Cortical hemisphere preference and brainstem ear asymmetry reflect experience-dependent functional modulation of pitch

Abstract

Temporal attributes of pitch processing at cortical and subcortical levels are differentially weighted and well-coordinated. The question is whether language experience induces functional modulation of hemispheric preference complemented by brainstem ear symmetry for pitch processing. Brainstem frequency-following and cortical pitch responses were recorded concurrently from Mandarin and English participants. A Mandarin syllable with a rising pitch contour was presented to both ears with monaural stimulation. At the cortical level, left ear stimulation in the Chinese group revealed an experience-dependent response for pitch processing in the right hemisphere, consistent with a functional account. The English group revealed a contralateral hemisphere preference consistent with a structural account. At the brainstem level, Chinese participants showed a functional leftward ear asymmetry, whereas English were consistent with a structural account. Overall, language experience modulates both cortical hemispheric preference and brainstem ear asymmetry in a complementary manner to optimize processing of temporal attributes of pitch.

1. Introduction

Voice pitch plays an important role in the perception of speech, language and music. It provides an excellent window for studying experience-dependent shaping of pitch processing at both brainstem and cortical levels of a well-coordinated, hierarchical pitch processing network. There is compelling empirical evidence to support the notion that neural representation of pitch-relevant information at both brainstem and cortical levels of processing is influenced by one’s experience with language and/or music. Electrophysiological studies have demonstrated enhanced brainstem neural activity in individuals with long-term linguistic experience (Krishnan & Gandour, 2009, 2014; Krishnan et al., 2012; Krishnan, Xu, Gandour, & Cariani, 2005), and musical training (Bidelman, Gandour, & Krishnan, 2011; Bidelman & Krishnan, 2009; Musacchia, Sams, Skoe, & Kraus, 2007; Wong, Skoe, Russo, Dees, & Kraus, 2007). In the music domain, brainstem neural representations of pitch, timing and timbre cues and cortical response timing are shaped in a coordinated manner through corticofugal modulation of subcortical afferent circuitry (Musacchia, Strait, & Kraus, 2008). In the language domain, cortical and brainstem responses reveal different patterns of relative changes in pitch representation along salience (Krishnan, Gandour, & Suresh, 2016) and acceleration continua (Krishnan, Suresh, & Gandour, 2017). Such findings point to differential processing of temporal attributes of pitch at subcortical and cortical levels along the pitch processing hierarchy.

Concurrent recording of the frequency-following response (FFR) and the cortical pitch response (CPR)—representing pitch-relevant neural activity at brainstem and cortical levels, respectively—provide an effective physiologic window to evaluate the hierarchical organization of pitch processing along the auditory pathway. The FFR reflects sustained phase-locked activity in a population of neural elements within the rostral brainstem (see Chandrasekaran & Kraus, 2010; Krishnan, 2007, for reviews). It has revealed that experience-dependent plasticity enhances neural representation of pitch in native speakers of a tone language (Krishnan et al., 2009; Krishnan et al., 2005) and individuals with extensive music experience (Bidelman et al., 2011a; Wong et al., 2007). Pitch-relevant information preserved in the FFR is strongly correlated with perceptual pitch measures (Bidelman & Krishnan, 2011; Krishnan, Bidelman, & Gandour, 2010; Krishnan & Plack, 2011; Parbery-Clark, Skoe, Lam, & Kraus, 2009), suggesting that acoustic features relevant to pitch are already emerging in representations at the level of the brainstem. The pitch-specific CPR, a correlate of the MEG-derived pitch specific response (Gutschalk et al., 2002; Krumbholz, Patterson, Seither-Preisler, Lammertmann, & Lutkenhoner, 2003), is characterized by multiple transient components (Na:125–150 ms, Pb: 200–220, Nb: 270–285 ms) that index different temporal pitch attributes that are subject to language experience-dependent enhancement (for reviews, Gandour and Krishnan, 2016; Krishnan and Gandour, 2014). Source analysis of MEG-derived pitch response (Gutschalk et al., 2002; Krumbholz et al., 2003)—corroborated by human depth electrode recordings (Griffiths et al., 2010; Schonwiesner and Zatorre, 2008)—and source analysis of the EEG-derived Na (Bidelman & Grall, 2014) indicates that the pitch-specific CPR is localized to the anterolateral portion of Heschl’s gyrus, the putative site of pitch processing (Johnsrude et al., 2000; Penagos, Melcher, & Oxenham, 2004).

Using diotic presentation of language-relevant pitch stimuli, we have consistently demonstrated a language-experience dependent RH dominance of the CPR components (Krishnan, Gandour, Ananthakrishnan, & Vijayaraghavan, 2014; Krishnan, Gandour, & Suresh, 2014; Krishnan, Gandour, & Suresh, 2015b). When native pitch contours of Mandarin have been presented in either speech or nonspeech contexts, the magnitude of CPR components in the RH of Mandarin speakers are larger than English speakers. Of relevance to the neurobiology of language, we infer that CPR components provide a metric of experience-dependent selective recruitment of the RH for optimal processing of pitch-relevant information, presumably at the early sensory level of processing in the auditory cortex.

