Elsevier

Journal of Physiology-Paris

Volume 102, Issues 1–3, January–May 2008, Pages 120-129
Journal of Physiology-Paris

Rhythmic processing in children with developmental dyslexia: Auditory and motor rhythms link to reading and spelling

https://doi.org/10.1016/j.jphysparis.2008.03.007Get rights and content

Abstract

Potential links between the language and motor systems in the brain have long attracted the interest of developmental psychologists. In this paper, we investigate a link often observed (e.g., [Wolff, P.H., 2002. Timing precision and rhythm in developmental dyslexia. Reading and Writing, 15 (1), 179–206.] between motor tapping and written language skills. We measure rhythmic finger tapping (paced by a metronome beat versus unpaced) and motor dexterity, phonological and auditory processing in 10-year old children, some of whom had a diagnosis of developmental dyslexia. We report links between paced motor tapping, auditory rhythmic processing and written language development. Motor dexterity does not explain these relationships. In regression analyses, paced finger tapping explained unique variance in reading and spelling. An interpretation based on the importance of rhythmic timing for both motor skills and language development is proposed.

Introduction

Children with language problems often have motor problems. In a review of the literature, Hill (2001) found that most children with a diagnosis of specific language impairment also had a diagnosis of developmental coordination disorder, namely movement difficulties out of proportion with their general development and intelligence. Children with written language problems can also show motor problems (Kaplan et al., 1998). Summarising his systematic research programme spanning over 20 years, Wolff (2002) noted that children with developmental dyslexia show motor sequencing difficulties when asked to tap to a metronome beat. Dyslexic students anticipate the beat by intervals 2–3 times as long as controls, showing atypical anticipation intervals both when tapping with their preferred index finger and when tapping bimanually. These children also show difficulty in recalibrating their timed responses when the metronome rate changes, and in reproducing the correct serial order of simple rhythmic patterns at much slower rates. Children with diagnosed speech and language impairments (SLI) also show difficulties in calibrating motor responses to a beat. Corriveau and Goswami (in press) asked children with SLI to tap in time with 3 different metronome rhythms, 1.5, 2 and 2.5 Hz. They found that the children with SLI were less accurate than both age-matched controls and younger language-matched controls at the slower rates of 1.5 Hz and 2 Hz. Corriveau and Goswami proposed that the children might have subtle neural impairments that affected both motor and language development. They suggested that a plausible candidate mechanism was that serving the perception and expression of rhythm and timing.

In fact, all children anticipate the beat when tapping in time with a metronome (McAuley et al., 2006). McAuley and colleagues studied 88 children aged 4–12 years, and measured their ability to tap in time to isochronous beats that varied from a fastest rate of 150 ms to a slowest rate of 1709 ms (logarithmic spacing of 8 rates in total). They reported that the range of accessible tapping rates widened during childhood, with older children able to tap accurately to a wider range of rates. Assessment of adults using the same paradigm showed that the range of accessible tapping rates narrowed again in old age: Very young children (<8 years) and older adults (>75 years) found it very difficult to tap at slower rates. The preferred spontaneous tempo between ages 8 and 12 years was an isochronous rate at around 500 ms (2 Hz). In this typically-developing sample of children, the main correlate of timing accuracy was non-verbal I.Q. (language data were not reported). However, in the study by Corriveau and Goswami (in press), language data were collected. Corriveau and Goswami found that composite measures of inter-tap interval variability and anticipation of the metronome beat accounted for unique variance in language, and also in phonology, reading and spelling in their sample of children with SLI. Inter-tap variability accounted for up to 13% of unique variance in phonological processing and 8% of unique variance in language development after controlling for non-verbal I.Q. Auditory rhythmic timing deficits had been observed in an earlier study of the same SLI population (Corriveau et al., 2007), and individual differences in the perception of auditory rhythmic timing cues were also strongly predictive of language, phonology, reading and spelling in this sample, accounting for up to 31% of unique variance in phonological processing and 20% of unique variance in language development.

There have been a number of theoretical proposals for why motor tapping should correlate with measures of literacy and language. Llinas (1993) proposed a physiological explanation based on a clocking or timing mechanism that mediated between central nervous system function and co-ordinated behaviour. He suggested that dyslexic children might experience dysfunction of this mechanism within a restricted time window, 40 Hz rhythmicity, which was suggested to interfere with the perception of stimuli presented at rapid rates (Llinas, 1993). Wolff (2002) argued that his data ruled out a physiological “dyschronia” mechanism, because anticipation times changed when both fingers were used for tapping, and because his dyslexic participants experienced inordinate difficulties at slower rates. He proposed instead that children with developmental dyslexia suffered from a temporal information processing deficit in reaction-anticipation transitions. He suggested that their perception–action system was disturbed, so that the latency before participants were aware of having moved their finger was much longer than normal, and therefore much more out of synchrony with the latency for the initiation of the movement itself than is typically found (Castiello et al., 1991). Here we explore the hypothesis that there is a developmental connection between auditory rhythmic timing, the accuracy of motor tapping and language and literacy because the perception and production of structured rhythmic and temporal patterns is a crucial part of language acquisition. This proposal is supported by a variety of developmental data.

