Abstract
Working memory is a hot research topic in psychology, education, and related fields. A literature search of the PsycINFO database (conducted on January 3, 2014) going back 5 years produced 2,819 results with “Working Memory” in the title, while a search of ERIC resulted in 409 articles. Baddeley and Hitch introduced the construct of working memory 40 years ago, but it is only in the past 10 to 20 years that we have begun to fully understand the crucial role working memory plays in nearly all aspects of cognitive functioning and learning. Baddeley’s tripartite model of working memory, with a “central executive” managing two modality-specific, short-term memory storage components, has withstood the test of time because it has been consistently validated.
Besides a well-supported model, there are several reasons for the current high interest in working memory. First, advances in neuroimaging technology have demonstrated that working memory functioning is not confined to the dorsolateral prefrontal cortex. Second, educational research has documented that working memory is highly predictive of academic learning. Third, there is a recent fascination with understanding and measuring executive functioning (working memory is considered an executive function). Fourth, the aging population has spurred interest in the role of working memory in cognitive decline. Fifth, the recent discovery that working memory training can change the brain has spawned interest in working memory interventions. Finally, the controversy over working memory’s relations with ADHD and inhibition is ongoing. Some researchers contend that poor inhibition accounts for both ADHD and working memory deficits, whereas others propose that a working memory deficit underlies ADHD.
The eight research articles in this special section contribute to the expanding empirical knowledge base, and they also provide important information that practitioners can apply when diagnosing and treating ADHD and comorbid working memory deficits. Four of the investigations include ADHD subjects. Van Ewijk et al. examine the neurocognitive relation between ADHD symptoms and visual-spatial working memory, whereas, Hudec et al. explore the functional relation between motor activity and working memory. Alloway and Cockcroft consider cultural differences in modality specific working memory deficits in ADHD subtypes. Gropper et al. round out the studies with ADHD subjects by assessing the efficacy of Cogmed training with college students. Three of the articles report factor analyses of measurement instruments. Both the Engel de Abreu et al. and Norman and Tannock studies analyze the Working Memory Rating Scale and recommend a five-item, brief version of it for screening working memory in the classroom. In the other measurement article, Dehn analyzes a broad cognitive processing scale and discovers an executive functions factor that includes working memory. Finally, two of the studies evaluated the effectiveness of computerized working memory training, Gropper et al. with college students and Wong et al. with Chinese children.
In the first article, Alloway and Cockcroft pursue a better understanding of the relation between ADHD and working memory. They begin by reminding readers that working memory impairments are usually associated with the inattentive and combined subtypes of ADHD, rather than the hyperactive-impulsive subtype. They also cite research that found ADHD children to demonstrate visuospatial working memory deficits while possessing normal verbal working memory. Given this background, Alloway and Cockcroft investigate the potential influence of cultural factors on the profile of working memory deficits in children with ADHD. They do so by comparing the performance of British and South African children on measures of both short-term and working memory. Only participants with the combined subtype of ADHD were included.
Relative to normal peers, ADHD participants from both cultures demonstrated deficits in both verbal and visuospatial tests of working memory. The results indicate that children with ADHD have similar working memory profiles across cultures. However, a potential cultural difference was discovered when a discriminant function analysis was performed. In the United Kingdom, visuospatial working memory was the best test for discriminating between ADHD and typically developing peers, but for the South African sample, verbal working memory was the best test to use for classification. The authors speculate that the distinction could be due to the fact that half of the South African participants were bilingual. The results of this study should encourage clinicians to include visuospatial working memory tests when conducting assessments with ADHD children.
Three of the articles explore recently developed rating scales that measure working memory and related processes. Normand and Tannock analyze the Working Memory Rating Scale (WMRS), a 20-item teacher rating scale dedicated to working memory assessment. Their investigation replicates and extends earlier factor-analytic research by sampling 524 Canadian children ages 6 to 9 and by having different sets of teachers rate the same participants in years 1 and 2 of the study. Original factor-analytic studies by the WMRS authors suggested that all of the items aligned with a general factor. However, when Normand and Tannock conducted confirmatory factor analysis they discovered that a short five-item model was a better fit with the scale’s working memory factor than the original 20 items. The study’s authors suggest that some of the 20 items are redundant and that reliable and valid assessment could be conducted by having teachers rate only five readily observable behaviors, such as “depends on neighbor to remind them of the current task.”
The study also examined internal consistency, convergent validity, and criterion-related validity with the proposed five-item scale. In addition to having excellent internal consistency and inter-rater reliability, this brief, alternative scale had significant correlations with reading and mathematics achievement scores. High levels of working memory difficulty were associated with reading and math deficiencies 6 and 18 months later. Furthermore, the alternative five-item scale generally had significant correlations with performance-based measures of working memory.
The study’s findings have important implications for the early identification of working memory impairments. A short screening instrument that could be completed by untrained school personnel within a few minutes is much more likely to be adopted and utilized by school systems than time intensive procedures. Also, when screening is conducted in pre-school or the early elementary years, evidence-based interventions can be implemented that might diminish future learning problems.
