writing (Karlep, 2000). We may assume that this phenomenon is at least partially caused by perception difficulties of the quantity degrees. Hypothesis 3: In conclusion, we expected group differences to be not only evident at T1 (see Themas et al., 2025a), but to also be observable at T2, particularly in the duration ratio change condition (with the standard stimulus in Q2 and the deviant in Q1). At T1, we found no brain-level discrimination between the standard and the Q3 deviant in either the DLD group or the TD group (see Themas et al., 2025a). This finding does not necessarily suggest that 4-6-year-old children do not differentiate these quantity degrees in everyday speech, but highlights the specific stimuli selected for the current experiment, along with presenting them without supporting linguistic context to have challenged the participants. At T2, we would now expect discrimination of these stimuli by the TD group, as most children will have reached the age (5-7-years) at which they begin written language acquisition, thereby advancing their phonological awareness abilities. During this process, they will also start to consciously learn about the three-way quantity system. This will enhance their capability to differentiate between quantity degrees and their ability to notice more subtle changes in syllable duration ratio and pitch contour. In the DLD group, on the other hand, we expect this ability to mature more slowly and to not be apparent at T2. Firstly, we anticipate this because research suggests children with DLD often develop DD later in life (Leonard, 2014), and following literacy lessons will be more difficult and time-consuming for them. Additionally, research suggests that neurophysiologically measured auditory processing is delayed in children with moderate to severe DLD (Kwok et al., 2018). Hypothesis 4: Thus, we expected the MMR for the pitch and duration change deviant (in Q3) to now be evident in the TD group, but less so in the DLD group. Hypothesis 5: Likewise, we expected both groups to be more accurate and respond faster in the behavioral discrimination tasks at T2 compared to T1, but that the DLD group will not reach the same level of accuracy as the TD group. During preschool and primary school years, ERP components in oddball paradigms, reflecting attentional processes, continue to mature (Karakaş, 2024), which is supported by behavioral studies showing rapid maturation in attention sustention abilities at this age (Suades-González et al., 2017). Furthermore, the impact of task load on reaction accuracy and time diminishes with age (Betts et al., 2006). Hypothesis 6: Taking this into account, we presumed that the more mature attention after one year will enable a more precise behavioral measurement of children’s auditory discrimination abilities, making these results align more closely with the findings in the EEG. Finally, conscious language learning through specialized methods has been shown to yield more favorable language development outcomes (Gillam et al., 2021), as several studies have demonstrated correlations between auditory processing abilities and overall language proficiency (Chen et al., 2016; Kautto et al., 2024; Kujala & Leminen, 2017). Additionally, targeted interventions aimed at auditory processing or phonological skills can effectively alter the morphology of the MMR/MMN and other ERPs (Dacewicz et al., 2018; Heim et al., 2016; Pihko et al., 2006). Hypothesis 7: We therefore anticipated the intensity of speech and language therapy to potentially be associated with the MMR in children with DLD. 2 Materials and Method The current research was preregistered on the Open Science Framework, including the study description, goals, hypotheses, and a condensed version of the methods section at https://doi.org/10.17605/OSF.IO/PCW98 reference Themas et al., (2025b). All procedures performed in this study involving human participants were conducted in accordance with the ethical standards of the institutional and/or national research committee and with the 2013 revision of the World Medical Association Declaration of Helsinki, as published on the website of the Journal of the American Medical Association. The study was approved by the Ethics Committee of Human Research of the University of Tartu (approval number: 383/M-12). Written informed consent was obtained from the parents or legal guardians of all participating children prior to their inclusion in the study. After an initial sampling phase, this longitudinal study consisted of two assessment time points (T1 and T2; see Figure 1). Psychometric assessments of language and intellectual abilities were conducted during the sampling phase. Children with language impairments meeting the diagnostic criteria were officially diagnosed with DLD and assigned to the clinical group. At T1 and after one year at T2, the same neurophysiological and behavioral measures of auditory discrimination were administered. 2.1 Participants Fifty monolingual Estonian-speaking children (27 males, 23 females) from the southern counties of Estonia participated in the present study (25 children with DLD and 25 age and sex matched TD peers). Recruitment lasted from autumn 2022 until spring 2024. Children could participate if they did not show any intellectual disabilities, sensory impairments, neurological and psychiatric diagnoses (except for language disorders). Children’s ages ranged from 4.5 to 6.5 years at T1 and from 5.5 to 7.5 years at T2. Two children were excluded from T1 analysis due to noisy EEG data, but they still were re-invited for the second assessment (T2). At T2 three participants did not follow the invitation (including one that already had been excluded from T1 due to noisy EEG data). All excluded participants belonged to the DLD group (see Figure 1). We could include 46 participants (DLD = 21; TD = 25) in ERP analysis and statistical modeling. We were able to include one additional participant in the behavioral data analysis because, although the participant lacked EEG data from T1, they completed all behavioral experiments at T1 and T2. 2.2 Psychometrical testing To assess all children’s language abilities as well as to assess DLD, tasks of the standardized language battery for 5-6-year-old Estonian speaking children (Padrik et al., 2013) testing phonology, vocabulary, grammar usage, pseudoword repetition, sentence repetition, and sentence comprehension, were conducted. The above-mentioned tasks were selected based on their sensitivity in differentiating atypical from age-appropriate language development, while ensuring coverage of all major language domains. To control for children’s intellectual abilities, the Wechsler Preschool and Primary Scale of Intelligence fourth edition of the United Kingdom version (WPPSI-IV UK; Wechsler, 2013) was used for DLD and TD children at T1. As children with DLD are disadvantaged by tasks requiring language abilities, we assessed only the non-verbal intelligence and used those subtests that cover domains of visuospatial abilities, working memory, processing speed and fluid reasoning (Wechsler, 2013). 2.3 Neurophysiological experiment at T1 and T2 2.3.1 Paradigm and stimuli The method and stimuli used in the current study were identical to those in our previous work (Themas et al., 2025a), which focused exclusively on data from T1. Below, we provide a brief description of the paradigm and stimuli for clarity. A passive Optimum-1 paradigm (see Figure 2) was applied, where the standard stimulus was an Estonian word [sɑːtɑ] ( ‘send!, imp. sg2’ ) which is pronounced in the second quantity degree (Q2) . For the first deviant (henceforth durQ1) we changed the quantity degree of the standard to be in Q1 [sɑtɑ] (syllable duration ratio change), which also resulted in the change in meaning ( ‘hundred, nom. sg’ ). For the second deviant (henceforth pitchdurQ3) we changed the quantity degree of the standard to be in Q3 [sɑːːtɑ] (both syllable duration ratio and pitch contour change), which also resulted in the change in meaning ( ‘to get, inf.’ ). Behavioral evidence suggests greater difficulty in discriminating between Q2 and Q3 (Leppik, Lippus & Asu, 2023; Karlep, 2000). Therefore, the durQ1 deviant was expected to be less challenging to discriminate, whereas pitchdurQ3 was expected to be more challenging. The stimuli were naturally spoken by a female native Estonian speaker (the first author of the paper) in a soundproof room. The test words were extracted from carrier phrases with PRAAT (version 6.1, Boersma & Weenink, 2021), where they were placed in the central position of the phrase (e.g., ütle … kõvasti ‘say … loudly’; ‘say a hundred loudly’) to ensure natural intonation. Further description of the stimuli can be found in Table 1 and Figure 3. In the context of a lager project, participants were presented with three different stimulus blocks (total duration 35 min). Here, we report the results from one of these blocks (duration 11 min), in which a total of 1,100 stimuli (see above for description) were delivered through headphones using MATLAB (The MathWorks, 2015; version 8.5.0). Standard (60%) and deviant (40%) stimuli were presented such that the number of standards between the deviants was either one or three (see Figure 2) to investigate habituation differences between children with DLD and their TD peers (results presented elsewhere). The inter-stimulus interval (ISI) alternated between 400, 425, or 450 ms. 2.3.2 EEG recording The EEG was recorded from 32 active electrodes (ActiveTwo system, BioSemi B.V., Amsterdam, Netherlands) placed at standard positions according to the 10-20 system, with the signal referenced to the CMS electrode and common mode rejection handled by the DRL electrode during data acquisition. Electrooculograms were recorded from two electrodes at the outer canthi of both eyes, one electrode on the orbital ridge under the right eye near the lower eye socket and one electrode located near the supraorbital ridge above the right eye, slightly lateral to the midline of the forehead. The EEG data were digitized at 512 Hz frequency with a bandpass filter of 0.16–100 Hz. 2.4 Behavioral tasks at T1 and T2 We used the same method and stimuli as described in Themas et al. (2025a) to study behavioral discrimination. For clarity, a brief description is provided below. 2.4.1 Quantity discrimination task To behaviorally test the children’s quantity discrimination abilities, the roving standard paradigm was used (Fong et al., 2020) at T1 and T2. At the beginning of the task, a random stimulus word gets presented 3-5 times. Due to the repeated presentation, the stimulus gets established as a standard. Then a new stimulus, differing in quantity degree here, is presented, also 3-5 times, until it becomes the new standard. Children were asked to actively listen to the stimuli and to press a button as soon as they heard a new word in the sequence of stimuli. Participants’ response accuracy and reaction times were recorded. Additional details regarding the task stimuli and procedure are available in Appendix A. 2.4.2 Lexical decision task In the lexical decision task children’s ability to differentiate between meaningful and meaningless words, depending on quantity, was tested at T1 and T2. Compared to the typical auditory lexical decision tasks (e.g., Crosbie et al., 2004), two adaptations were made to ensure accessibility for children with DLD: 1) Auditory stimuli were presented together with pictures and 2) children had to use one response button (not two) and they were asked to press it in the case of a real word (not in case of a pseudoword). Additional details regarding the task stimuli and procedure are available in Appendix B. 2.5 Parental questionnaire at T1 and T2 A parental questionnaire was used to gather information about the child´s family background, home language environment, caregivers emotional state, attitude towards language impairment and the attendance in speech and language therapy. The latter two were exclusively included in the questionnaires answered by parents of the DLD group. 2.6 Procedure As can be seen in Figure 1, psychometrical testing was conducted only at T1. For the DLD group, it was conducted at the University of Tartu Children’s Hospital by the speech and language pathologist and clinical psychologist of the hospital. The TD group underwent testing in the experimental psychology laboratory, administered by the first author of the present paper as well as by a psychologist from the research team with at least a master’s level education. At T1 and T2, both groups followed the same assessment procedure and were given identical tests. Testing was done across two sessions and lasted approximately 2.5 hours in total. At T1 and T2 neurophysiological recordings as well as behavioral testing were conducted with identical procedures. The interval between T1 and T2 was approximately 12 months for most participants (DLD group: M = 12.3 months, SD = 0.5; TD group: M = 12.2 months, SD = 0.7). All measurements together took about 1h and 40 min. Upon arrival at the laboratory, the parent and the child were introduced to the entire procedure and given a sticker board, where they could place a sticker after completing each task. For further motivation, the children were also shown the rewards (a small toy of their choosing and a gift card) they would receive at the end of the whole procedure. For the neurophysiological recording, the participant sat in a comfortable chair in front of a computer screen. Through all the EEG cap fitting and experimental procedure, the children watched a silent cartoon of their choice. The stimuli were presented via headphones (JBL, Tune290) with a comfortable volume level kept between 55-65 dB. Once the EEG study was concluded and the cap removed, children started the behavioral tasks. Throughout the whole procedure, a parent was always present in the laboratory. 