Grilli & McFarland

(2011)

United States

RCT Crossover

PEDro=4

N=12

Population: Mean age=49.42yr; Gender: Male=5, Female=7.

Intervention: Participants were either instructed to self-imagine participation in a memory trivia game or rehearse the information they wanted to remember out loud during memory training trials.

Outcome Measures: Prospective memory, neuropsychological functioning (executive functioning).

1.        There was a significant between groups difference, where self-imagination instruction improved prospective memory (p<0.01). However, the proportion of questions answered correctly did not differ significantly between groups.

2.        A Pearson correlations test showed that performance in the self-imagination condition was not significantly correlated to memory or executive functioning.

Bourgeois et al.

(2007)

USA

RCT

PEDro=2

N=38

Population: TBI patients: Mean Age: 41.5yr; Gender: male=24, female=14; Mean Time Post-Injury: 11.3yr.

Intervention: Participants were randomized to receive either Spaced Retrieval (SR) training (n=22) delivered over the telephone or didactic strategy instruction (DSI) (n=16). Participants in both groups identified three memory-related goals to master.

Outcome Measures: Goal Mastery, Cognitive Difficulties Questionnaire (CDS).

1.        Those in the SR group showed significant improvement in goal mastery compared to the SI group (p<0.05). This was maintained at the one-month post intervention.

2.        Results on the CDS showed both groups having fewer significantly difficulties following treatment (p<0.001; p<0.005).

3.        There were no significant between-groups differences in participant reports of generalized strategy use or reported memory problems at either time-point (p>0.05).

Kaschel et al.

(2002)

Germany

RCT

PEDro=6

N=27

Population: Pragmatic Group (N=15): Mean age=36.6yr; Gender: Male=12, Female=3. Imagery Group (N=11): Mean age=41.9yr; Gender: Male=9, Female=2.

Intervention: Individuals were assigned to either imagery-based training (experimental), or pragmatic-based training (control) for 10 weeks, 3 times a week. Individuals were assessed at baseline, immediately following treatment conclusions, and at 3-months follow-up.

Outcome Measures: Concentration endurance, Wechsler Memory Scale (WMS), Rivermead Behavioral Memory Test (RBMT), Appointments test, Memory Assessment Clinics Rating Scales (MAC).

1.        There was a significant effect of time for the assessment of concentration endurance with both groups significantly improving over time (p<0.05). No other significant measures were found for concentration endurance.

2.        There were no significant differences between groups on the WMS.

3.        For the RBMT, only a significant effect of time was observed (p<0.05). A specific subset of the RBMT for logical memory showed a significant group (p<0.01) and interaction effect (p<0.05) indicating that those in the imagery condition had improved logical memory.

4.        When assessing ability to recall multiple appointment times, a significant effect of group (p<0.05), and time (p<0.01) was observed with individuals in the imagery performing better.

5.        On the MAC scale for relative’s rating of memory problems, there were significant interactions at all time points (p<0.05), and a significant effect of time (p<0.05) indicating that the self-imagery group had greater gains in memory according to relative’s ratings compared to the pragmatic group.

Milders et al.

(1995)

Netherlands

RCT

PEDro=5

N=31

Population: Closed Head Injury (CHI); Strategy Group (n=15): Mean Age=42.4yr; Mean Time Post-Injury=10.4yr; Pseudo Group (n=8): Mean Age=35.6yr; Mean Time Post-Injury=12.4yr; No-Treatment Group (n=8): Mean Age=37.7yr; Mean Time Post-Injury=12.9yr; Healthy Control (n=13): Mean Age=41.1yr

Intervention: 4yr follow-up to Berg et al. (1991).

Outcome Measures: Four-choice Reaction Time Task, Distraction Reaction Time Task, 15-Words Test, Face-Name Learning, Shopping Lists.

1.        Standardized memory sum scores at long-term follow-up were significantly lower in the three patient groups than in the normal control group (p<0.05).