In a related investigation of the WMRS, Engel de Abreu et al. replicate Normand and Tannock’s factor-analytic findings, and they examine the utility of the WMRS with a diverse Latin American population. Engel de Abreu et al. used a Portuguese translation of the WMRS to collect data on 355 monolingual Brazilian children. Factor analysis of the Brazilian data confirmed that the five-item scale identified by Normand and Tannock in Canada (the WMRS-Br) was superior to the original 20-item scale developed in the United Kingdom. The psychometric properties of the Portuguese WMRS-Br were highly similar to the original 20-item English version, with good internal reliability and evidence for construct validity. Children who obtained poor scores on objective measures of working memory received higher (problematic) scores when rated by teachers with the WMRS-Br. Furthermore, WMRS-Br scores were associated with objective measures of selective attention, switching, and interference suppression but not with response inhibition. The study’s authors conclude that the WMRS has diagnostic utility in a Latin American context and with children from a broader range of socioeconomic backgrounds than those typically found in the United Kingdom.
In the third measurement article, Dehn conducts factor analyses of the Children’s Psychological Processes Scale (CPPS), as well as two concurrent validity studies. The CPPS is a 121-item teacher rating scale that samples 11 psychological processes. After a general factor, an executive functions factor that included the attention and working memory subscales was the second factor to emerge. This finding supports the theoretical distinction between cognitive processes and the executive functions that direct and manage cognitive processes. Interestingly, the working memory subscale aligned more strongly with the general cognitive processing factor than it did with the executive functions factor, highlighting working memory’s strong interrelations with cognitive processes. Moreover, this finding is consistent with this section’s articles on the relations between attention and working memory. Apparently, working memory is at least somewhat separable from executive functions and specifically from attention.
The CPPS executive functions factor and working memory subscale received further validation from high correlations with the composites and subscales of a prominent executive functions rating scale. Another validity study included in the article found most CPPS subscales to have moderate correlations with comparable, performance-based subtests of cognitive abilities. The assessment implication of these findings is that executive functions and cognitive processes can be reliably and validly assessed through rating scale methodology. Such an approach would not only be time-efficient but also might reduce the concerns regarding the ecological validity of performance-based measures of executive function.
Two of the studies in the special section investigate the efficacy of computerized working memory training. Wong, He, and Chan completed a working memory training program with a Hong Kong Chinese sample. Prior studies of computerized, adaptive training have reported improvements in working memory performance that are associated with increased activity in the working memory-related prefrontal and parietal cortices, along with increases in myelination and synaptic connectivity. A reduction in inattentive symptoms in ADHD children has also been reported. As to whether the working memory gains generalize to improved academic learning and performance, the results of previous studies have been equivocal.
The computerized training program employed by Wong, He, and Chan consisted of five visuospatial and three auditory tasks. Participants trained for 35 to 40 minutes per day for 20 to 25 days over 5 weeks. On the neuropsychological working memory measures, the experimental group gained significantly on both visuospatial and verbal working memory. Relative to the control group, there was significant improvement on parent (but not teacher) ratings of working memory, but no treatment effect was found with parent inattention or hyperactivity ratings. Moderate gains in working memory remained when the participants were re-tested 10 to 12 weeks after the immediate post-intervention assessment.
In one of the first studies of its kind with a college population, Gropper et al. pursue working memory training with college students who have either ADHD or a learning disability. In a randomized, controlled design, 62 students who were receiving student disability services completed 25 sessions of the adult version of Cogmed. Training related effects were greatest on a measure of visual-spatial working memory, although there was also a significant gain in verbal working memory (assessed with the WAIS-IV Digit Span subtest). Males and those with lower WAIS-IV Digit Span pre-test scores had the greatest gains in Digit Span scores after training. Moreover, the working memory training group reported significantly fewer ADHD symptoms and significantly fewer everyday cognitive failures. The gains in working memory and the reduction in cognitive failures were maintained for at least two months following the training. However, the improvements in working memory did not transfer to improved academic skills and performance. Nonetheless, Gropper et al.’s findings provide further support for the efficacy of computerized working memory training, as well as support for previous findings that Cogmed working memory training can reduce ADHD symptoms.
In an effort to more fully understand the working memory profile of individuals with ADHD, van Ewijk et al. study visual-spatial working memory (VSWM) in ADHD subjects. As predicted, the ADHD patients had impaired VSWM functioning, but their performance did not worsen with higher memory loads, signifying a central executive deficit rather than reduced visual-spatial storage capacity. Using a sample that ranged from 8-29 years, Van Ewijk et al. also discovered that VSWM deficits in ADHD individuals are relatively stable and pervasive into young adulthood. Comparing their ADHD subjects with unaffected siblings, van Ewijk et al. found that the unaffected siblings did not display impaired VSWM performance, suggesting that VSWM deficits in ADHD may not be a genetically mediated weakness. This finding led Ewijk et al. to conclude that VSWM deficits are a co-occurring problem in ADHD, rather than a cause of the behavioral symptoms. Consequently, the results of this study do not support the use VSWM as a neurocognitive endophenotype for ADHD.
In another study of the relations between working memory and ADHD-related behavior, Hudec et al. test the hypothesis that increased demands on working memory contribute to hyperactivity in adults with ADHD. Previous studies with ADHD children and adults have reported a functional relationship between motor activity and working memory. Hudec et al.’s study was the first to use actigraphy to more accurately measure motor activity, and it also was the first isolate the contributions of phonological storage and the central executive. Although the motor activity of healthy controls also increased during higher phonological and central executive demands, the group with ADHD exhibited even greater activity. The interesting implication is that potentially impairing hyperactivity in adults with ADHD (who exhibit minor fidgeting and restlessness during low working memory demands) may be the result of high working memory demands.