Mapping visual working memory models to a theoretical framework

Abstract

The body of research on visual working memory (VWM) -the system often described as a limited memory store of visual information in service of ongoing tasks -is growing rapidly. The discovery of numerous related phenomena, and the many subtly different definitions of working memory, signify a challenge to maintain a coherent theoretical framework to discuss concepts, compare models and design studies. A lack of robust theory development has been a noteworthy concern in the psychological sciences, thought to be a precursor to the reproducibility crisis (Oberauer & Lewandowsky, 2019). I review the theoretical landscape of the VWM field by examining two prominent debates -whether VWM is object-based or feature-based, and whether discrete-slots or variable-precision best describe VWM limits. I share my concerns about the dualistic nature of these debates and the lack of clear model specification that prevents fully determined empirical tests. In hopes of promoting theory development, I provide a working theory map by using the broadly encompassing Memory for Latent Representations model (Hedayati et al., 2022) as a scaffold for relevant phenomena and current theories. I illustrate how opposing viewpoints can be brought into accordance, situating leading models of VWM to better identify their differences and improve their comparison. The hope is that the theory map will help VWM researchers get on the same page -clarifying hidden intuitions and aligning varying definitions -and become a useful device for meaningful discussions, development of models, and definitive empirical tests of theories.

Just-in-Time Encoding Into Visual Working Memory Is Contingent Upon Constant Availability of External Information

“…When provided with such complexity, introducing any additional working memory load may introduce undesired noise (e.g., Bays, 2014;Oberauer & Lin, 2017;Schurgin et al, 2020), thereby making the task not only more effortful, but also more prone to mistakes. We here focused on manipulating the ability to sample just-in-time, but it is clear that the paradigm provides a rich environment in which to study different aspects of working memory and to place them within a formalized model (Ngiam, 2023).…”

Dimension-Based Retro-Cue Benefit in Working Memory Does Not Require Unfocused Dimension Removal

“…However, to discover the mechanism of dimension-based retro-cue benefit, we should first define the storage nature of dimensions in VWM. Ngiam (2023) proposed a Memory for Latent Representations Model (MRL) to reconcile controversial accounts for the storage of VWM. In terms of MRL model, dimension inputs are bound with context information and stored in the binding pool to fit flourished findings that dimensions can be stored and forgotten independently (D. Fougnie and Alvarez 2011; Markov, Tiurina, and Utochkin 2019).…”

More liberal and less sensitive: Individual differences in visual working memory capacity predicts the metacognitive assessment of representational accuracy

“…Consequently, there are as many estimates of VWM accuracy as the number of theories. Given that it is not our primary goal to adjudicate between different theories that often make hard-to-distinguish predictions (e.g., Ngiam, 2023;van den Berg & Ma, 2014), we chose to empirically measure the VWM representations free of any assumptions in the following manner (Sone et al, 2021). First, we computed the response offset (i.e., the angular difference between the original and recalled color) for each response and produced a frequency distribution of response offset from -179° to 180° degrees.…”