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

Institution: Karolinska Institute

Title: Computerized training of working memory in children with ADHD – A randomized, controlled, trial

Researcher(s): Klingberg T, Fernell E, Olesen P, Johnson M, Gustafsson P, Dahlstrom K, Gillberg CG, Forssberg H, Westerberg H.

Program: Cogmed RM

Published: Journal of American Academy of Child and Adolescent Psychiatry, February 2005

Funding: This study was supported by the Swedish Research Foundation (Vetenskapsrådet), the Wallenberg Global Learning Network, and Cogmed Cognitive Medical Systems AB. Training program, investigator training, and technical support provided by Cogmed. No funding provided by Cogmed.

There is a pressing need to develop treatments for ADHD in addition to medication and behavior therapy that have strong research support and a number of different reasons why this is important.

Regarding medication, not all children benefit from taking it, some experience intolerable side effects, and many continue to struggle even though medication may be somewhat helpful. Behavior therapy can be difficult for parents to consistently implement, and does not generally reduce behavior difficulties to normative levels.

Furthermore, even though both treatments can be extremely helpful in managing ADHD symptoms and reducing oppositional behavior, they do not induce lasting changes in the child that persist after treatments is discontinued. Finally, despite numerous studies documenting the short- and intermediate term benefits of these treatments, their impact on children’s long-term success remains to be clearly documented.

Because of these limitations, there have been numerous efforts to develop alternative treatments for ADHD that may enhance the benefits offered by medication and behavior therapy. In a future issue of Attention Research Update, I hope to provide an overview of these efforts. Today, however, I am excited to report results from a recently published study on the use of computerized training of working memory in children with ADHD (Klingberg, et al., 2005. Computerized training of working memory in children with ADHD – A randomized controlled trial,. Journal of the American Academy of Child and Adolescent Psychiatry, 44, 177-186.) Unlike many studies of alternative or complementary ADHD treatments, this study includes a number of experimental controls that allow for greater confidence in the findings.

Executive functioning deficits are believed to play an important role in ADHD. Executive functions refer to those mental operations that help to organize and direct complex behavior, and include such operations as goal setting, planning, reasoning, cognitive flexibility, and the ability to delay responding. Working memory (WM) – the ability to hold information in one’s mind for subsequent use – is a particularly important executive function because it may underlie other executive functions such as reasoning. WM deficits in individuals with ADHD have been demonstrated repeatedly and have been suggested to contribute to the academic struggles that many children with ADHD experience. Developing an intervention to enhance WM in children with ADHD could thus be extremely helpful.

The study reviewed below investigated whether systematic training of WM tasks during a 5-week period could improve WM, other executive functions, and reduce ADHD symptoms in 50 7-12 year old children – approximately 90% male – diagnosed with ADHD. None of the children were on medication when the study began, or at any time during the study. For reasons that are unclear, children with oppositional defiant disorder or depression in addition to ADHD were excluded, and only those with ADHD diagnoses exclusively were recruited into the sample.

Participants were randomly assigned to 1 of 2 conditions – a high intensity (HI) WM training condition and a low intensity (LI) WM training condition. The HI treatment consisted of performing WM tasks via a computer program developed for the study. These included visuospatioal WM tasks – remembering the position of objects on the screen – as well as verbal tasks – remembering sequences of letters, sounds, and digits. In all cases, children responded to the WM task by clicking on various choices with the computer mouse. Each training session provided exposure to 90 WM tasks and required about 45 minutes to complete. The difficulty level of the WM tasks was automatically adjusted to match the WM ability of the child by modifying the number of elements the child was required to hold in memory on a trial-by-trial basis. By this method, children were continually challenged to improve their WM ability by presenting them with more difficult tasks after they succeeded with easier ones. In addition to improving their working memory, children had to attend consistently and remain relatively inactive – remaining seated in front of the computer – to perform well. Thus, in addition to training memory, benefits in attention and reductions in activity level might be expected to occur as well.

The LI condition was identical to that described above except that the difficulty of the 90 WM trials remained at a low level throughout, i.e., the number of items children were required to recall never increased beyond 2 to 3. Thus, these children had the same experience as children in the high intensity treatment group, i.e., they spent the same amount of time engaging in computerized WM tasks, but they were not challenged to improve. As a result, children in this condition were not expected to show the same improvement in WM, other executive functions, or ADHD symptoms as children receiving the high intensity treatment.

During the 5-weeks of training, children needed to complete a minimum of 20 training sessions that were completed at home or wherever else children had access to a computer. (Note: Parents had been given a CD with the training software and simply had to install it on a computer their child had access to.) Training session results were uploaded by the researchers via the Internet so that they could maintain a database of each child’s performance. Parents were contacted by phone on a weekly basis to inquire about any technical difficulties and provide feedback about how many training sessions the child had completed that week. This was intended to insure that all children completed the required number of sessions.

