- Email: [email protected]
Mismatch negativity shows that 3–6-year-old children can learn to discriminate non-native speech sounds within two months
Neuroscience Letters 325 (2002) 187–190 www.elsevier.com/locate/neulet
Mismatch negativity shows that 3–6-year-old children can learn to discriminate non-native speech sounds within two months Marie Cheour a,b,*, Anna Shestakova b, Paavo Alku c, Rita Ceponiene b, Risto Na¨a¨ta¨nen b a
Language and the Developing Brain Laboratory, Centre for Cognitive Neuroscience, University of Turku, Turku, Finland b Cognitive Brain Research Unit, Department of Psychology, University of Helsinki, Helsinki, Finland c Laboratory of Acoustics and Audio Signal Processing, Department of Electrical and Communications Engineering, Helsinki University of Technology, Helsinki, Finland Received 7 March 2002; received in revised form 13 March 2002; accepted 13 March 2002
Abstract Using 3–6-year-old children as subjects, we describe the neural plasticity accompanying the concurrent learning of a foreign language in a natural environment. Children were monitored for 6 months as they either enrolled in schools or daycare centers where only Finnish was spoken (Control group) or as they joined a French school or a daycare center where French was spoken 50–90% of the time (Experimental group). Whereas mismatch negativity (MMN)—a brain’s electrical change-detection response—for a French speech contrast was initially absent or very small in both groups, it was conspicuous 2 months after Finnish children had joined a French kindergarten. Consequently, the data suggest that youngsters can learn to distinguish non-native speech sounds in natural language environment without any special training in just a couple of months. Accordingly, these data herald the vast potential MMN may entail for studying language learning, especially in situations where behavioral responses cannot be readily elicited. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Event-related potentials; Children; Language; Learning; Mismatch negativity; Non-native; Speech
When studying second language acquisition, many experiments have concentrated on investigating the learning of grammar, syntax, or vocabulary. To be able to study these factors, subjects need to have quite good knowledge of a second language already. In the early stage of a language learning, before a person is able to speak or understand a certain language, he has to learn to discriminate the speech sounds of this language. Thus, the recognition patterns that are specific for this language need to be built in the cortex. These language-specific memory traces that develop during the first year of life for native language [2,4,7] help one to perceive and discriminate native speech sounds [11,17]. By studying the formation of these memory traces, it is possible to reveal the learning of second language long before the person himself or people around him are able to report that
* Corresponding author. Language and the Developing Brain Laboratory, Centre for Cognitive Neuroscience, University of Turku, Turku, Finland. Tel.: 1358-2-3338774; fax: 1358-23338770. E-mail address: [email protected] (M. Cheour).
he has gained knowledge on a certain language or is able to discriminate certain speech sounds [5]. The present study attempts to capture some aspects of the neural ability involved in correctly perceiving foreign speech utterances. Most previous studies that investigate the discrimination of non-native speech sounds have been done either in the laboratory or by comparing subjects exposed to a foreign language at different ages [5,9,17]. Here we present longitudinal data that have been collected in a natural environment showing the neurobiological manifestation of the second language learning. Native monolingual Finnish children were monitored for 6 months as they joined either a French school or a daycare center where French was spoken 50–90% of the time. French vowels were used to determine how quickly and accurately children developed cortical memory traces and discrimination abilities for non-native speech sounds during this period. This was done by recording mismatch negativity (MMN), the brain’s automatic electrophysiological response elicited by infrequent changes in auditory stimulation [3,10]. Not long ago, behavioral methods, such as