2.7 Data analysis The raw EEG data that support the findings of this study are openly available in the University of Tartu Library DataDOI repository at https://doi.org/10.23673/re-525, reference Themas (2025).The code and data for modeling are available in the Zenodo repository at https://doi.org/10.5281/zenodo.15799644, reference Themas & Nickel (2025). 2.7.1 Offline processing of the EEG data The data was processed in MATLAB (The MathWorks, 2023; version 9.14) using a standardized pre- and post-processing pipeline specifically designed for developmental data – the Maryland analysis of developmental EEG (MADE – for a detailed description see Appendix C; Debnath et al., 2020). No participant had to be excluded due to severe artifacts. No group differences were observed at either T1 or T2 in the number of usable epochs (see Table 2). Grand averages of the ERPs to the standard and deviant stimuli as well as the difference waveform were plotted using EEGLAB (version 2023.1; Delorme & Makeig, 2004; see Figure 4 and 5). 2.7.2 The difference wave To identify time windows and electrodes where ERPs to the standard and deviant stimuli significantly differed, that is the MMR, cluster-based permutation tests (MATLAB, Fieldtrip toolbox; Oostenveld et al., 2011) were calculated for each time point between stimulus onset and epoch end (698 ms). Only the largest positive and negative clusters representing the deviant-standard difference were considered, as this approach controls for the family-wise error rate and minimizes the likelihood of false discoveries. For a more detailed description of the process along with the original topographical plots of standard–deviant clusters, see Appendix D. The method and the code are based on Meyer et al. (2021). 2.7.3 Linear-mixed models (LMMs) The statistical modeling strategy followed Murphy et al. (2022) and was implemented in R (version 4.3.2; R Core Team) using the lme4 package (version 1.1-35.3; Bates et al., 2015) and lmerTest (version 3.1-3; Kuznetsova et al., 2017), with visualizations created using ggplot2 (version 3.4.4; Wickham, 2016). We created an LMM for each deviant condition – durQ1 and pitchdurQ3 - separately. In the LMMs the MMR difference wave amplitudes (deviant – standard) served as a dependent variable. The time windows showing significant deviant-standard differences, as determined by cluster-based permutation tests, were divided into two segments: from the start to the midpoint, and from the midpoint to the end. Consequently, two mean amplitude values of the MMR were extracted from both T1 and T2 data, resulting in a total of four mean MMR values for each participant per condition. A random intercept was included to account for the correlation between repeated measures, and a random slope for time was added to capture anticipated variations in slopes over time among participants. To follow the first and second research question on how the MMR’s amplitude changes over the course of one year in children with DLD compared to their TD peers, measurement time (with levels: T1, T2) and group (levels: DLD and TD) were added as predicting variables. To follow the third research question on how results of the behavioral discrimination tasks are associated with the MMR amplitude, scores of the behavioral tests (quantity discrimination task, lexical decision task, standardized language test) were additionally added to the model. In addition, to account for participant’s general development, participant’s age at T1 was added as a predicting variable. A further predicting variable was time window (with levels: TW1, TW2). The models were fitted with maximum likelihood estimation and degrees of freedom were approximated using the Satterthwaite method. Models were created incrementally, by adding predictors one-by-one. The optimal model was chosen by comparing the model residual variances by using analyses of variances (ANOVA) and by evaluating the model fit with Akaike information criterion (AIC). Since we did not perform a priori power analysis for the study, we assessed confidence intervals, calculated effect sizes (standardized beta coefficients, β ), and examined marginal and conditional pseudo R-squared ( R² ) values to improve the interpretability of the models. For model diagnostics, we checked for potential multicollinearity, non-linearity between dependent and independent variables, heteroscedasticity in the model’s residual variance and normal distribution of the model’s residuals, random effects and random intercepts (see Appendix E). 2.7.4 Descriptive statistics and behavioral data analysis We used information from the parental questionnaire about the level of maternal education (basic education - completed 9 years of school; secondary education - completed 12 years of school; vocational education - providing specialized skills and knowledge for a particular trade or occupation; higher education - at least a bachelor’s degree) and children’s handedness to compare the two participant groups with Wilcoxon rank-sum test and Chi-squared test. Additionally, we applied these statistical tests to compare age, sex, standardized language test, and WPPSI-IV UK non-verbal IQ scores between the two participant groups. To follow our research question on how maturation manifests in the results of behavioral discrimination tasks and whether these results are associated with MMR amplitude - we calculated the sensitivity index ( d’ ) for each participant. This index indicates performance accuracy in the quantity discrimination and lexical decision tasks (as shown in the equation below). Z (Hit Rate) represents how many standard deviations the observed hit rate was from the mean of the standard normal distribution and the Z (False Alarm Rate) indicates how many standard deviations the observed false alarm rate was from the mean of the standard normal distribution. d’ = Z(Hit Rate) - Z(False Alarm Rate) Additionally, we calculated the mean reaction time (RT) for each participant. Only trials with correct responses and over 300 ms response time were included. To analyze group differences between children with and without DLD regarding d’ scores and RT, the Wilcoxon rank-sum test was used. To assess whether there were significant differences between the measurements at T1 and T2, we conducted Wilcoxon signed-rank tests for paired samples. Lastly, we wanted to describe how participants differentiated between different quantity contrasts and compared the following pairs: first quantity degree (Q1) versus second quantity degree (Q2; syllable duration difference), first quantity degree (Q1) versus third quantity degree (Q3; a large syllable duration difference and a pitch contour difference), and second quantity degree (Q2) versus third quantity degree (Q3; a small syllable duration difference and a pitch contour difference). In the quantity discrimination task, an error occurred when the participant failed to press the button in response to a change in quantity. In the lexical decision task, an error was defined as a button press where the participant incorrectly identified a word as correct, despite the word being pronounced in the wrong quantity degree. The errors in discriminating between specific quantity contrasts were counted, summed across participants, and group differences were calculated with the Wilcoxon rank-sum test. 2.7.5 Correlation analysis between therapy, MMR amplitude and behavioral data To follow our research question how attending speech and language therapy relates to MMR amplitude and the results of behavioral discrimination tasks—we examined the associations between speech therapy intensity, MMR amplitude at T2 and performance of the behavioral tasks. Speech therapy attendance was quantified based on the frequency of sessions per month and the typical duration of each session, as reported by the children’s parents in the parental questionnaire. By this, time spent in therapy (in minutes) between T1 and T2 could be calculated. We used the mean MMR value from T2 for the durQ1 deviant and pitchdurQ3 deviant separately. The mean MMRs were calculated for the time-windows and electrodes identified by the cluster-based permutation tests. For behavioral measures we used the d’s from T2. To determine the distribution of these variables, we used the Shapiro–Wilk test. As all variables were normally distributed, we then used Pearson’s correlation to investigate the association between the time spent in therapy, the MMRs amplitude at T2 and the d’s at T2. 3 Results 3.1 Description of the final sample Children with DLD did not differ significantly from their TD peers in age, sex or handedness, but they differed in maternal education level, standardized language test scores and WPPSI-IV UK nonverbal IQ scores (see Table 3). 3.2 Results of the cluster-based permutation tests 3.2.1 durQ1 condition Grand averages of the ERPs to the standard and deviant stimuli (durQ1) as well as of the difference waveform (deviant durQ1 – standard) at T1 and T2 separately for children with and without DLD can be found in Figure 4 and 5. At T1 the Cluster-based permutation tests (Figure 6; original plots in Appendix D) revealed a more positive amplitude in response to the deviant stimulus compared to the standard stimulus between 240 and 340 ms after stimulus onset at frontal, fronto-central and central electrodes (AF3, F3, FC5, F7, Fz, FC1, C3, FC6, F8). At T2 the Cluster-based permutation tests (Figure 6; original plots in Appendix D) revealed a more negative amplitude in response to the deviant stimulus compared to the standard stimulus between 390 and 490 ms after stimulus onset at frontal, fronto-central, central, centro-parietal and parietal electrodes (F3, F4, FC1, FC2, C4, Cz, CP2, Pz). The above-mentioned time-window and electrodes were further analyzed with LMMs. For further significant clusters, not related to the MMR, see Appendix D. Table 4 presents the mean and standard deviation of the difference wave amplitudes (deviant – standard) at T1 and T2 for the largest cluster in the durQ1 condition. 3.2.2 pitchdurQ3 condition Grand averages of the ERPs to the standard and deviant stimuli (pitchdurQ3) as well as of the difference waveform (deviant pitchdurQ3 – standard) at T1 and T2 separately for children with and without DLD can be found in Figure 4 and 5. Cluster-based permutation tests (Figure 6; original figures in Appendix D) revealed no time windows or electrodes where the waveforms of the standard and deviant stimuli were significantly different neither at T1 nor at T2. For exploratory LMM analyses, we chose a time-window between 340 and 440 ms after stimulus onset at F3, F4, Fz, FC1, FC2, C3, Cz and CP1 (see Figure 6). Table 4 presents the mean and standard deviation for this exploratory cluster at T1 and T2. 3.3 Results of the LMMs Two LMMs were calculated providing a total of 184 observations across T1 and T2 for each model. As the four participants who had missing values either at T1 or T2 were all children with DLD, data had to be assumed to not be Missing at Random (NMAR). Thus, we decided against using Multivariate Imputation by Chained Equations (Buuren & Groothuis-Oudshoorn, 2011) and excluded these participants from further analyses. Both optimal models passed the diagnostic testing mentioned in the method section. As the d’ scores from the behavioral quantity discrimination task and lexical decision task, along with the standardized language test score, were highly correlated and caused multicollinearity, they were not included together in the models during the selection of the optimal model. Instead, we modeled the main effects and first-level interactions separately for these variables, compared the model fits using AIC, and selected the variable with greater explanatory power. However, none of these variables proved to be strong predictors. Below we report the results of the LMMs which are relevant to our research questions. Other results from the models are described in Appendix F. 3.3.1 durQ1 condition In the durQ1 condition the model’s dependent variables were the mean values of the positive cluster (240-340 ms after stimulus onset at AF3, F3, FC5, F7, Fz, FC1, C3, FC6, F8) at T1 and the mean values of the negative cluster (390-490 ms after stimulus onset at F3, F4, FC1, FC2, C4, Cz, CP2, Pz) at T2. The optimal model (see Table 5) revealed a main effect of time indicating significantly more negative amplitudes at T2 compared to T1 in the first time window in the DLD group. The CIs were relatively wide showing moderate precision of the effect, but the effect size was large ( β = -0.54). The time:group interaction showed a stronger negative change in the TD group from T1 to T2 compared to the DLD group (see Figure 7). Based on the CIs this effect was not as precise as the others in the model, but it was large ( β = -0.70). The correlation between the random intercept and random slope was large ( r = -0.928), indicating that participants with more positive amplitudes at T1 (positive intercepts) showed a more negative change over time (negative slopes), compared to those with lower amplitudes at T1. The fixed effects alone accounted for approximately 42% of the variance (marginal R² = 0.42), while the full model, including both fixed and random effects, explained approximately 89% (conditional R² = 0.89). 