How well do we know our city? It turns out, much more poorly than we might imagine. We used declarative memory and eye-tracking techniques to examine people's ability to detect modifications to real-world landmarks and scenes in Toronto locales with which they have had extensive experience. Participants were poor at identifying which scenes contained altered landmarks, whether the modification was to the landmarks' relative size, internal features, or relation to surrounding context. To determine whether an indirect measure would prove more sensitive, we tracked eye movements during viewing. Changes in overall visual exploration, but not to specific regions of change, were related to participants' explicit endorsement of scenes as modified. These results support the contention that very familiar landmarks are represented at a global or gist level, but not local or fine-grained, level. These findings offer a unified view of memory for gist across verbal and spatial domains, and across recent and remote memory, with implications for hippocampal-neocortical interactions.

中文翻译：

---

知名地点发生的巨大变化却被忽视

我们对我们的城市了解多少？事实证明，情况比我们想象的要糟糕得多。我们使用陈述性记忆和眼球追踪技术来检查人们检测他们在多伦多地区拥有丰富经验的现实世界地标和场景的修改的能力。参与者很难识别哪些场景包含改变的地标，无论修改是对地标的相对大小、内部特征还是与周围环境的关系。为了确定间接测量是否更敏感，我们跟踪了观看过程中的眼球运动。整体视觉探索的变化，但与特定区域的变化无关，与参与者对修改后的场景的明确认可有关。这些结果支持这样的论点：非常熟悉的地标是在全局或要点级别上表示的，而不是在局部或细粒度级别上表示的。这些发现为跨语言和空间领域、跨近期和远期记忆的要点记忆提供了统一的观点，对海马-新皮质相互作用具有影响。