As far as we know, there is only one earlier study to report on ear asymmetries in the representation of FFRs that were elicited by monaural stimulation of both the left ear (LE) and right ear (RE) in response to a prototypical representation of Mandarin Tone 2 (T2, high rising) and a nonnative variant (T2’) in which the pitch contour was a mirror image of T2 with the same starting and ending frequencies (Krishnan, Gandour, Ananthakrishnan, Bidelman, & Smalt, 2011). A late region of interest in T2—characterized by a fast-rising pitch contour—is known to contribute most importantly to its tonal recognition (Whalen & Xu, 1992). With monaural RE stimulation, FFRs in this late region were larger in native T2 relative to T2’—the latter characterized by a slow-falling pitch contour (Krishnan et al., 2011, Fig. 1). This finding suggests that pitch encoding in the rostral brainstem is sensitive to rapid changes in native pitch contours. In this study, however, only brainstem responses were evaluated; only Mandarin speakers participated. Both T2 and T2’ pitch contours were overlaid on a mid-front rounded vowel [œ] that does not occur in Mandarin.

Figure 1.

(A) Stimulus paradigm consists of three segments: 250 ms pitch (red) preceded (750 ms) and followed (235 ms) by noise (black). (B) Temporal characteristics of the pitch segment show robust periodicity of Tone 2. (C) Spectrogram shows lower harmonics (h1-h5) of f0 with corresponding rising contour of Tone 2. The f0 rising contour (white) is superimposed on the speech segment.

Both structural and functional asymmetries along the auditory pathway influence hemispheric preferences for speech processing in the cerebral cortex. It is well established that structural asymmetry of the ascending auditory pathways (ipsilateral vs contralateral) is characterized by a dominant contralateral pathway. Consequently, this structural asymmetry produces greater activation in the auditory cortex contralateral to the ear of stimulation as measured by EEG (Khosla et al., 2003; Picton et al., 1999; Woldorff et al., 1999); MEG (Gutschalk & Steinmann, 2015, Jäncke, Wüstenberg, Schulze, & Heinze, 2002; Königs & Gutschalk, 2012; Langers, van Dijk, & Backes, 2005; Schönwiesner et al., 2007); and fMRI (Gutschalk & Steinmann, 2015; Jäncke et al., 2002; Langers et al., 2005; Schönwiesner et al., 2007). It is this structural dominance of contralateral pathways in auditory processing that is thought to drive the contralateral pattern of structural hemispheric preference.

In contrast, functional asymmetries are processing asymmetries induced by certain spectrotemporal features of auditory signal and/or their linguistic function (Meyer, 2008; Poeppel, Idsardi, & van Wassenhove, 2008; Zatorre & Gandour, 2008) that can modulate both ear asymmetries and hemispheric preference. Thus, they can potentially override or amplify the structural asymmetry (Hornickel, Skoe, & Kraus, 2009; Schönwiesner, et al., 2007). For example, ear asymmetries at the auditory brainstem (inferior colliculus) level, and hemispheric preferences at the cortical level can be modulated by experience shaped linguistic function to alter the fixed contralateral structural asymmetry (Krishnan & Gandour, 2014; Krishnan & Gandour, 2009). We are not aware of any other published reports specifically examining hemispheric preference and/or brainstem ear asymmetries for pitch processing using monaural stimulation.

The overarching theoretical issue pertains to functional modulation of hemispheric preference by language experience and the extent to which this modulation is driven by pitch-relevant functional ear asymmetries at the level of the midbrain. The current experiment involves a direct comparison of native speakers of Mandarin and monolingual English speakers by concurrent recording of both cortical CPR and brainstem FFR responses with monaural stimulation of both the LE and RE. The major question is to determine whether the RH preference for cortical pitch processing in Mandarin speakers is wholly confined to CPR responses or driven by pitch-relevant neural activity as reflected in scalp-recorded brainstem FFR responses. Our experimental results are brought to bear on two specific questions. At the level of the cerebral cortex, to what extent does the pattern of hemispheric lateralization per language group support a structural account and/or functional account? At the level of the brainstem, do ear asymmetry patterns—also influenced by language experience—provide evidence for the notion that brainstem ear asymmetries may be a driving influence of cortical hemispheric preferences? We hypothesize that pitch at both cortical (hemispheric preference) and brainstem (ear asymmetry) levels of processing will be modulated by language experience. We expect the Chinese group to provide evidence in support of functional asymmetries at both cortical and brainstem levels of processing. We expect hemispheric preference of the English group to be congruent with the structural asymmetry account, i.e., responses will be larger in the hemisphere contralateral to the ear of stimulation.

2. Methods

2.1. Participants

Twenty-six native speakers of Mandarin Chinese (6 male, 7 female) and American English (7 male, 6 female) were recruited from Purdue University to participate in the experiment. All participants exhibited normal hearing sensitivity at audiometric frequencies between 500 and 4000 Hz. They were closely matched in age (Chinese, 24.1 ± 3.1 years; English, 23.0 ± 2.6), formal education (Chinese, 17.2 ± 2.0 years; English, 15.9 ± 2.8), and strongly right-handed (Chinese, 91.7 ± 13.8%; English, 97.3 ± 4.8) as measured by the Edinburgh Handedness Inventory (Oldfield, 1971). All Chinese participants were born and raised in mainland China. None had received formal instruction in English before the age of nine (12.6 ± 3.8 years). Self-ratings of their English language proficiency on a 7-point Likert-type scale ranging from 1 (very poor) to 7 (native-like) for spoken English language proficiency were, on average, 4.75 ± 1.2 (Li, Sepanski, & Zhao, 2006). Their daily usage of Mandarin and English, in order, was reported to be 72% and 28%. As determined by a music history questionnaire (Gandour et al., 2007), all participants were closely matched in terms of years of musical training (Chinese, 0.56 ± 0.86 years; English, 0.83 ± 1.9) on any combination of instruments, except for one English participant with eight years of musical training. Each participant was paid and gave informed consent in conformity with the 2013 World Medical Association Declaration of Helsinki and compliance with a protocol approved by the Institutional Review Board of Purdue University.