3.3.2 pitchdurQ3 condition In the pitchdurQ3 condition, the dependent variables in the model were the mean values of the difference wave amplitudes of the exploratory cluster (340-440 ms after stimulus onset at F3, F4, Fz, FC1, FC2, C3, Cz, CP1). The same cluster was used for T1 and T2. The optimal model (see Table 6) revealed a main effect of time, showing a decrease in amplitude between the measurements (see Figure 8). The CIs were relatively wide, indicating a moderate precision of the effect. The effect size on the other hand was large ( β = -0.50). The correlation between the random intercept and slope was strong ( r = -0.77), suggesting that the participants with more positive MMRs at T1 showed a larger decrease towards a negativity over time. The fixed effects alone explained about 4.4% of the variance in the model (marginal R² = 0.044), while the full model, including both fixed and random effects, explained approximately 87% (conditional R² = 0.87). 3.4 Results of the behavioral quantity discrimination tasks 3.4.1. Results of Performance Accuracy and RT Progress: Comparison Between T1 and T2 In both of the behavioral discrimination tasks, participants performed significantly better at T2 compared to T1, with higher d’ scores in both the quantity discrimination ( V = 163, p = 2.25 × 10⁻⁵) and lexical decision tasks ( V = 257, p = 0.0012; see Table 7), and shorter reaction times in both tasks (quantity discrimination: V = 673, p = 8.07 × 10⁻⁵; lexical decision: V = 801, p = 0.011). 3.4.2 Group differences and errors in the quantity discrimination task and lexical decision task Group means differed significantly in d’ scores of the quantity discrimination task and lexical decision task at T1 and T2, but no differences in means were found in RTs (see Table 7). We further examined the discrimination of different quantity degrees. Table 7 shows the sum of errors made while discriminating specific quantity contrasts and group differences in means at T1 and T2. 3.5 Correlations between therapy, MMR amplitude and behavioral data For children with DLD, a significant correlation was found between therapy intensity and the mean MMR values in the durQ1 deviant condition ( r = 0.43, p = .047), and between therapy intensity and the mean MMR values in the pitchdurQ3 deviant condition ( r = 0.56, p = .007).We did not find any statistically significant correlations between the therapy intensity and the scores of behavioural tasks. 4 Discussion The aim of the current longitudinal study was to investigate the development of duration and pitch perception, the two main cues for distinguishing between quantity degrees, in Estonian children with DLD compared to their TD peers across one year behaviorally as well as neurophysiologically. At the neurophysiological level, we found a positive MMR in the duration change condition at T1 and a negative MMR at T2 with a significant change in amplitude across both groups of children. However, the change in amplitude was less pronounced in the DLD group compared to their TD peers. In contrast, in the duration and pitch change condition we did not find an MMR at T1, but after one year a negative MMR began to emerge with a significant change in amplitude across both groups from T1 to T2. Here, no significant group difference was found. Behavioral results indicated more precise and faster quantity discrimination in both participant groups at T2 compared to T1; however, group differences remained significant in terms of precision in favor of TD children. The highest number of errors occurred when participants had to differentiate duration and pitch simultaneously (Q2 vs Q3). No significant associations were found between neurophysiological and behavioral measures of auditory duration and pitch perception. Concerning the association between MMR and speech and language therapy, we observed a positive correlation between the amplitude of the MMR and the intensity of speech and language therapy. 4.1 Neurophysiological discrimination of duration and pitch contour In line with the hypothesis of a significant change in amplitude from T1 to T2 in the duration change condition in both groups, we found a significant shift from a positive to a negative MMR amplitude. These findings are in line with evidence that duration change detection is still maturing in children at preschool age. For example, Linnavalli et al., (2018) also found a significant increase in negativity in response to a shortening of a vowel in a consonant-vowel syllable in 5-6-year-olds that were followed longitudinally. Based on the assumption that the negative amplitude represents a more mature brain response to auditory changes (see e.g., Themas et al., 2023), we concluded that, in the present sample, the maturation period concerning duration change detection occurred between the ages of 5 to 7. This coincides with the period when Estonian children start to acquire written language and gain introductory knowledge about changes in phoneme length. While it is possible that conscious learning processes contributed to the observed shift toward a more mature MMR, it is equally plausible that this development reflects a primarily spontaneous maturation process, with formal instruction playing a supportive or coincidental role. In contrast to this assumption and our findings, other studies found duration change detection to be stable from the age of 6 years on (Engström et al., 2021; Putkinen et al., 2014). However, in these studies, simple tone stimuli were used, which is in contrast to the present study, where we used more complex stimuli, namely meaningful words. As several studies have shown differences in the developmental trajectories of the MMR to non-speech and speech stimuli (Chen et al., 2022; Kuuluvainen et al., 2016; Paquette et al., 2015), differences between our findings and those reported by Engström et al. (2021) and Putkinen et al. (2014) are most likely caused by differences in nature and complexity of stimuli. Even though both groups showed a significant shift from positivity to negativity in response to duration changes, we also found differences between children with and without DLD, which is line with our hypothesis that group differences remain stable from T1 (see Themas et al., 2025a) to T2 in the duration change condition. Specifically, we found a smaller reduction of MMR amplitude from T1 to T2 and less negative amplitudes at T2 in children with DLD compared to their TD peers. While Uwer et al. (2002) did not find any differences in duration change discrimination between children with and without DLD that were, however, older than our participants, studies with school children with DD showed attenuated negative MMR amplitudes in children with DD compared to their TD peers (e.g., Männel et al., 2017; Schaadt & Männel, 2019). Interestingly, LMM predictions in our previous study suggested differently, as we found decreased MMR amplitudes (i.e., more negative, mature MMRs) in older children with DLD, but increased MMR amplitudes (i.e., more positive, less mature MMRs) in older TD children (Themas et al., 2025a). This discrepancy to the findings of this longitudinal study might be explained by observations suggesting the developmental trajectory of the MMR to not be linear (Werwach et al., 2022). It could be that the group of TD children have gone through a change towards positivity that was measured at T1 and then started to increase again resulting in a negative MMR at T2. Even if the latter indicates that during childhood one year between repeated measures might be too long of a period to capture nuances in the developmental trajectory of the MMR, the longitudinal approach used in the current study is superior and more robust compared to single measurements, allowing us to draw more general conclusions, namely that the amplitude of the MMR became more mature in both groups, but less so in children with DLD. In contrast to our findings on auditory duration discrimination and in contrast to our hypothesis that the MMR to the pitch and duration contrast emerges at T2, at least in TD children, cluster-based permutation tests showed no significant deviant-standard differences either at T1 or T2. This finding might not be a true null result, because the underlying effect can be small and hard to detect given our relatively small sample size and the inherent noisiness and individual variability of developmental EEG data. Exploratorily computing LMM, we found significantly more negative MMR amplitudes at T2 in both groups. Further, the model revealed that individual differences played a substantial role: both the initial MMR responses (random intercept) and how they changed over time (random slope) varied significantly between children. This could mean that the detection of changes in duration and pitch simultaneously is still developing in 7-year-old children with DLD as well as in their TD peers. However, to draw more specific conclusions, an additional assessment when children are 8 years old would be required. When looking more closely at the latency of the MMRs to the duration change condition, we found an increase in latency from T1 to T2 (T1 positivity: 240-340 ms vs. T2 negativity: 390-490 ms) in both groups, which is in contrast to studies showing an MMR latency decrease with age (Choudhury and Benasich, 2011; Morr et al., 2002; Shafer et al., 2000; Shafer et al., 2010; Themas et al., 2023). However, some other studies, though with younger children, also found increases in MMR (and other ERP) latencies with age (Alatorre-Cruz et al., 2023; Čeponienė et al., 2005; Choudhury & Benasich, 2011; Jing & Benasich, 2006; Virtala et al., 2022), possibly related to more sophisticated and at the same time more time-consuming processing of the acoustic stimulus features in older children (Alatorre-Cruz et al., 2023; Čeponienė et al., 2005; Siegler, 2004). In line, it has been suggested that the different MMR polarities reflect distinct underlying neural processes (Rivera-Gaxiola et al., 2005), potentially linked to bottom-up (positive MMR) and top-down (negative MMR) processing pathways, which are thought to involve separate neural generators (see Schaadt et al., 2015). Thus, the negative MMR at T2 may reflect more sophisticated top-down processes that are, however, less mature and more time-consuming in comparison to adults. Again, to prove this assumption, an additional measurement would be needed to show that processes become more efficient, reflected in shorter MMR latencies, when children get older. An alternative explanation for the observed latency differences could be that the positive MMR observed at T1 was not detectable at T2 due to it crossing zero caused by an ongoing polarity shift (Marklund et al., 2019). Thus, instead the negative MMR at T2 may represent a Late Discriminative Negativity (LDN; Cheour et al., 2001) that decreased in latency from T1 (590-690 ms after stimulus onset, see Appendix D) to T2 (390-490 after stimulus onset, see Appendix D). The results on the latency decrease of LDN are contradictory so far. Some report reduction of latency with age (Bishop et al., 2011; Mueller et al., 2008), some report no latency change (David et al., 2020; Hommet et al., 2009) and some report increase in latency with age (Virtala et al., 2022). We argue that our findings align with David et al. (2020) and Hommet et al. (2009), because we also observed a small but significant deviant-standard difference cluster at T2, occurring at the same time-window (590-690 after stimulus onset) and over similar region as the LDN at T1 (Appendix D). Thus, the negativity observed between 390-490 ms at T2 is unlikely to reflect an LDN, but rather an MMR. Instead, the LDN at T2 appeared later like at T1 and was smaller in amplitude compared to T1, possibly due to the participants being older at T2 (Bishop et al., 2011; Gumenyuk et al., 2004; Hommet et al., 2009; Linnavalli et al., 2018; Liu et al., 2014; Morales et al., 2022; Torppa et al. 2022). 