Training in maintenance and cognitive control: The complex working memory span task

Another task that has been used in the working memory training literature is the complex working memory span task introduced by Daneman and Carpenter (1980). In the original version the subjects had to read sentences while remembering the final words of all the sentences for later serial recall. Since then many versions of the complex working memory task have been developed but core features are that these tasks jointly tap storage and processing in working memory. Moreover this task has served as the foundation of many studies linking working memory capacity to complex cognition (e.g. Unsworth, Heitz, & Engle, 2005), making it an interesting task as a target for training.

In a study by Borella and colleagues (2010) they offered three sessions of training to a group of older adults in a categorization working memory span task. Here, the participants were presented with lists of words and had to tap their hand if the presented word belonged to the category of animals (processing) while trying to remember the last word presented in each list (storage). The training increased in difficulty and was adaptive to the performance of the subject. No feedback was provided. A contact control group was used, filling in questionnaires related to autobiographical memory and well-being. The test-battery included near- and far-transfer tests tapping mental speed, short-term memory, working memory, inhibition, and fluid intelligence. Follow-up assessment was carried out eight months after completion of training. The results were impressive, showing transfer effects to all tasks administered to assess short-term memory, working memory, inhibition, mental speed, and fluid intelligence. These effects were still present for fluid intelligence and mental speed eight months later. One factor raised by the authors as critical for obtaining the positive effects was, besides focusing on critical WM processes, that the training was variable and flexible (Borella et al., 2010).

A slightly different version of the complex span task was used in a recent study by Richmond and colleagues (2011). Here they provided 20 sessions of training in one verbal and one visuo-spatial complex span task to a group of older adults. The training was adaptive to the participant’s performance level and feedback was provided. Also an active control group, engaged in solving trivia quizzes, was used. The results indicated performance gains in both trained complex span tasks and near transfer to a similar complex working memory span task (reading span). Far transfer was only seen in an episodic memory task where fewer repetitions of recalled items were made after training (Richmond et al., 2011). Moreover the results of the older trained group were also compared to younger adults who had been part of an earlier study receiving a very similar training protocol (see Chein & Morrison, 2010). In line with the above results, older adults performed at a lower level but the magnitude of gain as function of training was similar for young and old adults.

Finally, a study by Buschkuehl and colleagues (2008) used a training protocol with three tasks to train working memory—one simple span task and two complex span tasks. They also included two simple reaction time tasks mainly to make the training more variable and attractive. The participants were 80 years old and took part in 23 sessions. As for the above studies they used an active control group receiving light physical training. The results only showed a near-transfer effect to an untrained simple span task similar to the task used in training but no far-transfer effects to episodic memory. A one-year follow-up assessment did not reveal any maintenance effects. Hence the results indicate slim transfer effects with limited durability in older samples.

First of all, these studies enjoy several positive design characteristics by including active control groups, adaptive training protocols, and the use of more extensive transfer batteries to evaluate training effects. However, given that all three studies offered training in rather similar complex working memory span tasks in healthy older adults, the finding of impressive far-transfer effects after three practice sessions reported in the Borella et al. (2010) study is intriguing in relation to the more humble near-transfer effects seen after 20 sessions or more of training in the studies by Richmond et al. (2011) and Buschkuehl et al. (2008) respectively. As mentioned above, Borella et al. (2010) offered two tentative explanations for their positive far-transfer effects, namely that the training was challenging and variable and that the short time frame of the training phase had a facilitating effect on performance. However, the training protocols used in both the Richmond et al. (2011) and Buschkuehl et al. (2008) studies were also challenging and variable, in that several training tasks were used and the training was adaptive to the performance of the subjects. Hence training variability does not seem like a strong candidate for driving the effects. Also it is hard to reconcile that a few practice sessions should have a more facilitating effect on performance over several sessions (see Jaeggi et al., 2008). Clearly more research is needed on the factors important to foster far-transfer effects as well as the replication and extension of successful training protocols.

< Prev   CONTENTS   Source   Next >