Musical chills study claims new view of reward system…not so!
An already influential paper by Salimpoor et al. in Nature Neuroscience, suggests that we need to rethink how the reward system works, most notably the neural distinction between expectation and experience of rewards. By using PET scanning to probe the binding of dopamine in the (subcortical) brain, and using fMRI to learn more about the brain regions involved, Salimpoor had subjects listen to music that were very pleasurable, and indeed produced “shivers down their spines”, or musical chills.
The researchers report something going against what most other studies seem to suggest: that both the expectation and the experience of reward is found in the striatum. Simplified, the results from brain activation and receptor (raclopride) binding analyses suggest that the anticipation phase engages the caudate nucleus while experiencing the chills (i.e. reward of listening to favoured music) engaged the ventral striatum (aka nucleus accumbens, although neither the resolution in PET or 3T fMRI can determine with certainty that it really is the nucleus accumbens). Dopamine function – as measured by raclopride binding – when analysed through regions defined by the fMRI analysis, further supported this finding.
In other words: dopamine function seems to be engaged in both anticipation and experience of rewards, and there was a functional divide between anticipation (caudate) and experience (ventral striatum). In Salimpoor’s own words:
And the main results from that paper:
But the findings are at odds with many other studies, that have already reported two things. First, the ventral striatum is repeatedly implicated in expectation, anticipation and prediction. Second, many studies of pleasure experience that now so often report orbitofrontal cortex (OfC) activation, do not report finding ventral striatum activation.
So here is an alternate explanation of the Salimpoor et al results:
First, the PET binding used – raclopride – is especially targeted at the striatum and subcortical regions, while it has low/no affinity in other important regions, such the OfC. IOW, we cannot learn about the engagement of the OfC in this task, nor the role of dopamine in this region.
Second, the fMRI results only focus on the striatum, and Salimpoor et al. do not report activation in other regions. Here, it would be good to have confirmed that experiencing chills generate activation in the OfC. If not, why would it be different from so many other studies? It would serve as a good benchmark.
Finally, there is an alternate explanation (that is still interesting). Building on the research by Kent Berridge, it has been suggested that the (ventral) striatum is involved in driving motivated behaviours (“wanting”). Thus, during the musical chills epochs in the Salimpoor study, one can contend that the ventral striatum activation reported actually represents the motivation to “keep listening” to that rewarding tune. Thus, the activation and binding found only relates to the motivated behaviour, and NOT the experience itself. Here, again, it would have been nice to see some OfC activation. Basically, it is a fallacy to equate all activation during the muscal chills epoch with the musical chills experience.
Nevertheless, it is interesting to note that anticipation/expectation was more represented in the caudate nucleus, while other studies tend to suggest that ventral striatum. The caudate involvement seems to resonate well with some of my own recent fMRI analyses and upcoming articles. In addition, many of those other studies tend to use the term “ventral striatum” rather loosely, leading to confirmation bias. But that’s another story.