Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Nov 27;9(1):17700.
doi: 10.1038/s41598-019-54060-x.

Scaling Behaviour in Music and Cortical Dynamics Interplay to Mediate Music Listening Pleasure

Affiliations
Free PMC article

Scaling Behaviour in Music and Cortical Dynamics Interplay to Mediate Music Listening Pleasure

Ana Filipa Teixeira Borges et al. Sci Rep. .
Free PMC article

Abstract

The pleasure of music listening regulates daily behaviour and promotes rehabilitation in healthcare. Human behaviour emerges from the modulation of spontaneous timely coordinated neuronal networks. Too little is known about the physical properties and neurophysiological underpinnings of music to understand its perception, its health benefit and to deploy personalized or standardized music-therapy. Prior studies revealed how macroscopic neuronal and music patterns scale with frequency according to a 1/fα relationship, where a is the scaling exponent. Here, we examine how this hallmark in music and neuronal dynamics relate to pleasure. Using electroencephalography, electrocardiography and behavioural data in healthy subjects, we show that music listening decreases the scaling exponent of neuronal activity and-in temporal areas-this change is linked to pleasure. Default-state scaling exponents of the most pleased individuals were higher and approached those found in music loudness fluctuations. Furthermore, the scaling in selective regions and timescales and the average heart rate were largely proportional to the scaling of the melody. The scaling behaviour of heartbeat and neuronal fluctuations were associated during music listening. Our results point to a 1/f resonance between brain and music and a temporal rescaling of neuronal activity in the temporal cortex as mechanisms underlying music appreciation.

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
Scheme of the investigation. (a) Stimuli example—excerpt of the sound signal from the Sonata no. 62 Allegro (Haydn) and partial score. (b) (Left) Correspondent loudness, pitch and rhythm time series representing respectively the audio envelope, successive dominant note-frequency changes and note intervals. (Right) Approximate linear relationship between log t and the average fluctuation log F(t) for t ∈ [3, 15]s reveals a fractal scaling characteristic of how the musical features unfold in time. (c) Scaling exponents obtained for each of the music dimensions form a gradient from near randomness (0.5) to smooth and highly correlated fluctuations (>1). (d) Broadband EEG trace and the different timescales (Empirical Mode Decomposition) analysed. (e) Channels (dots) and regions (colour-coded) analysed (cf. Methods for details and Table 1 for a glossary of the main experiment variables). (f) Heartbeat and interbeat intervals (NN) obtained from ECG signals. Music sheet is a courtesy of Musopen. Brain and heart images in (d,f) by Sinisa Maric and Marcus Hartmann, under Pixabay license.
Figure 2
Figure 2
Music-induced decrease in the scaling of the envelope fluctuations in most frequency ranges relative to the rest, default-state. Head surface maps of the scaling exponent of neuronal components (γh − δ) during rest (a) during a music listening task (b) and of the difference between the latter two. (c) The decreases are accentuated in the parietal and occipital regions. Channels marked with dark blue dots display significant differences (p < 0.05, uncorrected), purple dots signal significant differences after FDR correction (q = 0.05), minimum p = 0.005 (α), p = 0.011 (β)).
Figure 3
Figure 3
Behaviour induced by music listening and its relationship to musical features. (a) Individual pleasure, familiarity and concentration ratings for each participant (x-axis) and piece (y-axis) and their average (s¯) per individual/piece (dark colour) and SD (shade in light colours) (lateral plots). (b) Relationship between s¯piece and the scaling exponent (αmusic) of each piece for all music dimensions; only s¯piece of Familiarity shows a significant association with the dynamics of the stimuli musical dimensions.
Figure 4
Figure 4
The scaling behaviour of neuronal activity during baseline and its induced change during listening capture the individual pleasure experienced with the music. (a) Head surface mappings of the associations between Pleasure (s¯individual) and the scaling exponents of the components (γh − δ) during baseline (rest), during a music listening task, and with the induced change in scaling between the latter. The channels marked in dark blue indicate a nominally significant correlation (Spearman coefficient rs, p < 0.05), the channels in purple show significant association after FDR correction (q = 0.1, minimum p-values at this FDR spanned between 2.82 × 10−4 and 0.025). (b) Scatter plots of the highlighted channels (green star in (c) exemplify the individual values (n = 28), a locally weighted regression line was added to aid visualising the relationship and the shadowed area represent the confidence interval.
Figure 5
Figure 5
Correlation analysis reveals a link between the scaling behaviour of the music and the scaling of neuronal activity in the α, β and γh-components. (a) Headplots of the correlation between the scaling exponent of the pitch series (αpitch) and the average scaling exhibited by the multiscale neuronal activity (α¯piece for γh − δ) for the music pieces; the dark blue dots indicate a channel with significant correlation (Spearman coefficient rs, p < 0.05), the purple ones show significant association after FDR correction (q = 0.2, minimum p = 0.017 (γh, α) and p = 0.006 (β)). (b) Scatterplots portray the association between the scaling of neuronal activity in the γh, β and α components, for the channels highlighted with a green star in (a) and the scaling of pitch successions; the error bars denote the standard error of the mean.
Figure 6
Figure 6
Heart rate dependency on musical dimensions of the music stimuli and 1/f resonance between brain and heart dynamics. (a) Average heart rate (AVNN) during music listening shows a significant increase relative to baseline. (b) Individual AVNN depends on music piece as shown by Violin plots with overlaid boxplots, the box limits show the 25th and 75th percentile together with the medians, whiskers extend 1.5 times the interquartile range (IQR) and the coloured polygons representing density estimates of AVNN. (c) Relationship between the scaling behaviour of the musical dimensions and the average AVNN for each piece; error bars indicate standard error of the mean. (d) The self-similarity in individual heart rate variability (α1) is associated with the neuronal scaling of both high- (γh, γ1 and β activity) and low-frequency oscillations (δ). (e) Correlation between brain and α1 is strengthened during music relative to baseline suggesting music may facilitate an 1/f resonance between brain and heart. Brain and heart images in (d) by Sinisa Maric and Marcus Hartmann. Clef and chair images in (e) by rawpixel and Pettycon. All under Pixabay license.

Similar articles

See all similar articles

References

    1. McDermott J, Hauser M. The origins of music: Innateness, uniqueness, and evolution. Music. Percept. 2005;23:29–59. doi: 10.1525/mp.2005.23.1.29. - DOI
    1. Janata P. Brain networks that track musical structure. Ann NY Acad Sci. 2005;1060:111–124. doi: 10.1196/annals.1360.008. - DOI - PubMed
    1. Levitin DJ, Tirovolas AK. Current advances in the cognitive neuroscience of music. Ann NY Acad Sci. 2009;1156:211–31. doi: 10.1111/j.1749-6632.2009.04417.x. - DOI - PubMed
    1. Koelsch S. Brain correlates of music-evoked emotions. Nat Rev Neurosci. 2014;15:170–180. doi: 10.1038/nrn3666. - DOI - PubMed
    1. Salimpoor VN, Benovoy M, Larcher K, Dagher A, Zatorre RJ. Anatomically distinct dopamine release during anticipation and experience of peak emotion to music. Nat Neurosci. 2011;14:257–262. doi: 10.1038/nn.2726. - DOI - PubMed
Feedback