THE NEUROSCIENCES AND MUSIC II I: DISORDERS AND PLASTICITY

Musical Experience Promotes SubcorticalEfficiency in Processing Emotional

Vocal SoundsDana L. Strait,a,b Nina Kraus,b,c,d,e Erika Skoe,b,c

and Richard Ashleya,f

aBienen School of Music, bAuditory Neuroscience Laboratory, Departments ofcCommunication Sciences, dNeurobiology and Physiology, eOtolaryngology, and

f Cognitive Science at Northwestern University, Evanston, Illinois, USA

Tounderstandhowmusical experience influences subcortical processing of emotionallysalient sounds, we recorded brain stem potentials to affective vocal sounds. Our resultssuggest that auditory expertise engenders subcortical auditory processing efficiencythat is intricately connected with acoustic features important for the communicationof emotion. This establishes a subcortical role in the auditory processing of emotionalcues, providing the first biological evidence for musicians enhanced perception of vo-cally expressed emotion.

Key words: auditory brain stem response; brain; emotion; musicians; plasticity

Perception of emotion in speech and musicrelies on shared acoustic and neural mech-anisms,1 suggesting that extensive experiencein one domain may lend perceptual benefitsto the other. Accordingly, musical experienceenhances perceptual sensitivity to emotion inspeech.2,3

Musical experience also shapes subcorticalsound transcription (see the paper by Krauset al. in this volume4). Because of its fidelityin representing spectral and temporal acous-tic features, the auditory brainstem response(ABR) provides a mechanism for exploringmusicians subcortical sensitivity to acousticfeatures contributing to the perception of emo-tion.5 ABRs may also advance our understand-ing of subcortical function in the processing ofemotionally charged auditory cues.

To better understand musical experiencesinfluence on neural processing of affective

Address for correspondence: Dana L. Strait, Auditory NeuroscienceLaboratory, Northwestern University, 2240 Campus Drive, Evanston, IL60208. Voice: 847-491-2465. Dana.Strait@u.northwestern.edu

bFor more information about the Auditory Neuroscience Labora-tory and the work presented herein, please visit http://www.brainvolts.northwestern.edu.

speech-related events, we recorded ABRs to anemotionally charged vocal soundan infantsunhappy cry. We aimed to provide a biologi-cal basis for musicians enhanced perception ofemotion in speech by investigating the contri-bution of subcortical mechanisms to processingvocally communicated emotion.

Methods

Subjects were 30 normal-hearing adultsaged 1935 years, with musicians grouped ac-cording to two criteria: Musicians by OnsetAge (MusAge, n = 11, began musical train-ing 7 yr) and Musicians by Years (MusYrs,n = 15, 10 years of consistent musical train-ing).Nonmusicians (NonMus)were categorizedby failure to meet those criteria.

ABRswere elicited by a complex vocal soundderived from an emotional auditory scene fromthe Center for the Study of Emotions and At-tention (University of Florida, Gainesville; file278).6 For further recording and data process-ing parameters see Strait et al.7

The Neurosciences and Music III: Disorders and Plasticity: Ann. N.Y. Acad. Sci. 1169: 209213 (2009).doi: 10.1111/j.1749-6632.2009.04864.x c 2009 New York Academy of Sciences.

209

210 Annals of the New York Academy of Sciences

Figure 1. Stimulus and grand average response waveforms. The boxes correspond tothe periodic and complex portions, respectively. Figures adapted from Strait et al.7 withpermission. (In color in Annals online.)

We divided the stimulus into two segments(periodic and complex) that were acousti-cally contrastive and internally consistent. Neu-ral responses to the complex region resulted ina series of peaks that aligned with amplitudebursts in the stimulus (Fig. 1). Peak latencies andamplitudes were recorded for the largest andmost replicable peaks, whereas rectified meanamplitudes (RMAs) provided a gross measureof response amplitude.

We extracted spectral components of theneural responses using the fast Fourier trans-form. Amplitudes were recorded for spectralpeaks corresponding to the stimuluss funda-mental frequency and spectral componentsrepresenting the maxima within given fre-quency ranges (F0: 280305; H2: 470485; H3:570585Hz), interpreted to correspond to rep-resentations of pitch (F0) and timbre (H2, H3).

Results

Regression analyses supported the groupingof musicians into two subgroups: whereas theMusYrs grouping was predicted best by ampli-tude and latencymeasures (P < 0.001;MusAgeP < 0.06), MusAge was predicted best by fre-quency encoding measures (P < 0.05; MusYrsP < 0.40).

