The role of working memory in the search of long-term memory

Đokić, Ratko. (2017). The role of working memory in the search of long-term memory. PhD Thesis. Filozofski fakultet u Zagrebu, Department of Psychology.
(Poslijediplomski doktorski studij psihologije) [mentor Domijan, Dražen].

Preview

PDF (Croatian) Download (3MB) | Preview

Abstract

Strategic search of long-term memory is based on retrieval cues by which the search is directed to narrower, well-defined search sets. By a successive switching of the search from one search set to another, the probability of finding target information increases compared to the situation in which long-term memory is searched in a non-strategic fashion, within one broader category. Generating the retrieval cues is considered to be a function of the working memory; in that way, working memory determines the efficiency of retrieval. However, there are opposite views on other functions of the working memory necessary for the process of strategic search. According to Rosen and Engle (1997), a persistent activity of the previouslyretrieved items produces a proactive interference which hinders generation of new retrieval cues that are necessary for continuation of the search; therefore, controlled inhibition of the previously-activated items is necessary for a successful continuation of the search. On the contrary, pursuant to the general hypothesis of the current research, and supported by the results of Unsworth and Engle (2007) and Unsworth, Spillers, and Brewer (2011), the release from proactive interference in the process of strategic search occurs automatically, independently of the individual’s working memory capacity. Namely, with each redirection of the search to the new semantic subcategory, items from the previous subcategory are being excluded from the current search set; by that means, the proactive interference produced by these items automatically fades away, as well. The above-stated hypothesis is tested within two consecutive studies. In Study 1, the process of strategic search of long-term memory by successive semantic subcategories was simulated by proactive interference build-up task based upon the Brown-Peterson paradigm. In this task, participants switched from learning and retrieving items from one subcategory, e.g. (forest) mammals, to learning and retrieving items from another subcategory, e.g. (nonprey) birds. Study 1 demonstrated that the prolonged learning and retrieving within one subcategory produces the proactive interference effect. While Experiment 1 (N = 67) did not show significant correlation between this effect and working memory capacity, introduction of additional cognitive load in Experiment 2 showed that this effect was more pronounced in the experimental condition (n = 59), compared to the control condition (n = 52). However, a switch to the new semantic subcategory, within one superior category (Experiment 1), as well as a switch to the new semantic subset, even within the single subcategory (Experiment 2), produced the proactive interference release effect. In accordance with the current hypothesis, this proactive interference release effect occurred automatically, i.e. independently of the subjects’ (available) working memory capacity. In addition, a greater proactive interference effect, as well as smaller proactive interference release effect in conditions in which subjects learned and retrieved less familiar names of birds could be understand as further indications of the importance of working memory in specifying efficient retrieval cues. Study 2 demonstrated that the proactive interference release effect also occurs automatically in memory tasks in which the search process is entirely internally organized, such as verbal fluency task. In Study 2, immediately after completing verbal fluency task, subjects engaged in lexical decision task; in that way, the level of inhibition of items activated in the course of long-term memory search was measured directly. Experiment 3 (N = 74) and Experiment 4 (N = 78) confirmed that the subjects with higher working memory capacity on their disposal (i.e. subjects with no additional cognitive load) implement more strategicallyorganized search of their long-term memory; as a consequence, these subjects generated more clusters and made more and faster switches between different semantic contexts which all resulted in their superior general fluency. At the same time, quantitative, not qualitative differences, in the manner of retrieval between the two groups of subjects (no load vs. load), suggest that individuals with high and low working memory capacity differ in the level of strategic organization of their search, but not in the very nature of the search. Accordingly, all subjects, independently of the presence of additional cognitive load during their performance of the verbal fluency task, accomplished the fastest reaction times in response to the most frequent stimuli presented in the lexical decision task; this pattern was consistent, irrespective of whether these stimuli during the search were activated experimentally (Experiment 3) or internally (Experiment 4). In conclusion, these results question the assumption of Rosen and Engle (1997) on the controlled inhibition of the previously-activated and currently-interfering items as a necessary prerequisite for the implementation of the strategic search of long-term memory. Instead, the current results affirm hypothesis on the automatic release of proactive interference as a beneficial side-effect of the strategic search.

Actions (login required)

View Item