Trained vs. non-trained tasks cont.

3b. Trained vs. Non-trained tasks
During Cogmed Working Memory Training, users practice on both visuo-spatial and verbal working memory (WM) exercises for 5 days per week for 5 weeks. Given the adaptive nature of these exercises – becoming increasingly/decreasingly more difficult depending on user performance – the user becomes quite skilled at and familiar with these exercises. The user is expected to improve on the specific exercises or “trained tasks” that they have been training on for 5 weeks. However, how can one be sure that improvement from training is not limited to only these trained tasks and instead, impacts all different types of working memory taxing activities?

One way of ensuring that a user’s WM capacity has been improved in a broader – less context dependent manner – is to assess the user with tasks that have not appeared during training but that still tap the underlying WM ability. Non-trained tasks differ in presentation, administration and response mode. Let’s go back to our 4×4 grid example from “adaptive vs. non-adaptive training “discussion. If the user trains on the computerized 4×4 grid for the duration of training – holding in mind increasingly/decreasingly difficult sequences of dots lighting up – they will probably become pretty familiar with how to use their WM in this type of task.

In order to gauge whether training on this task has impacted the underlying WM ability, one should assess the user with a “non-trained task”. An example of a non-trained task in this situation might be a display of wooden blocks where an administrator points to a certain sequence of blocks with his/her hand and the user must repeat this span sequence in backwards order by pointing to the blocks. In this way, the task is different in presentation (blocks rather than dots), administration (real objects rather than computerized graphics), and response mode (pointing with the finger rather than clicking with the mouse). The user must also repeat the sequence in backwards order, forcing them to use their WM in a manner distinct from training – where they were required to hold in mind and repeat a sequence of dots in forward order. In improving on non-trained tasks, the user has provided evidence that the underlying WM ability has increased.