0304-3940/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 2) 00 26 9- 0
188
M. Cheour et al. / Neuroscience Letters 325 (2002) 187–190
forced-choice button pressing, were the only tools to investigate the development of speech perception in children. In recent years, however, MMN has been shown to be a promising tool for investigating the neural bases of speech perception. For example, it enables one to demonstrate the existence of language-specific memory traces in the adult human cortex [11]. Thereafter, it has been shown that these native-language memory traces develop in human infants between 6 and 12 months [2] or perhaps even earlier [4]. Our experimental group consisted of 17 monolingual 3– 6-year-old native Finnish children (four males) with no prior exposure to French. Similarly, the control group consisted of 17, 3–6-year-old native Finns (four males) enrolled in schools or daycare centers where only Finnish was spoken. Finnish children exposed to French were tested three times during the period of 4 months. Electrical recordings were obtained from the children upon joining a French school, following 6–8 weeks (mean, 53 days), and following 12–16 weeks (mean, 105 days), respectively. The control group, exposed to no French, was investigated twice: initially; and 20 weeks (mean, 150 days) after the first session. We recorded the brain responses of all subjects to the French vowels /i/ (‘standard’; P ¼ 0:8), /e/, and /1/ (‘deviants’; P ¼ 0:1 for each). Prior to the experiment, the adult, native Finnish listeners were asked to rate the category goodness of the stimuli used in this experiment among other French and Finnish vowels. None of these subjects perceived French stimuli as best exemplars of Finnish vowels. Thus, they were not considered as phoneme prototypes in Finnish (Fig. 1) [6,7]. This is why it was reasonable to expect that Finnish subjects with no prior exposure to French would possess no permanent memory traces for these vowels [2,4]. These vowels of 65 dB SPL, lasting 207 ms (with 10-ms rise and fall times) were presented through two loudspeakers located about 1.5 m from the subjects’ ears. The (onset-to-onset) inter-stimulus-interval was 907 ms. Speech sounds were generated using a male voice in a semi-synthetic method. This produces natural sounding vowel stimuli by exciting an artificial vocal-tract model with an excitation waveform extracted from a real utterance [1]. These stimuli were randomly presented in blocks of 200 vowels, 10–12 blocks per subject, in every session. During the experiment, subjects watched selfselected silent movies. Electroencephalogram (EEG, sampling rate at 250 Hz) was recorded with silver/silver chloride electrodes using a NeuroScan PC-3.0 based system. All experiments were conducted in an acoustically and electrically shielded room. Raw data were averaged and digitally filtered (bandpass, 0.1–30 Hz; 24 dB points) at eight scalp sites: F3; F4; C3; C4; T3; T4; P3; and P4. Eye movements were monitored with electrooculogram (EOG) electrodes attached below and at the outer canthus of the right eye. The right/ left-hemisphere electrodes were referred to the right/left mastoid, respectively. EOG was referenced to the right mastoids. Exclusion criteria included EEG exceeding 200
mV at any one electrode, epochs for the first three stimuli of each block, and responses to each standard immediately following a deviant stimulus. For any one child, there were at least 200 acceptable EEG epochs for each deviant stimulus in each session. The baseline for the waveforms was defined as the mean amplitude between 2100 and 0 ms relative to stimulus onset. The MMN amplitude was determined from the F3, F4, C3, C4, P3 and P4 electrodes (excluding T3 and T4 electrodes where MMN did not usually significantly differ from 0 mV). MMN was measured by separately computing difference waves that were obtained by subtracting the average standard-stimulus response from the average deviant-stimu-
Fig. 1. Stimuli used in the present experiments. The French vowels are (indicated by dark gray) interspersed amongst the Finnish vowels (light gray). The location of each vowel, /i/, /e/, and /1/, is shown in the F1–F2 space by upward arrows. The first value in the pair of each tag corresponds to F2 and the second to F1.
