Effects of a computerized working memory training program on working memory, attention, and academics in adolescents with severe LD and comorbid ADHD; a randomized controlled trial

Publication: The Journal of Child Psychology and Psychiatry

Institution: University of Toronto, SickKids Hospital, Jewish Vocational Services (JVS) Toronto

Investigator(s): Sarah A. Gray, Peter Chaban, Rhonda Martinussen, Rachel Goldberg, Howell Gotlieb, Reena Kronitz, M. Hockenberry, & Rosemary Tannock

Program: Cogmed RM

Background & Aim: Working memory (WM) is below average in populations with co-occurring learning disability (LD) and ADHD. Youths with LD/ADHD, which comprise up to 4% of clinical samples, are at high risk for poor psychological and occupational outcomes as adults. However, the underlying cognitive deficits in areas such as WM are currently not well targeted by current interventions for either LD or ADHD. Thus, the aim of this study was to determine the feasibility of implementing WM training (Cogmed) in a residential school setting for a group of treatment resistant adolescents with combined LD and ADHD. Secondly, the study investigated whether WM training could improve WM, inattentive symptoms as rated by parents and teachers, and academic outcomes in the sample. To serve as an active control, adolescents in WM training were compared with students who took part in an intervention targeting math skills (Academy of Math). Both programs included similar features including built in reinforcement and individually based algorithms and were completed in 45 minutes training sessions, 4 to 5 days per week, for 5 weeks.

Population & Sample Size: N = 60 children with severe LD & comorbid ADHD, ages 12 – 17 years

• n = 32 children in adaptive Cogmed training group
• n = 20 children in math intervention group
• n = 8 children withdrew

* 82% of participants scored below the 25th percentile on reading, spelling, and math (WJ-ACH)
* 72% of participants scored below the 25th percentile on working memory (WISC-IV, WMI)
* Majority of sample received long-acting stimulant medication throughout the study

Design: Randomized, blinded, active controlled, test-retest

T1 = baseline, T2 = post-test


I.WM training group improved significantly over math intervention group on:

1) Trained verbal WM (Backward Digit Span; WISC-IV)
– 36% more improvement than math training

2) Non-trained visuo-spatial WM (Spatial Span Backwards & Forwards; CANTAB)
– 28% more improvement than math training

II No significant group differences on:

1) Strategy skills and WM capacity (Spatial Working Memory, CANTAB)

2) WM classroom behaviors (Working Memory Rating Scale)

3) Individual attention and concentration (D2 Test of Attention)

4) Teacher and parent rated attention behaviors (SWAN)

5) Word reading, sentence comprehension, spelling, mathematics (WRAT-4)

III. Significant negative correlation between:

1) Cogmed Training Index (measure of performance on trained tasks) and parent rated inattentive/impulsive/overactive and oppositional behaviors at home (IOWA Conners scale). Adolescents with the most improvement on Cogmed tasks were rated as having lower symptoms by parents

IV. Students in both groups attending the intensive remedial school combined with stimulant medication treatment showed gains over the study period in cognitive attention, reading, math, and parent rated behavior.

Summary and Implications: This was first study evaluating the effectiveness of WM training in adolescents with severe LD and comorbid ADHD. Students who had previously been unresponsive to typical interventions (special education, medication) and who were highly symptomatic attended a specialized school for one year in effort to improve academic, social, and behavioral outcomes. Children remained on stimulant medication while at the school.

WM deficits in these adolescents were pervasive at baseline, with a mean score of 85 on the WMI and 57% of students scoring below the 16th percentile on verbal WM measures (Digit Span Forward and Backwards; WISC-IV). For visuo-spatial WM, 34% scored below average (Spatial Span, CANTAB). Following the study, participants in WM training (Cogmed) improved on verbal WM measures similar to those practiced on in training (16% improving from below to at or above the 16 percentile). Importantly, they also improved on a visuo-spatial measure distinct from the training tasks, with 28% increasing their scores to at or above the 16 percentile. Children in math training did not improve significantly on WM tasks.

Unlike previous WM training research using Cogmed, this study did not find transfer to other measures of WM, attention, or academic attainment. The author suggested that longer or more extensive training may have been required to ameliorate the severe difficulties faced by the sample and that long term follow up, which was not included in this research, may have revealed further aspects of transfer. Another limitation to this study was a small sample size where participants were not categorized based on their type LD, a possible predictor variable in future research.

The authors concluded that, for WM training (Cogmed), there is accumulating evidence that aspects of WM can be improved and that this work supported the premise that WM shows “remarkable neuroplasticity” across wide age range. WM training is feasible in a school setting for severely LD adolescents with comorbid ADHD, a treatment resistant group. They suggested that future developments to the program may be needed to promote transfer to other cognitive, behavioral, and academic measures.

Funding: Funding for this project was provided by the Ontario Provincial Centre for Excellence for Child and Youth Mental Health at CHEO, the Canada Research Chairs Program, and the Social Science and Humanities Research Council