Elsevier

Intelligence

Volume 40, Issue 6, November–December 2012, Pages 531-542
Intelligence

Working memory training does not improve intelligence in healthy young adults

https://doi.org/10.1016/j.intell.2012.07.004Get rights and content

Abstract

Jaeggi and her colleagues claimed that they were able to improve fluid intelligence by training working memory. Subjects who trained their working memory on a dual n-back task for a period of time showed significant improvements in working memory span tasks and fluid intelligence tests such as the Raven's Progressive Matrices and the Bochumer Matrices Test after training compared to those without training. The current study aimed to replicate and extend the original study in a well-controlled experiment that could explain the cause or causes of such transfer if indeed the case. There were a total of 93 participants who completed the study, and they were assigned to one of three groups—passive control group, active control group and experimental group. Half of the participants were assigned to the 8-day condition and the other half to the 20-day condition. All participants completed a battery of tests at pre- and post-tests that consisted of short timed tests, a complex working memory span and a matrix reasoning task. Although participants' performance on the training task improved, results from the current study did not suggest any significant improvement in the mental abilities tested, especially fluid intelligence and working memory capacity, after training for 8 days or 20 days. This does not support the notion that increasing one's working memory capacity by training and practice could transfer to improvement on fluid intelligence as asserted by Jaeggi and her colleagues.

Highlights

► We replicate a study that claims to improve intelligence by training working memory. ► Participants in our study train their working memory for 8 days or 20 days. ► Working memory training does not improve verbal, perceptual or spatial abilities. ► Working memory training does not improve intellectual ability.

Introduction

Part of the nature versus nurture debate is the issue of the malleability of intelligence (Wahlsten, 1997)—can the environment modify intellectual ability? Interests and efforts to raise intelligence as well as other cognitive abilities have been around for more than a century (Spitz, 1986). The idea that it may be possible to manipulate intelligence has been very appealing to researchers in education and the behavioral sciences, and the large body of research focused on aspects of the treatment of intellectual impairment provides an excellent example of these efforts.

Long term intervention programs to improve intelligence such as Head Start and the Abecedarian Project have not been successful. At the completion of the Abecedarian Project, results showed that there was substantial improvement in IQ scores in the experimental group compared to the control group but the superior performance quickly decreased when the project ended (Spitz, 1986). A large number of studies involving short-term interventions have also been conducted. In general, short-term intervention programs have not significantly improved latent ability but may have only increased task specific variance. These studies suggested that training or repeated practice on a task with instructional aid improved performance on that specific task but rarely did the improvement transfer to other general cognitive abilities (Belmont and Butterfield, 1977, Ferrara et al., 1986).

The past several years have seen a proliferation of research on cognitive training, especially working memory training (Morrison & Chein, 2011). Many studies have reported training and transfer effects as a result of working memory, executive function, and attention type training, most of which were done in young children or the older adult population. Klingberg, Forssberg, and Westerberg (2002) and Klingberg et al. (2005) observed improvements in matrix reasoning tasks besides reduced inattentive symptoms in children with Attention Deficit Hyperactivity Disorder (ADHD). Others found improvements in fluid reasoning after training on working memory in the older adult population (Borella et al., 2010, Schmiedek et al., 2010), task switching training in three different age groups (Karbach & Kray, 2009), executive control/planning training in the older adult population (Basak, Boot, Voss, & Kramer, 2008) and attention training in children (Rueda, Rothbabrt, McCandliss, Saccomanno, & Posner, 2005). Studies done on old adults that employed working memory training (Buschkuehl et al., 2008) and strategic training (Carretti, Borella, & De Beni, 2007) impacted memory performance, a near transfer effect, and these studies did not report significant improvements on gf tasks. Van der Molen, Van Luit, Van der Molen, Klugkist, and Jongmans (2010) found short-term memory improvement but no IQ improvements in adolescents with mild intellectual disability. Minear and Shah (2008) employed a task-switching training paradigm and did not report improvements on IQ. Li et al. (2008) reported no far transfer effects from working memory training to complex span tasks and did not report improvements in IQ.

