Department of Brain and Cognitive Sciences, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul, 08826, Republic of Korea.
Sci Rep. 2023 Oct 24;13(1):18162. doi: 10.1038/s41598-023-45505-5.
Our perception is often attracted to what we have seen before, a phenomenon called 'serial dependence.' Serial dependence can help maintain a stable perception of the world, given the statistical regularity in the environment. If serial dependence serves this presumed utility, it should be pronounced when consecutive elements share the same identity when multiple elements spatially shift across successive views. However, such preferential serial dependence between identity-matching elements in dynamic situations has never been empirically tested. Here, we hypothesized that serial dependence between consecutive elements is modulated more effectively by the spatial correspondence in relative space than by that in absolute space because spatial correspondence in relative coordinates can warrant identity matching invariantly to changes in absolute coordinates. To test this hypothesis, we developed a task where two targets change positions in unison between successive views. We found that serial dependence was substantially modulated by the correspondence in relative coordinates, but not by that in absolute coordinates. Moreover, such selective modulation by the correspondence in relative space was also observed even for the serial dependence defined by previous non-target elements. Our findings are consistent with the view that serial dependence subserves object-based perceptual stabilization over time in dynamic situations.
我们的感知常常会被我们之前见过的东西所吸引,这一现象被称为“序列依赖”。在环境中存在统计规律的情况下,序列依赖有助于维持对世界的稳定感知。如果序列依赖具有这种假定的效用,那么当连续的元素具有相同的身份时,在多个元素在连续的视图中空间移动时,应该会出现明显的序列依赖。然而,在动态情况下,身份匹配元素之间的这种优先序列依赖关系从未经过实证检验。在这里,我们假设,在相对空间中的空间对应关系比在绝对空间中的空间对应关系更有效地调节连续元素之间的序列依赖关系,因为相对坐标中的空间对应关系可以保证身份匹配在绝对坐标的变化中保持不变。为了验证这一假设,我们开发了一项任务,其中两个目标在连续的视图中同时改变位置。我们发现,序列依赖关系受到相对坐标中对应关系的显著调节,但不受绝对坐标中对应关系的调节。此外,即使对于由先前非目标元素定义的序列依赖关系,也观察到了对相对空间中对应关系的这种选择性调节。我们的发现与以下观点一致,即在动态情况下,序列依赖有助于基于对象的感知随时间的稳定。