Spectral Amplitude Modulation in EEG: Potential Correlations with Musical Stimuli

Abstract

The study of neural responses to musical stimuli offers valuable insights into how the brain encodes complex auditory information. This paper compares spectral amplitude modulations from various EEG electrodes, time-locked to attended music, to explore potential correlations. Instead of using traditional 2D TF plots, we utilized a 3D time-frequency-amplitude (TFA) fabric over a 500 ms window, synchronized with the audio of the song. This approach was applied to five songs, examining the TFA fabric across different electrodes. The TFA fabric refers to the continuous manifold created by the frequency-amplitude (FA) curve over time. This manifold, or fabric, is characterized by its inherent smoothness and contiguity across frequency, time, amplitude, and brain location (an inherent result of using complex Morlet wavelet convolution with fine temporal and frequency resolution). Both excitation and inhibition contribute to pulling the FA curve away from an approximately 1/f baseline, effectively causing perturbations that determine its complex modulation. In particular, bursts (and sometimes dips) in beta/gamma power (amplitude here) often seem to correlate with transients in music, especially at electrode Fz. Also, by visual inspection only, songs with greater rhythmic emphasis appeared to show stronger driving of Fz activity (the song "Smoke" by PVRIS showed greatest correlation).