2.2. Stimulus

A monosyllabic, speech stimulus in Mandarin Chinese was chosen to represent a lexical tone with time-varying, rising pitch contour (/yi²/‘aunt’) in citation form. Fig. 1A displays the three segments of the stimulus paradigm used to acquire cortical and brainstem responses: (i) a preceding 750 ms noise segment; (ii) a 250 ms pitch segment (red); and (iii) a following 235 ms noise segment. The evoked response to the entire three-segment (noise-pitch-noise) stimulus is characterized by obligatory components (P1/N1) corresponding to the onset of energy in the precursor noise segment of the stimulus followed by several transient CPR components (Na, Pb, Nb) occurring after the onset of the pitch-eliciting segment of the stimulus and an offset component (Po) following the offset of the last noise segment in the stimulus. The noise precursor segment enabled us to isolate the pitch-specific response from the obligatory onset response (cf. Krumbholz et al., 2003; Krishnan et al., 2014). Similarly, the noise segment that follows the pitch segment serves to disentangle pitch offset response from sound offset response. The pitch segment (Fig. 1B) shows robust periodicity and shorter periods with a rapid increase in rising f0. Fig. 1C displays a waveform with rising spectral bands corresponding to lower f0 harmonics (h1-h5) and a dynamic, rising f0 contour (white) superimposed on the speech segment.

The overall RMS level of each segment was equated such that a continuous stimulus was created with no discernible difference in amplitude among the three segments. To avoid evoked potentials to amplitude change, the transitions from noise-to-pitch, and from pitch-to-noise segments were crossfaded using 7ms cos² ramps to ensure no amplitude differences at transitions between the three segments. Thus, overall stimulus duration of 1221 ms; overall level of the stimulus was 80 dB SPL. Monaural left and right stimulus conditions were presented through magnetically-shielded tubal insert earphones (ER-3A; Etymotic Research, Elk Grove Village, IL, USA) at a repetition rate of 0.56/s (ISI = 550 ms) using a fixed onset polarity (rarefaction) for the pitch-eliciting stimulus. The stimulus presentation order was randomized both within and across participants. All stimuli were generated and played out using an auditory evoked potential system (SmartEP, Intelligent Hearing Systems; Miami, FL, USA).

2.3.1. EEG acquisition to extract cortical and brainstem responses

Participants reclined comfortably in an electro-acoustically shielded booth. They were instructed to relax, refrain from extraneous body movement to minimize myogenic artifacts, and to ignore the stimuli as they watched a silent video of their choice (minus subtitles) throughout the recording session. The EEG was acquired continuously (5000 Hz sampling rate; 0.3–2500 Hz analog bandpass) through the ASA-Lab EEG system (ANT Inc., The Netherlands) using a 32-channel amplifier (REFA8–32, TMS International BV) and WaveGuard electrode cap (ANT Inc., The Netherlands) with 32-shielded sintered Ag/AgCl electrodes configured in the standard 10–20-montage. The high sampling rate of 5 kHz was necessary to recover the brainstem frequency-following responses in addition to the relatively slower cortical pitch components. While EEG acquisition included 9 electrode locations (Fpz, AFz, Fz, F3, F4, Cz, T7, T8, M1, M2), response extraction and analysis were limited to T7 and T8 since the focus of this study was to compare cortical pitch responses over the left and right auditory cortices. The AFz electrode served as the common ground. The common average of all connected unipolar electrode inputs served as the default reference for the REFA8–32 amplifier. An additional bipolar channel with one electrode placed lateral to the outer canthus of the left eye and another electrode placed above the left eye was used to monitor artifacts introduced by ocular activity. Inter-electrode impedances were maintained below 10 kΩ. For each stimulus condition, EEGs were acquired in two sessions of 750 sweeps each. The experimental protocol took about two hours to complete.

2.3.2. Extraction of cortical and brainstem evoked response

To extract the CPR components, EEG data were first down-sampled from 5000 Hz to 1024 Hz. They were then digitally band-pass filtered (2–25 Hz, Butterworth zero phase shift filter with 24 dB/octave rejection rate) to enhance the transient components and minimize the sustained component. Sweeps containing electrical activity exceeding ±75 mV, including ocular artifacts were rejected automatically. Subsequently, averaging was performed on unipolar electrode locations T7 and T8 using a 500 ms analysis epoch (from 650 ms to 1150 ms) including the 100 ms pre-stimulus baseline. This analysis epoch was chosen to capture only the CPR components Na, Pb, and Nb. Using the common reference, we were able to compare CPR components at the right (T8) and left (T7) temporal electrode sites to evaluate hemispheric preference. Our aim here was to characterize the relative difference in the pitch-related neural activity over the widely separated left and right temporal electrode sites. Indeed, our previous crosslanguage CPR studies have consistently demonstrated robust differences in CPR neural activity over the T7 and T8 electrode sites that reflect a functional, experience-dependent RH preference (Krishnan, Gandour, & Suresh, 2014; Krishnan, Gandour, & Suresh, 2015a; Krishnan et al., 2015b).