4.2 Behavioral discrimination of duration and pitch contour In line with the hypothesis of both groups to be more accurate and to respond faster in the behavioral discrimination tasks, we found increased accuracy in discriminating between quantity degrees and decreased reaction times at T2 compared to T1. Also, in line with the hypothesis, we found children with DLD to perform less accurate compared to their TD peers. These findings are consistent with the findings on the MMR, indicating maturation of quantity change detection in both groups. More specifically, we found the emergence of negative MMRs and significantly larger d’ scores in behavioral data at T2. In addition, behavioral results suggest duration changes to be easier to detect (discrimination of Q1 vs. Q2) compared to duration and pitch changes (discrimination of Q2 vs Q3) and TD children to show less difficulties in both conditions, compared to children with DLD. Similarly, we found significant MMRs in the duration change condition across both groups, while TD children showed more mature MMRs at T2. We found no significant MMRs in the duration and pitch change condition either in children with or without DLD. However, there were substantial individual differences in the duration and pitch contrast, suggesting it to remain difficult for many participants at the age of 7,5 years, which is in line with the behavioral results. In contrast to our hypothesis of an association between the d’ prime sensitivity indices from the behavioral discrimination tasks and the MMR amplitude, the LMMs did not show any association neither between d’ scores and MMR amplitude, nor between language test scores and the MMR amplitudes at both assessments, which is in line with several studies failing to identify these association (Ahmmed et al., 2008; Bishop et al., 2010; Halliday et al., 2014; McArthur & Bishop, 2005; Pihko et al., 2006; Uwer et al., 2002). In contrast, however, other studies found associations between behavioral and neurophysiological markers of auditory discrimination and language abilities (Chen et al., 2016; Kautto et al., 2024; Kujala & Leminen, 2017; Kwok et al., 2018; McArthur & Bishop, 2005; Nan et al., 2018; Pihko et al., 2006). Different explanations might elucidate these discrepancies. First, some studies have examined associations between MMR amplitudes or latencies and language abilities across specific domains. Chen et al. (2016) for example, reported significant correlations with receptive vocabulary and syntax production scores, while Kushnerenko et al. (2013) found links with receptive language test scores. Additionally, Harwood et al. (2017) found an association between ERP (named as an early negative component in the study) latency and language abilities. In contrast, we focused solely on the overall language score and its relationship with MMR amplitude, not latency, which may explain the lack of significant associations. Second, the relation between ERPs and behavioral data might be more complex and influenced by development, such that simple linear measurements are insufficient to capture it (Ahmmed et al., 2008; Clarke & Adams, 2007). Third, even though we anticipated the impact of task load to diminish with age (Betts et al., 2006), attention sustention or working-memory demands may have still influenced behavioral results, but not ERP results, such that no association between MMR amplitude and d’ scores could have been observed. 4.3 Association between MMR amplitude and speech and language therapy attendance Even though we found a significant correlation between the intensity of speech and language therapy and the amplitude of the MMR, this should not be interpreted as evidence of a causal relationship. Interestingly, the direction of the association was unexpected: children who attended therapy more frequently showed more positive (i.e., less mature) MMRs. This contrasts with the evidence suggesting that structured, conscious language learning through specialized methods leads to more favorable language development outcomes (Gillam et al., 2021). Moreover, targeted interventions focusing on auditory processing or phonological skills have been shown to modulate the morphology of the MMR or MMN (Dacewicz et al., 2018; Heim et al., 2016; Pihko et al., 2006). Thus, we would have expected children with higher attendance at speech and language therapy to show more mature, namely more negative MMRs. In contrast, we observed more positive MMRs in children with more intense therapy. Possibly, children with more positive MMRs are more severely affected by DLD and thus, were reported to attend speech and language therapy more intensely. In addition, it needs to be discussed critically that correlational analysis was only performed on 22 children with DLD and the estimation of therapy intensity lacks precision, as it was based on average minutes of attendance per year and did not account for periods when the child was absent due to illness or holidays. Finally, we did not have information about the quality or content of therapy, factors that should be considered in future research. 4.4 Limitations As already pointed out, conclusions drawn in the present study are based on a relatively small sample size, due to difficulties in recruiting participants with DLD fulfilling strict diagnostic criteria in a country with a small and dispersed population, such as Estonia. This limited sample may reduce the generalizability of the findings and impact their statistical reliability. Additionally, the small sample size may have hindered the detection of small effects in statistical models. For example, in the duration change condition (durQ1), the LMM did not show a significant main effect of group, but the confidence intervals suggested that a true effect size is likely small, and a large sample may be needed to detect such subtle group differences. Related to this, we did not perform a priori power analysis, even though we acknowledge its importance for avoiding Type II error. However, due to the high specificity of the stimuli used in the present study and as the properties of the MMR are highly dependent on the nature of the stimuli (e.g., Themas et al., 2023), extracting an effect size from studies using different stimuli or study population, would have led to an erroneous estimation of the effect size. Given that our study was the first EEG study conducted with Estonian children, performing a power analysis would have required speculation, potentially leading to inaccurate conclusions about the required sample size. 5 Conclusion In conclusion, our findings suggest that MMRs in response to duration and pitch contour changes — within the context of Estonian three-way quantity degrees — are still maturing at preschool age, with slower maturation in DLD participants in the duration change condition, suggesting a delay in auditory processing development. Even though our findings should be interpreted with caution due to the limited sample size and the inclusion of only two longitudinal measurement points, our study offers valuable insights into the maturation of auditory change detection mechanisms and contributes to a deeper understanding of how neurophysiological and behavioral measures interact in both typically developing and language-impaired Estonian children. Conflict of interest statement The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Declaration of Generative AI and AI-assisted technologies in the writing process During the preparation of this work the author(s) used ChatGPT version 4 in order to correct the use of English language. After using this tool, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication. Additionally, ChatGPT was used to find mistakes in code and generate some parts of code which was used for data analysis. Before use the generated code was reviewed by the author(s). References Ahmmed, A. U., Clarke, E. M., & Adams, C. (2008). Mismatch negativity and frequency representational width in children with specific language impairment. Developmental Medicine & Child Neurology , 50 (12), 938–944. https://doi.org/10.1111/j.1469-8749.2008.03093.x Alatorre-Cruz, G. C., Andres, A., Gu, Y., Downs, H., Hagood, D., Sorensen, S. T., Williams, D. K., & Larson-Prior, L. J. (2023). Impact of feeding habits on the development of language-specific processing of phonemes in brain: An event-related potentials study. Frontiers in Nutrition, 10, 1032413. https://doi.org/10.3389/fnut.2023.1032413 Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 67(1). https://doi.org/10.18637/jss.v067.i01 Beattie, R. L., & Manis, F. R. (2013). Rise Time Perception in Children With Reading and Combined Reading and Language Difficulties. Journal of Learning Disabilities, 46(3), 200–209. https://doi.org/10.1177/0022219412449421 Betts, J., Mckay, J., Maruff, P., & Anderson, V. (2006). The Development of Sustained Attention in Children: The Effect of Age and Task Load. Child Neuropsychology, 12(3), 205–221. https://doi.org/10.1080/09297040500488522 Bishop, D. V. M., Hardiman, M. J., & Barry, J. G. (2011). Is auditory discrimination mature by middle childhood? A study using time-frequency analysis of mismatch responses from 7 years to adulthood. Developmental Science, 14(2), 402–416. https://doi.org/10.1111/j.1467-7687.2010.00990.x Bishop, D. V. M., Hardiman, M. J., & Barry, R. J. (2010). Lower-Frequency Event-Related Desynchronization: A Signature of Late Mismatch Responses to Sounds, Which Is Reduced or Absent in Children with Specific Language Impairment. Journal of Neuroscience, 17, 15578–15584. Bishop, D. V. M., Hardiman, M., Uwer, R., & Von Suchodoletz, W. (2007). Atypical long-latency auditory event-related potentials in a subset of children with specific language impairment. Developmental Science , 10 (5), 576–587. https://doi.org/10.1111/j.1467-7687.2007.00620.x Bishop, D. V. M., & McArthur, G. M. (2004). Immature cortical responses to auditory stimuli in specific language impairment: Evidence from ERPs to rapid tone sequences. Developmental Science , 7 (4), F11–F18. https://doi.org/10.1111/j.1467-7687.2004.00356.x Bishop, D. V. M., Snowling, M. J., Thompson, P. A., Greenhalgh, T., & and the CATALISE-2 consortium. (2017). Phase 2 of CATALISE: A multinational and multidisciplinary Delphi consensus study of problems with language development: Terminology. Journal of Child Psychology and Psychiatry , 58 (10), 1068–1080. https://doi.org/10.1111/jcpp.12721 Boersma, P., & Weernink, D. (2021). Praat: Doing phonetics by computer (Version 6.1) [Computer software]. Buuren, S. V., & Groothuis-Oudshoorn, K. (2011). mice: Multivariate Imputation by Chained Equations in R. Journal of Statistical Software, 45(3). https://doi.org/10.18637/jss.v045.i03 Calet, N., Martín-Peregrina, M. Á., Jiménez-Fernández, G., & Martínez-Castilla, P. (2021). Prosodic skills of Spanish-speaking children with developmental language disorder. Int J Lang Commun Disord , 56 , 784–796. https://doi.org/DOI: 10.1111/1460-6984.12627 Campos, A., Tuomainen, J., & Tuomainen, O. (2023). Mismatch Responses to Speech Contrasts in Preschoolers with and without Developmental Language Disorder. Brain Sciences , 14 (1), 42. https://doi.org/10.3390/brainsci14010042 Cantiani, C., Riva, V., Piazza, C., Bettoni, R., Molteni, M., Choudhury, N., Marino, C., & Benasich, A. A. (2016). Auditory discrimination predicts linguistic outcome in Italian infants with and without familial risk for language learning impairment. Developmental Cognitive Neuroscience , 20 , 23–34. https://doi.org/10.1016/j.dcn.2016.03.002 Čeponienė, R., Alku, P., Westerfield, M., Torki, M., & Townsend, J. (2005). ERPs differentiate syllable and nonphonetic sound processing in children and adults. Psychophysiology, 42(4), 391–406. https://doi.org/10.1111/j.1469-8986.2005.00305.x Čeponienė, R., Cummings, A., Wulfeck, B., Ballantyne, A., & Townsend, J. (2009). Spectral vs. temporal auditory processing in specific language impairment: A developmental ERP study. Brain and Language , 110 (3), 107–120. https://doi.org/10.1016/j.bandl.2009.04.003 Chen, A., Peter, V., & Burnham, D. (2022). Development of neural discrimination of pitch across speech and music in the first year of life, a mismatch response study. Language, Cognition and Neuroscience, 37(9), 1153–1168. https://doi.org/10.1080/23273798.2022.2051571 Chen, Y., Tsao, F.-M., & Liu, H.-M. (2016). Developmental changes in brain response to speech perception in late-talking children: A longitudinal MMR study. Developmental Cognitive Neuroscience, 19, 190–199. https://doi.org/10.1016/j.dcn.2016.03.007 Cheng, Y.-Y., Wu, H.-C., Shih, H.-Y., Yeh, P.-W., Yen, H.-L., & Lee, C.-Y. (2021). Deficits in Processing of Lexical Tones in Mandarin-Speaking Children With Developmental Language Disorder: Electrophysiological Evidence. Journal of Speech, Language, and Hearing Research , 64 (4), 1176–1188. https://doi.org/10.1044/2021_JSLHR-19-00392 Cheng, Y.-Y., Wu, H.-C., Tzeng, Y.-L., Yang, M.-T., Zhao, L.-L., & Lee, C.-Y. (2015). Feature-specific transition from positive mismatch response to mismatch negativity in early infancy: Mismatch responses to vowels and initial consonants. International Journal of Psychophysiology , 96 (2), 84–94. https://doi.org/10.1016/j.ijpsycho.2015.03.007 Cheour, M., Korpilahti, P., Martynova, O., & Lang, A.