Erp Component Analysis Study

Event-related potentials (ERPs) are brain potentials evoked by such as sensory, motor, or cognitive events. As an important tool for brain function reasearch, ERP technique has been used extensively in cognitive neuroscience. Good experimental design is the entrance for problem study, however, without efficient and reliable data procesing method, any experiment is meaningless. If we can not get the actual ERP components reflecting the real cognitive activity, any explanation of ERPs is unreliable. In a word, ERP extraction is a quite important bridge that links the problem and the solution.In a reaction time task, the stimulus and response both evoke the ERPs respectively. Because the short interval between the stimulus onset and response onset, the stimulus-locked component and response-locked component are temporally overlapped, resulting in cross-contamination. How to decompose the overlapping components and get the pure ERP waveform is the aim of this work, the main contents are as follows:1. Considered in detail the ERP temporal decomposition model(for short,TD model) it is developed by Takeda et al, based on the temporal event triggers and the ERP components information. We found the factor that made unstable.2. Based on the TD model, we proposed a new components decoposition method, and its efficiency was confirmed by simulation study.3. For an ERP experiment, three events are also possible, in this situation, the overlapping is more serious. We generalize the new algorithm to three events situation,and the algorithm is also stable.4. We applied the new method to real inhibition of return(IOR) EEG data, the results confirmed on the feasibility of the new method in actual reaserch. and the resulted ERPs provide more reasonable interpretation for specific brain activity.