Picking up on the first lecture's questioning of the adaptive value of music (Darwin, 1871), the last lecture focused on music as a vector of social cohesion through the emotion it arouses.
We began by situating music among activities considered to generate pleasure, based on an empirical classification of secular categories of pleasure (Dubé, Le Bel, 2003), which highlights the differentiated character of music as an emotional pleasure forming a sub-category of its own in the hierarchy of pleasures. We then discussed, illustrating with sound examples, music's capacity to arouse various forms of emotion (gaiety, sadness, anxiety, anger, serenity, etc.), a striking capacity that justifies Kant's definition: "Music is the language of emotions". We then turned our attention to the neural bases of musical emotion, illuminated over the last twenty years by functional brain imaging studies. Two articles caught our attention: one, published in 1999 by Alan Evans' group, concerns the identification by positron emission tomography of brain centers involved in the perception of dissonance (Blood et al., 1999); and the other, published in 2009 by Robert Zatorre's group, concerns the identification by functional MRI of brain regions activated by intense musical emotion (Blood, Zatorre, 2001; Salimpoor et al., 2011). A major conclusion emerges from these studies: the perception of intense musical emotions involves "reward neuronal circuits" and notably implicates dopamine. The structural organization and mode of response of these circuits was the subject of a detailed presentation.
Finally, we briefly discussed the "reward error prediction" hypothesis (according to which behavioral adjustments involved in learning are based on a representation of the difference between the reward expected following an action and the reward actually obtained), a general adaptive learning strategy that would govern the functioning of reward circuits. We mentioned the possible application of this paradigm to the development and resolution of musical "tensions", proposed by Koelsch on the example of Bach's cantatas (Koelsch, 2014). Functional imaging data thus suggest an unexpected link between musical listening and learning, which brings us back to Darwin's original question.