-H. (2001). Mismatch Negativity and Late Discriminative Negativity in Investigating Speech Perception and Learning in Children and Infants. Audiology and Neuro-Otology, 6(1), 2–11. https://doi.org/10.1159/000046804 Choudhury, N., & Benasich, A. A. (2011). Maturation of auditory evoked potentials from 6 to 48 months: Prediction to 3 and 4 year language and cognitive abilities. Clinical Neurophysiology, 122(2), 320–338. https://doi.org/10.1016/j.clinph.2010.05.035 Clarke, E. M., & Adams, C. (2007). Binaural interaction in specific language impairment: An auditory evoked potential study. Developmental Medicine & Child Neurology, 49(4), 274–279. https://doi.org/10.1111/j.1469-8749.2007.00274.x Corriveau, K., Pasquini, E., & Goswami, U. (2007). Basic Auditory Processing Skills and Specific Language Impairment: A New Look at an Old Hypothesis. Journal of Speech, Language, and Hearing Research, 50(3), 647–666. https://doi.org/10.1044/1092-4388(2007/046) Coughler, C., Hamel, E. M., Cardy, J. O., Archibald, L. M. D., & Purcell, D. W. (2021). Compensation to Altered Auditory Feedback in Children With Developmental Language Disorder and Typical Development. Journal of Speech, Language, and Hearing Research, 64(6S), 2363–2376. https://doi.org/10.1044/2020_JSLHR-20-00374 Crosbie, S. L., Howard, D., & Dodd, B. (2004). Auditory lexical decisions in children with specific language impairment. British Journal of Developmental Psychology, 22, 103–121. Cumming, R., Wilson, A., & Goswami, U. (2015). Basic auditory processing and sensitivity to prosodic structure in children with specific language impairments: A new look at a perceptual hypothesis. Frontiers in Psychology, 6. https://doi.org/10.3389/fpsyg.2015.00972 Dacewicz, A., Szymaszek, A., Nowak, K., & Szelag, E. (2018). Training-Induced Changes in Rapid Auditory Processing in Children With Specific Language Impairment: Electrophysiological Indicators. Frontiers in Human Neuroscience, 12, 310. https://doi.org/10.3389/fnhum.2018.00310 Datta, H., Shafer, V. L., Morr, M. L., Kurtzberg, D., & Schwartz, R. G. (2010). Electrophysiological Indices of Discrimination of Long-Duration, Phonetically Similar Vowels in Children With Typical and Atypical Language Development. Journal of Speech, Language, and Hearing Research, 53(3), 757–777. https://doi.org/10.1044/1092-4388(2009/08-0123) David, C., Roux, S., Bonnet‐Brilhault, F., Ferré, S., & Gomot, M. (2020). Brain responses to change in phonological structures of varying complexity in children and adults. Psychophysiology, 57(9). https://doi.org/10.1111/psyp.13621 Davids, N., Segers, E., van den Brink, D., Mitterer, H., van Balkom, H., Hagoort, P., & Verhoeven, L. (2011). The nature of auditory discrimination problems in children with specific language impairment: An MMN study. Neuropsychologia, 49(1), 19–28. https://doi.org/10.1016/j.neuropsychologia.2010.11.001 Debnath, R., Buzzell, G. A., Morales, S., Bowers, M. E., Leach, S. C., & Fox, N. A. (2020). The Maryland analysis of developmental EEG (MADE) pipeline. Psychophysiology, 57(6). https://doi.org/10.1111/psyp.13580 Delorme, A., & Makeig, S. (2004). EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. Journal of Neuroscience Methods , 134(1), 9–21. https://doi.org/10.1016/j.jneumeth.2003.10.009 Eichhorn, K., & Juur, A. (2024). Perception of degrees of quantity in preschoolers with developmental language disorder and typical language development (Master’s thesis). University of Tartu, Estonia. Engström, E., Kallioinen, P., Nakeva von Mentzer, C., Lindgren, M., Sahlén, B., Lyxell, B., Ors, M., & Uhlén, I. (2021). Auditory event-related potentials and mismatch negativity in children with hearing loss using hearing aids or cochlear implants – A three-year follow-up study. International Journal of Pediatric Otorhinolaryngology, 140, 110519. https://doi.org/10.1016/j.ijporl.2020.110519 Fong, C. Y., Law, W. H. C., Uka, T., & Koike, S. (2020). Auditory Mismatch Negativity Under Predictive Coding Framework and Its Role in Psychotic Disorders. Frontiers in Psychiatry, 11, 557932. https://doi.org/10.3389/fpsyt.2020.557932 Friedrich, M., & Friederici, A. D. (2006). Early N400 development and later language acquisition. Psychophysiology , 43 (1), 1–12. https://doi.org/10.1111/j.1469-8986.2006.00381.x Friedrich, M., Herold, B., & Friederici, A. D. (2009). ERP correlates of processing native and non-native language word stress in infants with different language outcomes. Cortex , 45 (5), 662–676. https://doi.org/10.1016/j.cortex.2008.06.014 Friedrich, M., Weber, C., & Friederici, A. D. (2004). Electrophysiological evidence for delayed mismatch response in infants at-risk for specific language impairment. Psychophysiology , 41 (5), 772–782. https://doi.org/doi: 10.1111/j.1469-8986.2004.00202.x. Gillam, S. L., Holbrook, S., & Kamhi, A. G. (2021). Developmental Language Disorder. In J. S. Damico, N. Müller, & M. J. Ball (Eds.), The Handbook of Language and Speech Disorders (1st ed., pp. 171–191). Wiley. https://doi.org/10.1002/9781119606987.ch9 Gumenyuk, V., Korzyukov, O., Alho, K., Escera, C., & Näätänen, R. (2004). Effects of auditory distraction on electrophysiological brain activity and performance in children aged 8–13 years. Psychophysiology, 41(1), 30–36. https://doi.org/10.1111/1469-8986.00123 Guttorm, T. K., Leppänen, P. H. T., Poikkeus, A.-M., Eklund, K. M., Lyytinen, P., & Lyytinen, H. (2005). Brain Event-Related Potentials (ERPs) Measured at Birth Predict Later Language Development in Children with and Without Familial Risk for Dyslexia. Cortex , 41 (3), 291–303. https://doi.org/10.1016/S0010-9452(08)70267-3 Halliday, L. F., Barry, J. G., Hardiman, M. J., & Bishop, D. V. (2014). Late, not early mismatch responses to changes in frequency are reduced or deviant in children with dyslexia: An event-related potential study. Journal of Neurodevelopmental Disorders, 6(1), 21. https://doi.org/10.1186/1866-1955-6-21 Harwood, V., Preston, J., Grela, B., Roy, D., Harold, O., Turcios, J., Andrada, K., & Landi, N. (2017). Electrophysiology of Perception and Processing of Phonological Information as Indices of Toddlers’ Language Performance. Journal of Speech, Language, and Hearing Research, 60(4), 999–1011. https://doi.org/10.1044/2016_JSLHR-L-15-0437 Heim, S., Choudhury, N., & Benasich, A. A. (2016). Electrocortical Dynamics in Children with a Language-Learning Impairment Before and After Audiovisual Training. Brain Topography, 29(3), 459–476. https://doi.org/10.1007/s10548-015-0466-y Hommet, C., Vidal, J., Roux, S., Blanc, R., Barthez, M. A., De Becque, B., Barthelemy, C., Bruneau, N., & Gomot, M. (2009). Topography of syllable change-detection electrophysiological indices in children and adults with reading disabilities. Neuropsychologia, 47(3), 761–770. https://doi.org/10.1016/j.neuropsychologia.2008.12.010 Jing, H., & Benasich, A. A. (2006). Brain responses to tonal changes in the first two years of life. Brain and Development, 28(4), 247–256. https://doi.org/10.1016/j.braindev.2005.09.002 Karakaş, S. (2024). A Review of Childhood Developmental Changes in Attention as Indexed in the Electrical Activity of the Brain. Brain Sciences, 14(5), 458. https://doi.org/10.3390/brainsci14050458 Karlep, K. (2000). WRITING DISABILITIES OF ESTONIAN CHILDREN. Trames. Journal of the Humanities and Social Sciences, 4(1), 53. https://doi.org/10.3176/tr.2000.1.03 Kautto, A., Railo, H., & Mainela-Arnold, E. (2024). Low-Level Auditory Processing Correlates With Language Abilities: An ERP Study Investigating Sequence Learning and Auditory Processing in School-Aged Children. Neurobiology of Language, 5(2), 341–359. https://doi.org/10.1162/nol_a_00129 Kujala, T. (2007). The Role of Early Auditory Discrimination Deficits in Language Disorders. Journal of Psychophysiology , 21 (3–4), 239–250. https://doi.org/10.1027/0269-8803.21.34.239 Kujala, T., & Leminen, M. (2017). Low-level neural auditory discrimination dysfunctions in specific language impairment—A review on mismatch negativity findings. Developmental Cognitive Neuroscience , 28 , 65–75. https://doi.org/10.1016/j.dcn.2017.10.005 Kushnerenko, E., Tomalski, P., Ballieux, H., Potton, A., Birtles, D., Frostick, C., & Moore, D. G. (2013). Brain responses and looking behavior during audiovisual speech integration in infants predict auditory speech comprehension in the second year of life. Frontiers in Psychology, 4. https://doi.org/10.3389/fpsyg.2013.00432 Kuuluvainen, S., Alku, P., Makkonen, T., Lipsanen, J., & Kujala, T. (2016). Cortical speech and non-speech discrimination in relation to cognitive measures in preschool children. European Journal of Neuroscience, 43(6), 738–750. https://doi.org/10.1111/ejn.13141 Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2017). lmerTest Package: Tests in Linear Mixed Effects Models. Journal of Statistical Software, 82(13). https://doi.org/10.18637/jss.v082.i13 Kwok, E. Y. L., Joanisse, M. F., Archibald, L. M. D., & Cardy, J. O. (2018). Immature Auditory Evoked Potentials in Children With Moderate–Severe Developmental Language Disorder. Journal of Speech, Language, and Hearing Research, 61(7), 1718–1730. https://doi.org/10.1044/2018_JSLHR-L-17-0420 Lee, C.-Y., Yen, H., Yeh, P., Lin, W.-H., Cheng, Y.-Y., Tzeng, Y.-L., & Wu, H.-C. (2012). Mismatch responses to lexical tone, initial consonant, and vowel in Mandarin-speaking preschoolers. Neuropsychologia. https://doi.org/10.1016/j.neuropsychologia.2012.08.025 Leonard, L. B. (2014). Children with specific language impairment (Second edition). The MIT Press. Leonard, L. B., & Kueser, J. B. (2019). Five overarching factors central to grammatical learning and treatment in children with developmental language disorder. International Journal of Language & Communication Disorders , 54 (3), 347–361. https://doi.org/10.1111/1460-6984.12456 Leppik, K., Lippus, P., & Asu, E. L. (2023). The perception and production of Estonian vowels and vocalic quantity contrasts by Spanish L1 learners. Ampersand, 11, 100147. https://doi.org/10.1016/j.amper.2023.100147 Leppänen, P. H. T., Guttorm, T. K., Pihko, E., Takkinen, S., Eklund, K. M., & Lyytinen, H. (2004). Maturational effects on newborn ERPs measured in the mismatch negativity paradigm. Experimental Neurology , 190 , 91–101. https://doi.org/10.1016/j.expneurol.2004.06.002 Linnavalli, T., Putkinen, V., Huotilainen, M., & Tervaniemi, M. (2018). Maturation of Speech-Sound ERPs in 5–6-Year-Old Children: A Longitudinal Study. Frontiers in Neuroscience, 12, 814. https://doi.org/10.3389/fnins.2018.00814 Lippus, P., Asu, E. L., Teras, P., & Tuisk, T. (2013). Quantity-related variation of duration, pitch and vowel quality in spontaneous Estonian. Journal of Phonetics , 41 (1), 17–28. https://doi.org/10.1016/j.wocn.2012.09.005 Liu, H.-M., Chen, Y., & Tsao, F.-M. (2014). Developmental Changes in Mismatch Responses to Mandarin Consonants and Lexical Tones from Early to Middle Childhood. PLoS ONE, 9(4), e95587. https://doi.org/10.1371/journal.pone.0095587 Männel, C., Schaadt, G., Illner, F. K., Van Der Meer, E., & Friederici, A. D. (2017). Phonological abilities in literacy-impaired children: Brain potentials reveal deficient phoneme discrimination, but intact prosodic processing. Developmental Cognitive Neuroscience, 23, 14–25. https://doi.org/10.1016/j.dcn.2016.11.007 Marklund, E., Schwarz, I.-C., & Lacerda, F. (2019). Amount of speech exposure predicts vowel perception in four- to eight-month-olds. Developmental Cognitive Neuroscience, 36, 100622. https://doi.org/10.1016/j.dcn.2019.100622 Maurer, U., Bucher, K., Brem, S., & Brandeis, D. (2003). Development of the automatic mismatch response: From frontal positivity in kindergarten children to the mismatch negativity. Clinical Neurophysiology , 114 (5), 808–817. https://doi.org/10.1016/S1388-2457(03)00032-4 McArthur, G. M., & Bishop, D. V. M. (2005). Speech and non-speech processing in people with specific language impairment: A behavioural and electrophysiological study. Brain and Language . Meyer, M., Lamers, D., Kayhan, E., Hunnius, S., & Oostenveld, R. (2021). Enhancing reproducibility in developmental EEG research: BIDS, cluster-based permutation tests, and effect sizes. Developmental Cognitive Neuroscience, 52, 101036. https://doi.org/10.1016/j.dcn.2021.101036 Morales, S., Bowers, M. E., Leach, S. C., Buzzell, G. A., McSweeney, M., Yoder, L., Fifer, W., Elliott, A. J., & Fox, N. A. (2022). Development of auditory change-detection and attentional capture, and their relation to inhibitory control. Psychophysiology, 60(4). https://doi.org/10.1111/psyp.14211 Morr, M. L., Shafer, V. L., Kreuzer, J. A., & Kurtzberg, D. (2002). Maturation of Mismatch Negativity in Typically Developing Infants and Preschool Children: Ear and Hearing, 23(2), 118–136. https://doi.org/10.1097/00003446-200204000-00005 Mueller, V., Brehmer, Y., von Oertzen, T., Li, S.