提出了一個新的選擇性注意腦電信號分解模型WGVVua+=，并應(yīng)用總體最小二乘法(TLS)求解該模型方程。根據(jù)TLS求得的G和W，注意效應(yīng)被分解成了兩個方面，注意引起的腦神經(jīng)活動W和注意對非注意情況腦活動的門限調(diào)控作用G。將該模型應(yīng)用于空間注意ERP數(shù)據(jù)，進(jìn)一步提出了選擇性注意的一個新的理論框架，即在刺激后的200 ms內(nèi)出現(xiàn)的調(diào)制放大(G>1.0)和新活動成分(W)主要起改善信噪比的作用，而在200 ms后出現(xiàn)的調(diào)制抑制(G<1.0)和新活動成分(W)則旨在對選擇性增強(qiáng)的信號成分的特定認(rèn)知處理。
關(guān) 鍵 詞 注意; 事件相關(guān)電位; 信號模型; 調(diào)制作用

Abstract Presented is an analytic signal model with and corresponding to the event-related potentials (ERPs) evoked by attended and unattended visual stimuli, respectively. A total least square algorithm is utilized to get the G and W, then the effect of attention is decomposed into two aspects: attention induced activities (W) and a gain control (G) of unattended. Applied to a spatial attention ERP data, an integrated concept of selective attention is suggested that the early gain amplification (G>1.0) and newly involved neuron activities (W) before 200 ms after the stimulus onset mainly aim at obtaining an improved signal/noise ratio, the late gain supression (G<1.0) and newly involved neuron activities (W) after 200 ms aim at specific cognitive processing of the selectively enhanced signal. WGVVua+=aVuV
Key words attention; ERP; signal model; modulation