Saturday, August 20, 2005

Working memory and Gf - is complexity necessary?

More on Gf and working memory (I think I'm perseverating this morning).

Below is an interesting article that suggests that highly Gf/g predictive working memory tasks do NOT need to require complex cognitive processing. Rather, simple working memory tasks are as strongly related to higher-level cognitive processing if they are designed to place significant demands on executive-controlled attention. This finding is consistent with the exectuive-controlled attention working memory model of Engle and Kane (see prior post for brief explanation and link to their home page)

Lepine, R. & Barrouillet, P. What makes working memory spans so predictive of high-level cognition? Psychonomic Bulletin & Review 2005, 12 (1), 165-170

  • Working memory (WM) span tasks involving a complex activity performed concurrently with item retention have proven to be good predictors of high-level cognitive performance. The present study demonstrates that replacing these complex self-paced activities with simpler but computer-paced processes, such as reading successive letters, yields more predictive WM span measures. This finding suggests that WM span tasks evaluate a fundamental capacity that underpins complex as well as elementary cognitive processes. Moreover, the higher predictive power of computer-paced WM span tasks suggests that strategic factors do not contribute to the relationship between WM spans and high-level cognition.

Article highlights
  • A popular hypothesis regarding the strong working memory (Gsm-MW) and Gf/g relationship has been that working memory measures assesses a “fundamental capacity required by complex activities, which is conceived of as a capacity to control attention (Engle, Kane, & Tuholski, 1999) or to supervise and coordinate multiple-system functioning (Baddeley, 1990). Consequently, the activities included as processing components in WM span tasks are usually selected from those thought to require a high level of executive control (e.g., problem solving, reading comprehension, reasoning, mental calculation). The underlying idea is that more controlled and complex activities provide better WM span measures because complex activities tap the limited pool of cognitive resources sufficiently to disrupt maintenance and permit an accurate measure of WM capacity.”
  • The current investigation sought to ascertain if “complexity of the processing component of WM span tasks necessary to disrupt maintenance and accurately assess WM capacity?” The authors cite research that has shown that “very simple activities included as processing components in WM span tasks have an equally detrimental effect on recall as do complex activities, provided that they are not self-paced but computer-paced…the authors accounted for this effect by proposing that simple but time-constrained activities capture a sufficient amount of attention to disrupt the maintenance of items to be recalled.
  • The current study suggested that highly predictive working memory tasks do not necessarily need to require complex processing. The author’s state:
    • A first conclusion, therefore, is that the complexity of the processing component in most of the traditional WM span tasks (reading span, operation span, alphabet recoding, and ABCD) is a superfluous characteristic. In fact, self paced WM span tasks require complex activities to induce the necessary time pressure that is inherent to their structure.
    • The second conclusion is that the predictive value of the traditional WM spans does not stem from their capacity to assess an ability to strategically cope with the demands of complex span tasks—which would also be involved in any complex activity—because when the possibility of dealing strategically with the task is reduced by computer-paced presentation, the predictive value is increased. This does not mean that mnemonic strategies for encoding and maintaining memory items are unimportant in complex spans (McNamara & Scott, 2001). However, our results show that complexity and strategies do not contribute to the predictive value of WM spans but rather introduce more noise than information into their relationship with high-level cognition.
    • This conclusion has theoretical and practical implications. The predictive power of span tasks with simple processing components suggests that WM tasks measure some fundamental and general capacity involved in both elementary and complex cognitive processing.
    • The current study supports the controlled-attention view of working memory. The results are consistent with “models of WM that conceive of cognitive resources as a kind of mental energy required to produce activation (Engle, Cantor, & Carullo, 1992; Just & Carpenter, 1992; Lovett, Reder, & Lebière, 1999). This result is also in line with the time-based resource-sharing model (Barrouillet et al., 2004), which assumes, following Anderson (1993) and Cowan (1995), that this fundamental capacity is attentional.

No comments: