The Mechanism of Involuntary Visual Spatial Attention Revealed by a New Linear Computation Model

Abstract

EEG reveals brain electrical activities with high temporal resolution. Yet, multiple implicit variables may be involved in limited event related potential (ERP) measures. Special computation techniques are needed to recover these parameters. In the study of involuntary visual spatial attention, we may obtain the ERP in valid cued (V), invalid cued (I) and neutral cued (N) conditions. Usually, the effect of involuntary attention is computed by the subtraction model with the assumption that V/I and N are independent. Yet, they should be related. Treating V/I as a function of N, a linear model V(I) = W + GN is assumed, where W and G are implicit in the ERP measures. G is the gain control on the neutral function. Provided G and W are constant over a local brain region, we may use the Total Least Square (TLS) algorithm to compute their values. The values of W and G computed from an involuntary attention experiment data show that multiple implicit variables are involved in obtained ERPs. Here G acts as a "top-down" sensory modulator on the neutral ERPs and W is related to possible newly involved neural activities. The parameters derived from the new linear model also suggest that there are different mechanisms involved in involuntary attention and voluntary allocation of attention.