Both MusYrs and MusAge musicians re-sponses exhibited enhancements and econ-

omy connected with time-varying acousticfeatures of the stimulus. Enhancements (largertime- and frequency-domain response magni-tudes) were apparent in musicians responsesto the complex portion of the sound, witheconomy (smaller amplitudes) seen in their re-sponses to the periodic portion.

Figure 1 shows both the stimulus and aver-age responses for MusYrs and NonMus, withboxes defining acoustically distinct sections.The first portion of the stimulus is character-ized by greater periodicity, whereas the secondis characterized by greater complexity (devia-tion in pitch from the F0: 0.5% in the periodicportion and 2.7% in the complex, harmonicjitter: 1.27% and 2.03%, and signal-to-noiseratio: 13.46 and 6.82 dB).

The response RMAs are plotted inFigure 2A, for which an ANOVA revealed aninteraction between group and response por-tion (F = 6.04, P < 0.02). This indicatesthat MusYrs and NonMus have differentiatedresponses to the two sections. Whereas theMusYrs within-group RMAs differ betweenthe periodic and complex portions (t = 4.70,P < 0.0001), NonMus do not (t = 0.025,P < 0.99). Peak amplitudes confirmMusYrs tohave larger responses to the complex portionthan NonMus (peak 1: F = 10.25, P < 0.003;peak 2: F = 4.88, P < 0.03).

Peak amplitudes within the complex por-tion correlated with years of musical practice

Strait et al.: Musical Experience and Neural Efficiency 211

Figure 2. Interactions between responses to theperiodic and complex stimulus portions. (A) Group portion interaction between MusYrs and NonMusresponse RMAs and (B) MusAge and NonMus F0.(C) MusAge also show enhanced encoding of fre-quencies above the F0 in responses to the complexportion. (In color in Annals online.)

across all individuals with musical experience(n = 20; Fig. 3; peak 1: r = 0.454, P < 0.04).Timing-related enhancements were specificallyobserved in MusYrs responses, even as early asthe onset (Fig. 4; onset peak:F = 4.82,P < 0.04;peak 1: F = 8.72, P < 0.006). Earlier latenciesin musicians reflect faster synchronous neuralresponses to the timing characteristics of thestimulus.

Compared to NonMus, MusAge showed en-hanced representations of frequencies impor-tant for the perception of pitch and timbre in re-sponses to the complex stimulus portion. These

differences were connected with acoustic char-acteristics of the auditory input. AnANOVAre-vealed an interaction between group and F0 en-coding for the two response portions (F = 7.04,P < 0.01), indicating that MusAge and Non-Mus have differentiated F0 responses to the twosections. MusAge showed smaller representa-tions of the F0 in responses to the periodic por-tion of the stimulus thanNonMus, but larger F0amplitudes to the complex portion (Fig. 2B; pe-riodic: F = 7.04, P < 0.01; complex: F = 5.04,P < 0.03). F0 amplitudes in responses to thecomplex portion correlated with the age thatmusical training began, with individuals whobegan at an earlier age showing larger F0 rep-resentations (Fig. 3: r = 0.500, P < 0.03).MusAge also demonstrated enhanced repre-sentations of spectral peaks H2 and H3 in re-sponses to the complex portion of the stim-ulus compared with NonMus (Fig. 2C; H2:F = 7.95, P < 0.01; H3: F = 6.16, P < 0.02).

Conclusions

We suggest that musical experience has per-vasive effects on the auditory system, resultingin fine neural tuning to acoustic features im-portant for vocal communication. Musical ex-perience sharpens subcortical auditory process-ing, with the behavioral relevance and relativecomplexity of the stimulus playing a prominentrole in subcortical malleability. This sharpen-ing is likely mediated by the corticofugal sys-tem known to shape receptive field propertiesin the primary auditory cortex, thalamus, andauditory brain stem.8

The interplay between response enhance-ment and economy may engender musiciansenhanced perceptual capabilities of emotionalcues in speech.2,3 Musicians responses to theperiodic stimulus section demonstrate smallermagnitudes, reflecting neural efficiency (re-cruitment of fewer resources) in processing sim-pler acoustic features. Such observations havebeen interpreted to reflect domain-general ex-pertise.9 MusYrs faster responses indicate that

212 Annals of the New York Academy of Sciences

Figure 3. Correlations between musical experience and subcortical responsecharacteristics.

Figure 4. Peak latencies for MusYrs and NonMus. (In color in Annals online.)

long-term musical experience contributes toenhanced subcortical timing.

