Institution: Karolinska Institute

Title: Training of working memory in children with ADHD

Researcher(s): Klingberg, T., Forsberg, H., Westerberg, H., & Hirvikoski, T.

Program: Cogmed RM

Published: Journal of Clinical & Experimental Neuropsychology, November 2002

Funding: This work was funded by Jeansson Stiftelse, The Swedish Medical Research Foundation, HjaÈrnfonden, Svenska DyslexifoÈreningen, Sven Jerrings Stiftelse, Frimurarna Barnahuset, and SaÈllskapet BarnavaÊrd. Training program, investigator training, and technical support provided by Cogmed. No funding provided by Cogmed.

Overview
Is Working Memory (WM) a skill that can be improved with training? This is the fundamental question addressed in the initial study of WM training published by Torkel Klingberg and his colleagues in 2002.

This was a two-part study that examined WM training in both children and adults. In Study 1, participants were 14 children between the ages of 7 and 15 who were diagnosed with ADHD. These children were randomly assigned to one of two conditions – a high intensity (HI) WM training condition and a low intensity (LI) WM training condition.

Children in the HI treatment received WM training via a computer program developed for the study. These included visuospatial WM tasks – remembering the position of objects on the screen – as well as verbal tasks – remembering sequences of letters and digits. Difficulty level of the tasks was adjusted to match the child’s WM ability by frequently modifying the number of elements that had to be recalled. For example, if the child correctly recalled a string of three digits on two successive trials the number of digits to be recalled on the next trial increased to four. By adjusting the difficulty level to match the child’s performance, children were continually challenged to improve their WM ability; it was hypothesized that this would lead to gains in WM capacity over time.

The LI condition was identical to that described above except the difficulty level remained constant throughout the training, i.e., the number of items children had to recall did not increase to match the child’s performance. In the LI condition, children also completed fewer trials. As a result, no gains in WM capacity were expected.

The impact of training was evaluated using several measures of WM and other cognitive functions that were completed before and after training. To avoid biasing the results, examiners who were blind to the type of training that children received administered these measures. The measures included: 1) a test of visuospatial WM similar to one that had been trained; 2) a second measure of visuospatial WM that was different from what children encountered in the training; 3) Raven’s Progressive Matrices, a measure of nonverbal intelligence; 4) the Stroop Test, a widely used measure of response inhibition; and, 5) a measure of head movements acquired while children were completed a computerized attention test. This was collected using an infrared camera and is believed to provide an objective measure of ‘fidgetiness’.

Results
Results confirmed the researchers predictions in that children receiving HI training showed significant gains in nearly all the measures administered. These gains were substantial in magnitude in addition to being statistically significant. Not only did the overall group average improve, but every single child who received HI training made gains in both the trained and untrained measure of WM as well as on Ravens Progressive Matrices. These children also showed an average reduction in head movements during the CPT of nearly 75% and all seven children showed a reduction.

In contrast, no significant overall gains made by children assigned to the LI training condition on any measure. On the untrained measure of WM and on the Ravens, several children made higher scores after training but the majority either stayed the same or declined. Head movements, a measure of ‘figitiness’ also increased for 6 of the 7 children.

The final analysis conducted was restricted to the HI group and looked at the relationship between WM improvement during training and improvement on the other outcome measures. Results indicated that WM improvement was significantly correlated with gains in WM on the nontrained task, as well as on the Ravens, a portion of the Stroop test, and with reductions in head movements. The magnitude of these correlations ranged from .42 to .85, and three were above .70. These are large relationships and suggest that improvement in WM contributed to improvement in other aspects of cognitive functioning.

Study 2
The second portion of this study examined WM training in four healthy young adults. All adults received the HI training over a five-week period; unlike Study 1, there was no control group. The goal was to learn whether young adults without any known attention or WM deficits would realize improvements similar to found with ADHD children in Study 1. To assess training benefits, pre- and post-measures of WM and other cognitive functions similar to those described in Study 1 were employed.

Results indicated that all four adults demonstrated significant improvement on the training tasks over the five-week period. In addition, pre-post testing indicated that all four made gains on non-trained measures of WM as well as on the Ravens Progressive Matrices Test. As noted earlier, the Ravens is regarded as a reliable measure of non-verbal intelligence and one for which improvement on retesting is not typically found.

Summary and Implications
This is an important study in that it is the first demonstration that an individual’s WM capacity – something that has been believed to be fixed – can be increased with intensive training. Importantly, improvements in WM capacity were not limited to tasks on which training occurred, but were also observed for non-trained tasks as well as for measures of other cognitive functions. Furthermore, similar gains were observed among children diagnosed with ADHD as well as among healthy young adults. For both groups, the improvement found on the Ravens, a reliable measure of non-verbal intelligence, was striking. These gains may reflect the fact that complex reasoning skills depend on WM and that training induced improvement in WM facilitated improvements in participants’ reasoning abilities.

While these are exciting results, the study has limitations that are important to recognize. First, the sample size in both studies was small, and replication with larger samples would thus be important. Second, the researchers did not examine whether children with ADHD were observed to show reduced ADHD symptoms by their parents and teachers. Although the gains made in several measures of cognitive functioning are noteworthy, it should not be assumed that similar gains in behavioral functioning and classroom performance would also occur. As the researchers note, this would be important to document in subsequent research.

Finally, the improvements observed were all based on testing that occurred immediately after training. The durability of training effects is thus unknown and determining this will require a new investigation that includes a longer-term follow-up.

J Clin Exp Neuropsych 24:781-791