For example, babbling is a core activity in language acquisition, and babbling is a rhythmic activity. Vocal babbling follows the rhythm, timing and stress patterns of natural language prosody. If adults are played tape recordings of French, Cantonese and Arabic babies babbling, they successfully identify the “language” that is being babbled on the basis of rhythmic prosodic cues (de Boysson-Bardies et al., 1984). Human newborns are very sensitive to perceptual rhythmic cues, and can identify human languages from different rhythmic classes (such as Dutch versus Japanese) at 1–4 days of age (Mehler et al., 1986). This appears to be an evolutionarily-driven skill, as rats can also distinguish languages like Dutch and Japanese on the basis of prosodic cues (Toro et al., 2003), as can cotton-top tamarins (Ramus et al., 2000). Regarding production, apes produce calls with pure tonal notes, repetition, rhythm and phrasing which are like the multi-syllabic babbling sounds produced by babies (ape “singing”, Masataka, 2007). Adult gibbons can produce well-coordinated duets, characterised by a vocal matching absent in juveniles, suggesting that the perception and production of such vocalizations are inter-dependent (as in humans). Masataka (2007) proposes that a shared mechanism for representing the rhythmic characteristics of the input might underlie both human linguistic development and ape signing.

Further evidence that the perception and production of structured rhythmic and temporal patterns is crucial for language acquisition comes from evidence that deaf babies “babble” with their hands. Deaf babies born to deaf parents who communicate with them by signing duplicate the rhythmic timing and stress of hand shapes in natural signs (Petitto et al., 2004). Deaf infants born to hearing parents who talk to them do not show hand babbling, suggesting that this rhythmic motor behaviour is indeed specifically linguistic. In a recent review of studies of early language development in hearing children, Nittrouer (2006) pointed out the developmental importance of acoustic changes that arise from the slow modulations of the vocal tract which are those produced first by infants (de Boysson-Bardies et al., 1986). Noting that listeners can recover linguistic structure when all of the traditional cues to phonemic identity are eliminated from the signal, she argued for the importance of amplitude envelope information in language learning. Temporal segmentation of the continuous acoustic signal at the syllabic level is facilitated by tracking the rate of change of the amplitude envelope at onset, that is, by tracking amplitude envelope rise time.

The possibility that rise time perception is important for language development, particularly phonological development, is supported by a series of studies examining rise time perception in children with developmental dyslexia. Children with developmental dyslexia are characterised by deficits in phonological representation and difficulties with written language (Snowling, 2000, Swan and Goswami, 1997). Rise time is important for phonological representation because it carries information about syllable structure and about the vowel in any syllable (Scott, 1998; the vowel is the syllabic nucleus). Periodicity in speech is related to the onsets of the vowels in stressed syllables (Allen et al., 1972). When we intentionally produce rhythmic speech, we time our motor production of the rise time of the vowels in stressed syllables (Scott, 1998). It has been suggested that the “phonological grammar” of a particular language is built upon the characteristic rhythms of the language at the time scale of syllables (Port, 2003). Individual differences in phonological processing indeed seem to be linked to individual differences in the auditory perception of rise time (Goswami et al., 2002, Muneaux et al., 2004, Richardson et al., 2004).

For example, Goswami et al. (2002) found that children with developmental dyslexia (and poor phonological skills) were significantly less sensitive to amplitude envelope rise time compared to age-matched controls, whereas precocious readers (with superior phonological skills) were significantly more sensitive to rise time compared to their age-matched controls. Rise time sensitivity was assessed through a “beat detection” task in which sequences of amplitude-modulated tones had either sharp rise times (e.g., 15 ms, yielding a strong perception of a beat), or slower rise times (up to 300 ms). A further study with a different sample of dyslexic children assessed rise time discrimination thresholds (Richardson et al., 2004). Children were either presented with two amplitude modulated tones and asked which sound had the sharper beat (shorter rise time, 2IFC task), or were presented with three amplitude modulated tones and asked to select the tone that matched the standard (AXB task). Group differences in rise time sensitivity between dyslexic children and their age-matched peers were found for both measures, and both measures predicted unique variance in measures of phonological processing and reading. A measure of sensitivity to auditory duration also showed significantly different thresholds in the two groups, whereas a measure of sensitivity to sound intensity did not. As all the auditory measures used the same psychoacoustic procedures, the difficulties in perceiving rise time and duration appear to be specific ones. Duration is also an important cue to rhythm.

