Main content



Loading wiki pages...

Wiki Version:
Earworms often occur when a tune resides in the head and cannot be easily shaken. The present experiment studies which melodic features lead to earworms and how the brain responds to those features. The experiment employed Eigenvector Scaling of five short identical-length 8-bar melodies composed to share the same timbre, key, meter, tempo and volume. 3 sets of melodies were presented in all possible pairs ((5*4)/2 =10 pairs) for 30 randomized presentations). Participants judged which melody was most likely to result in an earworm and selected a number to represent that likelihood. Eigenvectors were calculated on each square matrix of data to produce ratio-scale measures that represent whether melodies are likely to result in an earworm. EEG measures were taken at four sites: Fz, P3, T3 and T4. A 64-channel Neuroscan E-Series Digital EEG measured upper alpha and theta bands that were then processed using Matlab software to normalize and compute 10-sec moving averages. Differences “Fz minus P3” and “T3 minus T4” were formed and plotted to show brain activity. Eigenvector results show that melodies with rhythmic variation, greater note range and a “Toch wave” may become Earworms. Most interesting was the finding that melodies without dotted rhythms or melodic complexity are processed between T3 (left brain) and T4 (right brain), whereas rhythmic and melodic variation cause the brain to shift between Fz (Working Memory) and P3 (supporting WM)