Computerized Training of Working Memory in Children With ADHD���A Randomized, Controlled Trial TORKEL KLINGBERG, M.D., PH.D., ELISABETH FERNELL, M.D., PH.D., PERNILLE J. OLESEN, M.SC., MATS JOHNSON, M.D., PER GUSTAFSSON, M.D., PH.D., KERSTIN DAHLSTROM, ��� M.D., CHRISTOPHER G. GILLBERG, M.D., PH.D., HANS FORSSBERG, M.D., PH.D., AND HELENA WESTERBERG, L.P., PH.D. ABSTRACT Objective: Deficits in executive functioning, including working memory (WM) deficits, have been suggested to be important in attention-deficit/hyperactivity disorder (ADHD). During 2002 to 2003, the authors conducted a multicenter, randomized, controlled, double-blind trial to investigate the effect of improving WM by computerized, systematic practice of WM tasks. Method: Included in the trial were 53 children with ADHD (9 girls 15 of 53 inattentive subtype), aged 7 to 12 years, without stimulant medication. The compliance criterion (20 days of training) was met by 44 subjects, 42 of whom were also eval- uated at follow-up 3 months later. Participants were randomly assigned to use either the treatment computer program for training WM or a comparison program. The main outcome measure was the span-board task, a visuospatial WM task that was not part of the training program. Results: For the span-board task, there was a significant treatment effect both post- intervention and at follow-up. In addition, there were significant effects for secondary outcome tasks measuring verbal WM, response inhibition, and complex reasoning. Parent ratings showed significant reduction in symptoms of inattention and hyperactivity/impulsivity, both post-intervention and at follow-up. Conclusions: This study shows that WM can be improved by training in children with ADHD. This training also improved response inhibition and reasoning and resulted in a reduction of the parent-rated inattentive symptoms of ADHD. J. Am. Acad. Child Adolesc. Psychiatry, 2005 44(2):177���186. Key Words: attention-deficit/hyperactivity disorder, intervention, working memory, response inhibition. Attention-deficit/hyperactivity disorder (ADHD) af- fects 3% to 5% of school-age children with serious impairments in both academic performance and social functioning. Many of these problems persist into adult- hood (Biederman et al., 2000 Rasmussen and Gillberg, 2000). Deficits in executive functioning, including working memory (WM), response inhibition, and tem- poral processing, have been suggested to play an impor- tant role in ADHD (Barkley, 1997 Castellanos and Tannock, 2002 Rapport et al., 2000). Executive func- tions is a broad concept that includes, among other func- tions, the ability to inhibit a prepotent response, planning, reasoning, and WM. WM is the ability to retain informa- tion during a delay and then to make a response based on that internal representation. Furthermore, WM is often regarded as a more fundamental function, underlying other executive functions such as reasoning. WM deficits in ADHD have been demonstrated repeatedly (Dowson Accepted September 14, 2004. Drs. Klingberg, Fernell, Forssberg, and Westerberg and Ms. Olesen are with the Unit of Neuropediatrics, Department Women and Children���s Health, Karo- linska Institute, Stockholm Institute, Stockholm Drs. Johnson and Gillberg are with the Department of Child and Adolescent Psychiatry, Goteborg ��� University, Sweden Dr. Gustafsson is with the Division of Child and Adolescent Psychiatry, Faculty of Health Sciences, Linkoping ��� University, Sweden Dr. Dahlstr��� om is with the Department of Neuropediatrics, Huddinge University Hospital, Sweden. This study was supported by the Swedish Research Foundation (Vetenskaps- ra��det), the Wallenberg Global Learning Network, and Cogmed Cognitive Med- ical Systems AB. The authors thank Anna-Karin Adler, Gunilla Berglund, and Sven ��� Ostlund for performing the neuropsychological testing, Jonas Beckeman and David Skoglund for their significant contribution to task design, Maria Andersson for managing the patient calls, and Johan Bring for statistical analyses. Reprint requests to Dr. Klingberg, Unit of Neuropediatrics, Department of Women and Children���s Health, Astrid Lindgren���s Children���s Hospital, Q2:07, Karolinska Institute, 171 76 Stockholm, Sweden e-mail: torkel.klingberg@kbh.ki.se. 0890-8567/05/4402���0177��2005 by the American Academy of Child and Adolescent Psychiatry. J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 44:2, FEBRUARY 2005 177

