The role of distractor strength in visual working memory

Tomić, Ivan. (2019). The role of distractor strength in visual working memory. PhD Thesis. Filozofski fakultet u Zagrebu, Department of Psychology.
(PDS Psihologije) [mentor Ivanec, Dragutin].

Preview

PDF (Croatian) Download (7MB) | Preview

Abstract

Introduction Our environment is overloaded with visual information, with only a fraction of them necessary for an ongoing task. What determines the success of performing everyday tasks in such an environment? Visual working memory (VWM) is considered a vital component of most complex behaviours, but previous studies provided evidence of its susceptibility to irrelevant visual information, i.e. distractors. When is VWM impervious to distractors? VWM is best described as a highly limited resource that is flexibly shared among items in a visual scene. As the number of items increases, the amount of resources allocated to each item decreases, leading to a decline in strength (i.e. quality) of memory representations, and consequently to less precise recall of each item. This finding is consistent with a resource model of VWM and has been highlighted as a hallmark observation in VWM studies. An alternative view is provided by the influential "slot" model of VWM which claims that VWM is limited with a fixed maximum number of items that can be held in memory at one time. Moreover, according to this model, an item is either represented in its entirety in a memory slot or not stored at all. Strength of memory representation is therefore almost completely neglected in this type of model. However, investigation of the role of representation strength, besides the well known set size effect, has been limited even in studies motivated by the resource model. It is an intuitive prediction that memorandum strength should influence task performance. For example, our ability to maintain relevant information in the presence of distracting visual input should depend on the strength of memoranda, with stronger representations suffering less from irrelevant visual input. Here we thoroughly investigated the roles of strength of representations in VWM and distractor strength. Method To this end, we conducted four experiments (N = 64) in which we systematically manipulated the strength of VWM representations and the strength of distractors. We employed a delayedestimation task with continuous report, wherein subjects memorized colour stimuli. We manipulated strength of representation by: manipulating set size and encoding time (experiment 1), prioritizing one item for recall (experiment 2a), refreshing a representation of a single item during maintenance (experiment 2b), or interrupting memory phases before and after a stable representation was formed (experiment 3). After showing a memory array but before recalling one of the memorized items, irrelevant visual stimuli were shown. Simultaneously with memoranda strength, we manipulated distractor strength (no distractors, weak distractors, strong distractors). Results In all experiments we consistently found evidence that strength of memoranda serves a protective role against visual distractors. Regardless of the method used to manipulate memoranda strength, recall of stronger memoranda was less vulnerable to distractors. On the other hand, the effect of distractor strength showed a complex pattern which differed between experiments and depended on the manipulation of VWM representation strength. Next, we fit the data with a mixture model which assumes that the recall error distribution is a mixture of target recall, swap errors, intrusions, and guesses. This model captured the data well and showed better fit than the alternative normal + uniform model. When analyzing parameters of this model the most consistent manipulation dependent changes were observed on the target recall parameter. Intrusions were observed in all experiments. Interestingly, they depended more on VWM representation strength than on distractor strength. Conclusion This study revealed that strength of memoranda in VWM serves a protective role against visual distractors, making any addition of mechanisms of memoranda protection or distractor inhibition unnecessary. However, we showed that distractors, regardless of their strength, are able to penetrate VWM and lead to a decrease in recall precision. These findings are consistent with a resource conceptualisation of VWM where representational strength (i.e. the amount of allocated resources) plays a crucial role in ability to perform a task.

Actions (login required)

View Item