-C., & Lindenberger, U. (2008). Electrophysiological correlates of selective attention: A lifespan comparison. BMC Neuroscience, 9(1), 18. https://doi.org/10.1186/1471-2202-9-18 Mueller, J. L., Friederici, A. D., & Männel, C. (2012). Auditory perception at the root of language learning. Proceedings of the National Academy of Sciences , 109 (39), 15953–15958. https://doi.org/10.1073/pnas.1204319109 Murphy, J. I., Weaver, N. E., & Hendricks, A. E. (2022). Accessible analysis of longitudinal data with linear mixed effects models. Disease Models & Mechanisms, 15(5), dmm048025. https://doi.org/10.1242/dmm.048025 Näätänen, R. (2001). The perception of speech sounds by the human brain as reflected by the mismatch negativity (MMN) and its magnetic equivalent (MMNm). Psychophysiology , 38 (1), 1–21. https://doi.org/10.1111/1469-8986.3810001 Näätänen, R., Gaillard, A. W. K., & Mantysalo, S. (1978). EARLY SELECTIVE-AmENTION EFFECT ON EVOKED POTENTIAL REINTERPRETED. Acta Psychologica , 42 (4), 313–329. https://doi.org/10.1016/0001-6918(78)90006-9 Näätänen, R., Kujala, T., Escera, C., Baldeweg, T., Kreegipuu, K., Carlson, S., & Ponton, C. (2012). The mismatch negativity (MMN) – A unique window to disturbed central auditory processing in ageing and different clinical conditions. Clinical Neurophysiology , 123 (3), 424–458. https://doi.org/10.1016/j.clinph.2011.09.020 Näätänen, R., Paavilainen, P., Rinne, T., & Alho, K. (2007). The mismatch negativity (MMN) in basic research of central auditory processing: A review. Clinical Neurophysiology , 118 (12), 2544–2590. https://doi.org/10.1016/j.clinph.2007.04.026 Näätänen, R., Sussman, E. S., Salisbury, D., & Shafer, V. L. (2014). Mismatch Negativity (MMN) as an Index of Cognitive Dysfunction. Brain Topogr , 27 (4). https://doi.org/DOI: 10.1007/s10548-014-0374-6 Nan, Y., Liu, L., Geiser, E., Shu, H., Gong, C. C., Dong, Q., Gabrieli, J. D. E., & Desimone, R. (2018). Piano training enhances the neural processing of pitch and improves speech perception in Mandarin-speaking children. Proceedings of the National Academy of Sciences , 115 (28), E6630–E6639. https://doi.org/10.1073/pnas.1808412115 Norton, E. S., MacNeill, L. A., Harriott, E. M., Allen, N., Krogh-Jespersen, S., Smyser, C. D., Rogers, C. E., Smyser, T. A., Luby, J., & Wakschlag, L. (2021). EEG/ERP as a pragmatic method to expand the reach of infant-toddler neuroimaging in HBCD: Promises and challenges. Developmental Cognitive Neuroscience , 51 , 100988. https://doi.org/10.1016/j.dcn.2021.100988 Oostenveld, R., Fries, P., Maris, E., & Schoffelen, J.-M. (2011). FieldTrip: Open Source Software for Advanced Analysis of MEG, EEG, and Invasive Electrophysiological Data. Computational Intelligence and Neuroscience, 2011, 1–9. https://doi.org/10.1155/2011/156869 Padrik, M., Hallap, M., & Mäll, R. (2013). 5-6 aastatse laste kõne test. Studium Publishers. Paquette, N., Vannasing, P., Tremblay, J., Lefebvre, F., Roy, M.-S., McKerral, M., Lepore, F., Lassonde, M., & Gallagher, A. (2015). Early electrophysiological markers of atypical language processing in prematurely born infants. Neuropsychologia, 79, 21–32. https://doi.org/10.1016/j.neuropsychologia.2015.10.021 Pihko, E., Mickos, A., Kujala, T., Pihlgren, A., Westman, M., Alku, P., Byring, R., & Korkman, M. (2006). Group Intervention Changes Brain Activity in Bilingual Language-Impaired Children. Cerebral Cortex, 17(4), 849–858. https://doi.org/10.1093/cercor/bhk037 Putkinen, V., Tervaniemi, M., Saarikivi, K., Ojala, P., & Huotilainen, M. (2014). Enhanced development of auditory change detection in musically trained school-aged children: A longitudinal event-related potential study. Developmental Science, 17(2), 282–297. https://doi.org/10.1111/desc.12109 R Core Team. (2023). R: A language and environment for statistical computing (Version 4.3.2) [Software]. R Foundation for Statistical Computing. https://www.R-project.org/ Richards, S., & Goswami, U. (2019). Impaired Recognition of Metrical and Syntactic Boundaries in Children with Developmental Language Disorders. Brain Sciences , 9 (2), 33. https://doi.org/10.3390/brainsci9020033 Rivera‐Gaxiola, M., Silva‐Pereyra, J., & Kuhl, P. K. (2005). Brain potentials to native and non‐native speech contrasts in 7‐ and 11‐month‐old American infants. Developmental Science, 8(2), 162–172. https://doi.org/10.1111/j.1467-7687.2005.00403.x Schaadt, G., & Männel, C. (2019). Phonemes, words, and phrases: Tracking phonological processing in pre-schoolers developing dyslexia. Clinical Neurophysiology, 130(8), 1329–1341. https://doi.org/10.1016/j.clinph.2019.05.018 Schaadt, G., Männel, C., van der Meer, E., Pannekamp, A., Oberecker, R., & Friederici, A. D. (2015). Present and past: Can writing abilities in school children be associated with their auditory discrimination capacities in infancy?. Res Dev Disabil, 47, 318–333. https://doi.org/DOI: 10.1016/j.ridd.2015.10.002 Schaadt, G., Werwach, A., Obrig, H., Friederici, A. D., & Männel, C. (2023). Maturation of consonant perception, but not vowel perception, predicts lexical skills at 12 months. Child Development , 94 (3). https://doi.org/10.1111/cdev.13892 Shafer, V. L., Morr, M. L., Kreuzer, J. A., & Kurtzberg, D. (2000). Maturation of Mismatch Negativity in School-Aged Children. Ear & Hearing, 21, 242–251. Shafer, V. L., Yu, Y. H., & Datta, H. (2010). Maturation of Speech Discrimination in 4- to 7-Yr-Old Children as Indexed by Event-Related Potential Mismatch Responses. Ear and Hearing , 31 (6), 735–745. https://doi.org/10.1097/AUD.0b013e3181e5d1a7 Siegler, R. S. (2004). U-Shaped Interest in U-Shaped Development-and What It Means. Journal of Cognition and Development, 5(1), 1–10. https://doi.org/10.1207/s15327647jcd0501_1 Squires, N. K., Squires, K. C., & Hillyard, S. A. (1975). Two varieties of long-latency positive waves evoked by unpredictable auditory stimuli in man. Electroencephalography and Clinical Neurophysiology , 38 (4), 387–401. https://doi.org/10.1016/0013-4694(75)90263-1 Suades-González, E., Forns, J., García-Esteban, R., López-Vicente, M., Esnaola, M., Álvarez-Pedrerol, M., Julvez, J., Cáceres, A., Basagaña, X., López-Sala, A., & Sunyer, J. (2017). A Longitudinal Study on Attention Development in Primary School Children with and without Teacher-Reported Symptoms of ADHD. Frontiers in Psychology, 8, 655. https://doi.org/10.3389/fpsyg.2017.00655 Sundström, S., Lyxell, B., & Samuelsson, C. (2019). Prosodic aspects of repetition in Swedish-speaking children with developmental language disorder. International Journal of Speech-Language Pathology , 21 (6), 623–634. https://doi.org/10.1080/17549507.2018.1508500 The MathWorks Inc. (2015). MATLAB (version: 8.5.0, 2015a) [Software]. Natick, Massachusetts: The MathWorks Inc. https://www.mathworks.com The MathWorks, Inc. (2023). MATLAB (Version 9.14.0, R2023a) [Software]. Natick, MA: The MathWorks, Inc. https://www.mathworks.com Themas, L. (2025). Auditory mismatch response EEG data for Estonian children with DLD [Data set]. University of Tartu Library DataDOI. https://doi.org/10.23673/re-525 Themas, L., Nickel, E. S. (2025). DLD-Estonian-Quantity/second_round_analysis: Code and Data for Longitudinal EEG Study on Auditory Mismatch Response in Estonian Children with DLD (v1.1). Zenodo. https://doi.org/10.5281/zenodo.15799644 Themas, L., Lippus, P., Padrik, M., Kask, L., & Kreegipuu, K. (2023). Maturation of the mismatch response in pre-school children: Systematic literature review and meta-analysis. Neuroscience & Biobehavioral Reviews, 153, 105366. https://doi.org/10.1016/j.neubiorev.2023.105366 Themas, L., Nickel, E. S., Lippus, P., Padrik, M., & Kreegipuu, K. (2025a). Exploring pitch and length perception: An EEG study on quantity discrimination in preschoolers with developmental language disorder. Acta Psychologica , 258 , 105157. https://doi.org/10.1016/j.actpsy.2025.105157 Themas, L., Schaadt, G., Lippus, P., Padrik, M., & Kreegipuu, K. (2025b). Maturation of Auditory Pitch and Duration Perception in Children with DLD: Insights from a Longitudinal Study [Preregistration]. OSF. https://doi.org/10.17605/OSF.IO/PCW98 Torppa, R., Kuuluvainen, S., & Lipsanen, J. (2022). The development of cortical processing of speech differs between children with cochlear implants and normal hearing and changes with parental singing. Frontiers in Neuroscience, 16, 976767. https://doi.org/10.3389/fnins.2022.976767 Uwer, R., Albrecht, R., & Suchodoletz, W. (2002). Automatic processing of tones and speech stimuli in children with specific language impairment. Developmental Medicine & Child Neurology, 44(8), 527–532. https://doi.org/10.1111/j.1469-8749.2002.tb00324.x Vandewalle, E., Boets, B., Ghesquière, P., & Zink, I. (2012). Auditory processing and speech perception in children with specific language impairment: Relations with oral language and literacy skills. Research in Developmental Disabilities, 33(2), 635–644. https://doi.org/10.1016/j.ridd.2011.11.005 Virtala, P., Putkinen, V., Kailaheimo-Lönnqvist, L., Thiede, A., Partanen, E., & Kujala, T. (2022). Infancy and early childhood maturation of neural auditory change detection and its associations to familial dyslexia risk. Clinical Neurophysiology, 137, 159–176. https://doi.org/10.1016/j.clinph.2022.03.005 Weber, C., Hahne, A., Friedrich, M., & Friederici, A. D. (2005). Reduced stress pattern discrimination in 5-month-olds as a marker of risk for later language impairment: Neurophysiologial evidence. Cognitive Brain Research , 25 (1), 180–187. https://doi.org/10.1016/j.cogbrainres.2005.05.007 Wechsler, D. (2013). Wechsler Preschool & Primary Scale of Intelligence—Fourth UK Edition (WPPSI-IV UK). Werwach, A., Männel, C., Obrig, H., Friederici, A. D., & Schaadt, G. (2022). Longitudinal trajectories of electrophysiological mismatch responses in infant speech discrimination differ across speech features. Developmental Cognitive Neuroscience, 56, 101127. https://doi.org/10.1016/j.dcn.2022.101127 Wickham, H. (2016). ggplot2: Elegant graphics for data analysis . Springer-Verlag. https://ggplot2.tidyverse.org Tables Table 1. Characteristics of the standard and two deviant stimuli. standard ST Q2 460 270 200 durQ1 DEV 1 Q1 405 190 200 pitchdurQ3 DEV 2 Q3 458 316 150 Note. Stimulus name – names that the stimuli will be referred to onward; Quant. degree – quantity degree; Q1 – first quantity degree; Q2 – second quantity degree; Q3 – third quantity degree. Total duration – total duration of the stimulus in milliseconds; 1st syllable duration – the duration of the first syllable in the stimulus in milliseconds; Pitch fall – the approximate start of the fall of the pitch contour after stimulus onset in milliseconds. Table 2. Means, minimum values, maximum values and standard deviation of the usable epochs by group and sex at T1 and T2. T1 T2 T1 T2 T1 T2 T1 T2 DLD 578 578 11 125 734 718 176 146 TD 578 578 417 273 706 729 73 133 Male 610 613 11 349 734 729 137 118 Female 593 558 70 125 704 725 136 159 Tabel 3. Sample statistics. T1 N = 23 N = 25 Mode Mode (χ², df) p Sex male (N = 13) male (N =14) (0, 1) 1 Mat. edu. sec. (N = 8) high. (N = 22) (18, 3) .00036 Hand. right (N = 22) right (N = 22) (7.2 × 10⁻³², 1) 1 Mean SD Mean SD (W) p Age 69 6.0 69 6.2 (306) .92 Lang. test 20 6.9 36 4.7 (535) 3.0 × 10⁻⁸ Nonverbal IQ 96 8.3 108 13 (421) .0019 T2 N = 22 N = 25 Mode Mode (χ², df) p Sex male (N = 12) male (N = 14) (0, 1) 1 Mat. edu. sec. (N = 8) high. (N = 22) (16, 3) .00096 Hand. right (N = 21) right (N = 22) (0.15, 1) .070 Mean SD Mean SD (W) p Age 82 6.0 81 5.8 (453) .64 Note. Mat. edu. - maternal education level: sec. – secondary education, high. – higher education; Hand. – handedness; Age is reported in months; Lang. test at T1- the scaled overall score of the language test; Nonverbal IQ at T1 - the overall scaled score of the WPPSI-IV UK nonverbal IQ score. Table 4. Means and standard deviations of the amplitudes of the MMRs at T1 and T2. (r, df) p Mean SD Mean SD T1 T2 T1 T2 T1 T2 T1 T2 Q1 (-0.10, 44) .50 1.01 -0.83 1.94 0.96 1.34 -1.4 1.56 0.86 Q3 (-0.0073, 44) .96 -0.56 -1.4 2.28 2.1 -0.18 -1.0 1.48 1.7 Note. Q1 – MMR in the duration change condition (durQ1). Q3 – MMR in the duration and fundamental frequency contour change (pitchdurQ3). The values are presented in microvolts (µV). Pearsons’s product-moment correlation was calculated between the mean values of MMR-s at T1 and T2. Table 5. The results of the MMR linear mixed model of the duration deviant condition (durQ1). Intercept 3.2 1.3 52 2.5 -0.16, 0.64 .016* TW 2 0.59 0.12 92 4.8 0.20, 0.47 5.3 × 10⁻⁶ *** Time 2 -0.95 0.42 50 -2.3 -1.0, -0.07 .028* Group TD 0.60 0.48 46 1.2 -0.20, 0.87 .22 Age 1 -0.04 0.02 46 -2.2 -0.25, -0.02 .030* TW 2: Time 2 -0.94 0.17 92 -5.4 -0.72, -0.34 4.7 × 10⁻⁷ *** Time 2: Group TD -1.2 0.56 46 -2.2 -1.3, -0.08 .031* Significance Codes: *** p < 0.001, ** p < 0.01, * p < 0.05, . p < 0.1, p ≥ 0.1 Note. TW – time-window 1 and 2; Time 1 and 2; Group – DLD (developmental language disorder) and TD (typically developing); Age 1 – age at T1 and reported in months. The intercept is TW1, Time 1 and Group DLD. Table 6. The results of the MMR linear mixed model of the duration and pitch deviant condition. Intercept -0.26 0.29 52 -0.89 -0.10, 0.50 .38 TW 2 -0.01 0.14 92 -0.05 -0.91, -0.09 .96 Time 2 -0.95 0.40 53 -2.4 -0.15, 0.14 .020* TW 2 : Time2 0.41 0.20 92 2.1 0.01, 0.43 .043* Significance Codes: *** p < 0.001, ** p < 0.01, * p < 0.05, . p < 0.1, p ≥ 0.1 Note. TW – time-window 1 and 2; Time 1 and 2. The intercept is TW 1 and Time 1. Tabel 7. Means and group differences in d’ and RT values. Sum of errors in