Lundqvist 2010.

Study characteristics
Methods	Design: cross‐over RCT (wait group) Duration of trial: not reported Unit of randomisation: adult outpatients with WM impairments after acquired brain injury Recruitment and allocation: 38 rehabilitation outpatients from 76 people with acquired brain injury met inclusion criteria; 21 accepted and were randomised by drawing of lots to 2 groups; IG: 10; CG: 11. 13 (62%) of these had stroke: IG: 5; CG: 8.
Participants	Setting: outpatient hospital rehabilitation department Country: Sweden Sample size: 13 adults with stroke, 54% men; IG: 5; CG: 8 Inclusion criteria: aged 20–65 years, time since injury/illness ≥ 1 year, self‐reported WM impairments and a significantly impaired Working Memory Index compared to the index for verbal comprehension or index for perceptual organization (or both) or a Working Memory Index < 80 using the WAIS III (Wechsler 1997), and a reported motivation for training. Exclusion criteria: IQ ≤ 70 based on WAIS III, depression according to DSM‐IV, perceptual or motor impairments that made computerised training impossible, extensive cognitive impairments and completion of other rehabilitation programmes. Age: mean: IG: 48.4 (SD 8.2) years; CG: 43.25 (SD 9.97) years Time since onset of stroke: mean: IG: 51.4 (SD 46.39) months; CG: 51.13 (SD 36.24) months Types of stroke and sites of lesion: not reported
Interventions	Intervention group Brief name: WM training, QM, now called Cogmed Recipients: outpatient adults with WM impairments after acquired brain injury Why: to improve WM function using individualised and intense computerised training that occurred in a clinic, at least initially, so that participants received the benefit of coaching, meeting with other participants, and training in a calm, quiet environment. What (materials): QM (formerly called ReMemo and now called CogMed) for adults, a WM training computerised system with visuo‐spatial and verbal WM tasks developed at the Karolinska Institute and Cogmed Cognitive Medical System AB, Sweden (www.cogmed.com/healthcare). See also Klingberg 2002; Westerberg 2004; Westerberg 2007 for details of the program; personal computer. What (procedures): participants performed their WM training program on a personal computer in pairs in the presence of 1–3 certified coaches who provided special feedback once per week beside the continuous statistics, which the participants could follow themselves on the computer. The training program specified different visuo‐spatial and verbal WM tasks. Each task was introduced by a speaker voice and the person responded by localising and remembering the stimuli. Participants used a computer mouse to respond: visuo‐spatial WM tasks required the participant to remember the position of stimuli in a 4 × 4 grid and then they are asked to reproduce stimuli in the same order, in the reverse order, or in the grid after the grid had been rotated; verbal WM tasks requires the participant to remember sequences of letters and digits forwards or backwards, or both. Who provided: 3 certified coaches; authors confirmed that these coaches were occupational therapists. How: face‐to‐face in pairs in the presence of 1–3 certified coaches Where: in a separate quiet room at the Department of Rehabilitation When and how much: 45–60 minutes of intense training per day, 5 days per week, for 5 weeks (up to 25 hours), weekly coach feedback Tailoring: the difficulty level of each training task automatically adjusted according to each participant's progress, increasing the WM load according to each participant's performance levels. The coach provided feedback once per week in addition to ongoing statistics provided to the participant on the computer. Modifications: not reported How well (planned): none reported How well (actual): not reported Comparator group None (received no training during the same period) 4 weeks after the IG participants had had their training, WM functions were assessed in both groups. Then, CG received the same training for 5 weeks.
Outcomes	Primary: none Secondary COPM (Satisfaction and Performance) PASAT 2.4 Listening Span Task Picture Span Block‐Span‐board forwards and backwards from WAIS R‐NI CWIT 4 Working Memory Improvement Index Other EQ‐5D health‐related quality‐of‐life Methods of data collection: a neuropsychologist assessed all participants for each time point. Data collection time points: before training (baseline) and 4 and 20 weeks after training
Notes	Funding: yes Conflict of interest: none Published trial protocol: no Trial registration: no Ethics approval: yes Author Contact: further information and stroke‐specific data from authors requested and supplied.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "… randomized into two groups at the beginning of the study, through 'drawing of lots' ".
Allocation concealment (selection bias)	Unclear risk	Not reported.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Certified coaches provided the intervention, including tailored feedback to participants.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "A neuropsychologist assessed all subjects before training (baseline) and at 4 and 20 weeks after training".  Comment: COPM semi‐structured interview was administered by an occupational therapist; unclear if either of these people were one of the certified coaches providing the intervention.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Personal correspondence reported no participants lost to follow‐up and provided data for same.
Selective reporting (reporting bias)	Low risk	No protocol/trial registry reported or located but reports on significant and non‐significant outcome measures.
Other bias	Low risk	No other identifiable bias.