Why a Song Gives You Chills
Musical pleasure has something almost nothing else in psychology has: a proper causal test. It also has an objection the causal test cannot answer, and a group of people who feel none of it.
Key takeaways
- In a 2013 Science study, activity in the nucleus accumbens while listening to unfamiliar music predicted what people would then pay for it at auction.
- The predictive signal was not activity in any single region but the strength of communication between auditory and valuation regions and the reward system.
- The proposed mechanism runs through prediction. Pleasure tracks the relationship between what a listener expected and what arrived, which is why a song wears out.
- A 2019 PNAS study gave people a dopamine precursor, a dopamine antagonist and a placebo. The two drugs moved musical pleasure and motivation in opposite directions, which is a rare causal test.
- Kent Berridge's wanting and liking distinction means that test cannot separate enjoying the music from wanting more of it, and people with specific musical anhedonia respond normally to money and not at all to music.
There is a moment in a piece of music you know well, usually a resolution you have heard a hundred times and could hum from memory, where something goes down the back of your neck and the hairs on your arm lift. It is involuntary. It is over in about a second. It happens to most people, it happens on the same passage every time, and it survives knowing exactly what is coming. The standard explanation is that it is a [reward-system event](https://thewhyweproject.com/articles/is-the-lizard-brain-real-in-marketing), which is true, well evidenced, and does much less explanatory work than it appears to. Explaining a feeling by pointing to an ancient part of the brain does even less.
Our position, stated first: the brain's reward machinery is genuinely involved in musical pleasure, and there is a proper causal test behind that claim, more than most psychological findings ever get. What the chemistry does not do is explain the experience. And the most interesting constraint on the whole account comes from people who feel none of it.
Salimpoor's auction
In 2013 Science published the study that turned this into a magazine subject. Participants heard music they had never heard before while being scanned, then bid real money for the tracks in an auction.
Activity in the mesolimbic striatum, the nucleus accumbens in particular, was the best predictor of what a person would pay. The auditory cortex, the amygdala and the ventromedial prefrontal regions all became more active when listening required a judgment, but their activity level predicted nothing about value. What predicted value was how strongly those regions were in communication with the accumbens as the music became more rewarding.
Valorie Salimpoor led that work with Robert Zatorre, and the framing matters: aesthetic reward turned up in an interaction between systems, not at an address.
Zatorre's account
Zatorre and Salimpoor set the account out in a review the following year, and it runs through prediction. Cortical loops store tonal patterns and generate expectations about what is coming. Subcortical systems assign reward and value. Pleasure, on this account, emerges where the two meet, which is to say it tracks the relationship between what you expected and what arrived, not the sound itself.
It explains a few things you already know. Why a song stops working somewhere around the fortieth listen. Why music with no surprises in it is pleasant and instantly forgettable. Why the passage that gets you is almost always a resolution, a delay or a break in a pattern the piece spent two minutes teaching you.
Why nobody can sell it
It also explains why the account is so hard to turn into advice. If the reward tracks a relationship between expectation and event, then it depends on everything you have already heard, which is a different set for every listener and changes with every listen. There is no property of the waveform to isolate. The streaming services have spent a decade and a great deal of money looking for one.
Ferreri's drugs
Claims of this kind are almost never tested causally, because testing them means altering somebody's neurochemistry and watching what moves. Laura Ferreri and colleagues did it, in PNAS in 2019. Each participant received a dopamine precursor, a dopamine antagonist and a placebo in separate sessions, while listening to music. The two drugs pushed in opposite directions. The precursor raised both the hedonic experience and the motivational response. The antagonist lowered both.
The paper concludes that dopamine plays a causal role in musical pleasure. Nothing else in the area comes close to it as evidence, and it is worth being clear why that matters. Almost everything else in this literature is a correlation between a brain measure and a self-report, taken at the same moment, from people who knew they were in a music study. Intervening on the chemistry and watching the report move in both directions is a different order of claim.
Berridge's objection
The objection is older than the music work and it concerns what dopamine does at all. Kent Berridge has argued for two decades that dopamine carries incentive salience, the wanting, and not the hedonic impact, the liking, and that the two dissociate under experiment. In one demonstration, mutant mice carrying elevated synaptic dopamine worked harder for a sweet reward while showing no increase in the affective reactions that index enjoying it.
Apply that to the drug study and an ambiguity appears that the design cannot settle. The precursor raised pleasure ratings and motivation together. It is not possible to tell from those data whether the drug made the music more enjoyable or made people want more of it and report the wanting as enjoyment. The result stands. The interpretation is carrying weight the evidence does not supply.
This is the same distinction that undoes most popular writing about dopamine, and it has a whole article of its own in this publication.