discriminating quantity contrasts and group differences at T1 and T2. Quant. diff. T1 T2 T1 T2 T1 (W) p T2 (W) p d’ 1.7 2.3 2.6 3.4 (427) .0012 (346) .00046 RT 1697 1136 1555 1229 (187) .93 (206) .14 Q1 vs Q2 100 45 45 22 (96) .00014 (169) .019 Q1 vs Q3 42 33 17 4 (182) .035 (72) 2.2 × 10⁻⁶ Q2 vs Q3 111 55 76 36 (163) .016 (192) .070 Lex. des. T1 T2 T1 T2 T1 (W) p T2 (W) p d’ 1.6 2.3 2.6 3.0 (403) .0067 (430) .00089 RT 1738 1582 1590 1492 (201) .12 (237) .43 Q1 vs Q2 42 22 10 8 (128) .00071 (160) .0061 Q1 vs Q3 10 7 2 3 (207) .034 (202) 0.18 Q2 vs Q3 51 45 26 16 (161) .012 (146) 0.0040 Note. Quant. diff. - quantity discrimination task; Lex. des. - lexical decision task; Group diff. - group differences. Figures Figure 1. The stages of the longitudinal study. Sampling phase - participant selection; N – number of participants; T1 – experimental measurement at time one; T2 – experimental measurement at time two. Figure 2. Structure of the stimulus presentation. ST – standard stimulus; DEV – deviant stimulus. Figure 3. The wave form, spectrogram, pitch contour (blue line on spectrogram) and transcription of the stimuli. Figure 4. ERPs to the standard and deviant stimuli at T1 and T2 separately for children with and without DLD. Depicted are the mean grand averages and their corresponding confidence intervals across all electrodes showing significant effects in the cluster-based permutation test. The gray-shaded areas indicate the time windows of statistically significant differences between ERPs to the standard and deviant stimuli in the durQ1 condition; while in the pitchdurQ3 condition, it denotes the location of the time-window analyzed in the exploratory analysis. For visualization purposes, the waveforms were smoothed using a 15-point moving average (equivalent to 30 ms at 500 Hz sampling rate). Figure 5. Difference waves (deviant – standard) at T1 and T2 separately for children with and without DLD. Depicted are the grand averages of the differences waves across all electrodes showing significant effects in the cluster-based permutation test. The gray-shaded areas indicate the time window of statistically significant differences between ERPs to the standard and deviant stimuli in the durQ1 condition; while in the pitchdurQ3 condition, it denotes the location of the time-window analyzed in the exploratory analysis. Figure 6. Contrasts of the responses to the deviant and standard stimuli based on the cluster-based permutation tests across both groups (children with and without DLD). White asterisks on the topographic plots indicate significant differences between the response to the standard and deviant stimuli. Figure 7 . Changes in Individual MMR Values from T1 to T2 with Mean Estimates from the LMM. Dots - individual raw MMR amplitude values at T1 and at T2. Thin lines - connect each participant’s mean amplitude at T1 to their corresponding value at T2. Thick line – connects the mean LMM estimate of the MMR value at T1 to the corresponding estimate at T2. Figure 8 . Changes in Individual MMR Values from T1 to T2 with Mean Estimates from the LMM. Dots - individual raw MMR amplitude values at T1 and at T2. Thin lines - connect each participant’s mean amplitude at T1 to their corresponding value at T2. Thick line – connects the mean LMM estimate of the MMR value at T1 to the corresponding estimate at T2. List of Figure Legends Figure 1. The stages of the longitudinal study. Sampling phase - participant selection; N – number of participants; T1 – experimental measurement at time one; T2 – experimental measurement at time two. Figure 2 . Structure of the stimulus presentation. ST – standard stimulus; DEV – deviant stimulus. Figure 3 . The wave form, spectrogram, pitch contour (blue line on spectrogram) and transcription of the stimuli. Figure 4. ERPs to the standard and deviant stimuli at T1 and T2 separately for children with and without DLD. Depicted are the mean grand averages and their corresponding confidence intervals across all electrodes showing significant effects in the cluster-based permutation test. The gray-shaded areas indicate the time windows of statistically significant differences between ERPs to the standard and deviant stimuli in the durQ1 condition; while in the pitchdurQ3 condition, it denotes the location of the time-window analyzed in the exploratory analysis. For visualization purposes, the waveforms were smoothed using a 15-point moving average (equivalent to 30 ms at 500 Hz sampling rate). Figure 5. Difference waves (deviant – standard) at T1 and T2 separately for children with and without DLD. Depicted are the grand averages of the differences waves across all electrodes showing significant effects in the cluster-based permutation test. The gray-shaded areas indicate the time window of statistically significant differences between ERPs to the standard and deviant stimuli in the durQ1 condition; while in the pitchdurQ3 condition, it denotes the location of the time-window analyzed in the exploratory analysis. Figure 6. Contrasts of the responses to the deviant and standard stimuli based on the cluster-based permutation tests across both groups (children with and without DLD). White asterisks on the topographic plots indicate significant differences between the response to the standard and deviant stimuli. Figure 7 . Changes in Individual MMR Values from T1 to T2 with Mean Estimates from the LMM. Dots - individual raw MMR amplitude values at T1 and at T2. Thin lines - connect each participant’s mean amplitude at T1 to their corresponding value at T2. Thick line – connects the mean LMM estimate of the MMR value at T1 to the corresponding estimate at T2. Figure 8 . Changes in Individual MMR Values from T1 to T2 with Mean Estimates from the LMM. Dots - individual raw MMR amplitude values at T1 and at T2. Thin lines - connect each participant’s mean amplitude at T1 to their corresponding value at T2. Thick line – connects the mean LMM estimate of the MMR value at T1 to the corresponding estimate at T2. Appendices Appendix A A.1.The task and stimuli The quantity discrimination task consisted of three blocks in which the stimuli alternated between three words. All of the words chosen had a meaning and were widely used in Estonian language, thus probably familiar to children. The recording and preparing the stimuli was done the same way as for the neurophysiological experiment. The stimuli in the three blocks are as follows: Block 1: vala (Q1, ‘pour! imp. 2sg’, duration 333 ms), vaala (Q2, ‘whale, gen. sg’, duration 421 ms), vaala (Q3, ‘whale, part.sg’, duration 453 ms). Block 2: keda (Q1, ‘who, part. sg’, duration 395 ms), keeda (Q2, ‘boil, imp. 2sg’, duration 445 ms), keeda (Q3, ‘to boil, infinitive’, duration 491 ms). Block 3: pole (Q1, ‘isn’t’, duration 402 ms), poole (Q2, ‘half, gen. sg’, duration 465 ms), poole (Q3, ‘towards’, duration 483 ms). ISI is 1000 ms, and altogether there are approximately 105 stimuli. The duration of the task was 4 and a half minutes. A.2.Procedure of the quantity discrimination task Prior to the research task, a two-step teaching phase for the subjects is conducted, following the procedure of studies by Crosbie et al., (2004) and Edwards & Lahey, (1996). Firstly, the researcher presented the stimuli verbally. They instructed the child to listen to the same word four times and then confirmed that it was indeed the same word. Following this, the researcher informed the child that a different word would be presented. The initial word was repeated four times again, and then a fifth word was introduced, differing in quantity from the previous ones. The researcher then asked the child whether they had heard a different word and explained that the child should press the button in the upcoming task if they heard a different word. The practicing began and the researcher provided feedback on the child’s performance. There was no set duration for the first teaching phase. The researcher decides when the child responded correctly and consistently enough to move on to the next stage. Secondly, the teaching occurred using the same computer program would be used to collect results for the research. A traffic light with a red signal appeared on the screen, and when the light turned green, the training phase begone. Next, a fixation cross appeared on the screen, materializing the audible auditory stimulus and helping to maintain the subject’s attention on the task. 100 ms after the cross appeared, the child heard a word from headphones. Then, the next cross appeared on the screen, and the child heard the next word, which may be the same as or different from the previous one. At this stage, the computer program provided visual feedback on the correctness of the answers, appearing on the screen after pressing the button. After a correct button press the child saw a smiling face and after an incorrect button press, they saw a sad face. The researcher heard the same stimuli as the child and helps them to press the button at the right time giving extra feedback. Similarly to the previous stage, there was no predetermined number of practice times. The researcher decided when the subject is ready to move on to the research phase. Lastly, the research stage followed, which was conducted similarly to the second teaching phase. During this stage, the child was encouraged to complete the task, and no other help was given. Different auditory stimuli were used in this stage compared to the two teaching phases. Part of the task process is presented in Figure S1. Figure A1. Crosbie, S. L., Howard, D., & Dodd, B. (2004). Auditory lexical decisions in children with specific language impairment. British Journal of Developmental Psychology , 22 , 103–121. Edwards, J., & Lahey, M. (1996). Auditory Lexical Decisions of Children With Specific Language Impairment. Journal of Speech and Hearing Research , 39 , 1263–1273. Appendix B B.1. The task and stimuli In a classical auditory lexical decision task, the subject hears either a real word or a pseudoword. A pseudoword has no meaning in the specific language but complies with the phonotactic rules of the language. The subject’s task is to press the ”yes” button if they hear a real word and press the ”no” button if they hear a pseudoword. We added some adaptations to this classical paradigm. Firstly, following the example of the study by Friedrich & Friederici (2006) visual material was added to the experiment. A picture appeared on the screen, and after 300 ms, a word was played through headphones. The auditory stimulus may have matched the image if it was a real word, or not match if it was a pseudoword. The child was required to press a button if the word and picture match. Secondly, the method of pressing two different buttons (a ”yes” and a ”no” button) was replaced with pressing only one button: the ”yes” button (known as the go/no-go procedure). Moret-Tatay & Perea (2011) have highlighted several problems with the methodology of pressing two buttons. Specifically, classical lexical decision task is difficult for even the TD children, them making many mistakes and have longer reaction times. This may be because children must remember which button to press for real words and which for pseudowords during the task. However, by using the method of pressing only one button, the number of mistakes decreases, and the reaction time shortens. This has been found in studies of both adults (Perea et al., 2002, 2003) and children (Moret-Tatay & Perea, 2011). At the same time, the methodology of pressing one button is as sensitive for investigating the recognition of real words and pseudowords as the methodology of pressing two buttons (Perea et al., 2002, 2003, 2010). The stimuli for the lexical decision task consisted of 15 nouns that are familiar to children and have the same syllable structure (CV(::)CVC), with five words belonging to the same quantity category. Pseudowords, where the meaning of the word is lost due to a change in the quantity, were created for the selected stimuli by a native speaker pronouncing these words in a wrong quantity degree. Pairs of pseudowords were formed for the real words so that contrasts of all quantity degrees were represented. For instance, a real word with Q1 was paired with a pseudoword in Q2, and another real word in Q1 was paired with a pseudoword in Q3. Similar to the stimuli for the neurophysiological experiment and the first behavioural task, the change in quantity followed a vowel-centric pattern. Each stimulus word was heard by the participant once in the study, but each picture was seen twice, because a real word is presented with the same picture once and a pseudoword another time. B.2.Procedure of the lexical decision task Before the experimental task, a two-step teaching phase was carried out, modelled after the procedures in studies by Edwards & Lahey (1996) and Crosbie et al. (2004) The instruction given to a child was as follows: “You will hear some words that sound correct and match the picture. You will also hear some words that sound wrong and do not match the picture. Press the button when you hear a word that matches the picture.” Firstly, the researcher presented the words and pseudowords verbally and showed A5-sized pictures. The participant was encouraged to press the space bar, marked with a sticker, on the computer keyboard upon hearing a real word. The researcher provided feedback on the correctness of the responses. There was no set number of words or pseudowords presented to the participant. The researcher decided when the child responded correctly and consistently enough to move on to the next teaching phase. Secondly, the teaching is conducted using the same computer program as for collecting the results for the research. A colourful drawing of a specific object or animal appeared on the screen. 