The reliability and internal consistency of one-shot and flicker change detection for measuring individual differences in visual working memory capacity

Abstract

We investigated the psychometric properties of the one-shot change detection task for estimating visual working memory (VWM) storage capacity-and also introduced and tested an alternative flicker change detection task for estimating these limits. In three experiments, we found that the one-shot whole-display task returns estimates of VWM storage capacity (K) that are unreliable across set sizes-suggesting that the whole-display task is measuring different things at different set sizes. In two additional experiments, we found that the one-shot single-probe variant shows improvements in the reliability and consistency of K estimates. In another additional experiment, we found that a one-shot whole-display-with-click task (requiring target localization) also showed improvements in reliability and consistency. The latter results suggest that the one-shot task can return reliable and consistent estimates of VWM storage capacity (K), and they highlight the possibility that the requirement to localize the changed target is what engenders this enhancement. Through a final series of four experiments, we introduced and tested an alternative flicker change detection method that also requires the observer to localize the changing target and that generates, from response times, an estimate of VWM storage capacity (K). We found that estimates of K from the flicker task correlated with estimates from the traditional one-shot task and also had high reliability and consistency. We highlight the flicker method's ability to estimate executive functions as well as VWM storage capacity, and discuss the potential for measuring multiple abilities with the one-shot and flicker tasks.

Visual working memory capacity increases between ages 3 and 8 years, controlling for gains in attention, perception, and executive control

“…2 A) Schematics of arrays used in Feature Search and Conjunction Search tasks. B) Response times observed in both tasks decreased as a function of age storage limit of VWM (K), using an approach that we have used for adult participants (Pailian & Halberda, 2015).…”

“…Reinforcing the idea that the One-Shot task may be strongly affected by such differences, we have found in our own work with adults that that the One-Shot paradigm produces highly variable capacity estimates even within a single participant. One-Shot VWM capacity estimates varied significantly within each individual observer depending on the parameters used, and large numbers of trials were required to return reliable estimates (Pailian & Halberda, 2015). Such differences might help to account for the differences in VWM storage capacity observed by Cowan and colleagues (2005), Riggs andcolleagues (2006), andSimmering (2012), especially considering that our work in adults (Pailian & Halberda, 2015) suggests that the small number of trials presented to each participant (as is almost unavoidable when testing children) may have led to widely varying capacity estimates.…”

“…For example, although the brief presentation of items in the One-Shot task is intended to minimize the verbal recoding of information and to disrupt any engagement of complex strategies, this short viewing time may, by preventing participants from attending each item individually, lead participants to recruit mechanisms of global scene processing (e.g., ensemble statistics) (Brady & Alvarez, 2015;Pailian & Halberda, 2015).…”

“…In the Flicker task, by measuring an observer's reaction time to find the changing target across trials and by manipulating set size to estimate search rate, we obtain an estimate of VWM storage capacity. In our previous work with adults (Pailian & Halberda, 2015), we found that Flicker estimates of VWM storage capacity are highly correlated with those produced by the One-Shot task (r = 0.70, p < 0.001). However, exploration of the psychometric properties of these two tasks revealed that the Flicker task produced more reliable and stable estimates of VWM storage capacity in adults than the One-Shot task and that performance in the Flicker task was less sensitive to changes in task parameters (Pailian & Halberda, 2015).…”

“…To meet these goals, we developed a new "Flicker" task (Halberda, Pailian, Wetherhold, & Simons, 2006;Pailian & Halberda, 2015) that uses displays like those in the One-Shot task (e.g., to-be-remembered array of items, followed by a probe array in which one item has changed). Although the One-Shot task allows only a single viewing of the to-beremembered array, the Flicker task presents the two arrays in continuous alternation, whereas the observer searches through these alternations until they find the changing target (a changing target is present on every trial).…”