Author Archives: Christopher Prugh

It’s Not Just a Phase, Mom- How Sad Music is More Enjoyable than Happy Music

In light of the recent music festival and the fact that music can be heard regularly throughout the Métro and streets of Paris, I decided to look into the effects of music on the brain. In my experience I’ve always found that, in a public setting such as the Métro, I prefer to hear people playing calming instrumental music, like an acoustic guitar, rather than an entire band playing an upbeat song. While this may just be based on personal opinion, I wanted to know if there was a neurobiological process that governed this reaction. Obviously, examining every sentiment or bias towards music is beyond the scope of one or two studies, so I refined my question: what brain processes drive us to form opinions of music that is perceptually happy or sad?

Figure 1: Map of all the Fête de la Musique major events in Paris- there’s music for all tastes!

My first inquiry led me to an article by Brattico et al. (2011) that aimed to show the difference in activity of certain brain regions from music that is happy or sad, and with or without lyrics. They hypothesized that songs with lyrics will activate the left fronto-temporal language network, while songs without lyrics would activate right-hemispheric brain structures. Also, they expected to observe activation of left-hemispheric auditory areas by happy music (which is richer in fast transitions) and of right-hemispheric areas by sad music (with slower “attacks” and tempos). They used fifteen subjects who were told to bring in 16 familiar music pieces: four sad and four happy pieces from favorite music, and four sad and four happy pieces from disliked or even hated music. The music, within the four categories, was then computer-analyzed to average the attack slope (sharpness of musical events, for example, most percussion would result in a high attack slope) and spectral centroid (brightness and frequency balance of the music, similar to timbre), as well as tempo and mode (major or minor chord quality) (Figure 2).

Figure 2: Differences of attacks slope, spectral centroid, tempo, and mode in the four music categories.

The subjects listened to 18 second excerpts of the music they brought while their brain activity was monitored. After the excerpts, they were asked if they liked or disliked the music, as well as if they thought the music was happy or sad. Their findings of the difference in activated areas between categories are shown in Figure 3. Although the researchers described in detail what each activated brain region meant in correlation with its usual information processing, I’ll only mention a few interesting points that relate to my original question:

  • Sad music induced activity in the right caudate head and the left thalamus. Interestingly, the left thalamus is one of only a few brain structures that is involved in processing sadness in faces, suggesting a link between emotions evoked by visual or auditory stimuli.
  • Also, sad music led to activity in both the subcortical stratial region, which is involved in judging musical and physical beauty.
  • Happy music without lyrics more strongly activated structures associated with perception and recognition of basic emotions, like the left anterior cingulate cortex and the right insula, than happy music with lyrics.
  • However, sad music led to wider brain activity during music with lyrics than without, such as emotion-related areas like the right claustrum and left medial frontal gyrus.

Figure 3: Differences in activation of brain regions due to music emotion and presence of lyrics. For example, “Lyrics > Instrumental” signifies the regions that were activated in lyrical music, but not in intrumental. ITG stands for inferior temporal gyrus, ACC stands for anterior cingulate cortex, Cau for caudate, Cun for cuneus, CG for cingulate gyrus, Dec for cerebellar declive, ITG stands for inferior temporal gyrus, Put for putamen, STG for superior temporal gyrus, TTG for transverse temporal gyrus, and Thal for thalamus.

Based off these results, the researchers concluded that instrumental music is efficient in conveying positive emotions, while sad emotions are reinforced when lyrics are present. They suggest that vocal cues in sad music activate deep emotion-related structures which produce mental associations with negative emotional experiences, as shown in activation of limbic and paralimbic regions. This activation causes people to have “moving” experiences.

Below are two pieces of music I think Brattico et al. would suggest have high emotional impact- a familiar sad song with vocals and a familiar instrumental happy song. How do they make you feel? (Songs are Hallelujah by Jeff Buckley and Canon in D by Pachelbel, I own no rights to these)

Overall, I thought the study provides a thorough analysis of the brain regions that are differentially activated during happy or sad music, and even considers the effect of lyrics. The only aspect of the experimental procedure that confused me was their decision to have the subjects bring in their own music. Although the researchers say that the subjects had similar familiarity with the music, there was probably some differing in familiarity throughout the categories. I, for one, would probably have a more difficult time finding four sad songs that I hate but know well, than I would in finding four happy songs that I like.

Something that I still didn’t understand fully was their mention of the underlying feelings of being “moved” from music. So, I looked at another article, this time by Vuoskoski and Eerola (2017), that examined the effect of perceptions of music, such as beauty, on this sentiment. They hypothesized that sadness in music has a positive association with beauty, and mediates the feeling of being moved, which in turn causes a sense of enjoyment or pleasure.