M. Cheour et al. / Neuroscience Letters 325 (2002) 187–190
lus response for each subject, electrode, and session. The mean amplitude, that is, the mean voltage over the 50-ms period time window, was centered at the latency of largest peak occurring between 150 and 350 ms. Using two-tailed t-tests, we tested the presence of MMN. The differences in the MMN amplitude and MMN latency between the groups, stimuli, session, and electrode were tested with four-way analysis of variance (ANOVA): Group (Experimental Subjects vs. Control Subjects) £ Session (First, Second, Last) £ Stimulus (/e/, /1/) £ Electrode (F3, F4, C3, C4, P3, P4). The session effect for the MMN amplitude and latency was separately tested for the Experimental Subjects with three-way ANOVA: Session (First, Second, Last) £ Stimulus (/e/ vs. /1/) £ Electrode (F3, F4, C3, C4, P3, P4), to monitor the dynamics of French-specific phoneme discrimination more precisely. Brain responses to French speech sounds are presented in Fig. 2. These responses clearly implicate the development of non-native memory traces in Finnish children who were exposed to French. The Session effect on MMN amplitude was tested separately for the experimental subjects using a three-way ANOVA: Session (First, Second, Last) £ Stimulus (e,1) £ Electrode (F3, F4, C3, C4, P3, P4). This analysis was highly significant (Fð2;32Þ ¼ 18:12, P , 0:000005). Post-hoc testing revealed that the MMN amplitude for the experimental group was significantly increased between the first and second sessions (P , 0:000046). Hence, in these children, the MMN ampli-
Fig. 2. Development of French-specific discrimination ability (indexed by MMN) in 3–6-year-old Finnish children for French phonemes /e/ and /1/ during the period of approximately 4 months. This figure presents the difference waves (obtained by subtracting the average standard /i/ stimulus response from the average deviant /e/ or deviant /1/ stimulus response) of 17 Finnish children who began to attend to French daycare are shown in the beginning of the test period, after a period of 1.5–2 months, and after a period of about 3–4 months. The difference waves of 17 monolingual Finnish speaking children who were not exposed to French are also shown in the beginning of the test period and after about 3–4 months.
189
tude increased strikingly in only a couple of months in a natural environment without any special training. No significant improvement in the MMN amplitude could be obtained between the second and the third sessions. The Session effect on MMN latency was tested exactly like that on MMN amplitude. This analysis was also significant (Fð2;32Þ ¼ 3:51, P , 0:041654). The least significant difference (LSD) post-hoc test for MMN latency demonstrated that this result was due to the MMN in the last session (204.54 ms) being significantly shorter than that in the beginning session (230.74 ms; P , 0:012712). A four-way ANOVA with the Group as a grouping factor revealed a significant Group £ Session interaction when comparing the experimental vs. control subjects (Fð1;32Þ ¼ 11:96, P , 0:001560). LSD post-hoc testing showed that the MMN amplitude was significantly larger in the experimental study group at the last session as compared with the first (P , 0:000008). Moreover, the MMN amplitude in the last session was significantly larger in the experimental group as compared with the control group (P , 0:000432). Importantly, no significant increase in the MMN amplitude was found in the control group between the first and second recordings, nor were there any significant MMN-amplitude differences between the experimental and control groups in the first session. Hence, the laboratory experiments themselves were not enough for the children to learn to discriminate foreign phonemes and, apparently, to develop memory traces for them. Numerous behavioral and event-related potential studies have demonstrated that normal language learning occurs primarily or exclusively within childhood. This so-called critical period hypothesis that was first introduced by Lenneberg [8] concerned only the first language acquisition, however. Whether this superior capacity of young children to acquire languages also involves second language is still under debate, however. When investigating late stages of learning, most of the studies show superiority of early learners for phonology, syntax, and grammar [12,13,15]. Interestingly, when early stages of learning have been investigated, most studies have reported that adults seem to be moving toward second language proficiency more quickly than children [13,15]. The present results implicate, however, that the memory representations of foreign phoneme system develop remarkably quickly in the cortex of young children, as compared with the results of previous adult studies [17]. Hence, although young children may have reduced some of their proficiency to differentiate foreign speech sounds once they have learned to differentiate their native language sounds better [2,4,7,14–16], they may regain their non-native contrast discrimination abilities more quickly than older children or adults. We hypothesize that this fast non-native contrast learning is likely to happen when youngsters are still in the process of learning to discriminate their native language sounds when concurrently their ability to differ-