2.        Pseudo-rehab group improved significantly (p<0.05) in memory from post-training to long-term follow-up; such improvements were not seen in any other groups. 75% of patients in the pseudo group improved compared to 20% in the strategy group and 37.5% in the no-treatment group.

3.        Reaction time scores did not differ significant between groups at follow-up (P=0.08).

Twum and Parente

(1994)

USA

RCT

PEDro=3

N=60

Population: TBI; Mean Age=21yr; Time Post Injury>6mo.

Intervention: Patients were randomized into one of four treatment groups: 1) No Imagery/No Verbal Labeling (control); 2) No Imagery/ Verbal Labeling; 3) Imagery/No Verbal Labeling; and 4) Imagery /Verbal Labeling. Verbal labeling and imagery instructions were given through Verbal Paired Associated (VerPA) and Visual Paired Associated (VisPA) tasks, respectively.

Outcome Measure: VerPA and VisPA tasks.

Berg et al.

(1991)

Netherlands

RCT

PEDro=5

N=39

Population: Closed Head Injury (CHI); Strategy Group (n=17): Mean Age=36yr; Mean Time Post Injury=5.3yr. Pseudo Group (n=11): Mean Age=33yr; Mean Time Post Injury=6.3yr. No-Treatment Group (n=11): Mean Age=35yr; Mean Time Post-Injury=6.8yr.

Intervention: Individuals were randomly assigned to one of three groups: strategy rehabilitation, pseudo-rehabilitation, or no-treatment. The strategy rehabilitation group had individualized training targeting to the identified memory problems (1hr, 3x/wk for 6wk). Daily homework was administered to augment the benefits of rehabilitation. The pseudo-rehabilitation (“drill and practice”) group participated in sessions consisted of memory tasks and games that were practiced in the laboratory and at home. The no-treatment group received no training.

Outcome Measures: Four-choice Reaction Time Task, Distraction Reaction Time Task, 15-Words Test, Face-Name Learning, Shopping Lists.

2.        No single effect of strategy training was found with respect to reaction time tasks post-training.

3.        While no significant effect of pseudo-training was found, strategy training had significant positive effects on all memory performance measures (memory sum score: p=0.011; acquisition score: p=0.038; delayed recall score: p=0.004), particularly at the final follow-up.

O’Neil-Pirozzi et al.

(2010b)

USA

Prospective Control Trial

N=94

Population: TBI; Experimental Group: Mean Age=47.3yr; Mean Time Post-Injury=11.8yr; Control Group: Mean Age=47.0yr; Mean Time Post-Injury=13.4yr

Intervention: Participants in the experimental group were trained to use Internal Memory Strategies (I-MEMS; n=54); the intervention consisted of 12 90-min sessions, held 2×/wk for 6 wk. It included memory education and emphasized internal strategy acquisition to improve memory function from encoding, storage and retrieval perspectives; the control group (n=40) consisted of a convenience sample.

Outcome Measure: Hopkins Verbal Learning Test-Revised (HVLT-R), Rivermead Behavioural Memory Test II (RBMT II). Patients were assessed on Week 1 (pretest), Week 7 (posttest 1), and Week 11 (posttest 2).

1.        Pretesting revealed a significant difference between experimental and control groups on the HVLT-R only (p=0.02).

2.        Individuals who had a severe TBI performed more poorly on the HVLT-R than those with moderate injuries.

3.        Although those with a severe injury did not improve as much as those with a mild or moderate injury, they did improve more than those in the control group at both posttest 1 (p=0.0002) and posttest 2 (p<0.0001).

4.        Similar to what was found with HVLT-R assessments, severe injury predicted worse RMBT II scores than moderate injury.

5.        RBMT II scores in the I-MEMS groups revealed significant improvements at both posttest 1 (p=0.045) and posttest 2 (p=0.0013) relative to control.

6.        Overall memory performance was improved for all those in the experimental group compared to the control group.

Manasse et al.

