doi:10.1016/j.neuropsychologia.2006.11.009 
Copyright © 2006 Elsevier Ltd All rights reserved.
Deviant neurophysiological responses to phonological regularities in speech in dyslexic children
Milene L. Bonte
, a, 
, Hanne Poelmansa and Leo Blomerta
aDepartment of Cognitive Neuroscience, Faculty of Psychology, University of Maastricht, The Netherlands
Received 26 October 2005;
revised 18 October 2006;
accepted 16 November 2006.
Available online 21 December 2006.
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Abstract
Developmental dyslexia is strongly associated with a phonological deficit. Yet, implicit phonological processing (in)capacities in dyslexia remain relatively unexplored. Here we use a neurophysiological response sensitive to experience-dependent auditory memory traces, the mismatch negativity (MMN), to investigate implicit phonological processing of natural speech in dyslexic and normally reading children. In a modified passive oddball design that minimizes the contribution of acoustic processes, we presented non-words that differed by the degree of phonotactic probability, i.e. the distributional frequency of phoneme combinations in a given language. Overall morphology of ERP responses to the non-words indicated comparable processing of acoustic–phonetic stimulus differences in both children groups. Consistent with previous findings in adults, normally reading children showed a significantly stronger MMN response to the non-word with high phonotactic probability (notsel) as compared to the non-word with low phonotactic probability (notkel), suggesting auditory cortical tuning to statistical regularities of phoneme combinations. In contrast, dyslexic children did not show this sensitivity to phonotactic probability. These findings indicate that the phonological problems often reported in dyslexia relate to a subtle deficit in the implicit phonetic–phonological processing of natural speech.
Keywords: Developmental dyslexia; ERP; Mismatch negativity; Phonological deficit; Phonotactic probability
Fig. 1. Pitch and intensity contours of the non-word stimuli used in the experiment: notsel [
] and notkel [
], which have a high versus low phonotactic probability. We used four utterances of each non-word.
Fig. 2. ERP activity elicited by notsel and notkel for both normally reading and dyslexic children. (A) Grand average ERPs elicited by standards of both non-word stimuli. (B) ERP difference waves (notsel–notkel) indicating enhanced ERP responses for notsel versus notkel in both subject groups. Note that auditory stimulus deviation occurred 
160 ms after stimulus onset.
Fig. 3. ERP activity elicited by standard and deviant stimuli for both normally reading and dyslexic children. (A) Grand average ERPs elicited by standard and deviant stimuli indicating mismatch effects for HPP and LPP non-words. (B) ERP difference waveforms of normally reading and dyslexic children for HPP (notsel deviant–notsel standard) and LPP (notkel deviant–notkel standard) non-words. Note that auditory stimulus deviation occurred 160 ms after stimulus onset.
Fig. 4. Mean (SEM) latency and amplitude measures of MMN activity in normal readers and dyslexic children. Peak latency and mean amplitude (50 ms window around the peak latency) were determined individually for each subject. Error bars represent standard error of the mean. Asterisks indicate significant differences between conditions (post hoc t comparisons).
Fig. 5. Correlation between non-word reading scores and individual MMN mean amplitude difference scores (notsel–notkel) for normally reading (grey circles) and dyslexic children (black circles).
Table 1.
Descriptive data for normal readers and dyslexic children
Age, IQ scores (mean (range)) and performance on language tests (mean (SD)).
a All children showed normal or above normal IQ scores.
b Age-appropriate norms (standardized scale; mean = 5,
SD = 2).
c Percentage correct.
Table 2.
Frequency counts of phonemes and phoneme combinations in our non-word stimuli (log-transformed)
Frequency counts were weighted for word frequency and based on the Celex corpus with 42 million Dutch words (Baayen et al., 1995). [··] indicates the phoneme(s) for which counts are given, c = consonant, v = vowel. For example, the third column: cv[cc]vc, gives frequency counts for /ts/ and /tk/ respectively. The frequency counts of the single phonemes /s/ and /k/ in the second column only include syllable-initial occurrences. “–” indicates a frequency of 0.
Table 3.
Mean (range) latency and amplitude characteristics of MMN activity in normally reading and dyslexic children
Peak latency, peak amplitude and mean amplitude (50 ms window around the peak latency) were determined individually for each subject. *p < .05, **p < .01.
Table 4.
Pearson's correlations between MMN difference scores (notsel–notkel) of latency, peak amplitude and mean amplitude (50 ms window around the peak latency), and age, IQ and behavioural measures for all subjects
*p < .05.
a Age-appropriate norms (standardized scale; mean = 5,
SD = 2).
b Percentage correct.
Abstract
Purchase PDF (591 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (5)
Purchase PDF (451 K)
