Prokopenko 2013.

Study characteristics
Methods	Design: parallel RCT Duration of trial: not reported Unit of randomisation: inpatients with cognitive impairments after a hemisphere stroke Recruitment and allocation: not reported in detail; inpatient stroke survivors were approached within 2 weeks after hemisphere stroke and 43 of those with cognitive impairments and who did not have epilepsy or significant speech pathology were randomised by 'method of letters' to 2 groups, IG (n = 24) and CG (n = 19).
Participants	Setting: inpatient facility Country: Russia Sample size: 43 adults, 53% men, IG: 24; CG: 19 Exclusion criteria: unable to give informed consent, not fluent in Russian, significant speech pathology, e.g. aphasia, severe cognitive deficits (MMSE < 20), medically unstable, epilepsy Age: median (quartiles): IG: 61 (25th 57, 75th 69) years; CG: 66 (25th 61, 75th 69) years Time since stroke onset: not reported except people were approached within 2 weeks of stroke Types of stroke and site of lesion: not reported
Interventions	Intervention group Brief name: computer training Recipients: inpatient adults with cognitive impairments from a hemisphere stroke Why: to restore cognitive function using 2‐part computer training focusing on attention and visual and spatial gnosis with built‐in feedback and a help option. Part 1 focused on 4 aspects of attention (sustained, selective, divided, and alternating); part 2 focused on figure‐background activities with gradually decreasing intensity of "background noise". Tasks were not aimed at evaluation of cognitive functions, but rather at training of these functions; though task performance speed in the attention task was measured in time and fed back to the participant, the feedback served only as a reference point for improvement. What (materials): computer programs: Schulte's Tables for attention training Figure‐background test for visual and spatial gnosis training Other computerised tasks: "remembering a sequence of symbols", "arranging the clock hands", and "the serial count" What (procedures): in addition to standard treatment at the inpatient rehabilitation department: attention training: Schulte's Table (a 5 × 5 square grid containing numbers from 1 to 25 in a random order) was presented in full screen mode on the computer monitor. The participant was timed while locating numbers from 1 to 25 in ascending order by clicking on the corresponding number with the mouse. Cues were provided after a fixed time period, e.g. the number pulsated or changed colour. The time taken to complete the task was displayed at the end of the session. Training of visual and spatial gnosis Figure ground: 1 picture image with decreasing intensity of "background noise" was presented on the computer screen. At the top of the screen, several different images, such as objects or letters without background were shown, including a picture of the image shown in the task. The patient needed to identify the picture that corresponded to the image presented in the task with the noise as soon as possible by clicking the mouse cursor on the corresponding image on the top part of the screen. Speed of recognition was assessed on scale of 0–10. Correctness of performance was marked by applause or a signal "incorrect". Position memory: a 5 × 5 grid with a gradually increasing number of objects (images of books) was used to train remembering of the position of images. The pictures were first shown, then hidden, and the participant clicked on the cells where they remembered the pictures were located until they made 2 mistakes. The paper provides figures with examples of the computer programs. Who provided: not clearly stated in paper; however, the authors confirmed sessions were conducted by occupational therapists. The paper stated that the approach could possibly be used independently by a participant without involvement of the medical personnel. How: face‐to‐face and individually Where: inpatient rehabilitation department When and how much: daily for 30 minutes per day for 2 weeks (up to 15 hours) Tailoring: Schulte's Tables training: difficulty level could be adjusted by changing the time allowed for the participant to find the number before the hint appeared. Visual and spatial gnosis training: figure‐ground training: there was a gradual reduction of 'noise' intensity until it completely disappeared. Position memory training: an extra object was added with correct answers and training continued until 2 mistakes were made. Information appeared on the screen about the speed and correctness of answers and highest amount of information memorised. Modifications: none reported How well (planned): use of a training protocol How well (actual): none reported Comparator group Brief name: standard treatment Recipients: inpatient adults with cognitive impairments from a hemisphere stroke Why: to provide standard rehabilitation. What: no details provided Who provided: inpatient rehabilitation department staff How: not described but presumably face‐to‐face Where: inpatient rehabilitation department When and how much: not reported Tailoring: none reported Modifications: not reported How well (planned): not reported How well (actual): not reported
Outcomes	Primary: none Secondary IADL scale MMSE МоСА FAB Schulte's Tables Other Clock Drawing Test HADS SS‐QOL Methods of data collection: cognitive assessments were conducted by a trained assessor blinded to randomisation. Data collection time points: baseline and postintervention (14–16 days after baseline)
Notes	Funding: yes Conflict of interest: none Published trial protocol: no Trial registration: no Ethics approval: yes Author contact: further details about intervention implementation were provided by the authors. Data: median and interquartile ranges converted to means and SDs
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation was performed using 'method of letters', which was not explained and authors did not reply to email request for explanation of the method.
Allocation concealment (selection bias)	Unclear risk	Quote: "Randomization was performed with the use of method of letters. Cognitive assessment was conducted on the day of inclusion into the study; assessments were repeated on day 14–16 by a trained assessor blind to randomization". Comment: authors did not specifically state that baseline assessment occurred prior to randomisation.
Blinding of participants and personnel (performance bias) All outcomes	High risk	As for allocation concealment above and also the intervention was delivered "in addition to standard treatment at the in‐patient rehabilitation department", so it appeared the staff and participants delivering the computer program intervention would have known it was additional to standard treatment.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "… assessments were repeated on day 14–16 by a trained assessor blind to randomization".
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up could be detected.
Selective reporting (reporting bias)	Low risk	No protocol/trial registry reported or located but reports on all outcome measures.
Other bias	Low risk	No other identifiable bias.