Elsevier

NeuroImage

Volume 34, Issue 1, 1 January 2007, Pages 416-425
NeuroImage

The processing of prosody: Evidence of interhemispheric specialization at the age of four

https://doi.org/10.1016/j.neuroimage.2006.09.009Get rights and content

Abstract

Beyond its multiple functions in language comprehension and emotional shaping, prosodic cues play a pivotal role for the infant’s amazingly rapid acquisition of language. However, cortical correlates of prosodic processing are largely controversial, even in adults, and functional imaging data in children are sparse. We here use an approach which allows to experimentally determine brain activations correlating to the perception and processing of sentence prosody during childhood. In 4-year-olds, we measured focal brain activation using near-infrared spectroscopy and demonstrate that processing prosody in isolation elicits a larger right fronto-temporal activation whereas a larger left hemispheric activation is elicited by the perception of normal language with full linguistic content. Hypothesized by the dual-pathway-model, the present data provide experimental evidence that in children specific language processes rely on interhemispheric specialization with a left hemispheric dominance for processing segmental (i.e. phonological) and a right hemispheric dominance for processing suprasegmental (i.e. prosodic) information. Generally in accordance with the imaging data reported in adults, our finding underlines the notion that interhemispheric specialization is a continuous process during the development of language.

Introduction

The lateralization of cortical systems to serve language function is the best established and clinically most widely applied neuropsychological asymmetry of the human cortex. While in the left hemisphere processing of syntactic, semantic, and phonological information converges to specific areas, the relevance of the right hemisphere for prosodic processing of spoken language has been postulated and a number of studies have provided evidence for a dynamic interaction between the hemispheres in response to regular speech of full segmental and suprasegmental content (Ross et al., 1997, Baum and Pell, 1999, Grimshaw et al., 2003, Friederici and Alter, 2004).

Beyond phonological, syntactic, and semantic content, prosodic suprasegmental information is required to determine the full meaning of a spoken sentence or phrase. Prosody is characterized by a specific intensity, pitch contour, and duration of spoken words, phrases, or sentences. Notably, prosodic cues can carry both linguistic and emotional information. As an example, pitch contour as a prosodic parameter can determine if someone is asking a question or making a statement but can also convey whether the speaker is appreciating or disapproving of the statement uttered. The present study focuses on the cerebral basis of linguistic prosody, which may be considered more relevant for the acquisition of language due to its potential to chunk the auditory stream. For a review on emotional prosody, see e.g. Baum and Pell (1999).

Prosodic cues, semantic information, and syntactic structures can interact and influence the interpretation of a sentence (e.g. Cutler et al., 1997, Steinhauer et al., 1999, Astesano et al., 2004, Friederici and Alter, 2004, Eckstein and Friederici, 2005, Magne et al., 2005, Frazier et al., 2006). Thus, the understanding of language requires decoding of both segmental (phonological) and suprasegmental (prosodic) information. As compared to the much more devastating loss of semantic and syntactic processing in the classical aphasias, a loss of prosodic abilities may be of lesser importance after a cerebral lesion. However, it has been hypothesized that prosodic, suprasegmental cues may be of supreme relevance during language acquisition in infants (e.g. Gleitman and Wanner, 1982, Jusczyk, 1997).

Infant language acquisition is thought to greatly rely on prosodic cues to segment the incoming auditory stream into linguistic units, thereby also facilitating the acquisition of syntactic structures and the recognition of words in sentences (e.g. Jusczyk et al., 1992, Jusczyk, 1999, Gout et al., 2004). In most languages, infant-directed speech contains an accentuated prosody such as exaggerated intonation, stretching of vowels or longer pauses, all improving the distinctiveness of speech categories. This may serve tuning the perceptual mechanisms in specialized brain areas to language specific requirements aiding the fascinating velocity in which language is acquired (Fernald et al., 1989, Jusczyk, 1997, Kuhl et al., 1997, Kuhl, 2004). The importance of prosodic cues for language acquisition is mirrored in the infant’s early sensitivity to prosodic information as was shown in behavioral as well as in electrophysiological studies (e.g. Mehler et al., 1988, Jusczyk et al., 1993, Jusczyk et al., 1999, Friederici et al., 2002, Thiessen and Saffran, 2003, Weber et al., 2004; for an overview, see Weissenborn and Hohle, 2001, Kuhl, 2004, Friederici, 2005).