The experimental procedure consisted of having 19 music students listen to 27 short film excerpts. The participants then rated the perceived emotion of the music based on six scales: sad/melancholic, moving/touching, tender/warm, peaceful/relaxing, scary/distressing, and happy/joyful, as well as if they liked it or not. The correlation between the qualities is shown in Figure 4. As the table shows, beauty was shown to have a positive correlation with sadness and a high correlation with liking. Also, the perception of being moved was the most highly correlated with beauty and sadness. Overall, Vuoskoski and Eerola found that the indirect effect via movingness on liking was twice the magnitude of that via beauty, which suggests that perceived movingness acts as the largest link between sadness and liking. In other words, the sadder a song is, the more you will be “moved”, and the more you will enjoy it. It is important to point out that this is not saying that happy songs are unlikeable- there was still a positive correlation between happiness and liking, but it was slightly lower than that of sadness.

Figure 4: Correlation values between different emotional qualities of music and liking.

This study gives convincing, albeit initially difficult to understand, connections between sadness and enjoyment of music through the sentiment of “being moved”. The only downside of the study is that the participants were asked about how they thought the music sounded, not how it made them feel. Although it might only be a slight difference in wording, it could play a larger role in terms in relating the feelings to regional activation in the brain, like in the study by Brattico et al. However, by accepting that perception and feeling are inherently linked, we can conclude that the largest enjoyment can be obtained from sad music with vocals, as it strongly activates the regions of the brain that cause the listeners to be emotionally moved. I think an interesting future direction would be to see the effect of human interaction on enjoyment from sad music- I would assume that there would be less enjoyment out of listening to sad music in a group setting, as cultural norms would start to play a larger role.

Even though the findings indicate that sadness gives a higher level of enjoyment, I find it hard to believe that this would not differ between people. What do you think? Do you find that you have a more emotional or pleasurable experience when listening to sad music than happy music?


Brattico E, Alluri V, Bogert B, Jacobsen T, Vartiainen N, Nieminen S, Mari Tervaniemi M (2011) A Functional MRI Study of Happy and Sad Emotions in Music with and without Lyrics. Frontiers in Psychology. 2:308.

Vuoskoski JK, Eerola T (2017) The Pleasure Evoked by Sad Music Is Mediated by Feelings of Being Moved. Frontiers in Psychology. 8:439.

Figure 1 was found through Creative Commons

Figure 2 and 3 were taken from the article by Brattico et al.

Figure 4 was taken from the article by Vuoskoski  and Eerola

Videos were taken from YouTube

The Baby Schema Scheme

Coming off the Metro on my first day in Paris, one of the most immediate sights was that of a woman and her two children sitting on the ground and holding a sign that read, “famille Syrienne”. Throughout the rest of the week, I saw countless homeless people and families, many with children under the age of three. Not only did the homeless often have children, but a large amount also had one or two dogs. While walking to class one day, I even saw a man with a puppy that couldn’t be over two months old. This sparked a question in me- does the appearance of a baby or puppy increase the chance of charitable giving from others?

Figure 1: Homeless Syrian woman with her baby

I believe most people would think the answer to that question is obvious; if given the option to donate to homeless person with a baby or a homeless person without, the logical decision, in terms of effectiveness of the donation, would lean towards the family. However, if we put aside logic-based decision making and focus on spur-of-the-moment choices, would having a baby or puppy make a difference?

Before I did research on experiments from the past, I conducted my own small observational study. At the Bastille Métro Station (Figure 2), I observed the number of people who gave money to both a woman with a baby (Figure 1) and a woman without for five minutes each. Out of 63 people who passed adjacent to the woman with a baby, 4 gave her change, for a percentage of 6.35%. Out of 56 people who passed adjacent to the woman without a baby, only 1 person gave change, for a percentage of 1.79%. Although this observation cannot be statistically analyzed to imply much, as it was a very short study with very few variables controlled, it seems as though the presence of the baby had helped to increase the chance of a donation.

Figure 2: Location of Bastille Station in Paris

In order to find out more information on the neurobiological processes involved in this difference, I read through a study performed by Glocker et al. (2009) on how the “baby schema” modulates the reward system in nulliparous women (women who have never given birth). The baby schema is the physical features of babies, such as a round face and big eyes, that motivates caretaking behavior and attracts attention. This short article modified different aspects of baby schema and observed the levels of activation in associated brain regions in 16 women in their twenties. Glocker et al. hypothesized that an increase in the baby schema “cuteness rating” would cause an increase in blood oxygenation level-dependent (BOLD) fMRI brain activity in the mesocorticolimbic system, which is comprised of the dopaminergic midbrain, nucleus accumbens, amygdala, and ventromedial prefrontal cortex.

