Abstract: Whether the capacity of visual consciousness is rich or sparse is a foundational issue in consciousness studies. The controversy originates primarily from the separation between the subjective experience and the objective performance of observers. For example, if an image flashes by, we can experience a rich and detailed visual scene, but a little can be reported. The disconnect between subjective experience and objective observations can be verified as a question to be explained. For example, observers believe that the rich visual experience is largely uncertain. Specifically, visually conscious content may exist in probability and may not necessarily be full of detail. However, it is subjectively considered as rich. Most studies on visual consciousness follow binary alternative-forced chosen procedures, limiting the likelihood that participants report probability consciousness. Therefore, we used a new Multiple Object Awareness Paradigm to allow participants to be multi-responses. An array containing eight stimuli arranged in a circle was presented quickly (the stimuli’s positions are randomly distributed in each trial). Then all stimuli were covered by gray circles, and a target stimulus was placed at a distance of 7° above the cross. The target of each trial was randomly generated. Then, participants were instructed to click on gray circles where they had seen the target appear until they got it correct. Importantly, participants needed to find the target with as few clicks as possible. Thus, even if participants did not find a target on the first click, they could try the second, third, or more until they found the target. Once the target was found, feedback was provided. We calculated the visual consciousness capacity via three methods. First was the traditional estimation method, which was based on the first click was correct or wrong. The second and third were based on total clicks, which were guessed from N stimuli randomly and guessed from N-K stimuli randomly when participants did not find targets on the first click. Results showed that the capacity estimated by the clicks guessing from N was significantly greater than that estimated by the first-attempt accuracy; however, the clicks guessing from N-K were significantly smaller than that estimated by the first-attempt accuracy. Further analysis found that if observers could not locate the target in their first attempt, they were more likely to click closer to the target. In addition, we found that even when observers used the same number of clicks to find a target (2 or 3 clicks), the average distance was shorter when observers reported high-level subjective visibility. This study provides evidence for the controversy of whether visual consciousness capacity is rich or sparse. These results fit the visual ensembles and summary statistics hypothesis. In this view, we have a rich visual experience that is somewhat uncertain or imprecise, as the content of visual awareness may be probabilistic or not necessarily full of details. In summary, we were less concerned with the participants’ correct choices and more concerned with how much information was obtained from their “wrong responses.” Driven by methodological advances in measuring, we can increasingly recognize the flexibility of our visual system.

Key words: multiple object awareness paradigm, visual consciousness, visual consciousness capacity, probabilism, visual ensembles and summary statistics hypothesis

摘要： 采用一种新的多目标觉知范式检验视觉意识容量。实验中首先迅速呈现包含8个刺激的圆形阵列(刺激位置随机分布), 然后所有刺激被灰色的圆覆盖, 并随机选取阵列中的一个刺激作为目标, 要求被试用尽可能少的点击次数找到目标。结果发现, 即使被试不能一次击中目标, 其也可能拥有对目标不精确的视觉信息。这些不精确的视觉信息能够帮助被试用更少的点击次数和更短的点击距离找到目标, 且被试对这些不精确的视觉信息是有意识的。研究结果表明视觉意识能够以概率性的表征存在, 并支持视觉整体统计假说, 也能够进一步解释视觉意识容量的主观测量和客观测量之间存在的冲突。

关键词: 多目标觉知范式, 视觉意识, 视觉意识容量, 概率性, 整体统计假说