190
M. Cheour et al. / Neuroscience Letters 325 (2002) 187–190
entiate non-native speech has not reached the nadir seen in older children and adults. In conclusion, the present data provide objective evidence to suggest that young children can learn to discriminate nonnative speech sounds in a natural environment within 2 months. Thus, although the MMN latency decreased gradually throughout all three sessions, the MMN amplitude significantly increased already between the first two (out of three) sessions, and did not show any growth after that. Hence, these results suggest that during a relatively short period of time, young Finnish children who were exposed to French developed the ability to discriminate different French phonemes as indexed by the MMN amplitude increase concurrently with a latency decrease. Those children not exposed to French exhibited no change in the MMN characteristics. The authors would like to thank Dr Veijo Vihanta for technical assistance, and Drs Amir Raz and Olli Aaltonen for constructive comments on the manuscript. Moreover, the authors wish to thank the Center for International Mobility (CIMO), The Jenni and Antti Wihuri Foundation, and The Academy of Finland for support. Finally, the authors are indebted to the children, parents, and personnel of Lyce´ e Franco-Finlandais d’Helsinki, Jardin d’enfants Jasmin, and the Ruskeasuo Daycare: Merci beaucoup! [1] Alku, P., Tiitinen, H. and Na¨ a¨ ta¨ nen, R., A method for generating natural-sounding speech stimuli for cognitive brain research, Clin. Neurophysiol., 110 (1999) 1329–1333. [2] Cheour, M., Ceponiene, R., Lehtokoski, A., Luuk, A., Allik, J., Alho, K. and Na¨ a¨ ta¨ nen, R., Development of language-specific phoneme representations in the infant brain, Nat. Neurosci., 1 (1998) 351–353. [3] Cheour, M., Leppa¨ nen, P. and Kraus, N., Mismatch negativity (MMN) as a tool for investigating auditory discrimination and sensory memory in infants and children, Clin. Neurophysiol., 111 (2000) 4–16. [4] Dehaene-Lambertz, G. and Baillet, S., A phonological representation in the infant brain, NeuroReport, 9 (1998) 1885– 1888.
[5] Kraus, N., McGee, T., Carrell, T., King, C., Tremblay, K. and Nicol, T., Central auditory system plasticity associated with speech discrimination training, J. Cogn. Neurosci., 7 (1995) 27–34. [6] Kuhl, P., Innate predispositions and the effects of experience in speech perception: the native language magnetic theory, In B. de Boyson-Bardies, S. de Schonen, P. Jusczyk, P. McNeilage and J. Morton (Eds.), Developmental Neurocognition: Speech and Face Processing in the First Year of Life, Kluwer Academic Publishers, Dordrecht, 1993, pp. 259–274. [7] Kuhl, P.K., Williams, K.A., Lacerda, W.F., Stevens, K.N. and Lindblom, B., Linguistic experiences alter phonetic perception in infants by 6 months of age, Science, 255 (1992) 606– 608. [8] Lenneberg, E., Biological Foundations of Language, Wiley, New York, 1967. [9] MacKain, K., Best, C. and Strange, W., Categorical perception of English /r/ and /l/ by Japanese bilinguals, Appl. Psycholinguist., 2 (1981) 369–390. [10] Na¨ a¨ ta¨ nen, R. and Escera, C., Mismatch negativity: clinical and other applications, Audiol. Neurootol., 5 (2000) 105– 110. [11] Na¨ a¨ ta¨ nen, R., Lehtokoski, A., Lennes, M., Cheour, M., Huotilainen, M., Iivonen, A., Vainio, M., Alku, P., Ilmoniemi, R.J., Luuk, A., Allik, J., Sinkkonen, J. and Alho, K., Languagespecific phoneme representations revealed by electric and magnetic brain responses, Nature, 358 (1997) 432–434. [12] Olson, L. and Samuels, S., The relationship between age and accuracy of foreign language pronunciation, J. Educ. Res., 66 (1973) 263–267. [13] Oyama, S., A sensitive period for the acquisition of a nonnative phonological system, J. Psycholinguist. Res., 5 (1976) 251–258. [14] Patkowski, M., The sensitive period for the acquisition of syntax in a second language, Lang. Learn., 30 (1980) 449–472. [15] Polka, L. and Werker, J.F., Developmental changes in perception of nonnative vowel contrasts, J. Exp. Psychol. Hum. Percept. Perform., 20 (1994) 421–435. [16] Werker, J.F. and Tees, R.C., Cross-language speech perception: evidence for perceptual reorganization during the first year of life, Infant Behav. Dev., 7 (1984) 49–63. [17] Winkler, I., Kujala, T., Tiitinen, H., Sivonen, P., Alku, P., Lehtokoski, A., Czigler, I., Cse´ pe, V., Ilmoniemi, R.J. and Na¨ a¨ ta¨ nen, R., Brain responses reveal the learning of foreign language phonemes, Psychophysiology, 36 (1999) 638–642.