Despite its relevance, the functional–anatomical correlates of prosodic processing are controversial. The dynamic dual-pathway-model of auditory language comprehension posits that syntax and semantics are predominantly processed by the left hemisphere, while prosody recruits a more dynamic network. In isolation, it is processed by the right hemisphere, however, the left hemisphere becomes the more engaged the more linguistic the task or the stimulus’ content (Friederici and Alter, 2004). Based on lesion and functional imaging studies, the model suggests that the right hemisphere has a pivotal role for prosodic processing in adults, however, findings are inconsistent with respect to a concise functional–anatomical localization (Baum and Pell, 1999, Friederici and Alter, 2004). A widespread neural network processing linguistic prosody has been demonstrated by functional imaging studies in adults. Meyer and colleagues found a right hemispheric dominance when only suprasegmental information of a sentence is available in the input (Meyer et al., 2002, Meyer et al., 2003). In another study, active processing of low-pass filtered speech material only containing prosodic contour elicited greater right frontal brain activation in adult subjects when compared to normal speech (Plante et al., 2002). Also the processing of high vs. low degrees of prosody in connected speech resulted in greater right hemispheric involvement (Hesling et al., 2005). Furthermore studies in adults have demonstrated that the lateralization of prosodic processing depends on the functional relevance of prosody in the respective language studied. A left hemispheric dominance could thus be shown for native speakers of languages in which lexical information highly depends on prosodic processing, such as Mandarin Chinese, when compared to the same stimuli presented to English speaking subjects (e.g. Gandour et al., 2004, Tong et al., 2005).

In children, functional imaging data on prosodic processing are sparse. Some studies, however, investigated basic issues concerning the left–right lateralization of language. In sleeping newborns, greater left hemispheric activation for infant-directed speech was shown by Pena et al. (2003). Using optical imaging, a larger increase in cerebral blood volume was demonstrated over left temporal brain regions during forward compared to backward speech. While the authors conclude that at this very early age the left hemisphere is already specialized for language processing differences in acoustic properties and potentially also an attentional difference between the stimulus types must be respected. In fact a functional magnetic resonance imaging (fMRI) study in 2- to 3-month-old infants demonstrated a robust left hemispheric activation for both types of stimuli which may stress the importance of e.g. the presence of fast transitions in the auditory material as an additional decisive feature beyond linguistic properties (Dehaene-Lambertz et al., 2002). The latter study furthermore reported a greater right frontal brain activation for forward as compared to backward speech in awake but not in sleeping infants. Taken together, these pioneering studies use a control stimulus that contains irregular phonemes and irregular prosody. Thus the effect of basic acoustic properties and differential attention towards the natural as compared to the unnatural stimulus type must be taken into account (see e.g. Zatorre and Belin, 2001, Zatorre et al., 2002, Poeppel, 2003, Boemio et al., 2005, Schonwiesner et al., 2005). It thus remains unclear whether the results reflect lateralization due to linguistic or acoustic processing since discrimination between acoustic properties of different languages has even been demonstrated in animals (e.g. Ramus et al., 2000). Another recent optical imaging study in sleeping 3-month-old infants demonstrated a larger right temporo-parietal response for normal when compared to flattened, non-prosodic sentences (Homae et al., 2006). The findings can be interpreted as first evidence of a specialization of the right hemisphere for the processing of prosody. However, the study did not investigate stimulus material which contained prosodic i.e. suprasegmental information in isolation, because in the stimulus material used segmental information was always present.

