A point process modeling approach for investigating the effect of online brain activity on perceptual switching

When watching an ambiguous figure that allows for multiple interpretations, our interpretation spontaneously switches between the possible options. Such spontaneous switching is called perceptual switching and it is modulated by top-down selective attention. In this study, we propose a point process modeling approach for investigating the effects of online brain activity on perceptual switching, where we define online activity as continuous brain activity including spontaneous background and induced activities. Specifically, we modeled perceptual switching during Necker cube perception using electroencephalography (EEG) data. Our method is based on the framework of point process model, which is a statistical model of a series of events. We regard perceptual switching phenomenon as a stochastic process and construct its model in a data-driven manner. We develop a model called the online activity regression model, which enables to determine whether online brain activity has excitatory or inhibitory effects on perceptual switching. By fitting online activity regression models to experimental data and applying the likelihood ratio testing with correction for multiple comparisons, we explore the brain regions and frequency bands with significant effects on perceptual switching. The results demonstrate that the modulation of online occipital alpha activity mediates the suppression of perceptual switching to the non-attended interpretation. Thus, our method provides a dynamic description of the attentional process by naturally accounting for the entire time course of brain activity, which is difficult to resolve by focusing only on the brain activity around the time of perceptual switching.