Mismatch negativity shows that 3–6-year-old children can learn to discriminate non-native speech sounds within two months Marie Cheour a,b,*, Anna Shestakova b, Paavo Alku c, Rita Ceponiene b, Risto Na¨a¨ta¨nen b a
Language and the Developing Brain Laboratory, Centre for Cognitive Neuroscience, University of Turku, Turku, Finland b Cognitive Brain Research Unit, Department of Psychology, University of Helsinki, Helsinki, Finland c Laboratory of Acoustics and Audio Signal Processing, Department of Electrical and Communications Engineering, Helsinki University of Technology, Helsinki, Finland Received 7 March 2002; received in revised form 13 March 2002; accepted 13 March 2002
Abstract Using 3–6-year-old children as subjects, we describe the neural plasticity accompanying the concurrent learning of a foreign language in a natural environment. Children were monitored for 6 months as they either enrolled in schools or daycare centers where only Finnish was spoken (Control group) or as they joined a French school or a daycare center where French was spoken 50–90% of the time (Experimental group). Whereas mismatch negativity (MMN)—a brain’s electrical change-detection response—for a French speech contrast was initially absent or very small in both groups, it was conspicuous 2 months after Finnish children had joined a French kindergarten. Consequently, the data suggest that youngsters can learn to distinguish non-native speech sounds in natural language environment without any special training in just a couple of months. Accordingly, these data herald the vast potential MMN may entail for studying language learning, especially in situations where behavioral responses cannot be readily elicited. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Event-related potentials; Children; Language; Learning; Mismatch negativity; Non-native; Speech
When studying second language acquisition, many experiments have concentrated on investigating the learning of grammar, syntax, or vocabulary. To be able to study these factors, subjects need to have quite good knowledge of a second language already. In the early stage of a language learning, before a person is able to speak or understand a certain language, he has to learn to discriminate the speech sounds of this language. Thus, the recognition patterns that are specific for this language need to be built in the cortex. These language-specific memory traces that develop during the first year of life for native language [2,4,7] help one to perceive and discriminate native speech sounds [11,17]. By studying the formation of these memory traces, it is possible to reveal the learning of second language long before the person himself or people around him are able to report that
* Corresponding author. Language and the Developing Brain Laboratory, Centre for Cognitive Neuroscience, University of Turku, Turku, Finland. Tel.: 1358-2-3338774; fax: 1358-23338770. E-mail address: [email protected] (M. Cheour).
he has gained knowledge on a certain language or is able to discriminate certain speech sounds [5]. The present study attempts to capture some aspects of the neural ability involved in correctly perceiving foreign speech utterances. Most previous studies that investigate the discrimination of non-native speech sounds have been done either in the laboratory or by comparing subjects exposed to a foreign language at different ages [5,9,17]. Here we present longitudinal data that have been collected in a natural environment showing the neurobiological manifestation of the second language learning. Native monolingual Finnish children were monitored for 6 months as they joined either a French school or a daycare center where French was spoken 50–90% of the time. French vowels were used to determine how quickly and accurately children developed cortical memory traces and discrimination abilities for non-native speech sounds during this period. This was done by recording mismatch negativity (MMN), the brain’s automatic electrophysiological response elicited by infrequent changes in auditory stimulation [3,10]. Not long ago, behavioral methods, such as
0304-3940/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 2) 00 26 9- 0
188
M. Cheour et al. / Neuroscience Letters 325 (2002) 187–190
forced-choice button pressing, were the only tools to investigate the development of speech perception in children. In recent years, however, MMN has been shown to be a promising tool for investigating the neural bases of speech perception. For example, it enables one to demonstrate the existence of language-specific memory traces in the adult human cortex [11]. Thereafter, it has been shown that these native-language memory traces develop in human infants between 6 and 12 months [2] or perhaps even earlier [4]. Our experimental group consisted of 17 monolingual 3– 6-year-old native Finnish children (four males) with no prior exposure to French. Similarly, the control group consisted of 17, 3–6-year-old native Finns (four males) enrolled in schools or daycare centers where only Finnish was spoken. Finnish children exposed to French were tested three times during the period of 4 months. Electrical recordings were obtained from the children upon joining a French school, following 6–8 weeks (mean, 53 days), and following 12–16 weeks (mean, 105 days), respectively. The control group, exposed