Jack Bradley N, Widmann Andreas, O'Shea Robert P, Schröger Erich, Roeber Urte
Discipline of Psychology, School of Health and Human Sciences, Southern Cross University, Coffs Harbour, Australia.
Institute of Psychology, University of Leipzig, Leipzig, Germany.
Psychophysiology. 2017 May;54(5):755-763. doi: 10.1111/psyp.12831. Epub 2017 Feb 20.
Predictive coding explains visual perception as the result of an interaction between bottom-up sensory input and top-down generative models at each level of the visual hierarchy. Evidence for this comes from the visual mismatch negativity (vMMN): a more negative ERP for rare, unpredictable visual stimuli-deviants, than for frequent, predictable visual stimuli-standards. Here, we show that the vMMN does not require conscious experience. We measured the vMMN from monocular luminance-decrement deviants that were either perceived or not during binocular rivalry dominance or suppression, respectively. We found that both sorts of deviants elicited the vMMN at about 250 ms after stimulus onset, with perceived deviants eliciting a bigger vMMN than not-perceived deviants. These results show that vMMN occurs in the absence of consciousness, and that consciousness enhances the processing underlying vMMN. We conclude that generative models of visual perception are tested, even when sensory input for those models is not perceived.
预测编码将视觉感知解释为视觉层级结构各层面上自下而上的感官输入与自上而下的生成模型之间相互作用的结果。这一观点的证据来自视觉失配负波（vMMN）：与频繁出现、可预测的视觉刺激（标准刺激）相比，罕见、不可预测的视觉刺激（偏差刺激）会引发更负的事件相关电位。在此，我们表明vMMN并不需要有意识的体验。我们分别在双眼竞争优势或抑制期间，测量了单眼亮度递减偏差刺激的vMMN，这些偏差刺激在相应过程中要么被感知到，要么未被感知到。我们发现，两种偏差刺激均在刺激开始后约250毫秒时引发了vMMN，其中被感知到的偏差刺激引发的vMMN比未被感知到的偏差刺激更大。这些结果表明，vMMN在无意识状态下也会出现，且意识会增强vMMN背后的处理过程。我们得出结论，即使那些模型的感官输入未被感知到，视觉感知的生成模型也会受到检验。
