Author

Year

Country

Research Design

PEDro

Sample Size

Yip & Man

(2013)

Hong Kong

RCT

PEDro=5

N=37

Population: ABI. Treatment Group (TG, n=19): Mean Age=37.83yr; Gender: Male=12, Female=7; Mean Time Post Injury=145.13d. Control Group (CG, n=18): Mean Age=38.53yr; Gender: Male=12, Female=6; Mean Time Post Injury=167.53d.

Intervention: Participants were randomized to receive virtual-reality (VR) prospective memory (PM) training (TG) or control (CG). VRPM training consisted of event-based tasks, time-based tasks, ongoing tasks, and recall tasks in both visual and auditory formats. Control training consisted of reading and games. Both were received in 30min sessions 2/wk for a total of 6wk. Outcomes were assessed at baseline and after treatment.

Outcome Measures: VR-based PM test (VRPMT); Real life behavioural PM test (RLPMT); Cambridge Prospective Memory Test–Chinese Version (CAMPROMPT-CV); Hong Kong List Learning Test (HKLLT); Frontal Assessment Battery (FAB); Word Fluency Test–Chinese Version (WFT-CV); Colour Trails Test (CTT); Community Integration Questionnaire–Chinese Version (CIQ-CV); Self-efficacy questionnaire (SEQ).

1.        In the TG, VRPMT showed significant improvements after treatment on immediate recall of tasks (p<0.05), number of time checks (p<0.001), and performance of event-based (p<0.001), time-based (p<0.001), and ongoing (p<0.01) tasks compared to baseline. No significant difference was found on delayed recall of tasks or total time lapsed.

2.        In the TG, RLPMT showed significant improvements after treatment in event-based (p<0.01) and time-based (p<0.01) tasks, but not ongoing tasks, compared to baseline.

3.        In the TG, significant improvements were found after treatment on CAMPROMPT-CV (p<0.05), FAB (p<0.01), WFT-CV (p<0.01), and SEQ (p<0.01) compared to baseline. No significant difference was found on HKLLT, CTT, or CIQ-CV.

4.        In the CG, no significant difference was found after treatment on any outcome measure compared to baseline.

5.        After treatment, a significant difference was found between groups on event-based tasks of RLPMT (p<0.05), FAB (p<0.01), WFT-CV (p<0.05), and CTT (p<0.05). No significant difference was found between groups on VRPMT, CAMPROMPT-CV, HKLLT, CIQ-CV, or SEQ.

Grealy et al.

(1999)

Scotland

RCT

PEDro=1

N=13

Population: TBI patients: Age Range: 19-64; Gender: male=8, female=5.

Intervention: Crossover design: patients were allocated to a 4-week intervention of receiving Virtual reality (VR) exercise or a no-exercise control condition.

Outcome Measure: Tests measuring attention, information processing, learning, memory, and reaction and movement times.

1.        Intervention group (n=13) performed significantly better than control group (n=320) on digit symbol (p<0.01), verbal (p>0.01) and visual (p<0.05) learning tasks.

2.        Reaction (p<0.01) and movement (p<0.05) times improved significantly after a single VR session.

Dahdah et al.

(2017)

USA

Pre-Post

N_Initial=21

N_Final=15

Population: CVA=6, TBI=5, Tumor=2, Anoxia brain injury=2; Mean Age=40.3yr; Gender: Male=12, Female=3.

Treatment: Participants received the virtual reality (VR) intervention sessions (apartment and classroom) twice per week for a 4wk period. Sessions 1 and 8 included all types of distractors, sessions 2 and 3 included no distracting stimuli, sessions 4 and 5 included only auditory distracting stimuli, and sessions 6 and 7 included only visual distracting stimuli.

Outcome Measure: Woodcock-Johnson, 3^rd edition (WJ-III pair cancellation subtest), Delis-Kaplan Executive Function System (D-KEFS Color-Word Interference subtest), Automated Neuropsychological Assessment Metrics (ANAM Go/No-Go and unimodal Stroop subtests), VR Stroop task (apartment and classroom).

1.        No statistically significant performance differences were found from baseline to conclusion of the study for the VR apartment Stroop or D-KEFS Stroop test.

2.        For the VR classroom, participants’ shortest response time on the word-reading condition was significantly reduced by session 8 (p=0.0383). All other VR classroom Stroop variables did not show significant differences.

3.        No significant differences from session 1 to session 8 were found for all pair cancellation subtest scores.

4.        From session 1 to 8, the ANAM Stroop word-reading percentage of items with a correct response (p=0.0293), ANAM Stroop word-reading number of correct responses per minute (p=0.0321), and ANAM Go/No-Go number of impulsive/bad responses (p=0.0408) significantly increased. All other ANAM variables did not show significant differences.

Sorita et al.

2013

France

PCT

N=27

Population: TBI. Treatment Group (TG, n=14): Mean Age=31.1; Gender: Male=12, Female=2; Mean Time Post Injury=4.67yr; Mean GCS=5.8. Control Group (CG, n=13): Mean Age=31.1; Gender: Male=13, Female=0; Mean Time Post Injury=6.77yr; Mean GCS=6.7.

Treatment: Participants engaged in the same route-learning task in either a real urban environment (CG) or a virtual simulation of that environment (TG). After a learning phase, participants repeated the task twice in a row and >24h later. Outcomes were assessed after each repetition and a series of tests was completed after the last repetition.

Outcome Measures: Route-learning task; Sketch map test; Map recognition test; Scene arrangement test.

1.        On the task, mean error rates for immediate and delayed recall were higher in the TG than in the CG, but this difference was not significant (p=0.42).

2.        On the task, mean scores were higher on the second (immediate) recall and the third (delayed) recall compared to the first (immediate) recall in both groups (p<0.001).

3.        On the task, mean scores were higher on the second recall than on the third recall in both groups, but the difference was not significant (p=0.44).

4.        No significant interactions between recall and environment were found.

5.        Mean scores on the scene arrangement test were significantly higher in the CG than in the TG (p=0.01).

6.        Mean scores on the sketch mapping test were higher in the CG than in the TG, but this difference was not significant (p=0.07).

7.        Mean scores on the map recognition test were the same in both groups (p=0.83).