These studies with children suggest that the perception of auditory cues to temporal organisation at the syllable level of language is important for phonological development. Regarding the motor production of structured rhythmic patterns, Thomson et al. (2006) investigated timed finger tapping in adults with developmental dyslexia. They used the 2IFC rise time task from Richardson et al. (2004) with a university student dyslexic population, and added a range of motor tasks comprising both paced and unpaced motor tapping at 3 different rates plus a measure of motor dexterity (a peg moving task). Additional auditory threshold tasks were also administered assessing duration discrimination and intensity discrimination, with the latter not expected to be impaired. Thomson et al. (2006) reported that the students with developmental dyslexia showed greater inter-tap interval (ITI) variability compared to controls at rates of both 1.5 and 2 Hz when tapping to a metronome beat. In zero-order correlations, metronome ITI variability in the presence of a cued beat was associated with measures of reading, spelling, phonological awareness (specifically, phoneme deletion), rise time sensitivity and duration discrimination. ITI variability in the absence of a beat was associated with auditory rise time sensitivity, phonological short-term memory and paced ITI variability. No group differences were found on the pegboard task, which tested manual motor skills in the absence of a rhythmic component. Partial correlations controlling for non-verbal I.Q. showed that ITI variability in synchronised tapping was related to reading development in this adult sample, and also to duration perception. Unpaced tapping was related to rise time perception and to digit span, but not to literacy.

Given the novel correlational evidence for associations between auditory and motor rhythmic timing and phonological and literacy skills found in this sample of young adults, the study reported here attempted to explore these relationships further in a sample of 10-year old children. Some of the children had a diagnosis of developmental dyslexia. A similar battery of tasks was used as in Thomson et al. (2006), to facilitate direct comparison between the adults and children. We hypothesised that relationships between motor skills and language tasks might be stronger earlier in the developmental trajectory. We were also interested in expanding the range of auditory measures used, to see whether the auditory relationships found would be unique to measures of rise time sensitivity and duration. We hence included extra measures of rise time sensitivity, and measures of frequency detection and tempi discrimination.

Section snippets

Participants

Forty-eight 10-year old children were studied. Twenty-five of the children (17 male; mean age 10 years, 8 months, s.d., 15.9 months) had received a diagnosis of specific reading difficulties from a qualified educational psychologist. All children’s first language was English and they attended schools in Southern England. In addition, the children had average or above average verbal and non-verbal intelligence (>80 I.Q. on the WISC III), normal sensory ability and no documented neurological

Phonological tasks

The means and standard deviations for performance on the phonological measures are shown in Table 2. One way ANOVAs to look at potential differences in terms of written language ability were carried out by reading group (Dyslexic, TD), taking each of the phonological measures as a dependent variable. Significant group differences were found for all phonological measures (for F values, significance values and effect sizes, see (Table 2).

Auditory processing tasks

The means and standard deviations for performance on the

Discussion

This study set out to explore the connection between auditory rhythmic and motor timing skills, and language and literacy development in 10-year old children. We were also interested in exploring potential relationships between auditory and motor rhythmic timing skills and phonological processing abilities. Significant differences for both auditory and motor rhythm measures were found in terms of the children’s development of written language skills. When speaking to a rhythm or perceiving

Acknowledgements

We would like to thank the staff and students of the participating schools in East Anglia, UK.

This study was possible through funding from the Economic and Social Research Council (ESRC), Grant RES-000-23-0475, awarded to Usha Goswami.

References (34)

  • C.D. Elliott et al.

    The British ability scales II

    (1996)
  • J.M. Findlay

    Estimates on probability functions: a more virulent PEST

    Perception and Psychophysics

    (1978)
  • U. Goswami et al.

    Amplitude envelope onsets and developmental dyslexia: a new hypothesis

    Proceedings of the National Academy of Sciences of the United States of America

    (2002)
  • E.L. Hill

    Non-specific nature of specific language impairment: a review of the literature with regard to concomitant motor impairments

    International Journal of Language and Communication Disorders

    (2001)
  • R. Llinas

    Is dyslexia a dyschronia?

    Annals of the New York Academy of Sciences

    (1993)
  • N. Masataka

    Music, evolution and language

    Developmental Science

    (2007)
  • Cited by (212)

    View all citing articles on Scopus
    1

    Present Address: Graduate School of Education, Harvard University, Appian Way, Cambridge, MA 02138, USA.

    View full text