to no French, was investigated twice: initially; and 20 weeks (mean, 150 days) after the first session. We recorded the brain responses of all subjects to the French vowels /i/ (‘standard’; P ¼ 0:8), /e/, and /1/ (‘deviants’; P ¼ 0:1 for each). Prior to the experiment, the adult, native Finnish listeners were asked to rate the category goodness of the stimuli used in this experiment among other French and Finnish vowels. None of these subjects perceived French stimuli as best exemplars of Finnish vowels. Thus, they were not considered as phoneme prototypes in Finnish (Fig. 1) [6,7]. This is why it was reasonable to expect that Finnish subjects with no prior exposure to French would possess no permanent memory traces for these vowels [2,4]. These vowels of 65 dB SPL, lasting 207 ms (with 10-ms rise and fall times) were presented through two loudspeakers located about 1.5 m from the subjects’ ears. The (onset-to-onset) inter-stimulus-interval was 907 ms. Speech sounds were generated using a male voice in a semi-synthetic method. This produces natural sounding vowel stimuli by exciting an artificial vocal-tract model with an excitation waveform extracted from a real utterance [1]. These stimuli were randomly presented in blocks of 200 vowels, 10–12 blocks per subject, in every session. During the experiment, subjects watched selfselected silent movies. Electroencephalogram (EEG, sampling rate at 250 Hz) was recorded with silver/silver chloride electrodes using a NeuroScan PC-3.0 based system. All experiments were conducted in an acoustically and electrically shielded room. Raw data were averaged and digitally filtered (bandpass, 0.1–30 Hz; 24 dB points) at eight scalp sites: F3; F4; C3; C4; T3; T4; P3; and P4. Eye movements were monitored with electrooculogram (EOG) electrodes attached below and at the outer canthus of the right eye. The right/ left-hemisphere electrodes were referred to the right/left mastoid, respectively. EOG was referenced to the right mastoids. Exclusion criteria included EEG exceeding 200
mV at any one electrode, epochs for the first three stimuli of each block, and responses to each standard immediately following a deviant stimulus. For any one child, there were at least 200 acceptable EEG epochs for each deviant stimulus in each session. The baseline for the waveforms was defined as the mean amplitude between 2100 and 0 ms relative to stimulus onset. The MMN amplitude was determined from the F3, F4, C3, C4, P3 and P4 electrodes (excluding T3 and T4 electrodes where MMN did not usually significantly differ from 0 mV). MMN was measured by separately computing difference waves that were obtained by subtracting the average standard-stimulus response from the average deviant-stimu-
Fig. 1. Stimuli used in the present experiments. The French vowels are (indicated by dark gray) interspersed amongst the Finnish vowels (light gray). The location of each vowel, /i/, /e/, and /1/, is shown in the F1–F2 space by upward arrows. The first value in the pair of each tag corresponds to F2 and the second to F1.
M. Cheour et al. / Neuroscience Letters 325 (2002) 187–190
lus response for each subject, electrode, and session. The mean amplitude, that is, the mean voltage over the 50-ms period time window, was centered at the latency of largest peak occurring between 150 and 350 ms. Using two-tailed t-tests, we tested the presence of MMN. The differences in the MMN amplitude and MMN latency between the groups, stimuli, session, and electrode were tested with four-way analysis of variance (ANOVA): Group (Experimental Subjects vs. Control Subjects) £ Session (First, Second, Last) £ Stimulus (/e/, /1/) £ Electrode (F3, F4, C3, C4, P3, P4). The session effect for the MMN amplitude and latency was separately tested for the Experimental Subjects with three-way ANOVA: Session (First, Second, Last) £ Stimulus (/e/ vs. /1/) £ Electrode (F3, F4, C3, C4, P3, P4), to monitor the dynamics of French-specific phoneme discrimination more precisely. Brain responses to French speech sounds are presented in Fig. 2. These responses clearly implicate the development of non-native memory traces in Finnish children who were exposed to French. The Session effect on MMN amplitude was tested separately for the experimental subjects using a three-way ANOVA: Session (First, Second, Last) £ Stimulus (e,1) £ Electrode (F3, F4, C3, C4, P3, P4). This analysis was highly significant (Fð2;32Þ ¼ 18:12, P , 0:000005). Post-hoc testing revealed that the MMN amplitude for the experimental group was significantly increased between the first and second sessions (P , 0:000046). Hence, in these children, the MMN ampli-
Fig. 2. Development of French-specific discrimination ability (indexed by MMN) in 3–6-year-old Finnish children for French phonemes /e/ and /1/ during the period of approximately 4 months. This figure presents the difference waves (obtained by subtracting the average standard /i/ stimulus response from the average deviant /e/ or deviant /1/ stimulus response) of 17 Finnish children who began to attend to French daycare are shown in the beginning of the test period, after a period of 1.5–2 months, and after a period of about 3–4 months. The difference waves of 17 monolingual Finnish speaking children who were not exposed to French are also shown in the beginning of the test period and after about 3–4 months.
