Elsevier

Clinical Neurophysiology

Volume 118, Issue 2, February 2007, Pages 412-419
Clinical Neurophysiology

Neonatal frequency discrimination in 250–4000-Hz range: Electrophysiological evidence

https://doi.org/10.1016/j.clinph.2006.10.008Get rights and content

Abstract

Objective

The precision of sound frequency discrimination in newborn infants in the 250–4000-Hz frequency range was determined using the neonatal electrophysiological mismatch response (MMR), the infant equivalent of adult mismatch negativity (MMN).

Methods

The electroencephalogram (EEG) was recorded in 11 full-term sleeping newborn infants mostly in active sleep (67% of the time). Pure tones were presented through loudspeakers in an oddball paradigm with a 800-ms stimulus onset asynchrony (SOA). Each stimulus block contained a standard (p = 0.76) of 250, 1000, or 4000 Hz in frequency (in separate blocks) and deviants with a frequency change of either 5% or 20% of the standard (p = 0.12 of each).

Results

A positive ERP deflection was found at 200–300 ms from stimulus onset in response to the 20% deviation from the 250, 1000, and 4000 Hz standard frequencies. The amplitude of the response in the 200–300 ms time window was significantly larger for the 20% than 5% deviation.

Conclusions

We observed in newborn infants automatic frequency discrimination as reflected by a positive MMR. The newborns were able to discriminate frequency change of 20% in the 250–4000-Hz frequency range, whereas the discrimination of the 5% frequency change was not statistically confirmed.

Significance

The present data hence suggest that the neonatal frequency discrimination has lower resolution than that in adult and older children data.

Introduction

The ability to tell apart sounds of different frequency is one of the important basic properties of the central auditory system. Both music and speech perception rely upon frequency discrimination. From the clinical perspective, the identification of abnormalities in frequency discrimination abilities may improve, for instance, the diagnostics of dyslexia (Baldeweg et al., 1999). Yet, the developmental course of the frequency-discrimination ability is poorly understood.

Behavioural studies in adult humans have shown that the ratio between the minimum distinguishable frequency interval (difference limen, DL) and the frequency is smallest in the 500–2000 Hz frequency range, becoming larger at higher and lower frequencies (Wier et al., 1977, Sek and Moore, 1995). As determined by otoacoustic emission studies (Abdala and Sininger, 1996), cochlear frequency resolution is fully mature by term birth; behavioural studies report sound-frequency discrimination in infants as young as 3 months (Werner and Gray, 1998). Early attempts (Leventhal and Lipsitt, 1964, Trehub, 1973) to find evidence for tonal frequency discrimination in newborns by using behavioural methods have failed. However, the sucking rate measurement and other methods indicate that newborns can discriminate the differences in the fundamental frequency of the human voice (DeCasper and Fifer, 1980, Mehler et al., 1988; for review, see Gerken and Aslin, 2005). At 3 months, DLs are smaller for low than for high-frequency tones (Olsho et al., 1987a), but the frequency discrimination at high frequencies develops faster, reaching the adult level by 6 months, whereas low-frequency DLs remain immature until the late childhood (Maxon and Hochberg, 1982).

However, the behavioural data obtained in infants younger than 5 months may provide incomplete and unreliable information about their frequency discrimination (Stapells and Kurtzberg, 1991). First, behavioural responses, such as head turning, often leave space for subjective interpretation (even though this could be overcome by introducing an additional observer, unaware of the paradigm, reporting any change in the infant’s behavior; Olsho et al., 1987b). Second, a failure in behavioural discrimination may be due either to genuine perceptual immaturity or to not fully developed attentional and memory mechanisms. This stresses the importance of the electrophysiological measures of frequency discrimination, which are independent of attention and enable direct measurement of auditory discriminative abilities (Kurtzberg and Vaughan, 1985, Stapells and Kurtzberg, 1991).

In adults, frequency discrimination can be assessed objectively by recording the mismatch negativity (MMN; Näätänen et al., 1978, Hari et al., 1984, Sams et al., 1985). The MMN correlates with the behavioural performance (Tervaniemi et al., 1993, Tiitinen et al., 1994, Amenedo and Escera, 2000, Novitski et al., 2004). It does not require the subject’s attention or task performance and can therefore be used in subject groups that are not able to cooperate with the experimenter. The MMN-like response to frequency change was found in newborns (Alho et al., 1990) and its magnetoencephalographic (MEG) analog (Hari et al., 1984) was recently discovered even in fetuses (Draganova et al., 2005, Huotilainen et al., 2005).