There are plenty of studies that found near transfer after working memory, executive functions and attention training but not so much on far transfer effects. Near transfer refers to changes in a domain caused by changes in another similar domain due to comparable ability or process, and far transfer effects refer to changes in domains caused by changes in a separate domain of different processes. Many of these studies were conducted on children ranging from 4 to 11 years of age, when their cognitive abilities were still developing and have not reached maturity (Fry & Hale, 2000). Bergman-Nutley et al. (2011) reported consistent near transfer effects in their study on 4-year old children. The authors observed improvements on reasoning tasks in groups that trained on reasoning skills, and they did not find transfer effects from working memory training to reasoning or fluid intelligence tasks. More examples of near transfer effect include a study by Mackey, Hill, Stone, and Bunge (2011), St. Clair-Thompson, Stevens, Hunt, and Bolder (2010) and Thorell, Lindqvist, Bergman-Nutley, Bohlin, and Klingberg (2009) that reported near transfer effects but neither far transfer effects nor IQ improvements. Holmes, Gathercole, and Dunning (2009) reported that IQ scores were unaffected by working memory training in children with ADHD, and Holmes et al. (2010) reported no boost in IQ performance in children with low working memory capacity. A recent review by Diamond and Lee (2011) on cognitive training conducted on children concluded that only core executive function—working memory, cognitive flexibility and inhibition—training is most beneficial to 4–12 year-olds, and most studies cited in this review reported near transfer effects (Diamond & Lee, 2011).

Nearly all the studies mentioned above that found transfer effects were studies that were conducted on children and older adults, when their cognitive development or decline was relatively malleable than young adults (Borella et al., 2010, Fry and Hale, 2000). Modifying cognitive abilities did not seem to be difficult during periods of growth when intervention could facilitate and perhaps accelerate development and maturity (Rueda et al., 2005). Modification also seemed possible in old adults when their cognitive decline could be delayed with intervention (Basak et al., 2008, Borella et al., 2010, Buschkuehl et al., 2008), consistent with the “disuse” hypothesis (Orrell & Sahakian, 1995). This hypothesis has been supported by animal and human studies that demonstrated considerable neuronal plasticity due to increased activities from experiential input and perceptual-sensory stimuli (Tranter & Koutstaal, 2008). However, studies that suggested possible modification of cognitive abilities in young adults, when general cognitive abilities are less malleable compared to childhood and aging periods were considerably fewer (Jaeggi et al., 2008, Jaeggi et al., 2010, Karbach and Kray, 2009, Schmiedek et al., 2010). More studies should be carried out on healthy, young adults before claiming with confidence that general cognitive abilities such as fluid reasoning could be improved with short periods of cognitive training.

Jaeggi et al. (2008) reported that they had improved fluid intelligence (gf) of young adults in a study by training their working memory (WM) through repeated practice with a dual N-back task. They argued that since WM and gf shared common variance (Ackerman et al., 2005, Colom et al., 2003, Fry and Hale, 1996, Jurden, 1995, Kane et al., 2005, Kyllonen and Christal, 1990, Oberauer et al., 2005, Stauffer et al., 1996, Tucker and Warr, 1996, Verguts and De Boeck, 2002), engaging neural circuits shared by WM and gf by training WM may transfer to improvements in gf. Studies in cognitive psychology and neuroscience that tried to explain the relationship between WM and gf such as Halford, Cowan, and Andrews (2007) suggested that WM and reasoning skills shared related capacity limits, and that the common thread between the two functions was the shared requirement to bind elements to slots of a hypothetical coordinate system in one's memory. The process of maintaining the bindings between elements required attention, which was essential to WM and reasoning abilities (Halford et al., 2007). Gray, Chabris, and Braver (2003) suggested that the relation between gf and WM was mediated by activities in the lateral prefrontal and parietal regions. Kane and Engle (2002) reported that the dorsolateral prefrontal cortex could have a role in WM especially related to attention control. Conway, Kane, and Engle (2003) supported the hypothesis by Gray et al. (2003) that WM span tasks activate regions in the prefrontal cortex when the executive-control mechanism is recruited to combat interference during the maintenance and manipulation of information.

The assumption that WM and gf may share the same neural network and mental resources was the theory behind Jaeggi et al.'s (2008) hypothesis that training to improve one's WM could transfer the improvement to gf. The adaptive nature of the dual N-back task used for training was intended to engage the executive attention at all times so that automatic responses could not develop. It was suggested that under consistent format and information conditions, practice would lead to automatic responses and less mental or attentional resources would be employed. However, in variable information or inconsistent contexts, controlled processing that utilized mental resources would still have been taken place even after practicing on the task for a considerable period of time (Ackerman, 1987). Jaeggi et al. (2008) were essentially targeting participants' executive attention in their WM training with the implementation of the adaptive feature of the training task. Therefore, an increase in WM performance at the end of training could mean an increase in attention span. If the effects of WM could be quantified through the analysis of goal management (Carpenter, Just, & Shell, 1990), improvements in WM could mean better and improved ability to manage representations of information. These effects could be measured by improved performance on fluid ability tests, such as the Raven's, and mental rotation tests where steps of abstractly manipulating the movements of three-dimensional objects must be actively managed in one's mind.