189
tude increased strikingly in only a couple of months in a natural environment without any special training. No significant improvement in the MMN amplitude could be obtained between the second and the third sessions. The Session effect on MMN latency was tested exactly like that on MMN amplitude. This analysis was also significant (Fð2;32Þ ¼ 3:51, P , 0:041654). The least significant difference (LSD) post-hoc test for MMN latency demonstrated that this result was due to the MMN in the last session (204.54 ms) being significantly shorter than that in the beginning session (230.74 ms; P , 0:012712). A four-way ANOVA with the Group as a grouping factor revealed a significant Group £ Session interaction when comparing the experimental vs. control subjects (Fð1;32Þ ¼ 11:96, P , 0:001560). LSD post-hoc testing showed that the MMN amplitude was significantly larger in the experimental study group at the last session as compared with the first (P , 0:000008). Moreover, the MMN amplitude in the last session was significantly larger in the experimental group as compared with the control group (P , 0:000432). Importantly, no significant increase in the MMN amplitude was found in the control group between the first and second recordings, nor were there any significant MMN-amplitude differences between the experimental and control groups in the first session. Hence, the laboratory experiments themselves were not enough for the children to learn to discriminate foreign phonemes and, apparently, to develop memory traces for them. Numerous behavioral and event-related potential studies have demonstrated that normal language learning occurs primarily or exclusively within childhood. This so-called critical period hypothesis that was first introduced by Lenneberg [8] concerned only the first language acquisition, however. Whether this superior capacity of young children to acquire languages also involves second language is still under debate, however. When investigating late stages of learning, most of the studies show superiority of early learners for phonology, syntax, and grammar [12,13,15]. Interestingly, when early stages of learning have been investigated, most studies have reported that adults seem to be moving toward second language proficiency more quickly than children [13,15]. The present results implicate, however, that the memory representations of foreign phoneme system develop remarkably quickly in the cortex of young children, as compared with the results of previous adult studies [17]. Hence, although young children may have reduced some of their proficiency to differentiate foreign speech sounds once they have learned to differentiate their native language sounds better [2,4,7,14–16], they may regain their non-native contrast discrimination abilities more quickly than older children or adults. We hypothesize that this fast non-native contrast learning is likely to happen when youngsters are still in the process of learning to discriminate their native language sounds when concurrently their ability to differ-
190
M. Cheour et al. / Neuroscience Letters 325 (2002) 187–190
entiate non-native speech has not reached the nadir seen in older children and adults. In conclusion, the present data provide objective evidence to suggest that young children can learn to discriminate nonnative speech sounds in a natural environment within 2 months. Thus, although the MMN latency decreased gradually throughout all three sessions, the MMN amplitude significantly increased already between the first two (out of three) sessions, and did not show any growth after that. Hence, these results suggest that during a relatively short period of time, young Finnish children who were exposed to French developed the ability to discriminate different