Working-memory training in younger and older adults: training gains, transfer, and maintenance

Publication: Frontiers in Human Neuroscience, March 2012

Institution: Karolinska Institute

Investigator(s): Yvonne Brehmer, Helena Westerberg, Lars Bäckman

Program: Cogmed RM

Background & Aim: Working memory (WM) is the ability to maintain and manipulate information over short periods of time, often in the face of distraction. WM is a key determinant of higher order cognitive abilities such as reasoning, fluid intelligence, problem solving and language comprehension. WM declines in late adulthood and is considered a main contributing factor to cognitive impairment in aging adults. As improving WM in older adults may be important for cognition in everyday life, the aim of this study was to investigate training gains, transfer effects, and 3-month maintenance effects of intensive, computerized WM training (Cogmed) in younger and older adults.

Population & Sample Size: N = 100 typical young adults (ages 20 -30 years) and older adults (ages 60-70 years)
• n = 29 younger adults in adaptive Cogmed training group
• n = 26 younger adults in non-adaptive Cogmed training group
• n = 26 older adults in adaptive Cogmed training group
• n = 19 older adults in non-adaptive Cogmed training group

Design: Randomized, placebo controlled, double-blinded, test-retest
T1 = baseline, T2 = post-test, T3 = 3 month follow-up



I. Adults in adaptive training improved significantly over adults in control group on:

1) Trained tasks of visuo-spatial and verbal working memory (Span Board Forward and Digit Span Backward)

2) Non-trained tasks of visuo-spatial and verbal working memory (Span Board Backward and Digit Span Forward)

3) Sustained attention (Paced Auditory Serial Addition Task; PASAT)

4) Self report of cognitive functioning (Cognitive Failures Questionnaire, CFQ)

II. No significant improvements on inhibition (Stroop), nonverbal reasoning task (Raven’s Standard Matrices) or episodic memory (Rey Auditory Verbal Learning Test, RAVLT)

III. Performance level reached after 5 weeks of training was maintained at 3 month follow-up


Summary and Implications: This study is a randomized, placebo controlled, double-blinded investigation of working memory training (Cogmed) in 100 typically functioning younger and older adults. It is the first known investigation in the aging domain to use an active control group, where individuals worked on the same material as the experimental group, with the only difference being task difficulty level (adaptive vs. non-adaptive training). The use of such a control group is an essential design element as it attenuates the impact of various factors that may influence the training effect including motivation, test familiarity, performance anxiety, expectations and stimulus-response mappings.

Overall, younger adults had higher baseline cognitive functioning scores and had larger differences in performance gains (trained and non-trained WM tasks) between the adaptive and non-adaptive groups than did than older adults. Younger adults in adaptive training had significantly larger gains on trained and non-trained measures of WM, sustained attention (PASAT) as well as, decreased self report of cognitive failures (CFQ) compared to younger adults in the non-adaptive (placebo) training group. Older adults in the adaptive training group had significantly larger gains on trained and non-trained measures of WM, sustained attention (PASAT) as well as, decreased self report of cognitive failures (CFQ) compared to older adults in the non-adaptive (placebo) training group. Improvements for younger and older adults in adaptive training were comparable for the CFQ and PASAT. These findings imply that the training paradigm is sensitive enough to impact even older adults and that training may influence one’s experience of everyday cognitive functioning. Importantly, gains in the younger and older adaptive groups were maintained at 3 month follow-up.

The adaptive and non-adaptive groups had comparable performance improvements on assessments of interference control (Stroop) and nonverbal reasoning (Raven’s Standard Matrices). The researchers interpreted these findings as a demonstration of general test-retest effects. It was also suggested that in the future, studies should include both active placebo and passive (no contact) controls so to disentangle the impact of the different levels of training and other performance influencing factors.

Thus, Brehmer et al. (2012) is a landmark working memory training study as it provides a conservative assessment of training and transfer gains in a healthy sample of younger and older adults through use of a well-designed placebo control. Younger and older adults can improve their working memory, as measured by tasks that differ from the task done during training (non-trained tasks) and these improvements are reflected in improved attention and self-reported cognitive functioning.

Funding: The Swedish Research Council for Working Life and Social Research supported this research through a FLARE post-doctoral grant announced by ERA-AGE. Other supporting grants include Hjärnfonden, Swedish Brain Power, and the Jochnick Foundation. Training program, investigator training, and technical support provided by Cogmed. No funding provided by Cogmed.