A circular diffusion model of continuous-outcome source memory retrieval: Contrasting continuous and threshold accounts

Abstract

A circular analogue of the diffusion model adapted for continuous response tasks is applied to a continuous-outcome source memory task. In contrast to existing models of source retrieval that attribute all of the variability in responding to memory, the circular diffusion model decomposes noise into variability arising from memory and from decision processes. We compared three models: (1) a single diffusion process with trial-to-trial variability in drift rate, (2) a mixture of two diffusion processes, one with positive drift that does not vary from trial-to-trial, and a second zero-drift process that represents discrete guessing, and (3) a hybrid model that mixed positive and zero-drift processes with trial-to-trial variability in the positive drift process. Comparison of model fits to joint response error and response-time (RT) data suggest that a memory strength threshold under which no information is retrieved appears to underlie responding in a continuous-report source memory task. Additionally, we also conditioned participants' source responding on their confidence in an old/new recognition task, ruling out the possibility that participant guessing was only due to unrecognized items. Overall, our findings support an all-or-none or some-or none view of source memory retrieval and pose a challenge to continuous models of source memory.

Local motion pooling is continuous, global motion perception is discrete.

“…However, using ROC curves to distinguish between continuous and discrete processing has serious limitations, whether they are constructed by manipulating response bias or by using confidence ratings (e.g., Bröder & Schütz, 2009; Rouder et al, 2021). More recently, formal modeling of confidence rating distributions (Province & Rouder, 2012) and reaction time (RT) distributions (Zhou et al, 2021) has been used to compare discrete and continuous theories. There is not a consensus on which of these approaches is the most appropriate for comparing discrete and continuous processing, and in most applications an entire cognitive process (e.g., working memory, recognition memory, perception) is determined to be mediated by either continuous or discrete processing, making it difficult to directly compare purportedly continuous and discrete cases.…”

The Spatiotemporal Gradient of Intrusion Errors in Continuous Outcome Source Memory

“…A number of studies that collected both item recognition and source memory ratings for the same items have found either a complete lack of source memory for unrecognized items (Malejka & Broder, 2016;Onyper et al, 2010;Bell et al, 2017) or at least very poor memory (Fox & Osth, 2020), suggesting that heavy-tailed distributions of responses may instead reflect a proportion of unrecognized items. To test this possibility, Zhou et al (2021) collected both item recognition and continuous outcome source memory judgments for the same items. While unrecognized items indeed showed no source memory, as evidenced by a flat distribution of source responses, when source recall was conditioned on recognized items heavy-tailed distributions were still found.…”

“…In the current study, we consider two additional sources of variability not considered in the Harlow and Donaldson (2013) work: the possibility that a nontarget item is reported instead of the target item and variability due to properties of the decision-making process which acts upon the information retrieved from memory to generate the observed response. To account for the contribution of decision processes to response variability, Zhou et al (2021) applied the circular diffusion model (Smith, 2016) to a source memory task using Harlow and Donaldson's (2013) paradigm. Unlike empirical characterizations of response error, like the one provided by the wrapped Cauchy model, the predicted distribution of response errors in the circular diffusion model is derived from an evidence accumulation model of the retrieval process.…”

“…Across-trial variability in drift rate is the circular diffusion model's counterpart of variable precision, as assumed in the successful model of visual working memory of van den Berg et al (2014). In a previous study, Zhou et al (2021) investigated whether this property of the diffusion decision model could account for the distribution of errors in source memory retrieval observed by Harlow and Donaldson (2013), without needing a threshold in the memory retrieval process. To do so, in Zhou et al (2021) we compared three different variants of the circular diffusion model: 1) a single diffusion process with across-trial drift rate variability, 2) a two-component mixture of a diffusion process with drift rate fixed across trials, and a zero-drift process (i.e.…”

“…However, under many conditions the predictions of the models have been found to mimic each other (Yonelinas & Parks, 2007;Klauer & Kellen, 2010;Wixted, 2007). This model mimicry has led researchers to test model predictions in continuous outcome source retrieval, in which heavy-tailed distributions of responses have been argued to reflect guesses from failed source retrieval (Harlow & Donaldson, 2013;Zhou, Osth, Lilburn, & Smith, 2021). The focus of the present work is in evaluating whether the heavy-tailed distributions are the reflection of failed source retrieval or are instead the result of intrusions from non-target responses.…”