Figure 3: Examples of high, unmanipulated, and low baby schema faces used in the study by Glocker et al.

Using adjusted images of infant faces, such as in Figure 3, they found a linear increase in activation due to baby schema in the left anterior cingulate cortex, left precuneus, left fusiform gyrus, and right nucleus accumbens (Figure 4). The researchers then went on to discuss their findings in relation to the functional properties of these regions, specifically the nucleus accumbens, precuneus, and fusiform gyrus.

Figure 4: (A) Results of fMRI BOLD testing by Glocker et al. Areas of interest include left anterior cingulate cortex (ACC), left precuneus (PCu), left fusiform gyrus (FG), and right nucleus accumbens (NAcc). (B) Increases in BOLD percent signal change due to increased baby schema.

They described the nucleus accumbens as being linked to reward-based behavior, and that its activation could release approach behavior towards infants. In addition, the nucleus accumbens is a part of the striatum, which has been associated with processes such as mutual cooperation, charitable donation, and social bonding. The activation of this region due to seeing a baby’s face could influence women into donating money. Another brain region of interest was that of the precuneus, which is commonly associated with attention, suggesting that baby schema brings and holds attention to an infant’s face. Finally, the fusiform gyrus plays a large role in facial perception, and may encode baby schema features to send along to the nucleus accumbens to appoint motivational value.

Overall, the study does a good job in identifying the regions of brain that are sensitive to baby schema. However, it was limited to women in their twenties who have never given birth. This category of people is only a small percentage of those who encounter homelessness, so it doesn’t fully answer my question. Despite its limited conclusions, Glocker et al. discusses how other studies have shown that, while women most likely are more responsive to the baby schema than men, they both process it similarly.

Although this article was informative on the effects of the human baby schema, I was interested in the subject of puppies as well. So, I read an article titled “Sweet Puppies and Cute Babies: Perceptual Adaptation to Babyfacedness Transfers across Species” by Golle et al. The researchers used a perceptual adaptation paradigm to test whether the evaluation of cuteness is species-specific or exists across multiple species. Their first experiment involved subjects rating 78 babies’ faces on a scale of 1-6. The 5 least cute and cutest babies were used as “adaptor” stimuli. All remaining faces were individually paired (one cute and one less cute) and morphed together. The subjects were then tested in three respective parts: rating the morphed faces in cuteness, looking at the adaptor faces carefully, and then rating the morphed faces again. In general, the subjects rated the babies as cuter during the second round of rating, after the adaptation phase. From this, it can be reasoned that the brain grows accustomed to a range of cuteness. During a second experiment, the researchers tested if a similar adaptation can occur when shown faces of dogs.

Figure 5: A homeless man with two dogs in Paris

Using the same procedure, but swapping the human infant adaptor stimuli with cute and less cute puppy faces, Golle et al. found that the adaptation of puppy faces similarly influenced the perception of baby faces to have an increased cuteness value during the second round of rating. From this data, the researchers concluded that facial cuteness adaptation transfers across species and induces the same “cuteness decoding” process (a.k.a. the effects of the baby schema found in the first study). They gather that human beings have a general instinct to take care of newborns of the same or different species- a desire that stems from the cuteness of the baby.

Figure 6: My dog, Buddy. What cuteness rating would you give him?

From these two studies, it can be concluded that both babies and puppies’ cuteness causes an activation in certain areas of the brain associated with caretaking, attention, and charitable giving. This in turn can lead to an increased influx of donations towards homeless with young children or dogs compared to those without. So, next time you give money to a homeless family, what might seem to be a simple altruistic decision might actually be a series of complicated facial analysis!


Glocker ML, Langleben DD, Ruparel K, Loughead JW, Valdez JN, Griffin MD, Sachser N, Gur RC (2009) Baby schema modulates the brain reward system in nulliparous women. Proceedings of the National Academy of Sciences of the United States of America. 106(22):9115-9119.

Golle J, Lisibach S, Mast FW, Lobmaier JS (2013) Sweet puppies and cute babies: perceptual adaptation to babyfacedness transfers across spepcies. PLoS ONE 8(3):e58248

Figures 1 and 6 were taken by me

Figures 2 and 5 were obtained from a search in Creative Commons:



Figures 3 and 4 were taken from the study by Glocker et al.