After decades of unsuccessful and inconclusive research and efforts to raise intelligence, a study that suggested otherwise in a sample population of individuals when their general cognitive ability is strongly suggested to have matured (Cattell, 1987) should be subjected to further examination through replication. Moody (2009) identified some weaknesses in the study by Jaeggi et al. (2008), and one of them was participants in the experimental group did not take the same IQ test. Those who trained the least (8 training sessions) did the Raven's Advanced Progressive Matrices (RAPM) and the rest of the participants who trained either 12, 17 or 19 days took the Bochumer Matrices Test (BOMAT). The researchers in the study reported significant group differences for the 12-day, 17-day and 19-day training groups. Participants who performed the RAPM in the 8-day group did not show significant improvement in their IQ scores after training for 8 days (Jaeggi et al., 2008). The reason behind this observation could either be that longer training produced more score gains, or the nature of the training task itself facilitated better test taking specifically for the BOMAT (Moody, 2009).

Both the IQ tests—the RAPM and BOMAT—employed in the study by Jaeggi et al. (2008) shared some similarities. Each item in both tests was a matrix of figures and a spot in the matrix was left blank. Items on RAPM consisted of 3 × 3 matrices, while items on BOMAT consisted of 5 × 3 matrices. The figures in the matrices were arranged according to a pattern, and test-takers would have to identify the pattern, or patterns, unique to each matrix in order to infer a solution from multiple possible answers given at the bottom of the matrices. Another similarity between the two tests was that the difficulty level of the questions increased as the test-taker progressed through either test. A high score would reflect the test-taker's ability to solve the difficult items. Test-takers would also be able to learn how to solve subsequent questions that were progressively harder based on the patterns they inferred from previous items that they had solved. Participants in the Jaeggi et al. (2008) study were not given the opportunity to attempt the more difficult questions because the researchers essentially removed the progressive nature of the tests by reducing the allotted time to take the test from 45 min to 10 min (Moody, 2009). When Jaeggi et al. (2010) replicated their results, they defended their time constraint testing protocol by citing studies by Salthouse (1993) and Unsworth and Engle (2005) that suggested no evidence for differential working memory effects for the various items on the RAPM. Jaeggi and her colleagues argued that these studies (Salthouse, 1993, Unsworth and Engle, 2005) provided justification that limiting participants to the first several items on the RAPM and BOMAT does not affect their measure of Gf (Jaeggi et al., 2010).

The present study aimed to replicate and extend the findings reported by Jaeggi et al. (2008), in addition to correcting the potential confounds identified by Moody (2009) and comparing results of WM training with well-controlled groups. In other words, the treatment group should differ from the control group only by the very element that affected the outcome. The objective of the current study was to determine if effects from WM training specifically, a core executive function (Diamond & Lee, 2011), would transfer to improvement in g and/or gf (producing a far transfer effect) in young adults. Though Jaeggi et al., 2008, Jaeggi et al., 2010 specifically targeted gf, this construct was suggested to be inadequate to capture the general intellectual ability, g (Johnson & Bouchard, 2005a). A hierarchical model of human intelligence has been widely accepted in the field of intelligence (Deary, 2001). The three-strata model of human intellectual abilities proposed by Carroll in 1993 (Carroll, 2003), and the practice of structural equation modeling in the field have been providing converging support that human intelligence could be described in terms of a hierarchical structure. Johnson and Bouchard (2005a) suggested a model of general intelligence, g, which influenced three factors—verbal, perceptual and mental rotation (VPR). The VPR model proposed has been repeatedly and empirically tested (Johnson and Bouchard, 2005a, Johnson and Bouchard, 2005b, Johnson et al., 2007) and offered an excellent example of the current direction and general acceptance of a hierarchical organization of human intelligence; thus this model would be applied in the present study as a measure of g using a theoretically based, comprehensive battery of cognitive ability tests to provide insight into specific mental processes influenced and not influenced by WM training. As Jaeggi et al. (2008) stated, “… tasks that measure Gf are picking up other cognitive skills as well, and perhaps the training is having an effect on these skills even if measures of capacity are not sensitive to them” (p. 6831). If WM training could improve intellectual performance, the improvement may be due to global effects or specific effects. The present study included a series of tests that supported the model postulated by Johnson and Bouchard (2005a), and none of the tests used possessed features similar to those in the working memory task used for training. The present study predicted that there would be no improvements in verbal and perceptual tests, but there could be improvements in spatial ability and matrix reasoning tests.