Mas-Herrero's exception
Some people get nothing from music, and they are not people with damaged reward systems. Ernest Mas-Herrero and colleagues documented the condition in Current Biology in 2014, and the paper's title states the finding: a dissociation between musical and monetary reward responses in specific musical anhedonia. Normal hearing. Normal music perception. Normal response to money. No musical pleasure.
Loui's wiring
Psyche Loui and colleagues then examined the wiring, comparing one severe case against a large control sample. Structural connectivity within the reward system was associated with how musically rewarded people were across that sample, and the connections between auditory and reward systems were unusual in the anhedonic case.
Be careful what that licenses, because it is easy to over-read and we nearly did. Connectivity findings of this kind are correlational, one case is one case, and specific musical anhedonia is rare, so none of it shows that musical pleasure is unusual or optional. What it does show is that the link between hearing something and being rewarded by it runs through anatomy that varies between people and can be absent entirely, which is not what a universal human response looks like from the inside.
Where that leaves it
The strong reading says musical pleasure is a dopamine phenomenon. It inherits Berridge's problem immediately, because it commits you to a chemical account of aesthetic experience that cannot yet distinguish loving a piece of music from wanting to hear it again.
The weak reading says the reward system is one participant in something the prediction machinery mostly drives. It keeps all the evidence and loses the headline, and it leaves you with no neurochemical account of why this song and not that one.
There is a position between them that the evidence probably supports better than either, and it is less quotable than both: dopamine is causally necessary and nowhere near sufficient, with wanting and liking both in play and no current method clean enough to weigh them separately in a human listening to music.
We hold the middle one. It is the least satisfying sentence in this article and it is the one the studies will bear.
Sources
Salimpoor, V. N., van den Bosch, I., Kovacevic, N., McIntosh, A. R., Dagher, A., & Zatorre, R. J. (2013). 'Interactions between the nucleus accumbens and auditory cortices predict music reward value.' Science, 340(6129), 216-219. Mesolimbic striatal activity, especially the nucleus accumbens, was the best predictor of what listeners would pay for previously unheard music in an auction; auditory, amygdala and ventromedial prefrontal activity did not predict reward value, which was instead predicted by increasing functional connectivity of those regions with the accumbens. doi:10.1126/science.1231059
Zatorre, R. J., & Salimpoor, V. N. (2013). 'From perception to pleasure: music and its neural substrates.' PNAS, 110(Suppl 2), 10430-10437. Proposes that pleasure in music arises from interactions between cortical loops that generate predictions and expectancies from sound patterns and subcortical systems responsible for reward and valuation. doi:10.1073/pnas.1301228110
Ferreri, L., Mas-Herrero, E., Zatorre, R. J., Ripollés, P., Gómez-Andrés, A., Alicart, H., Olivé, G., Marco-Pallarés, J., Antonijoan, R. M., Valle, M., Riba, J., & Rodríguez-Fornells, A. (2019). 'Dopamine modulates the reward experiences elicited by music.' PNAS, 116(9), 3793-3798. Within-participant administration of a dopamine precursor (levodopa), a dopamine antagonist (risperidone) and a placebo during music listening; levodopa increased the hedonic experience and music-related motivational responses against placebo, risperidone reduced both. doi:10.1073/pnas.1811878116
Berridge, K. C. (2007). 'The debate over dopamine's role in reward: the case for incentive salience.' Psychopharmacology, 191(3), 391-431. The argument that dopamine mediates incentive salience, the wanting, rather than hedonic impact, the liking. doi:10.1007/s00213-006-0578-x
Peciña, S., Cagniard, B., Berridge, K. C., Aldridge, J. W., & Zhuang, X. (2003). 'Hyperdopaminergic mutant mice have higher wanting but not liking for sweet rewards.' Journal of Neuroscience, 23(28), 9395-9402. Dopamine transporter knockdown mice with roughly seventy per cent elevated synaptic dopamine showed raised incentive motivation for a sweet reward without raised affective liking reactions to its taste. doi:10.1523/jneurosci.23-28-09395.2003
Mas-Herrero, E., Zatorre, R. J., Rodriguez-Fornells, A., & Marco-Pallarés, J. (2014). 'Dissociation between musical and monetary reward responses in specific musical anhedonia.' Current Biology, 24(6), 699-704. Cited here for the finding stated in its title; this publication has not obtained the abstract. doi:10.1016/j.cub.2014.01.068
Loui, P., Patterson, S., Sachs, M. E., Leung, Y., Zeng, T., & Przysinda, E. (2017). 'White matter correlates of musical anhedonia: implications for evolution of music.' Frontiers in Psychology, 8, 1664. Structural connectivity within the reward system predicted individual differences in musical reward across a large sample; connectivity between auditory and reward systems was atypical in a case of severe specific musical anhedonia. doi:10.3389/fpsyg.2017.01664