Memory resources recover gradually over time: The effects of word frequency, presentation rate, and list composition on binding errors and mnemonic precision in source memory.

Abstract

Normative word frequency has played a key role in the study of human memory, but there is little agreement as to the mechanism responsible for its effects. To determine whether word frequency affects binding probability or memory precision, we used a continuous reproduction task to examine working memory for spatial positions of words. In three experiments, after studying a list of five words, participants had to report the spatial location of one of them on a circle. Across experiments we varied word frequency, presentation rate and the proportion of low frequency words on each trial. A mixture model dissociated memory precision, binding failure and guessing rate parameters from the continuous distribution of errors. On trials that contained only low-or only high-frequency words, low-frequency words lead to a greater degree of error in recalling the associated location. This was due to a higher word-location binding failure and not due to differences in memory precision or guessing rates. Slowing down the presentation rate eliminated the word frequency effect by reducing binding failures for low-frequency words.Mixing frequencies in a single trial hurt high-frequency and helped low-frequency words. These findings support the idea that word frequency can lead to both positive and negative mnemonic effects depending on a trade-off between a HF encoding advantage and a LF retrieval cue advantage. We suggest that 1) low-frequency words require more resources for binding, 2) that these resources recover gradually over time, and that 3) binding fails when these resources are insufficient.

The Spatiotemporal Gradient of Intrusion Errors in Continuous Outcome Source Memory

“…A precise characterization of source memory retrieval requires an account of the effect of temporal contiguity (and other forms of contiguity explored in the following subsections) on source retrieval. Recently in the source memory literature, Popov et al (2021) investigated errors in a similar paradigm to ours and found that misbinding errors were most likely to come from neighboring serial positions, demonstrating a relationship between the probability of binding errors and serial position that can be explained as an effect of temporal contiguity. Building upon this finding, in our modeling, instead of freely estimating the probability of intrusions from each lag as Popov et al ( 2021) did, we assume that intrusion probability, like perceived similarity, decreases exponentially with increasing temporal distance (Howard & Kahana, 2002a;Murdock, 1997;Logan, 2021;Osth et al, 2018).…”

“…Specifically, we found that the temporal and spatial proximity of nontargets to targets at study affected the probability of intrusions. The temporal component of our model builds upon the Popov et al (2021) finding that intrusions are more likely to come from adjacent lags than distant lags. Rather than separately estimate intrusions from different lags, we constrained our temporal gradient model by assuming that intrusions follow an exponential decay function with directional asymmetry.…”

“…The link between swap errors in VWM and transposition errors in serial recall has been proposed to reflect a more general mechanism in memory by which items are bound to context dimensions (Oberauer & Lin, 2017;Schneegan et al, 2022). Our study aims to further extend the Popov et al (2021) findings by modeling how intrusion probability decreases with increasing distinctiveness in both the spatial and temporal domains.…”

“…Our modeling strategy was to start with a two-component mixture model like that of Zhang and Luck (2008), and then introduce successive elaborations of the intrusions model that represented different kinds of target-distractor similarity. First, we build upon the Popov et al (2021) finding of a temporal effect and introduce additional components in sequence. First, we extend the model of contextual similarity between items to include space as well as time, and then we add components that reflected the semantic and orthographic similarity between target and distractors, evaluating the improvement in model fit with each addition.…”

A Neural Index Reflecting the Amount of Cognitive Resources Available during Memory Encoding: A Model-based Approach

“…In the current work, we apply the SAC model of episodic memory to infer the amount of available resources during encoding, based on behavioral data during recall. The resource assumption in the SAC model has been supported by numerous behavioral findings (Kowialiewski et al, 2021;Mizrak & Oberauer, 2021;Popov et al, 2019;Popov & Reder, 2020;Popov et al, 2021). The inferred resource availability will then be used to validate the neural index of resource availability, obtained from the neural signal during encoding.…”