To extract the appreciably smaller amplitude brainstem FFR, EEG data were digitally band-pass filtered (75–1500 Hz, Butterworth zero phase filters with 24 dB/Octave rejection rate). Sweeps containing electrical activity exceeding ±40 mV were rejected automatically. Subsequently, time domain averaging was performed on three different re-referenced electrode montages (FPz-linked mastoids, Fz-linked mastoids, and Cz-linked mastoids) over an analysis window of 270 ms (from 743 to 1013 ms; 743 represents the onset of the pitch segment). Each FFR waveform represents the grand average of the FFRs derived from the three electrode montages in response to a total of 1500 sweeps obtained in two recordings of 750 sweeps each. These three channels were chosen because of the prominence of the FFRs in frontocentral locations—typical configurations used to record FFRs. The rationale for averaging across channels was to improve signal SNR.

2.3.3. Analysis of cortical and brainstem responses

The cortical pitch response is characterized by three transient components (Na,135–170 ms; Pb, 210–240 ms; and Nb, 270–310 ms) occurring after the onset of the pitch-eliciting segment. The peak-to-peak amplitude of Na-Pb was measured manually to characterize the effects of ear stimulation on the hemispheric preference of the pitch relevant activity in the auditory cortex. To improve the accuracy and consistency of manual peak picking for amplitude measurements from individual averaged responses, the grand-averaged response was used as a reference to facilitate response identification. Also, the amplitude measurement procedure was repeated independently by two members in the laboratory. They exhibited high interjudge reliability (95%). In addition to the analysis of the Na-Pb peak-to-peak amplitude recorded at the T7 and T8 electrode locations with stimulation of each ear, a measure of laterality index (LI) was obtained to more clearly examine the hemispheric laterality of the pitch response (LI= Na-Pb_T8 − Na-Pb_T7/Na-Pb_T8 + Na-Pb_T7). A value of zero indicates no hemispheric preference, whereas a positive or negative value signifies a preference for processing in the right or left hemisphere, respectively (cf. Krishnan et al., 2011).

For the sustained brainstem FFR reflecting pitch-relevant neural activity, neural periodicity strength was quantified by measuring the magnitude of the spectral peak at f0 for two temporal windows of the response (initial slow-changing pitch section (10–120 ms), and the other over the fast-changing pitch section (120–230 ms) from each response waveform. In native speakers of Mandarin, crosslanguage studies at the brainstem level have consistently shown that experience-dependent enhancement of pitch-relevant information is more robust over rapidly-accelerating vis-à-vis slowly-changing pitch sections (Krishnan, Swaminathan, & Gandour, 2009). In addition, this increased sensitivity to sections characterized by rapid changes in pitch is maintained even in severely degraded stimuli (Krishnan, Gandour, & Bidelman, 2010). Thus, experience-dependent brainstem pitch mechanisms are particularly sensitive to temporal attributes of pitch that provide high perceptual saliency. We therefore wish to determine whether neural representation of slow and rapid pitch sections differentially influence functional ear asymmetry at the brainstem level.

The spectrum of each response region was computed by obtaining the Fast Fourier Transform (FFT) of the time-windowed version of its temporal waveform (Gaussian window, 1 Hz resolution). For each subject, the magnitude of f0 was measured as the peak in the FFT relative to the noise floor. The noise floor was estimated by averaging the amplitude values over a 50 Hz band immediately below and above the f0 peak in the spectrum. All FFR data analyses were performed using custom routines coded in Matlab16 (The Math Works, Inc., Natick, MA, USA). To enhance the visualization of spectrotemporal changes in the FFR, a joint time-frequency analysis using a continuous wavelet transform was performed on the grand average waveforms.

2.4. Statistical Analysis

At the Fz-linked T7 and T8 electrode sites, a three-way (ear × hemisphere × group) mixed model ANOVA (SAS®; SAS Institute, Inc., Cary, NC, USA) was performed on peak-to-peak amplitude (Na–Pb). Monaural ear (LE, RE) and electrode site (T7 [left], T8 [right]) were treated as within-subject factors. Group (Chinese, English) served as the between-subjects factor; subjects nested within group as the random factor. A subsequent two-way (ear × hemisphere) ANOVA was performed on peak-to-peak amplitude by group. At the cortical level, a two-way ANOVA (ear, group) was also performed to evaluate a hemisphere laterality index derived from monaural stimulation. At the level of the brainstem, a two-way (ear × group) ANOVA was performed on log-transformed values of neural pitch strength (magnitude of the normalized autocorrelation peak) derived from FFRs over slow (10–120 ms) and fast (120–230 ms) acceleration sections of the Mandarin monosyllable. Post hoc multiple comparisons were adjusted with a Bonferroni significance level set at α = 0.05. Where appropriate, partial eta-squared (${\eta }_p^2$) values were reported to indicate effect size. Assumptions of normality of distribution and homogeneity of variance were confirmed before statistical inference.

3. Results

3.1. CPR waveforms (T7/T8)

Fig. 2 displays grand average CPR waveforms evoked by T2 for Chinese (red) and English (blue) groups per monaural ear of stimulation (LE, RE) and temporal electrode sites (T7, dashed line; T8, solid line) over the left and right hemispheres, respectively. CPR components are identifiable in both groups. For Chinese, CPR waveforms evoked by LE stimulation point to a RH preference (larger CPR responses over the right temporal electrode site (T8). No appreciable differences between hemispheres are observed in response to RE stimulation. For English, on the other hand, a hemispheric preference is observed that depends crucially on the ear of monaural stimulation. CPR waveforms evoked by RE stimulation yields a larger response over the left hemisphere (T7 > T8), whereas just the opposite occurs for CPR waveforms evoked by LE stimulation (T8 > T7). English effects reflect a structural asymmetry insomuch as they are consistent with the well-established contralateral dominance of auditory pathways; i.e., larger responses are expected in the hemisphere contralateral to the stimulated ear. For Chinese, on the other hand, LE stimulation shows a robust RH preference whereas no hemispheric preference is evoked by RE stimulation. Such a pattern reflects an experience-dependent functional effect that recruits optimal hemispheric resources for pitch processing regardless of the contralateral dominance of the auditory pathways.