A number of different measures were collected so that the impact of the computerized training of WM could be evaluated. These included several measures of WM, a measure of response inhibition, i.e., the ability to delay responding, and a measure of non-verbal reasoning ability. In addition, ratings of ADHD symptoms were obtained from parents and teachers. These measures were collected before treatment began, immediately following treatment, and 3 months after treatment ended. This enabled the researchers to determine whether children receiving high intensity treatment showed greater improvement immediately following the intervention, and whether these benefits remained evident 3 months later, even though no further training had occurred.

Before summarizing the results, it is important to highlight several important design features that make this study exceptionally strong. First, children were assigned at random to the HI or LI conditions. There is thus no reason to expect that pre-existing differences between children in the 2 groups could explain any treatment differences that emerged.

Second, parents, children, nor and teachers were not aware of which condition children had been assigned to. Parents and children were simply told that the child would be taking part in 1 of 2 different treatments, and that 1 of these treatments was expected to be more helpful than the other. It would not have been evident to parents or children, however, whether the child was in the condition that was expected to be more or less helpful. In addition, psychologists who performed the neuropsychological assessments were “blind” to the child’s condition. Thus, the outcome measures collected can be considered “unbiased” because no one responsible for providing or obtaining the data really knew whether the child would be expected to show improvement.

Finally, the experience of children in the HI and LI conditions was highly similar. They spent the same amount of time in training and completed training exercises that were similar in form. The only difference was in the difficulty level of the training tasks that were presented to them.

Collectively, these design elements help rule out the possibility that any treatment differences could be attributed to pre-existing differences between the groups, biased reports from raters, or non-specific differences in children’s experience during the study. Instead, any such differences that emerged are likely to reflect the actual WM training that children in the high intensity condition received.

To examine intervention effects, researchers compared results for the HI and LI groups immediately following the intervention and 3 months after treatment ended. In all analyses, baseline assessments were included as controls, which further assures that treatment differences could not be explained by pre-existing group differences. Results from this study were as follows: (Note – Participants included children with the combined and inattentive subtypes of ADHD. The results reported below apply to each subtype and no subtype differences in treatment response were observed.)

1) Immediately following the treatment, children in the HI group showed significantly better WM than children in the LI group. The gains made by children in the HI group were comparable to gains associated with medication treatment that had been demonstrated in prior studies. These benefits remained evident at the 3-month follow up and showed no decline in magnitude.

2) The HI group performed better on all other executive functioning assessments – non-verbal reasoning and response inhibition – at post treatment than the LI group, although the magnitude of the difference was not as great as for WM. The better performance was still evident at the 3-month follow up.

3) Parent ratings of inattentive and hyperactive-impulsive symptoms were lower at post-intervention and at the 3-month follow up for children in the HI group. The reductions in parent ratings of attention difficulties were substantial while reductions in ratings of hyperactive-impulsive symptoms were more moderate. Parent ratings of oppositional behavior were also significantly lower in the HI group at both time points.

4) Group differences in ADHD symptoms ratings completed by children’s teachers were not significant.

Summary and Implications

Results from this study provide strong evidence that approximately 20 hours of computerized WM training over a 5-week period produced gains in this important executive function and in other executive functions that were not the specific focus of training. These benefits were evident immediately following training and remained evident 3 months later, even though no further training had occurred. In addition, there were significant reductions in children’s ADHD symptoms according to parents, although no comparable benefits were evident in teachers’ ratings. The gains in WM and the reductions in attention difficulties reported by parents were large, and comparable in magnitude to effects obtained by medication.

The strong design of this study, which included random assignment and “blind” raters, provides a basis for attributing these gains to the WM training itself rather than to other confounding factors. As such, this would appear to be a very promising intervention for children with ADHD. It is especially noteworthy that benefits persisted for 3 months after treatment ended. This suggests that, unlike medication and behavior therapy, gains associated with computerized WM training may be more enduring. How long these gains might last is unclear, however, as is the need for continued training in order to sustain these gains over a longer period.

Despite the exciting nature of these findings, the authors themselves highlight several limitations to their study that will be important to address in future research.

First, despite the exciting nature of the findings, replication with a larger sample that includes more females will be important.

Second, because children with co-occurring oppositional defiant disorder and/or depression were excluded, the extent to which the results would generalize to children with ADHD and these other conditions is not known. Because many children with ADHD also have these other conditions, it will be especially important to learn whether this intervention is helpful to these children as well.

Third, the absence of teacher reported gains is concerning. Although the authors suggest that ratings from parents are more reliable because they were consistent with the executive functioning results, the basis for this suggestion is unclear. Because a goal of ADHD treatment is improving children’s functioning at school, demonstrating that this treatment produces this effect remains critically important. This would include assessment of “real-world” academic functioning, which was not part of the current study.

In summary, this study provides preliminary evidence that computerized training of WM may be an effective intervention for children with ADHD. It is important to recognize, however, that additional research will be required in order to conclusively document the benefits of this intervention as well as to understand its limitations. Lets hope that such research is currently underway and will become available shortly.

J American Academy of Child and Adolescent Psychiatry 44 (2):177-186

Link to abstract