Unlike in adults, in young infants the mismatch response was found to be in different studies of either negative (Alho et al., 1990, Cheour et al., 1998, Ceponiene et al., 2002) or positive polarity (Dehaene-Lambertz and Dehaene, 1994, Leppänen et al., 1997, Morr et al., 2002, Sambeth et al., 2006). In addition, in one study (Ceponiene et al., 2002) the negativity was present only in part of the subject group. The differences between individual newborns can be explained by different degrees of maturation: Leppänen et al. (2004) showed that less mature infants showed a tendency towards negativity and more mature ones towards positivity. The infant’s arousal state can also have an impact: 2-month-old infants showed negative responses when awake and positive responses when asleep (Friederici et al., 2002). Moreover, the inter-stimulus interval may also have some effect on polarity as the negative response was elicited with an 800-ms stimulus onset asynchrony (SOA) but not with a 450 or 1500-ms SOA (Cheour et al., 2002). Finally, the higher cutoff of high-pass filtering selectively reduced positive responses due to the longer duration of the positive response (Trainor et al., 2003, Weber et al., 2004). Despite the discrepancies in the polarity of the response, it is clear that the auditory discriminative abilities of newborns can be measured with ERP recordings. This is true not only for the physical features of a stimulus but also for the abstract rules governing an auditory stimulus stream (Ruusuvirta et al., 2003, Ruusuvirta et al., 2004, Winkler et al., 2003, Carral et al., 2005).

In the majority of the afore-mentioned studies, the frequency difference between the standard and the deviant was large: the minimum was 10%, while the most often used difference was 20%. Also, in tonal studies, the frequency range of 1000–2000 Hz has usually been applied whereas the discriminative response in other frequency ranges has not been investigated. In adults, a 5% frequency difference elicits a significant MMN in the frequency range from 250 to 4000 Hz (Novitski et al., 2004). The present study used the MMN paradigm to estimate automatic frequency discrimination in the 250–4000-Hz range.

Section snippets

Subjects and the experimental environment

Initially, 19 full-term newborn infants (mean age 2 days, mean gestational age at birth 41.2 ± 0.4 weeks, activity-pulse-grimace-appearance-respiration (APGAR) scores between 8 and 10) participated in the study. Their sleep stage during the recordings was determined in each experimental block by a qualified specialist on the basis of the EEG and behavioural patterns. Thereafter, the percentage of each of the three sleeping stages in the experiment was calculated asS1=(B1+B1,2/2+B1,3/2+B1,2,3/3)

Waveform description

The individual deviant-standard difference waves for 11 subjects in the 1000-Hz-standard condition are shown in Fig. 1. There was substantial individual variation in the shape of the waveform which had both negative and positive deflections. The grand-averaged ERPs of standards and deviants, the grand-average difference waves, and the voltage maps of the grand-average ERPs and difference waves at 288 ms in the 1000-Hz-standard condition are presented in Fig. 2. The most prominent deflection at

Discussion

The present results demonstrate that the electrophysiological response of healthy newborns to sound frequency changes is manifested as a positive deflection in the event-related potential in the 200–300-ms window after stimulus onset. Based on its automatic elicitation during sleep, the mismatch response (MMR) can be regarded as the newborn equivalent of the mismatch negativity (MMN).

This response was obtained for the 20% change in frequency around 250, 1000, and 4000 Hz. There was no conclusive

Acknowledgements

We thank Tarja Ilkka, RN for recording the data. We also thank Dr. Istvan Winkler, Dr. Elina Pihko, and Dr. Anke Sambeth for their helpful comments during the discussion of preliminary results of the present project and Dr. Ilkka Linnankoski for language editing. This research was supported by the Academy of Finland (projects 200522, 77322, and 73038).

References (47)

  • N. Novitski et al.

    Frequency discrimination at different frequency levels as indexed by electrophysiological and behavioral measures

    Cogn Brain Res

    (2004)
  • A. Sambeth et al.

    Newborns discriminate novel from harmonic sounds: a study using magnetoencephalography

    Clin Neurophysiol

    (2006)
  • M. Sams et al.

    Auditory frequency discrimination and event-related potentials

    Electorencephalogr Clin Neurophysiol

    (1985)
  • D.R. Stapells et al.

    Evoked potential assessment of auditory system integrity in infants

    Clin Perinatol

    (1991)
  • L. Trainor et al.

