A randomized control trial of the effects of home-based online attention training and working memory training on cognition and everyday function in a community stroke sample.

Participants were not chosen for having attention problems, and many had well-preserved cognition. On the main measure, around 70% had no spatial bias to begin with, so there was little room for the training to show an improvement. The weekly phone interviews that produced the one positive result were run by a researcher who knew which training each person had been given. Only the post-study interviews were carried out blind to group. The authors themselves cannot say whether the better self-ratings after attention training came from the training, from participants simply liking it more (they rated it more helpful and enjoyable than working memory training), or from progress being easier to see in that programme. Detailed information on lesion location was available for some participants only and is not reported. No group differences appeared at the 3-month follow-up, so nothing here speaks to lasting change.

The authors reported no potential conflict of interest. The text does not say who funded the work.

Do not read the improvement in participants' own ratings of memory, concentration, planning and mood as proof that the attention training works. The study's pre-chosen main measure and every standardised cognitive test showed no benefit over a waitlist, and the weekly ratings that did differ were collected by a researcher who knew which training each person had been given.