丛丰裕 教授、博士生导师
时间： 2015-11-16 10:00 - 12:00
地点： Room 1113 Wang Kezhen Building
This talk is devoted to the application of multi-way analysis of time-frequency representation of event-related potentials (ERPs) for the cognitive neuroscience. Conventionally, ERPs are usually produced through averaging single-trials of pre-processed EEG, and then, the interpretation of underlying evoked brain activities in the time domain is based on the ordinarily averaged EEG. Recently, the time-frequency analysis has been applied on each single-trial EEG, and then the time-frequency representations of multiple trials are averaged to obtain the induced brain activities in the time-frequency domain. No matter the average is applied in the time domain or in the time-frequency domain, we find that randomly fluctuating activities and artefacts can still present in the averaged EEG data, and that constant brain activities over single trials can overlap with each other in time, frequency and spatial domains. Therefore, this motivates us to develop a new group-level analysis approach for filtering time-frequency representation of averaged EEG in time, frequency and space domains to simultaneously obtain the pure ERP of interest. The key methodology includes wavelet transform and tensor (multi-way array) decomposition. Software of the proposed approaches is open-accessed via http://www.escience.cn/people/cong/index.html2015-11-16
2015-11-16