    Changes in auditory cortex and the development of mismatch negativity between 2 and 6 months of age

    Int J Psychophysiol

    (2003)
  • C. Weber et al.

    Discrimination of word stress in early infant perception: electrophysiological evidence

    Brain Res Cogn Brain Res

    (2004)
  • C. Abdala et al.

    The development of cochlear frequency resolution in the human auditory system

    Ear Hear

    (1996)
  • K. Alho et al.

    Processing of novel sounds and frequency changes in the human auditory cortex: magnetoencephalographic recordings

    Psychophysiology

    (1998)
  • E. Amenedo et al.

    The accuracy of sound duration representation in the human brain determines the accuracy of behavioural perception

    Eur J Neurosci

    (2000)
  • T. Baldeweg et al.

    Impaired auditory frequency discrimination in dyslexia detected with mismatch evoked potentials

    Ann Neurol

    (1999)
  • V. Carral et al.

    A kind of auditory ’primitive intelligence’ already present at birth

    Eur J Neurosci

    (2005)
  • R. Ceponiene et al.

    Event-related potential features indexing central auditory discrimination by newborns

    Brain Res Cogn Brain Res

    (2002)
  • M. Cheour et al.

    The auditory sensory memory trace decays rapidly in newborns

    Scand J Psychol

    (2002)
  • Cited by (57)

    • Sleep as a window to target traumatic memories

      2022, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      The main advantage of auditory stimulation is the presentation speed that enables phased-targeted TMR. The upwave part of a SO lasts 500 ms, while sounds as short as 0.25 ms have been shown to be perceived during sleep (Novitski et al., 2007). Another advantage of auditory stimulation, especially when using TMR to strengthen the positive outcome of EMDR, is that an auditory, distracting stimulus is a common feature of the EMDR treatment protocol.

    • Predictive processing of pitch trends in newborn infants

      2015, Brain Research
      Citation Excerpt :

      Because the responses elicited by violations of inter-tone contingencies have been found to be of rather low amplitude in adults and the signal-to-noise ratio of ERP measurement in neonates is substantially lower than that in adults, we chose to measure in neonates the response to sensory trend violations. Although it is difficult to establish a direct analogy between the adult MMN and the infant MMR (see Trainor, 2012), deviations from both simple and complex pitch regularities have been shown to elicit MMR in newborn infants: e.g., MMR has been elicited by deviations from a repeating pitch (Novitski et al., 2007) irrespective of timbre variance (Háden et al., 2009), by violations of the constancy of the direction (Carral et al., 2005) and size (Stefanics et al., 2009) of pitch change within tone pairs varying in absolute pitch, as well as by rare chords categorically differing from the majority of chords (Virtala et al., 2013). These previous studies established that neonates encode the direction and size of pitch steps.

    • Feature-specific transition from positive mismatch response to mismatch negativity in early infancy: Mismatch responses to vowels and initial consonants

      2015, International Journal of Psychophysiology
      Citation Excerpt :

      However, the infant MMN usually persists for a longer interval in a relatively late time window. Other studies have reported a positive MMR (P-MMR), rather than an MMN, to various speech and non-speech changes between 200 and 400 ms, mainly at a younger age (Dehaene-Lambertz and Baillet, 1998; Dehaene-Lambertz and Dehaene, 1994; Friederici et al., 2002; Jing and Benasich, 2006; Leppänen et al., 1997; Morr et al., 2002; Novitski et al., 2007). For example, Leppänen et al. (1997) observed a P-MMR peaking between 250 ms and 350 ms to pitch change of pure tones in newborns.

    • Detecting the temporal structure of sound sequences in newborn infants

      2015, International Journal of Psychophysiology
      Citation Excerpt :

      In general, we found that newborn infants have similar capabilities as adults for processing the cues that allow one to form a rough description of auditory objects. Although we have suggested in Introduction that such fundamental capabilities are required for infants for learning from others, the finding is still surprising on one sense: Research in young infants has consistently shown that when it comes to simple discrimination abilities, infantile capabilities are far from the adult level (see, e.g., pitch discrimination; Novitski et al., 2007; for a review, see Werner, 2007). Regarding temporal features, for example, the sensitivity of detecting changes in sound duration (Kushnerenko et al., 2001; Čeponienė et al., 2002; Cheour et al., 2002) or gaps between sounds is much lower than that in adults even at 6–7 months of age (Smith et al., 2006; Trainor et al., 2001, 2003; Werner et al., 1992).

    View all citing articles on Scopus
    View full text