The more, the better? Behavioral and neural correlates of frequent and infrequent vowel exposure

Sho Tsuji, Paula Fikkert, Yasuyo Minagawa, Emmanuel Dupoux, Luca Filippin, Maarten Versteegh, Peter Hagoort, Alejandrina CristiàDevelopmental 2017 · DOI

TL;DR

This study investigated whether the amount of exposure to speech sounds influences infants' ability to discriminate them, a key aspect of language acquisition. Using behavioral and neural measures, the researchers found that large differences in exposure frequency between native vowel sounds did not lead to differences in discrimination ability in 5-8-month-old infants. This suggests that once a certain threshold of exposure is met, additional frequency might not significantly enhance perceptual attunement.

What Problem Does This Paper Address?

A cornerstone of language acquisition theory is "perceptual attunement," the process by which infants become specialized in perceiving the speech sounds of their native language. A central assumption is that the more frequently an infant hears a particular speech sound contrast, the better they will become at discriminating it. While it's well-established that infants' perception shifts from being sensitive to all speech sounds to focusing on native language sounds, the specific role of exposure frequency within native language categories has been less directly tested. This paper aims to fill this gap by directly comparing infants' ability to discriminate native vowel contrasts that differ significantly in their frequency of occurrence in the input.

Methods

The study employed a two-pronged approach:

  • Experiment 1: Behavioral Visual Habituation: This experiment used a standard method for assessing infant speech perception. Infants were habituated to a specific vowel contrast (either frequent native, infrequent native, or non-native). Discrimination was then tested by presenting alternating trials (where the vowel changes) versus non-alternating trials (where the vowel stays the same). Looking time differences between these trial types served as the measure of discrimination.
  • Experiment 2: Near-Infrared Spectroscopy (NIRS): To gain a more sensitive measure of neural processing, NIRS was used. This technique measures changes in blood oxygenation in the brain, providing an index of neural activity. The NIRS experiment also used a habituation paradigm but measured brain responses to the frequent and infrequent native vowel contrasts. This allowed for a within-participant comparison, increasing statistical power.

Datasets

The study utilized existing corpora of spoken Dutch (CGN and IFA corpus) to determine the frequency of the native vowel sounds. The experimental stimuli were recorded by native speakers of Dutch and American English in an infant-directed register. No new datasets were introduced.

Key Insights and Ideas

  • Frequency Might Not Be Everything (Above a Threshold): The most striking finding is that large differences in the frequency of native vowel exposure did not translate into differences in discrimination ability for 5-8-month-old infants, as measured by both behavior and brain activity. This challenges the straightforward "more is better" assumption.
  • Saliency Over Frequency? The study suggests that the inherent acoustic properties (saliency) of the vowel contrasts might play a more significant role than frequency alone. The chosen native contrasts, despite differing in frequency, might have been acoustically distinct enough for infants to discriminate them equally well.
  • A Critical Minimum Exposure: It's possible that infants need a "critical" amount of exposure to a speech sound contrast, and once this threshold is met, further increases in frequency don't yield additional benefits for discrimination.
  • Behavioral vs. Neural Sensitivity: While Experiment 1 showed no difference between frequent and infrequent native contrasts, Experiment 2 (NIRS) did reveal a subtle difference in the timing of brain responses, with frequent stimuli eliciting earlier peaks. However, this did not translate to differences in the strength of discrimination responses.

Key Results

  • Experiment 1 (Behavioral): Infants discriminated both the frequent and infrequent native vowel contrasts equally well. They also successfully discriminated the native contrasts but not the non-native contrast, confirming the expected pattern of native language specialization.
  • Experiment 2 (NIRS): No significant differences were found in the overall neural discrimination responses (beta values) between the frequent and infrequent native vowel contrasts. However, a post-hoc analysis suggested that brain responses to frequent stimuli peaked earlier than those to infrequent stimuli.

Limitations and Future Work

  • Saliency of Contrasts: The authors acknowledge that the chosen native vowel contrasts might have been inherently salient due to factors like duration and diphthongization, potentially masking any frequency effects. Future research should explore contrasts with varying degrees of saliency.
  • Critical Minimum Exposure: The study raises the question of what constitutes a "critical" number of instances for speech sound acquisition. Future work could investigate the effects of very low exposure levels or different age groups to pinpoint this threshold.
  • Linguistic Level of Processing: While discrimination was not affected by frequency, the authors suggest that differences might emerge at later, more linguistic stages of processing, such as in the timing of hemispheric specialization. Further research could explore these later developmental effects.
  • Generalizability: The study focused on Dutch infants and specific vowel contrasts. Replicating these findings with different languages and phonetic contrasts is crucial for broader conclusions.

Summary updated

Abstract

A central assumption in the perceptual attunement literature holds that exposure to a speech sound contrast leads to improvement in native speech sound processing. However, whether the amount of exposure matters for this process has not been put to a direct test. We elucidated indicators of frequency-dependent perceptual attunement by comparing 5-8-month-old Dutch infants' discrimination of tokens containing a highly frequent [hɪt-he:t] and a highly infrequent [hʏt-hø:t] native vowel contrast as well as a non-native [hɛt-haet] vowel contrast in a behavioral visual habituation paradigm (Experiment 1). Infants discriminated both native contrasts similarly well, but did not discriminate the non-native contrast. We sought further evidence for subtle differences in the processing of the two native contrasts using near-infrared spectroscopy and a within-participant design (Experiment 2). The neuroimaging data did not provide additional evidence that responses to native contrasts are modulated by frequency of exposure. These results suggest that even large differences in exposure to a native contrast may not directly translate to behavioral and neural indicators of perceptual attunement, raising the possibility that frequency of exposure does not influence improvements in discriminating native contrasts.
BibTeX
@inproceedings{tsuji2017more,
  author = {Sho Tsuji and Paula Fikkert and Yasuyo Minagawa and Emmanuel Dupoux and Luca Filippin and Maarten Versteegh and Peter Hagoort and Alejandrina Cristià},
  title = {{The more, the better? Behavioral and neural correlates of frequent and infrequent vowel exposure}},
  booktitle = {Developmental},
  year = {2017},
  doi = {10.1002/dev.21534},
}
Figure