500 ms after the picture appeared, the child heard through headphones a real word that matched the picture or a pseudoword that had no meaning and thus did not match the picture. The child was encouraged to press a button upon hearing a real word. At this stage, the computer program provided visual feedback on the correctness of the responses, which appeared on the screen. After a correct button press the child saw a smiling face and after an incorrect button press, they saw a sad face. The researcher heard the same stimuli as the child and helped them to press the button at the right time giving extra feedback. Similarly to the previous stage, there was no fixed number of practice trials. The researcher decided when the participant is ready to move on to the research phase. Then the research stage followed, conducted similarly to the second teaching phase. The child was encouraged to complete the task, and no other help was provided. This stage used different pictures and auditory stimuli compared to the two teaching phases. The words and pseudowords were presented in a random order with 1000 ms ISI, but it was ensured that a word and its pseudoword form are not immediately following one another and vice versa. Part of the task process is presented in Figure B1. Figure B1. “w” denotes a real word and “pw” a pseudoword. Crosbie, S. L., Howard, D., & Dodd, B. (2004). Auditory lexical decisions in children with specific language impairment. British Journal of Developmental Psychology , 22 , 103–121. Edwards, J., & Lahey, M. (1996). Auditory Lexical Decisions of Children With Specific Language Impairment. Journal of Speech and Hearing Research , 39 , 1263–1273. Friedrich, M., & Friederici, A. D. (2006). Early N400 development and later language acquisition. Psychophysiology , 43 (1), 1–12. https://doi.org/10.1111/j.1469-8986.2006.00381.x Moret-Tatay, C., & Perea, M. (2011). Is the go/no-go lexical decision task preferable to the yes/no task with developing readers? Journal of Experimental Child Psychology , 110 (1), 125–132. https://doi.org/DOI: 10.1016/j.jecp.2011.04.005 Perea, M., Gómez, P., & Fraga, I. (2010). Masked nonword repetition effects in yes/no and go/no-go lexical decision: A test of the evidence accumulation and deadline accounts. Psychonomic Bulletin & Review , 17 (3), 369–374. https://doi.org/10.3758/PBR.17.3.369 Perea, M., Rosa, E., & Gómez, C. (2002). Is the go/no-go lexical decision task an alternative to the yes/no lexical decision task? Memory & Cognition , 30 (1), 34–45. https://doi.org/10.3758/BF03195263 Perea, M., Rosa, E., & Gómez, C. (2003). Influence of neighborhood size and exposure duration on visual-word recognition: Evidence with the yes/no and the go/no-go lexical decision tasks. Perception & Psychophysics , 65 (2), 273–286. https://doi.org/10.3758/BF03194799 Appendix C C.1.Detailed description of the MADE pipeline Here we present a detailed description of the EEG preprocessing steps using the MADE pipeline. As a first step to enhance interpretability, data was down-sampled from 512 to 500 Hz, following high (0.3 Hz, stopband of 0.1 Hz) and low (50 Hz, transition band of 10 Hz) pass filtering. Channels containing artifacts were identified and removed (later interpolated) using the FASTER plugin for EEGLAB (Nolan et al., 2010), if their Z-score exceeded 3. Participants with more than 10% of bad channels were excluded. To remove muscle and eye-movement artifacts, independent component analysis (ICA) was performed on a duplicate dataset high pass filtered at 1 Hz. After channels, containing excessive EMG or unusual high/low amplitudes for more than 20% of the recording were removed, calculated independent components were identified, using the Adjusted-ADJUST algorithm (Leach et al., 2020), saved, and applied to the data bandpass filtered at 0.3 to 50 Hz. The cleaned EEG dataset was then segmented into epochs (–100 to 698 ms), baseline corrected (–100 to 0 ms) and looped through to detect any residual artifacts exceeding a predetermined threshold of ±125 mV. The epochs, in which a non-ocular channel exceeded the voltage threshold, were interpolated using spherical spline interpolation. If more than 10% of the non-ocular channels exceeded the voltage threshold, the epoch was rejected completely. Before the final stage of the pipeline, the channels that were removed using FASTER at the beginning of the pipeline were interpolated using spherical spline interpolation. Lastly, the data was re-referenced using an average reference. Appendix D D.1.Detailed description of the cluster-based permutation tests procedure As a first step the within-subject deviant and standard conditions were compared with dependent sample t-tests. A t-value was computed for each time-point in each electrode in the observed data (our actual data). The datapoint comparisons (deviants and standard conditions) had to exceed a critical t-value, determined during the permutation process, to be selected for forming a cluster. Next, the two largest clusters were quantified using the sum of the t-values within this cluster (the cluster statistic). Then the condition labels were randomly shuffled across within-subject time-points and the same analysis that was done with the observed data (described above) was done with the shuffled data. This was repeated on 10,000 random shuffles employing the Monte Carlo method. Then, the largest cluster (positive and negative) in the observed data was compared with the largest clusters found in each random shuffle. The frequency of instances when the sum of t-values in the cluster in the observed data was larger than the sum of t-values in the clusters in the random shuffles was calculated. When the proportion of these instances was below the critical level (α ≤ 0.025, adjusting for a two-sided test), responses to deviant and standard stimuli were considered to be significantly different. The effect size was estimated by circumscribing a rectangular shape fitting tightly around a cluster and then calculating the Cohen’s d of the averaged data within this shape. Figure 1D. Contrasts of the responses to the deviant and standard stimuli based on the cluster-based permutation tests across both groups (children with and without DLD) in durQ1 condition at T1. White asterisks on the topographic plots indicate significant differences between the response to the standard and deviant stimuli. Figure 2D. Contrasts of the responses to the deviant and standard stimuli based on the cluster-based permutation tests across both groups (children with and without DLD) in pitchdurQ3 condition at T1. White asterisks on the topographic plots indicate significant differences between the response to the standard and deviant stimuli. Figure 3D. Contrasts of the responses to the deviant and standard stimuli based on the cluster-based permutation tests across both groups (children with and without DLD) in durQ1 condition at T2. White asterisks on the topographic plots indicate significant differences between the response to the standard and deviant stimuli. ‘ Figure 4D. Contrasts of the responses to the deviant and standard stimuli based on the cluster-based permutation tests across both groups (children with and without DLD) in pitchdurQ3 condition at T2. White asterisks on the topographic plots indicate significant differences between the response to the standard and deviant stimuli. Appendix E E.1.Detailed description of the linear mixed models’ diagnostics 1) Potential multicollinearity in the models, were examined with pair-wise correlations between all variables in the model to ensure that no correlation was excessively high (0.8; between continuous variables – Spearman’s rank correlation; between continuous and categorical variables – point-biserial correlation; between categorical variables Cramér’s V). 2) Potential non-linearity between the dependent and independent variables was assessed by plotting residuals against the fitted values of the optimal model and examining the LOESS line for any systematic patterns or deviations from a flat trend, which would indicate non-linearity. The latter was also done for each predictor separately. Based on visual inspection, no significant non-linearity was detected. 3) Potential heteroscedasticity in the model’s residual variance was assessed by plotting the residuals against the fitted values of the optimal model and examining the pattern for any signs of non-constant variance, which would indicate heteroscedasticity. The latter was also done for each predictor separately. 4) The Shapiro-Wilk test was used to test for the model’s residuals, random effects, and random intercepts normal distribution. In addition, Q-Q plots and histograms were examined. 5) We calculated Cook’s distance for each observation, to identify influential data points that could disproportionately affect the model’s results. Appendix F F.1.Results of the cluster-based permutation tests in the durQ1 condition – additional material At T1, an increased negativity in response to the deviant stimulus was observed at occipital and parietal sites (O1, Oz, O2, PO3, PO4, P7, P8, Pz, P4) between 290–340 ms post-stimulus. This was followed by a sustained positive shift at frontal sites (F7 and F8) between 340 to 690 ms after stimulus onset. Concurrently, a more negative deflection emerged in the deviant condition at a few centro-parietal electrodes (P4, CP1, CP2), culminating in a broader right-lateralized cluster (FC2, CP2, P4, CP1) between 590 and 690 ms. At T2 two smaller clusters were found: 1) 240-340 ms after stimulus onset at Pz and P2; 2) 540-690 ms after stimulus onset at F4, C4, FC2. F.2.Results of the LMM in the durQ1 condition – additional material Additionaly to what was reported in the article subsection 4.3.1. the optimal LLM revealed the main effect of group status at T1 not to be significant. The standardized beta coefficient showed a medium positive effect size ( β = 0.34). The model showed the main effect of time window at T1 in the DLD group, with the second time window showing significantly more positive amplitudes. The confidence intervals (CIs; Table 4) for this effect were relatively narrow, indicating a precise result with medium effect size ( β = 0.33). The main effect of age was significant, showing more negative difference waves in older children. The narrow CIs close to 0 suggests that the true effect is precise but likely small. Indeed, the standardized beta coefficients showed a small negative effect size ( β = -0.14). The interaction of time window and time showed that the decrease of the amplitude was more pronounced at the second time-window. Based on the CIs this effect is precise and large ( β = -0.53). F.3.Results of the LMM in the pitchdurQ3 condition – additional material Additionally, to what was reported in the article subsection 4.3.2. we found a significant time:time-window interaction, which shows that the decline in amplitude is less pronounced in the second time-window. The confidence intervals are relatively narrow, indicating that this is a precise and reliable effect with a small effect size ( β = 0.22). Supplementary Material File (figures.zip) Download 20.24 MB Information & Authors Information Version history V1 Version 1 28 July 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords child development eeg erp mmn speech perception Authors Affiliations Liis Themas 0000-0002-7615-0895 [email protected] University of Tartu View all articles by this author Gesa Schaadt Freie Universität Berlin View all articles by this author Pärtel Lippus University of Tartu View all articles by this author Marika Padrik University of Tartu View all articles by this author Liis Kask University of Tartu View all articles by this author Ulvi Vaher University of Tartu View all articles by this author Mairi Männamaa University of Tartu View all articles by this author Kairi Kreegipuu University of Tartu View all articles by this author Metrics & Citations Metrics Article Usage 282 views 186 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Liis Themas, Gesa Schaadt, Pärtel Lippus, et al. Auditory Mismatch Response to Pitch and Duration Changes in Children with Developmental Language Disorder: A Longitudinal Approach. Authorea . 28 July 2025. DOI: https://doi.org/10.22541/au.175367725.54753416/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.175367725.54753416/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9ffc3e527f0f1b23',t:'MTc3OTQ1NzMzOQ=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();