French phonemes as indexed by the MMN amplitude increase concurrently with a latency decrease. Those children not exposed to French exhibited no change in the MMN characteristics. The authors would like to thank Dr Veijo Vihanta for technical assistance, and Drs Amir Raz and Olli Aaltonen for constructive comments on the manuscript. Moreover, the authors wish to thank the Center for International Mobility (CIMO), The Jenni and Antti Wihuri Foundation, and The Academy of Finland for support. Finally, the authors are indebted to the children, parents, and personnel of Lyce´ e Franco-Finlandais d’Helsinki, Jardin d’enfants Jasmin, and the Ruskeasuo Daycare: Merci beaucoup! [1] Alku, P., Tiitinen, H. and Na¨ a¨ ta¨ nen, R., A method for generating natural-sounding speech stimuli for cognitive brain research, Clin. Neurophysiol., 110 (1999) 1329–1333. [2] Cheour, M., Ceponiene, R., Lehtokoski, A., Luuk, A., Allik, J., Alho, K. and Na¨ a¨ ta¨ nen, R., Development of language-specific phoneme representations in the infant brain, Nat. Neurosci., 1 (1998) 351–353. [3] Cheour, M., Leppa¨ nen, P. and Kraus, N., Mismatch negativity (MMN) as a tool for investigating auditory discrimination and sensory memory in infants and children, Clin. Neurophysiol., 111 (2000) 4–16. [4] Dehaene-Lambertz, G. and Baillet, S., A phonological representation in the infant brain, NeuroReport, 9 (1998) 1885– 1888.
[5] Kraus, N., McGee, T., Carrell, T., King, C., Tremblay, K. and Nicol, T., Central auditory system plasticity associated with speech discrimination training, J. Cogn. Neurosci., 7 (1995) 27–34. [6] Kuhl, P., Innate predispositions and the effects of experience in speech perception: the native language magnetic theory, In B. de Boyson-Bardies, S. de Schonen, P. Jusczyk, P. McNeilage and J. Morton (Eds.), Developmental Neurocognition: Speech and Face Processing in the First Year of Life, Kluwer Academic Publishers, Dordrecht, 1993, pp. 259–274. [7] Kuhl, P.K., Williams, K.A., Lacerda, W.F., Stevens, K.N. and Lindblom, B., Linguistic experiences alter phonetic perception in infants by 6 months of age, Science, 255 (1992) 606– 608. [8] Lenneberg, E., Biological Foundations of Language, Wiley, New York, 1967. [9] MacKain, K., Best, C. and Strange, W., Categorical perception of English /r/ and /l/ by Japanese bilinguals, Appl. Psycholinguist., 2 (1981) 369–390. [10] Na¨ a¨ ta¨ nen, R. and Escera, C., Mismatch negativity: clinical and other applications, Audiol. Neurootol., 5 (2000) 105– 110. [11] Na¨ a¨ ta¨ nen, R., Lehtokoski, A., Lennes, M., Cheour, M., Huotilainen, M., Iivonen, A., Vainio, M., Alku, P., Ilmoniemi, R.J., Luuk, A., Allik, J., Sinkkonen, J. and Alho, K., Languagespecific phoneme representations revealed by electric and magnetic brain responses, Nature, 358 (1997) 432–434. [12] Olson, L. and Samuels, S., The relationship between age and accuracy of foreign language pronunciation, J. Educ. Res., 66 (1973) 263–267. [13] Oyama, S., A sensitive period for the acquisition of a nonnative phonological system, J. Psycholinguist. Res., 5 (1976) 251–258. [14] Patkowski, M., The sensitive period for the acquisition of syntax in a second language, Lang. Learn., 30 (1980) 449–472. [15] Polka, L. and Werker, J.F., Developmental changes in perception of nonnative vowel contrasts, J. Exp. Psychol. Hum. Percept. Perform., 20 (1994) 421–435. [16] Werker, J.F. and Tees, R.C., Cross-language speech perception: evidence for perceptual reorganization during the first year of life, Infant Behav. Dev., 7 (1984) 49–63. [17] Winkler, I., Kujala, T., Tiitinen, H., Sivonen, P., Alku, P., Lehtokoski, A., Czigler, I., Cse´ pe, V., Ilmoniemi, R.J. and Na¨ a¨ ta¨ nen, R., Brain responses reveal the learning of foreign language phonemes, Psychophysiology, 36 (1999) 638–642.