Section snippets

Methods

Participants were students enrolled in Introduction to Psychology, Health Psychology or Quantitative Methods in Psychology at a private Midwestern university (N = 130). They were initially randomly assigned into one of 6 groups—2 experimental groups, 2 passive control groups and 2 active control groups. Participants assigned to the experimental or active control groups were given the option to skip the lab component completely and only return to post-test session only, thus they were allowed to

Results

There were a total of 130 participants who came to pre-test sessions but only 93 of them came back for post-test sessions—60 of whom were females. The following results were obtained from analyzing data provided from the 93 students who completed both pre- and post-tests. They averaged slightly younger than 20 years of age. There were more participants in the passive control group compared to the active control or experimental group because those assigned to the latter two groups were given the

Discussion

Results from the current study did not suggest improvement in general intelligence after repeated training on a challenging working memory task. Our prediction that spatial and reasoning abilities could be improved after working memory training was not supported. Paired t-tests, Analysis of Variance (ANOVA) and Analysis of Covariance (ANCOVA) did not show any significant changes between pre- and post-test scores across the experimental and control groups. This observation is in contrast to the

Role of the funding source

This work was supported by the Ministry of Higher Education, Malaysia as part of the corresponding author's dissertation. Besides financial contribution, the sponsor was not involved in any part of the preparation and execution of this work.

Acknowledgments

The authors would like to thank the following individuals for their support and contribution to the study: Professor Douglas Detterman, Dr. TJ McCallum, Professor H. Gerry Taylor, Michelle Blair, London Holt, Jasmine Core and Linda Lou.

References (61)

  • W. Johnson et al.

    Replication of the hierarchical visual-perceptual-image rotation model in de Wolff and Buiten's (1963) battery of 46 tests of mental ability

    Intelligence

    (2007)
  • T. Klingberg et al.

    Computerized training of working memory training in children with ADHD—A randomized, controlled trial

    Journal of the American Academy of Child and Adolescent Psychiatry

    (2005)
  • P.C. Kyllonen et al.

    Reasoning ability is (little more than) working-memory capacity?!

    Intelligence

    (1990)
  • D.E. Moody

    Can intelligence be increased by training on a task of working memory?

    Intelligence

    (2009)
  • P. Tucker et al.

    Intelligence, elementary cognitive components, and cognitive styles as predictors of complex performance

    Personality and Individual Differences

    (1996)
  • N. Unsworth et al.

    Working memory capacity and fluid abilities: Examining the correlation between Operation Span and Raven

    Intelligence

    (2005)
  • T. Verguts et al.

    On the correlation between working memory capacity and performance on intelligence tests

    Learning and Individual Differences

    (2002)
  • P.L. Ackerman

    Individual differences in skill learning: An integration of psychometric and information processing perspectives

    Psychological Bulletin

    (1987)
  • P.L. Ackerman et al.

    Working memory and intelligence: The same or different constructs?

    Psychological Bulletin

    (2005)
  • C. Basak et al.

    Can training in a real-time strategy video game attenuate cognitive decline in older adults?

    Psychology and Aging

    (2008)
  • J.M. Belmont et al.

    The instructional approach to developmental cognitive research

  • S. Bergman-Nutley et al.

    Gains in fluid intelligence after training non-verbal reasoning in 4-year-old children: A controlled, randomized study

    Developmental Science

    (2011)
  • E. Borella et al.

    Working memory training in older adults: Evidence of transfer and maintenance effects

    Psychology and Aging

    (2010)
  • M. Buschkuehl et al.

    Impact of working memory training on memory performance in old-old adults

    Psychology and Aging

    (2008)
  • P.A. Carpenter et al.

    What one intelligence test measures: A theoretical account of the processing in the Raven Progressive Matrices Test

    Psychological Review

    (1990)
  • B. Carretti et al.

    Does strategic memory training improve the working memory performance of younger and older adults?

    Experimental Psychology

    (2007)
  • R.B. Cattell

    Intelligence: Its structure, growth, and action

    (1987)
  • A. Diamond et al.

    Interventions shown to aid executive function development in children 4 to 12 years old

    Science

    (2011)
  • R.A. Ferrara et al.

    Children's learning and transfer of inductive reasoning rules: Studies of proximal development

    Child Development

    (1986)
  • A.F. Fry et al.

    Processing speed, working memory, and fluid intelligence: Evidence for a developmental cascade

    Psychological Science

    (1996)
  • Cited by (162)

    View all citing articles on Scopus
    View full text