To further elucidate the development of interhemispheric language specialization, it is thus essential to find out whether prosody in isolation will activate right hemispheric fronto-temporal language areas as it has been shown in adults (see above). Therefore, we here investigate how prosodic information is processed in children. As Plante et al. found a bilateral network of activation and a right frontal dominance for prosody using fMRI during a sentence-prosody-matching task in children from 5 to 18 years of age (Plante et al., 2006a), we studied the processing of pure prosodic information in 4-year-olds.

This research is motivated by the hypothesis that suprasegmental information guides language acquisition, and it is therefore required to disentangle the neural basis of specific processing of segmental and suprasegmental information. Electrophysiological studies (EEG) indicate that a number of phonological processes can indeed be differentiated during early language development: using odd-ball designs, an effective discrimination of phoneme duration and stress patterns by the age of 2–5 months has been shown (Friederici et al., 2002, Friedrich et al., 2004, Weber et al., 2004, Friederici, 2005). Aberrant patterns in the event related potentials (ERP) of infants at that age are moreover related to the risk for language impairments (Friedrich et al., 2004, Weber et al., 2005; see also Benasich et al., 2006). Concerning the issue of suprasegmental processing, a slow positive shift (Closure Positive Shift) in response to the end of an intonational phrase reflecting the processing of an intonational phrase boundary (Steinhauer et al., 1999) is present by the age of 8–9 months (Friederici, 2005, Pannekamp et al., 2006).

Despite the potential of ERPs to separate different processing stages, source localization is limited even as to the projection of the dipole to the left or right hemisphere. Therefore, we here use an optical imaging approach similar to the study by Homae et al. (2006) to examine the contribution of left and right hemispheric pathways for the perception of segmental and suprasegmental information during sentence processing. Based on results of previous functional imaging studies in adults and children, we hypothesized right hemispheric dominance for isolated prosodic processing and left hemispheric dominance for the processing of normal speech. The study also intends to further validate the versatility of functional optical imaging since its undemanding setup lends itself to studies in infants, thus allowing for longitudinal studies and the comparison with functional imaging data in adults.

Section snippets

Subjects

Fifty-one healthy children were investigated (age 4.06 ±  0.07 years [mean, SD], 30 boys, native and monolingual German speakers, normal hearing, no neurological disorders, normal language development). We acquired informed consent from both parents, the study protocol was approved by the local ethics committee (Charité, Berlin).

Material

Three kinds of sentences were presented: Normal sentences (child-directed speech) contained phonological, syntactic, semantic (segmental), and prosodic (suprasegmental)

Results

Generally the experiment was well tolerated by the young subjects. We did not have to exclude any of the children in whom the experiment was completed. The artifact rejection procedure yielded a minimum of 80% uncontaminated trials (rejection of 17.4 ± 3.0% across subjects). The rejected trials were balanced across conditions, thus not introducing a bias in signal to noise in one specific condition. Analysis of variance indicates no statistically significant main effect of hemisphere over all

Discussion

Our results in children are consistent with recent results in adults showing a right hemispheric dominance for processing prosodic information (e.g. Zatorre et al., 1992, Meyer et al., 2002, Gandour et al., 2004, Tong et al., 2005). For instance, Meyer and colleagues found greater left hemispheric responses to normal and syntactic speech as compared to prosodic speech devoid of syntactic and lexical–semantic information. Processing prosody in isolation resulted in greater right temporal and

Acknowledgments

Financial support of BMBF (Berlin NeuroImaging Center and Bernstein Center for Computational Neuroscience Berlin), EU (NEST 012778, EFRE 20002006 2/6), and International Leibniz Program are gratefully acknowledged. We thank H. Benav, J. Haselow, P. Koch, and C. Ruegen for their help with data acquisition and analysis and H.R. Heekeren, A. Pannekamp, and U. Toepel for their comments on the manuscript.

References (79)

  • F. Homae et al.

    The right hemisphere of sleeping infant perceives sentential prosody

    Neurosci. Res.

    (2006)
  • L. Jancke et al.

    Phonetic perception and the temporal cortex

    NeuroImage

    (2002)
  • P.W. Jusczyk

    How infants begin to extract words from speech

    Trends Cogn. Sci.

