[ CogSci Summaries home | UP | email ]

Rosen, V. M., & Engle, R. W. (1998). Working memory capacity and suppression. Journal of Memory and Language, 39, 418-436.

  author = 	 {Rosen, Virginia M. and Engle, Randall W.},
  title = 	 {Working memory capacity and suppression},
  journal = 	 {Memory and Language},
  year = 	 {1998},
  volume = 	 {39},
  pages = 	 {418--436}

Author of the summary: Leah MacQuarrie, 2011, lmacquar@connect.carleton.ca

Cite this paper for:

The actual paper can be found at http://www.psychology.gatech.edu/renglelab/Publications/1998/Working%20memory%20capacity%20and%20suppression.pdf

Engle, 1996: Working memory capacity is correlated with many higher level cognitive processes [418].

“Complex span” (ability to remember lists, as opposed to numbers or letters) has been found to be related to [418]:

  1. reading comprehension (Daneman & Carpenter, 1980, 1983)
  2. language comprehension (King & Just, 1991; MacDonald, Just, & Carpenter, 1992)
  3. vocabulary learning (Daneman & Green, 1986)
  4. following directions (Engle, Carullo, & Collins, 1991)
  5. reasoning (Kyllonen & Christal, 1990)
  6. complex learning (Shute, 1991; Kyllonen & Stephens, 1990).

Working memory capacity is related to ability to allocate focused attention to tasks (Conway & Engle, 1994; Engle, 1996; Rosen & Engle, 1997). [418]

Working memory capacity is important in task performance when: [419]

  1. information must be retained while facing distractions
  2. strategic search is necessary
  3. monitoring for errors is necessary
  4. suppression of errors is necessary
Baddeley & Hitch, 1974: The construct of the “central executive” is a place for storage and processing of stimuli. Rosen & Engle suggest that it is also responsible for suppressing intrusions on tasks. [419]

Rosen & Engle address two questions in the current study: [419]

  1. Is working memory capacity related to individuals’ vulnerability to intruding thoughts?
  2. Is working memory capacity related to individuals’ ability to suppress such thoughts?
Rosen and Engle used a paired-associates task, in which participants have to learn several lists of word pairs, to study the effects of individual differences in working memory capacity on interference [420].

Experiment 1: To assess their first question, the authors stressed fast response time during testing. This ensured that participants provided the first response that came to mind.
Results: participants with higher working memory capacity produced fewer intrusions (incorrect word pairs) during recall, and required fewer trials to reach the minimum number of correct responses. [424]
These findings suggest that the high memory capacity participants are less vulnerable to intrusions of previously learned word pairs during their test. [425]

Experiment 2: To assess their second question, the authors’ focus was on accuracy, not response time. This gave participants sufficient time to suppress incorrect responses before answering. [428]
Results: the participants with higher working memory capacity showed higher response times than their control group, which suggests that they were suppressing previously learned word lists in order to provide the correct answer. The low working memory capacity participants showed faster response times than their control group, suggesting that they did not experience any suppression of incorrect answers. [430]
These results indicate that a relationship exists between working memory capacity and ability to suppress intrusions. [430].

We can’t expect low span and high span individuals to behave the same way during testing. It is speculated that low working memory capacity participants experience more overall interference than participants with high working memory capacity. The authors refer to this as “general” interference, and it is due to elements outside of the experiment and stimuli. [431]

Rosen & Engle’s (1997) four category fluency model can be used to explain how information retrieval occurs during the paired-associates task. It consists of:

  1. Automatic spreading of activation, which occurs as a result of being presented with a cue. This does not require controlled attention, so the low and high span participants show no difference in this component. [432]
  2. Covert monitoring for errors, which occurs when presented with a new list. It is a controlled process, and so, requires sufficient time to be implemented. [432]
  3. Suppression of previously retrieved information, which occurs when participants are presented with an old cue that now matches a new response. High memory capacity participants perform better at this task than low memory capacity participants. It is possible that low span participants are unable to suppress intrusions because their attention can only accommodate one of the three controlled components at once. [433]
  4. Controlled search, which is necessary when correct responses are not readily available. It requires sufficient time to be implemented. The authors predict that the high span participants would perform this task better, but the current experiment did not test this hypothesis. [433]

Dempster, 1991, 1992: the ability to resist interrupting thoughts may be a source of individual differences, and this ability is related to frontal lobe functioning. [433]

Shimamura et. al, 1995: patients with frontal lobe deficits showed greater vulnerability to interference during a paired-associates task. [434]

Evidence that frontal lobe function is implicated in working memory capacity: Uhl et. Al, 1990, found evidence that the frontal lobe is active in normal participants during a paired-associates task. Rosen & Engle have found working memory capacity to be important for the same task. Due to this, the authors say “it seems only a short step to suggest that a relationship exists between working memory capacity and frontal lobe functioning". [434]

Engle, Tuholski, Laughlin, & Conway, in press (see summary author’s notes): a relationship exists between working memory capacity and general intelligence. This has been linked to frontal lobe functioning (Duncan, Emslie, & Williams, 1996). [434]

The authors recognize that in suggesting a relationship between working memory capacity and frontal lobe functioning, they seem to be comparing individuals with low working memory capacity with frontal lobe patients. However, they believe that both of these groups share a deficit in controlled attention, which decreases an individual’s ability to suppress interference. [434]

The ability to suppress intrusive thoughts and behaviours may demonstrate a link between working memory capacity and frontal lobe functioning. [434]

Summary author's notes:

Back to the Cognitive Science Summaries homepage
Cognitive Science Summaries Webmaster:
JimDavies (jim@jimdavies.org)