Investigating Memory Episodes in Location Probability Learning: Can Altering Response Features Reset Spatial Bias?

Abstract

When individuals are repeatedly exposed to a specific task, they store each occurrence as a distinct memory episode, which includes stimuli, responses, and outcomes. The accumulation of these episodes enables more efficient retrieval over time, particularly under consistent conditions, leading to quicker and more automatic responses. This mechanism likely underlies the statistical learning effect observed in tasks such as the probability cueing paradigm, where frequent target detection at a predictable location enhances performance through rapid, less conscious retrieval of relevant episodes. In this study, we explored the role of episodic retrieval in implicit location probability learning, focusing specifically on how changes in response features might impact the retrieval of memory episodes and, consequently, the learned attentional bias. Participants performed a visual search task, searching for a T among Ls, while unaware that the target was more often in one screen region. An attentional bias towards this region developed during training. In the testing phase, we examined whether changes in motor responses could negate the learned attentional bias, as such changes might prevent the retrieval of relevant memory episodes. Contrary to our initial hypothesis, two experiments showed that changing response features did not affect the attentional bias. This study expands our understanding of human statistical learning by examining the previously neglected yet increasingly recognized role of episodic retrieval in shaping implicit learning processes, thereby opening new avenues for understanding how our visual systems adapt to and learn from the environment.