Musics spectral and temporal complexitymakes it a powerful tool for engendering neuralplasticity during optimal periods of auditory de-velopment.Our data suggest thatmusical train-ing prior to the age of seven has an impact onsubcortical frequency representation, whereastiming-related enhancements are affected byduration of practice. This indicates an optimalperiod for the development of pitch and timbreencoding strategies. Deprivation studies pro-vide evidence for optimal periods in the acquisi-tion of tonotopic maps in the primary auditorycortex, with exposure to spectrally and tem-porally complex auditory input necessary forauditory evoked-response development.10,11

There does not seem to be a similar optimalperiod for the development of neural repre-sentations of timing. The discrepancy betweentiming- and frequency-related effects aligns

with evidence of distinct subcortical encodingmechanisms for different features of acousticstimuli.9,12

Fast subcortical-limbic pathways for emo-tional processing are established in the visualsystem,13 with analogous pathways in the au-ditory system indicated more recently.1416 Byshowing subcortical involvement in the encod-ing of acoustic features foundational to thevocal communication of emotion, our resultsprovide an advance in the study of human per-ception of biological states.

In conclusion, we found that musical train-ing engenders subcortical efficiency that isconnected with acoustic features integral tothe communication of emotion. Thus weprovide biological evidence for musiciansperceptual enhancements in detecting vo-cally expressed emotion and reveal profoundinteraction between cognitive and sensoryprocesses.4,9,1618

Strait et al.: Musical Experience and Neural Efficiency 213

Acknowledgements

This work was funded in part by NSF Grant0544846 and a Graduate Research Grant fromNorthwestern University. The authors wouldlike to thank Jennifer Krizman for assistance inpreparing final figures for this manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

References

1. Nair, D. et al. 2002. Expressive timing and perceptionof emotion in music: an fMRI study. In Proceedings ofthe 7th International Conference on Music Perception andCognition. C. Stevens et al., Eds.: 627630. CausalProductions. Adelaide, Australia.

2. Dmitrieva, E.S. et al. 2006. Ontogenetic featuresof the psychophysiological mechanisms of percep-tion of the emotional component of speech in musi-cally gifted children. Neurosci. Behav. Physiol. 36: 5362.

3. Thompson, W.F., E.G. Schellenberg & G. Husain.2004. Decoding speech prosody: do music lessonshelp? Emotion 4: 4664.

4. Kraus, N. et al. 2009. Experience-induced malleabil-ity in neural encoding of pitch, timbre and timing:implications for language and music. Ann. N. Y. Acad.Sci. The Neurosciences and Music III: Disorders andPlasticity. 1169: 543557.

5. Juslin, P.N. & P. Laukka. 2003. Communication ofemotions in vocal expression andmusic performance:different channels, same code? Psychol. Bull. 129:770814.

6. Bradley, M.M. & P.J. Lang. 2007. The InternationalAffective Digitized Sounds (2nd ed.; IADS-2): Affective Rat-ings of Sounds and Instruction Manual. NIMHCenter forthe Study of Emotion and Attention. University ofFlorida. Gainesville, FL.

7. Strait, D.L. et al. 2009.Musical experience and neuralefficiency: effects of training on subcortical processingof vocal expressions of emotion. Eur. J. Neurosci. 29:661668.

8. Luo, F. et al. 2008. Corticofugal modulation of initialsound processing in the brain. J. Neurosci. 28: 1161511621.

9. Grabner, R.H., A.C. Neubauer & E. Stern. 2006.Superior performance and neural efficiency: the im-pact of intelligence and expertise. Brain Res. Bull. 69:422439.

10. Trainor, L.J. 2005. Are there critical periods for mu-sical development? Dev. Psychobiol. 46: 262278.

11. Fallon, J.B., D.R.F. Irvine & R.K. Shepherd. 2002.Cochlear implants and brain plasticity. Br. Med. Bull.66: 321330.

12. Kraus, N. & T. Nicol. 2005. Brainstem origins forcortical whatand where pathways in the auditorysystem. Trends Neurosci. 28: 176181.

13. Johnson, M.H. 2005. Subcortical face processing.Nat. Rev. Neurosci. 6: 766774.

14. Bigand, E., S. Filipic & P. Lalitte. 2005. The timecourse of emotional responses to music. Ann. N. Y.Acad. Sci. 1060: 429437.

15. Brandao, M.L. et al. 2005. Gabaergic regulation ofthe neural organization of fear in the midbrain tec-tum. Neurosci. Biobehav. Rev. 29: 1299311.

16. Marsh, R.A. et al. 2002. Projection to the inferiorcolliculus from the basal nucleus of the amygdala.J. Neurosci. 22: 1044910460.

17. Banai, K. et al. 2009. Reading and subcortical au-ditory function. Cereb. Cortex. In press. doi: 10.1093/cercor/bhp024.

18. Tzounopoulos, T. & N. Kraus. 2009. Learning to en-code timing: mechanisms of plasticity in the auditorybrainstem. Neuron 62: 463469.