Figure 2.

Grand average time-waveforms of CPR components for the two language groups (Chinese, red; English, blue) recorded at electrode sites T7 (dashed) and T8 (solid) in response to monaural (LE, RE) stimulation of T2. As evidenced by monaural LE stimulation, CPR waveforms of the Chinese group—compared to the English— display a stronger rightward asymmetry in favor of the right temporal electrode site (T8). CPR, cortical pitch response; LE, left ear; RE, right ear; T7, temporal electrode site in the left hemisphere; T8, temporal electrode site in the right hemisphere.

3.2. CPR magnitude of Na-Pb per ear, hemisphere and group

Fig. 3 shows group (C, E) means of Na–Pb peak-to-peak amplitude elicited by T2 at T7/T8 electrode sites as a function of monaural stimulation (LE, RE) and hemisphere (LH, RH). A 3-way ANOVA revealed significant interaction effects between hemisphere × group (F_{1, 24} = 19.66, p = 0.0002, ${\eta }_p^2$ = 0.078) and ear × hemisphere (F_{1, 24} = 73.96, p = 0.0003, ${\eta }_p^2$ = 0.263). Pooling across ears, the hemisphere × group interaction revealed that response amplitude was larger in the RH (T8 > T7) for the Chinese group only. Pooling across groups, the ear × hemisphere interaction revealed that a LE stimulus elicits a RH advantage in peak-to-peak amplitude (RH > LH), whereas a RE stimulus elicits the opposite (LH > RH). Given the absence of a 3-way interaction (F_{1, 94} = 0.07, p = 0.7891), pooling across ears or groups made it impossible to extract experience-dependent functional effects associated with ear and hemisphere.

Figure 3.

Mean peak-to-peak amplitude of CPR component Na–Pb extracted from T7/T8 in the temporal lobe as a function of monaural stimulation (LE, RE) and hemisphere (LH, RH). The amplitude stimulated by the LE is larger in the RH (T8 > T7) for the Chinese group only. RE stimulation evoked no differences between language groups. Error bars = ±1 SE. LH, left hemisphere; RH, right hemisphere. CLH, Chinese LH; ELH, English LH; CRH, Chinese RH; ERH, English RH. Error bars represent ±1 SE.

Because mean amplitude data for the English group clearly show larger amplitude in the contralateral hemisphere (Fig. 3, right panel, RE)—indicating a hemispheric preference that simply reflects the contralateral dominance of the auditory pathway—we chose to further break the model down to 2-way ANOVAs (ear × hemisphere) by group to disentangle structural and functional effects. For Chinese, ANOVA revealed a significant main effect of hemisphere (F_{1, 48} = 14.72, p = 0.0004, ${\eta }_p^2$ = 0.235) and interaction of ear × hemisphere (F_{1, 48} = 8.55, p = 0.0053, ${\eta }_p^2$ = 0.151). In the RH, the simple effect comparison showed that LE displayed greater amplitude than the RE (t₄₈ = 2.70, p = 0.0095; Fig. 3, left panel). For English, ANOVA also revealed a significant interaction of ear × hemisphere (F_{1, 48} = 15.66, p = 0.0002, ${\eta }_p^2$ = 0.246). In contrast to the Chinese, however, simple effect comparisons exposed not only that RE amplitude was larger than LE in the LH (t₄₈ = −3.45, p = 0.0012; Fig. 3, right panel), but also that LE amplitude was larger than RE in the RH (t₄₈ = 2.14, p = 0.0373; Fig. 3, left panel). These findings suggest that hemispheric preferences of the English group (LE: RH > LH; RE: LH > RH) reveal a structural asymmetry, exactly what is expected from a dominant, contralateral auditory pathway. In contrast, the hemispheric preferences of the Chinese group suggest a language-dependent functional asymmetry that optimizes the processing of linguistically-relevant pitch information (LE: RH > LH).

3.3. Laterality index of hemispheric preference per ear and group

Functional ear asymmetry reflected by CPRs of Chinese and English varies depending on the linguistic status of a rising pitch contour in response to monaural LE and RE stimulation. ANOVA revealed significant main effects of ear (F_{1, 24} = 62.25, p < 0.0001, ${\eta }_p^2$ = 0.747) and group (F_{1, 24} = 30.85, p < 0.0001, ${\eta }_p^2$ = 0.562). For the Chinese group, the structural ear effect (LE > RE) reveals a larger degree of ear asymmetry in the RH (T8) with no significant ear asymmetry in the LH (RE = LE). In contrast, the English group shows an essentially equal contralateral effect (LH: RE > LE); RH: LE > RE). The experience-dependent group effect (C > E) is consistent with the linguistic status of the rising pitch contour (T2). These results reinforce differences in hemispheric preference between the two groups (cf. 3.2). The English group shows a laterality pattern wholly consistent with a structural account. In contrast, the Chinese expose a laterality pattern that points to a functional modulation overriding structural asymmetry of the auditory pathways.