    (1999)
  • P.W. Jusczyk et al.

    The beginnings of word segmentation in English-learning infants

    Cognit. Psychol.

    (1999)
  • S. Koelsch et al.

    Towards a neural basis of music perception

    Trends Cogn. Sci.

    (2005)
  • S. Koelsch et al.

    Bach speaks: a cortical “language-network” serves the processing of music

    NeuroImage

    (2002)
  • J. Mehler et al.

    A precursor of language acquisition in young infants

    Cognition

    (1988)
  • M. Meyer et al.

    Functional MR imaging exposes differential brain responses to syntax and prosody during auditory sentence comprehension

    J. Neurolinguist.

    (2003)
  • E. Plante et al.

    Dissociating sentential prosody from sentence processing: activation interacts with task demands

    NeuroImage

    (2002)
  • E. Plante et al.

    Prosodic processing by children: an fMRI study

    Brain Lang.

    (2006)
  • E. Plante et al.

    Sex differences in the activation of language cortex during childhood

    Neuropsychologia

    (2006)
  • D. Poeppel

    The analysis of speech in different temporal integration windows: cerebral lateralization as ‘asymmetric sampling in time’

    Speech Commun.

    (2003)
  • E.D. Ross et al.

    Lateralization of affective prosody in brain and the callosal integration of hemispheric language functions

    Brain Lang.

    (1997)
  • V.J. Schmithorst

    Separate cortical networks involved in music perception: preliminary functional MRI evidence for modularity of music processing

    NeuroImage

    (2005)
  • M. Tervaniemi et al.

    Lateralization of auditory-cortex functions

    Brain Res. Brain Res. Rev.

    (2003)
  • Y. Tong et al.

    Neural circuitry underlying sentence-level linguistic prosody

    NeuroImage

    (2005)
  • C. Weber et al.

    Discrimination of word stress in early infant perception: electrophysiological evidence

    Brain Res. Cogn. Brain Res.

    (2004)
  • C. Weber et al.

    Reduced stress pattern discrimination in 5-month-olds as a marker of risk for later language impairment: neurophysiological evidence

    Brain Res. Cogn. Brain Res.

    (2005)
  • P. Wobst et al.

    Linear aspects of changes in deoxygenated hemoglobin concentration and cytochrome oxidase oxidation during brain activation

    NeuroImage

    (2001)
  • R.J. Zatorre et al.

    Structure and function of auditory cortex: music and speech

    Trends Cogn. Sci.

    (2002)
  • S.R. Baum et al.

    The neural bases of prosody: insights from lesion studies and neuroimaging

    Aphasiology

    (1999)
  • A. Boemio et al.

    Hierarchical and asymmetric temporal sensitivity in human auditory cortices

    Nat. Neurosci.

    (2005)
  • A.S. Chan et al.

    Music training improves verbal memory

    Nature

    (1998)
  • M. Cope et al.

    System for long-term measurement of cerebral blood and tissue oxygenation on newborn infants by near infra-red transillumination

    Med. Biol. Eng. Comput.

    (1988)
  • A. Cutler et al.

    Prosody in the comprehension of spoken language: a literature review

    Lang. Speech

    (1997)
  • G. Dehaene-Lambertz et al.

    Functional neuroimaging of speech perception in infants

    Science

    (2002)
  • A. Fernald et al.

    A cross-language study of prosodic modifications in mothers’ and fathers’ speech to preverbal infants

    J. Child Lang.

    (1989)
  • P.T. Fox et al.

    Focal physiological uncoupling of cerebral blood flow and oxidative metabolism during somatosensory stimulation in human subjects

    Proc. Natl. Acad. Sci. U. S. A.

    (1986)
  • L. Frazier et al.

    Prosodic phrasing is central to language comprehension

    Trends Cogn. Sci.

    (2006)
  • Cited by (76)

    • Prosodic grouping at birth

      2016, Brain and Language
    View all citing articles on Scopus
    View full text