Affective Priming Music 451A Final Project
Affective Priming
Music 451A Final Project
The Question
● Music often makes us ‘feel’ a certain way. Does this feeling have semantic meaning like the words ‘happy’ or ‘sad’ do?
● Does music convey semantic emotional content?
Previous Studies: 2004● Koelsch investigates language and music. Found an N400 for music primes
and semantically unrelated target words.
Previous Studies: 2012● Goerlich study showed negligible results that could be due to the motor
response during the N400: mean RT’s = 600ms ○ Subjects had to respond as quickly as possible with congruent or incongruent.
Experimental Design
GoalTo compare N400 between congruent and incongruent primes when a prime is given major or minor, and a probe of happy or sad emoji is displayed.
HypothesisIncongruent stimuli (major+sad face, minor+happy face) elicit an N400 for all listeners. This could indicate melodic priming has semantic emotional content.
MethodsParticipants: 5 University students aged 20-29, 3 Female and 2 Male, 3 Musician and 2 Non-musician.
Stimuli and Tasks: A melody lasting ~3.5 seconds played on piano (27 C Major and 27 possible corresponding C Minor melodies), followed by a happy (3 possible) or sad (3 possible) emoji for 1 second. Some trials had a black screen instead of an emoji and participants were asked to respond with a button click. Trials were separated into 3 blocks, 12 minutes each, for a total recording time of 36 minutes.
EEG Recording Procedure: Participant was capped with a 67 electrode cap (standard 10-20 system), seated in a recording room with ear-buds, facing a computer screen showing stimuli. Participants were in the chamber for roughly 1 hour total, as they participated in two experiments. Order was randomized.
Grand Average Results
Congruent average: Major/happy and Minor/sad
Incongruent average: Major/sad and Minor/happy
Latency: 400ms
Latency: 400ms
Grand Average Results Difference
Incongruent Average - Congruent average: (Major/sad + Minor/happy) - (Major/happy + Minor/sad)
Small N400 in topo!But not visible in waveform
Latency: 400ms
Grand Average Results
Individual Results: Major
Barbara
Elena
Kunwoo
Michael
Trang
Major Melody Difference Waves
Individual Results: Minor
Barbara
Elena
Kunwoo
Michael
Trang
Minor Melody Difference Waves
Individual Results: BN
Congruence (Major/Happy) vs Incongruence (Major/Sad) Congruence (Minor/Sad) vs Incongruence (Minor/Happy)
Individual Results: TN
Congruence (Major/Happy) vs Incongruence (Major/Sad) Congruence (Minor/Sad) vs Incongruence (Minor/Happy)
Individual Results: EG
Congruence (Major/Happy) vs Incongruence (Major/Sad) Congruence (Minor/Sad) vs Incongruence (Minor/Happy)
Individual Results: KK
Congruence (Major/Happy) vs Incongruence (Major/Sad) Congruence (Minor/Sad) vs Incongruence (Minor/Happy)
Individual Results: MM
Congruence (Major/Happy) vs Incongruence (Major/Sad) Congruence (Minor/Sad) vs Incongruence (Minor/Happy)
Grand Average of EG, KK, MM
Congruence (Major/Happy) vs Incongruence (Major/Sad) Congruence (Minor/Sad) vs Incongruence (Minor/Happy)
Grand Average of EG, KK, MM
Incongruence (Major/Sad) - Congruence (Major/Happy) Topography at 400ms Latency
Incongruence (Minor/Happy) - Congruence (Minor/Sad)Topography at 400ms Latency
Discussion
● N400 in major (sad - happy) was greater than N400 in minor (happy - sad)● 4 out of 5 subjects showed N400 in major (happy/sad)● 4 out of 5 subjects showed N400 in minor (happy/sad)
○ The N400 for minor melody incongruence is low
DiscussionWhy low N400?
● Weak melodic primes○ Prime length of 3.5 seconds may not be enough to induce mood.○ 3rd (which helps distinguish major from minor) appeared infrequently or
too late (not enough priming occurred).○ All melodies were in the same key.○ Pitch was the only varying acoustic feature
● Visual response domination○ Happy and sad emojis may evoke significant responses that are mutually exclusive to auditory
primes
Exploration
What’s affecting the data?
1. Does the Happy/Sad Visual Response dominate Incongruence/Congruence?
2. Does the Major/ Minor Response dominate?
1. Happy/Sad Visual Response: Grand Average
Sad Grand Average: Major/sad and Minor/sad
Happy Grand Average: Major/happy and Minor/happy
1. Happy/Sad Visual Response: Grand Average Difference
Sad Average - Happy Average: (Major/sad and Minor/sad) - (Major/happy and Minor/happy)
The sad emoji evokes a slightly larger negative amplitude than the happy emoji, regardless of melody type (major or minor).
2. Major/Minor Response: Grand Average
Major Grand Average: Major/happy + Major/sad
Minor Grand Average: Minor/happy + Minor/sad
2. Major/Minor Response: Grand Average Difference
Minor Average - Major Average: (Minor/sad and Minor/happy) - (Major/happy and Major/sad)
A small amplitude difference around 250ms indicates the minor primes had a slightly stronger effect than major. This is less of a difference than the emojis.
Remediations
● Vary key so that subjects engage more with melodic context● Increase the number of acoustic features (tempo, timbre, harmony)● Increase length of melodies for better priming● Include more minor tonality (3rd, 6th, 7th)● Use soundtrack music instead to convey emotional content● Use human faces instead of emojis● Use pos/neg words associated with happy (e.g. “joy”) or sad (e.g.
“dark”)
References● Goerlich, K.S., Witteman, J., Schiller, N.O., Van Heuven, V.J., Aleman, A., &
Martens, S. (2012). The nature of affective priming in music and speech. Journal of cognitive neuroscience, 24(8), 1725–41.
● Koelsch, S., Kasper, E., Sammler, D., Schulze, K., Gunter, T., & Friederici, A.D. (2004). Music, language, and meaning: brain signatures of semantic processing. Nature Neuroscience, 7, 302-307.