3.4. FFR waveforms and spectrograms

Fig. 5 shows (A) grand-averaged waveforms and (B) spectrograms of the FFR component elicited by monaural stimulation (LE, RE) per language group (Chinese, left column; English, right column). (A) FFR waveforms for the Chinese appear to be larger in amplitude for LE stimulation, particularly over the faster accelerating pitch section. Waveforms for LE and RE stimulation in the English group are indistinguishable. Consistent with the waveform data, (B) spectrograms show stronger spectral bands of pitch-relevant neural activity at f0, 2f0, and 3f0 for monaural LE stimulation in the Chinese group (bottom left panel). In contrast, the spectral band in the English group (bottom right panel) appears to be relatively more robust with RE stimulation (top right panel). Overall, Chinese FFRs appear to be more robust than English in response to LE stimulation, particularly for the faster-accelerating pitch section (gray bar, 120–230 ms).

Figure 5.

Language group comparisons of brainstem FFR responses displayed in waveforms (A) and spectrograms (B). In contrast to the English group (A), FFR waveforms reveal that the Chinese show more robust periodicity encoding of T2 under monaural LE stimulation (dotted) relative to RE (solid). (B) Under monaural LE stimulation (bottom left panel), Chinese show more robust, clearer harmonics compared to English in the fast acceleration section (gray bar: 120–230 ms) compared to the slow acceleration section (white bar: 10–120 ms). FFR, frequency following response.

3.5. FFR f0 magnitude in slow and fast sections of T2 per ear and group

Fig. 6 shows the mean FFR f0 magnitude within the slow-changing (top row) and fast-changing (bottom row) acceleration sections of T2 per stimulus ear (LE, RE) and language group (C, E). Within the fast section of T2, a two-way (ear × group) ANOVA on log-transformed values of neural pitch strength yielded a significant ear × group interaction effect (F_{1, 24} = 13.94, p < 0.0010, ${\eta }_p^2$ = 0.367). In the Chinese group, the stimulus RE evoked a larger f0 magnitude than the LE; and in the RE, the Chinese exhibited larger f0 magnitude than English. Within the slow-changing pitch section, main effects of group (F_{1, 24} = 1.41, p = 0.2466) and ear (F_{1, 24} = 0.10, p = 0.7552) failed to reach significance. The group × ear interaction also failed to reach a robust level of significance (F_{1, 24} = 4.04, p = 0.0558). For the Chinese group, leftward ear asymmetry points to a RH preference at the cortical level that may, at least in part, be driven by experience-enhanced subcortical inputs from the right inferior colliculus. Otherwise, FFR responses do not show structural and/or functional ear asymmetries in pitch-relevant neural activity.

Figure 6.

Group comparisons of mean neural pitch strength derived from FFR responses within slow- (left panels) and fast-acceleration (right panels) regions of T2. As measured by the magnitude of the spectral component at f0 in the fast-acceleration region, Chinese pitch strength is larger than English when stimulated by the LE only (right panels). Error bars represent ±1 SE.

4. Discussion

Our findings demonstrate that patterns of hemispheric preference of pitch-relevant neural activity differ between Chinese and English groups as a function of the stimulus ear. At the cortical level, LE stimulation in the Chinese group evokes an experience-dependent RH preference consistent with a functional account. The English group, on the other hand, reveals a hemispheric preference consistent with a contralateral structural account. LE and RE stimulation evokes a RH and LH preference, respectively. At the brainstem level, the Chinese group shows an experience-dependent enhancement of neural representation of pitch-relevant information with LE stimulation as compared to RE stimulation, whereas the English group does not elicit an ear asymmetry. Taken together, these findings suggest that language-experience modulates structural and functional asymmetries at both brainstem and cortical levels to optimize processing of behaviorally-relevant attributes of pitch.

4.1. Auditory cortex

4.11. Structural versus functional shaping of hemispheric preference for pitch processing

4.1.1. Fixed structural determinant of hemispheric preference

For the English group, pitch-specific neural activity in the auditory cortex was dominant in the hemisphere contralateral to the ear of stimulation (Fig. 3). This finding is consistent with previous monaural studies of speech/nonspeech stimuli. They show greater activation in the auditory cortex contralateral to the ear of stimulation across different measures of neural activity: e.g., EEG (Khosla et al., 2003; Naatanen & Picton, 1987; Picton et al., 1999; Ponton et al., 2001; Woldorff et al., 1999); MEG (Gutschalk & Steinmann, 2015; Jäncke, Wüstenberg, Schulze, & Heinze, 2002; Kanno, Nakasato, Fujiwara, & Yoshimoto, 1996; Königs & Gutschalk, 2012; Langers, van Dijk, & Backes, 2005; Ross, Herdman, & Pantev, 2005; Schönwiesner, Krumbholz, Rübsamen, Fink, & von Cramon, 2007); and fMRI (Gutschalk & Steinmann, 2015; Jäncke et al., 2002; Langers et al., 2005; Schönwiesner et al., 2007). Collectively, these results reflect the structural dominance of contralateral pathways in auditory processing that drive this pattern of hemispheric preference.