et al., 2004 Karatekin and Asarnow, 1998 Kempton et al., 1999 Kuntsi et al., 2001 Mariani and Barkley, 1997 Westerberg et al., 2004, but see also Karatekin, 2004). This study investigated whether systematic training of WM tasks during a 5-week period would improve WM, improve other executive functions, and reduce the ADHD symptoms. Several studies have evaluated the effect of practice with various types of cognitive tasks in subjects with stroke (Sohlberg et al., 2000), in elderly subjects (Ball et al., 2002), and after traumatic brain in- jury (Salazar et al., 2000). The method evaluated in this study differs from that of previous ones in that it focuses entirely on training WM tasks. Moreover, the training is computerized, which makes it possible to automatically and continuously adapt the difficulty level to the perfor- mance of the child to optimize the training effect. The effect of WM training on brain activity was re- cently evaluated with functional magnetic resonance im- aging (Olesen et al., 2004). In that study, young, healthy adult subjects were scanned while performing a WM task and a control task before and after WM training. Training improved the WM performance of the subjects and resulted in increased brain activity in the dorsolat- eral prefrontal and parietal association cortices, indicat- ing plasticity of the neural systems underlying WM. These cortical areas partly overlap with the prefrontal regions implicated in ADHD pathology (Castellanos et al., 1996, 2002 Filipek et al., 1997), and this provides a neuroanatomical rationale for undertaking WM train- ing in children with ADHD. A previous preliminary study indicated that training of WM tasks can enhance executive functioning including WM, response inhibi- tion, and reasoning in children with ADHD (Klingberg et al., 2002b). A major shortcoming of that study was the low number of subjects (n = 7 in both the treatment and the comparison groups). Moreover, ratings of ADHD symptoms were not performed, only one clin- ical center was involved, and there was no follow-up measurement of both groups to estimate the extent to which training effects lasted. The current study was therefore conducted at four clinical sites evaluating the effects of practice of WM tasks in a randomized, controlled, double-blind design. Executive functions were measured and ADHD symptoms were rated be- fore, immediately after, and 3 months after intervention. In this study, we compared two similar versions of the same training program. In the treatment program, the children practiced WM tasks in which the difficulty level was adjusted to closely match the WM capacity of the child. This procedure was hypothesized to opti- mize the training effect. In the comparison condition, the same tasks were used, but the WM load (i.e., num- ber of items to be remembered) was low, thus resulting in easy tasks that were expected to result in only small training effects. By having two similar versions, we in- tended to control as much as possible for nonspecific effects of the training procedure, such as expectancy, passage of time, and maturation and specifically estimate the effect of improvement of WM. To evaluate the effect of training, we used tasks that were not part of the training program. METHOD Subjects Referral sources included pediatricians, child psychiatrists, and special teachers in schools. We included only nonmedicated children because they were thought to have more room for clinical improve- ment of ADHD symptoms than children on medication and there- fore give a better chance of detecting significant treatment effects. Furthermore, in Sweden, only a minority of children with ADHD receives medication. Inclusion criteria were (1) diagnosis of ADHD of either com- bined or predominantly inattentive subtype, (2) age between 7 and 12 years at inclusion, and (3) access to a personal computer with an Internet connection at home or in school. Exclusion criteria were (1) being treated with stimulants, atomoxetine, neuroleptic, or any other psychoactive drugs (2) fulfilling criteria for diagnosis of clin- ically significant oppositional defiant disorder, autistic syndrome, Asperger���s syndrome or depression (3) history of seizures during the past 2 years (4) IQ 80 (based on an IQ test or the physician���s clinical impression and school history) (5) motor or perceptual handicap that would prevent using the computer program (6) educational level and socioeconomic situation that made it unlikely that the family would be able to follow the treatment procedure and study requirements (the educational level of the parents was not specified in terms of academic degree) and (7) medical illness requiring immediate treatment. Of 56 patients attending the screening visit, 53 were included in the study and randomized to the treatment or comparison programs (Table 1, Fig. 1). Diagnostic assessment, including subtyping, had in most cases been made before the screening visit but was confirmed by the physician and based on global clinical impression taking DSM-IV rating scales (American Psychiatric Association, 1994) from parents and teachers into account. None of the children fulfilled criteria for conduct disorder or bipolar disorder. Two of the children had previously taken stimulant medication but stopped more than 1 year before the study. One child discontinued stimulant medication 1 week before the first measurements to be able to participate in the study. All other children had never been on medication for ADHD. The study was approved by the regional ethics committee at Karolinska Hospital and by the local ethics committees at the four KLINGBERG ET AL. 178 J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 44:2, FEBRUARY 2005