Ensemble clustering in visual working memory biases location memories and reduces the Weber noise of relative positions

Abstract

People seem to compute the ensemble statistics of objects and use this information to support the recall of individual objects in visual working memory. However, there are many different ways that hierarchical structure might be encoded. We examined the format of structured memories by asking subjects to recall the locations of objects arranged in different spatial clustering structures. Consistent with previous investigations of structured visual memory, subjects recalled objects biased toward the center of their clusters. Subjects also recalled locations more accurately when they were arranged in fewer clusters containing more objects, suggesting that subjects used the clustering structure of objects to aid recall. Furthermore, subjects had more difficulty recalling larger relative distances, consistent with subjects encoding the positions of objects relative to clusters and recalling them with magnitude-proportional (Weber) noise. Our results suggest that clustering improved the fidelity of recall by biasing the recall of locations toward cluster centers to compensate for uncertainty and by reducing the magnitude of encoded relative distances.

Chunking as a rational strategy for lossy data compression in visual working memory

“…Our work builds on two parallel lines of research. One has focused on how encoding and decoding of working memories are optimized under various statistical contingencies (Brady et al, 2009;Brady & Alvarez, 2011;Brady & Tenenbaum, 2013;Lew & Vul, 2015;Orhan & Jacobs, 2013;Sims et al, 2012), whereas the other has focused on understanding the nature of capacity limitations in visual working memory (Bays et al, 2009;Bays & Husain, 2008;van den Berg et al, 2012;Zhang & Luck, 2008;. Here, we explore how people optimize encoding in the same tasks that have formed the basis of our understanding of capacity limitations.…”

“…Thus, an open question with potentially broad implications is what should and do people store in memory during performance of the standard delayed recall tasks, and how do deviations from the standard assumptions affect our understanding of memory capacity? To this end, recent work has highlighted the ability of people to optimize memory encoding and decoding processes by pooling information across memoranda to enhance performance under different regimes (Brady & Alvarez, 2011;Brady, Konkle, & Alvarez, 2009;Lew & Vul, 2015;Orhan & Jacobs, 2013;Sims, Jacobs, & Knill, 2012;Wei et al, 2012). Specifically, people can integrate prior information to improve memory report precision (Bays et al, 2009;Brady & Alvarez, 2011) and, when stimuli are redundant, lossless compression strategies can be used to efficiently encode them (Bays et al, 2009;Brady et al, 2009;Zhang & Luck, 2008).…”

Compositional Inductive Biases in Function Learning

“…Markov random fields to encode information about object features; several other structural assumptions have been explored in the literature (Lew & Vul, 2015;Mathy & Feldman, 2012;Orhan & Jacobs, 2013). Here we explore the possibility that the representation of functions in short-term memory is compositional, leading to the prediction that change detection will be more accurate with compositional functions compared to with non-compositional functions.…”

The Whole Warps the Sum of Its Parts

“…A minimum sample size of 30 participants was chosen a priori on the basis of the sample sizes (ranging from 16 to 35 participants) used in the previous studies by Lew and Vul (2015) and Brady and Alvarez (2011). All procedures and protocols were in accordance with the guidelines of Bilkent University's ethical review board.…”