4.1.2. Functional modulation of hemispheric preference

In sharp contrast to the English group, the Chinese showed an enhanced RH preference with LE stimulation, but essentially symmetrical response with RE stimulation (Fig. 3). Both observations are congruent with a language experience-dependent functional modulation of hemispheric preference for pitch processing. We find no published reports that can be directly compared with our results for a cortical pitch-specific response. Nevertheless, a few studies have employed EEG (Hine & Debener, 2007; Picton et al., 1999; Yamazaki et al., 2018); MEG (Gabriel et al., 2004; Ross et al., 2005); and fMRI (Scheffler, Bilecen, Schmid, Tschopp, & Seelig, 1998). Across studies, a RH preference was reported for processing pure tones presented to the LE, and bilateral symmetrical activation when presented to the RE. Though their pattern of hemispheric preference is similar to our findings herein, we point out that cortical components (P1, N1, P2) evaluated with EEG and MEG and BOLD responses with fMRI do not specifically reflect the pitch-relevant neural activity (Gutschalk, Patterson, Scherg, Uppenkamp, & Rupp, 2004; Lutkenhoner, Seither-Preisler, & Seither, 2006; Yrttiaho, Tiitinen, Alku, Miettinen, & May, 2010).Exploiting CPRs, we have consistently demonstrated a language experience-dependent, functional RH preference for pitch processing of dynamic pitch contours presented diotically in speech (Krishnan, Suresh, & Gandour, 2019; Suresh, Krishnan, & Gandour, 2017), nonspeech (Krishnan et al., 2016) and both speech/nonspeech contexts (Krishnan, Gandour, & Suresh, 2014). Our findings of RH preference for processing of pitch-relevant information in the Chinese group converge with a large body of research literature that highlights the role of the RH in processing both linguistic and musical pitch (Friederici, 2009; Friederici & Alter, 2004; Gandour & Krishnan, 2016; Wong, 2002; Zatorre, Belin, & Penhune, 2002; Zatorre & Gandour, 2008).

Specific sources of EEG-derived CPR components are not known. However, source analyses of MEG-derived pitch onset response (Andermann, Günther, Patterson, & Rupp, 2021; Gutschalk, Patterson, Rupp, Uppenkamp, & Scherg, 2002; Gutschalk et al., 2004) and human depth electrode recordings (T.D. Griffiths et al., 2010; Schonwiesner & Zatorre, 2008) indicate that the cortical pitch response is localized to the anterolateral portion of Heschl’s gyrus. (T.D. Griffiths et al., 2010; Schonwiesner & Zatorre, 2008). It has been designated as the putative site of pitch processing (Bendor & Wang, 2005; Johnsrude, Penhune, & Zatorre, 2000; Patterson, Uppenkamp, Johnsrude, & Griffiths, 2002; Schonwiesner & Zatorre, 2008). Thus, we can infer that our CPR component likely reflects pitch-relevant activity anterolateral to the primary auditory cortex. This inference is consistent with the view that auditory processing is essentially symmetrical in the core (primary auditory cortex); asymmetries emerge beyond the primary auditory cortex (Poeppel, 2003; Poeppel, Idsardi, & van Wassenhove, 2008).

4.2. Auditory brainstem

4.2.1. Structural and functional shaping of ear asymmetry for pitch processing

The FFR reflects phase-locked activity in a population of neural elements in the rostral brainstem that preserves information relevant to the pitch of the stimulus (Krishnan, 2007; Krishnan & Gandour, 2014). The inferior colliculus (IC) at the midbrain level is thought to be the primary generator source for the EEG-derived FFR (Bidelman, 2015, 2018; Chandrasekaran & Kraus, 2010; Krishnan, 2007; White-Schwoch, Anderson, Krizman, Nicol, & Kraus, 2019). At the level of the auditory cortex, the MEG-derived FFR has also exposed a contribution for f0, albeit much weaker at frequencies above 100 Hz (Coffey, Herholz, Chepesiuk, Baillet, & Zatorre, 2016).

The English group did not show a significant ear asymmetry in f0 magnitude for either the slow or the fast portion of the pitch contour. This absence of ear asymmetry suggests that IC responses from either side of the brainstem are similar in magnitude and are dominated by contralateral stimulation. These results are consistent with neuroimaging recordings of both cortical and subcortical neural activity elicited by monaural stimulation (Boyen, de Kleine, van Dijk, & Langers, 2014; Gutschalk & Steinmann, 2015; Langers et al., 2005; Melcher, Sigalovsky, Guinan, & Levine, 2000; Schönwiesner et al., 2007; Suzuki et al., 2002). As predicted from the well-established structural account of contralateral auditory pathways, the IC on each side in the English group responds more strongly when the contralateral ear is stimulated.

In stark contrast, results from the Chinese group cannot be explained by the structural account. A robust leftward ear asymmetry is observed especially for the fast-accelerating portion of the pitch contour. This leftward asymmetry suggests that pitch-relevant neural activity is relatively greater in the right IC with stimulation of the contralateral left ear. Thus, long-term language experience not only enhances pitch-relevant activity but also induces a leftward functional ear asymmetry by modulating the activation strength in each IC. As native speakers of Mandarin, enhanced sensitivity to rapidly-changing sections of pitch contours is an important perceptual, time-varying dimension (Krishnan & Gandour, 2009; Krishnan, Gandour, & Bidelman, 2012; Krishnan et al., 2015a; Krishnan et al., 2017b). At the cortical level, the RH preference for processing pitch-relevant information in the Chinese group may reflect, at least in part, experience-dependent enhanced input from the right IC. Such experience-dependent enhancement likely reflects the result of reorganization of pitch mechanisms in the IC instantiated during language acquisition or learning and maintained by the top-down influence of the corticocollicular pathway.

