Eeg Study On The Perception Of Temporal Variations

Tempo is an extremely crucial element in music since it can influence how a piece sounds and feels. As the rate of beat, tempo can be slow, fast or in-between, and can change during a melody. Different tempo is related to different emotion evoked in music performance and appreciation. So the ability for discrimination of tempo change in music is critical not only for performers and composers, but also for listeners.This thesis focused on the detection of IOI deviations in an isochronous sequences. Based on scalp Electroencephalography (EEG) and some associated methods of signal processing, the following studies were presented and discussed.1. Methods of signal processing for EEG analysis in our experiments, such as the ERP (Event-Related Potential) waveform measurement in temporal analysis, the GBA in tempospectral analysis, the low resolution brain tomography algorithm (LORETA) estimation and the Laplacian technology in spatial analysis, were summarized firstly. Then the planar difference approximation approach for local Laplacian was extended to a spherical surface model. The 2nd and 4th order approximation were derived and a proportional coefficient difference was revealed between the planar and the spherical surface models.2. The effects of different tempos on the detection of temporal perturbation were investigated in 5-beat isochronous sequences. Modulations of gamma-band activity (GBA) were measured as subjects listened to theses 5-beat isochronous sequences with embedded inter-onset interval (IOI) variations. These sequences had three different base tempos: fast, modest or slow. Perturbations occurred at the last beat, which occurred early, on time, or late. The evoked (phase-locked) GBA peaks showed a smaller amplitude and earlier latency in fast sequences, and a larger amplitude and later latency in slow sequences. The N1 component had a larger amplitude and later latency in fast sequences, and a smaller amplitude and earlier latency in modest sequences. Also, the N1 amplitude was significantly different between the advanced and delayed target tones in modest sequences. Furtherly, the Laplacian current density mapping indicated that the brain activity evoked by targets in modest sequences was stronger than that in fast or slow sequences. This might be related with the perception resonance induced from the overlapping of auditory templates between the preceding tone and the target in modest sequences, implicating the higher discrimination sensitivity for temporal variations under modest simulation rate.3. This work was to investigate whether the pre-attentive perception of acceleration and deceleration were different when the basic tempo was at the behavioral indifference time. The standard stimulus was a 5-beat sequence with each inter-stimulus interval (ISI) of 300, 600, and 900ms for behavioral experiment, and only 600ms for ERP experiment. For the deviant stimulus, the ISI between the third and fourth beats was shortened or lengthened by 10% of the standard IOI. The behavioral data indicated that no perception bias was existed for the detection of the delayed and advanced targets in sequences with modest tempo. The ERP data showed that both deviants elicited a frontally mismatching negativity (MMN), with relatively earlier latency and greater magnitude for shorter IOI than for longer IOI. The LORETA source estimation showed an activity predominantly at the left prefrontal area. The results indicated that the temporal variation directions had an effect on the latency and magnitude of the MMNs even when the standard tempo was at the behavioral indifference time.4. The effect of intensity accent on the perception of temporal variations was examined in an Event-Related Potential (ERP) experiment. For a standard 5-beat isochronous sequence, the IOI before the fourth tone and the intensity of the fourth tone were modulated respectively (IOI: standard, shortened or lengthened by 15%; Intensity: standard, louder or softer by 4dB) to get 9 different stimulation sequences. Analyses of behavioral and ERP measures revealed that intensity accents had an asymmetric effect on the detection of IOI deviations in an isochronous sequence. The louder accents mainly reduced the discrimination sensitivity of the shorter and standard IOIs and thus made them much difficult to detect. The softer accents resulted in an earlier latency for P3 component. The suddenly intensity declination might introduce into a disturbance to attention, especially for the detection of the shorter IOIs, and then impede the entrainment of more high-level processing in the temporal perception. Importantly, the longer IOIs corresponded to larger P3 amplitude than the short IOIs in each of the three kinds of sequences, providing electrophysiological evidence for compensation hypothesis, which predicts that the longer IOIs were easier to detect than the shorter IOIs.