To compare current findings with our earlier study on ear asymmetries of FFRs (Krishnan et al., 2011), we exercise caution because of substantial differences in experimental design, participants and stimuli. Krishnan et al. (2011) observed a strong leftward asymmetry for the later slow-accelerating portion of a nonnative pitch contour (T2’); a relatively weak rightward asymmetry for the native pitch contour (T2). It appears that stimulus-specific differences in functional ear asymmetries may reflect differential sensitivity to temporal and spectral properties of the stimulus at the brainstem level that in turn may contribute to the hemispheric preference at the cortical level. Herein, only the Chinese group displayed a robust leftward ear asymmetry for the fast-accelerating portion of T2. Across studies (Gandour & Krishnan, 2016; Krishnan & Gandour, 2014; 2017), it has become clear that sensitivity to rapid changes in pitch represents crucial, perceptually-relevant features of lexical tones.

4.3. Top-down, bottom-up interactions and local mechanisms contribute to cortical hemispheric preference and brainstem ear asymmetry in pitch processing

Empirical evidence suggests that pitch-relevant information is processed in a well-coordinated, distributed hierarchical processing network that encompasses multiple cortical brain regions involving top-down and bottom-up connections—primary auditory cortex (sensory) as well as non-primary areas beyond HG (extrasensory) anteriorly and posteriorly (T. D. Griffiths & Hall, 2012; T.D. Griffiths et al., 2010; Puschmann, Uppenkamp, Kollmeier, & Thiel, 2010). At the cortical level, pitch-relevant neural activity is more labile and spatiotemporally broader (Warren & Griffiths, 2003; Winer, Miller, Lee, & Schreiner, 2005; Zatorre & Belin, 2001). In the brainstem, spectrotemporal representation of pitch information reflects sustained neural phase-locking to pitch-relevant periodicities in both stimulus envelope and temporal fine structure. Its activity represents a series of transient temporal neural events synchronized to certain temporal attributes of the pitch contour. Thus, experience-dependent enhancement of pitch representation represents a well-coordinated functional modulation of early sensory pitch mechanisms in the brainstem (ear asymmetry) and in the auditory cortex (hemispheric preference), plus extrasensory processes at higher hierarchical levels within the auditory cortex that interact with the brainstem.

Long-term experience shapes this adaptive process by engaging top-down connections including corticocortical (from HG to extrasensory cortical areas) and corticocollicular (from HG to brainstem). Both cortical and subcortical structures enhance neural responses to behaviourally-relevant attributes of the stimulus. These enhanced representations from brainstem pitch mechanisms are functionally reorganized by top-down influence during the critical period of language acquisition (Keuroghlian & Knudsen, 2007; Kral & Eggermont, 2007). Top-down influence mediated by the corticocollicular system is implicated in the enhancement of pitch-relevant information resulting from long-term linguistic experience (Krishnan & Gandour, 2014; 2017; Zatorre & Gandour, 2008). Selective differences in the strength of this top-down influence can alter the balance in strength and fidelity of pitch-relevant neural activity between left and right ICs. Corticocollicular inputs are primarily ipsilateral. We propose that reorganized brainstem pitch mechanisms, at least in part, may provide a bottom-up drive to the functional hemispheric pattern at the cortical level (Krishnan et al., 2012). In addition to top-down and bottom-up influences, experience-reorganized pitch mechanisms intrinsic to the brainstem and the auditory cortex could also contribute to modulation of ear asymmetry and hemispheric preference.

To advance future research on the interplay between the brainstem and auditory cortex, it will be necessary to employ (1) fMRI to optimize the spatial resolution of specific brain regions engaged for processing linguistic pitch and (2) EEG/MEG in order to optimize temporal resolution of pitch-relevant neural activity from the brainstem to auditory cortex. Both tools will be necessary to gain a better understanding of the spatiotemporal organization of the pitch processing network and how it is shaped by language experience.

5. Conclusion

The evidence presented herein supports the view that long-term language experience imparts a stimulus-dependent functional hemispheric preference at the cortical level as well as a functional ear asymmetry in the brainstem to enable optimal neural representation of behaviorally-relevant temporal attributes of pitch. We theorize that long-term effects are likely mediated by a well-coordinated top-down, local, and bottom-up interplay between sensory and extrasensory brain regions beyond the primary auditory cortex; and between the auditory cortex and midbrain by modulating the strength of top-down and bottom-up connections. The presence of an experience-dependent functional asymmetry at the brainstem level is complementary to cortical hemispheric preference. This interplay suggests that hemispheric preference for pitch processing seen at the cortical level may already be emerging in the brainstem.

Figure 4.

As measured by monaural ear stimulation (LE, RE) and hemisphere (LH, RH), a laterality index of functional hemisphere asymmetry reflects differences in the status of pitch in a tonal vis-à-vis nontonal language. Under LE or RE stimulation, Chinese laterality indices of hemisphere preference reveal much stronger functional activation than English in the RH. In contrast, English laterality indices reflect structural brain anatomy: monaural LE and RE stimulate activation in the RH and LH, respectively. Laterality index: LI=[Na-Pb_T8-Na-Pb_T7]/Na-Pb_T8+Na-Pb_T7]. Error bars represent ±1 SE.

HIGHLIGHTS.

• LE stimulation of Chinese evokes RH preference consistent with a functional account

• English contralateral hemispheric preference is in line with a structural account

• At the brainstem level, Chinese pitch FFRs are greater with LE than RE stimulation

• The English group, however, does not elicit an ear asymmetry at level of brainstem

• Functional effects reflect modulation of top-down, local, and bottom-up influences