(2005)

USA

Case Series

N=5

Population: TBI: Age Range: 29-48yr; Gender: male=3, female=2; Time Post-Injury: 1-29yr.

Intervention: Subjects were shown pictures of individuals they interacted with daily and asked to identify them.

Traditional treatment: To assist subjects in memory recall, pictures were paired with an imagery statement. There were 9 (3 weeklies over a 3-week period) one on one training sessions to assist the individuals with face name recognition.

Real-world treatment: Following the third week, “real-world” treatment was begun. During the next 15 days, 2 interactions were performed each day with 2 hours separating the interactions.  Researchers recorded the subjects’ spontaneous use and knowledge of the staff’s name.

Outcome Measures: Name recall.

1.        Traditional treatment: results indicate that 2 of the 5 subjects mastered 6 names during treatment, 1 of the 5 mastered 3 names and 4 of the 5 mastered one of the names.

2.        Real-world treatment: During the real-world cueing condition only 2 names were consistently used by each subject.

3.        Improved name learning was seen regardless of the cueing strategy.

Tailby & Haslam

(2003)

Australia

Pre-Post

N=24

Population: Head Injury=12, CVA=6, Hypoxia=3, Other=3. Severe memory impairment group (n=8): Mean Age=43yr. Moderate memory impairment group (n=8): Mean Age=43.8yr. Mild/no memory impairment group (n=8): Mean Age=37.5yr.

Intervention: 3 groups were formed based on Verbal Memory Index (VMI) on the Wechsler Memory Scale-III. All participants were tested in 3 conditions: errorful (EF), standard errorless (examiner-generated; EL-E) and self-generated errorless (EL-S). For training, 96 words of 5-6 letters were used over 2 sessions. Following the learning tasks, memory was tested explicitly and implicitly 5 and 30 min after the study phase, generating 6 scores for each learning condition.

Outcome Measures: Verbal Memory Index (VMI).

1.        Cued recall performance following EL-S learning was significantly better than standard errorless learning (EL-E) conditions (p<0.0001).

2.        Level of priming did not differ significantly between groups (p>0.05).

3.        Memory performance was significantly better following EL-E activity (p<0.0001) compared to EF.

4.        A significant effect of severity was found (p<0.005) for the standard EL-E conditions; mild and moderate groups performed significantly better than severe group (defined by VMI: p<0.0001); significant effect of severity was also found for the EF condition (p<0.001).

Sumowski et al.

(2014)

USA

Pre-Post

N=10

Population: Severe TBI=10; Mean Age=42.8 yr; Gender: Male=6, Female=4; Mean Time Post Injury=8.4 yr.

Intervention: Participants studied 48 verbal paired associates (VPAs) divided into 3 learning conditions: massed restudy (MR), spaced restudy (SR), and retrieval practice (RP). MR is similar to cramming, whereas SR is distributed learning. RP was similar to SR; however, re-exposure trials were framed as cued recall tests. Recall of VPAs was done at 30 min post intervention, and at 1 wk. Participants performed all 3 methods of learning.

Outcome Measure: Recall of VPAs.

1.        Participants recalled 46.3% of VPAs learned through RP compared with 12.5% through MR (p<0.0001), and 15% through SR (p=0.002).

2.        SR did not result in better memory than MR (p=0.0555).

3.        At 1wk, participants recalled 11.3% in the RP group compared to 0.0% in the MR (p=0.004), and 1.3% in SR (p=0.011). Again, SR and MR did not differ from each other (p=0.343).

Potvin et al.

(2011)

Canada

PCT

N=30

Population: TBI; Rehabilitation Group (n=10): Mean Age=35yr; Gender: Male=7, Female=3. Control Group (n=20): Mean Age=30.90yr; Gender: Male=11, Female=9.

Intervention: Participants were assigned to either prospective memory (PM) rehabilitation programme or the standard neuropsychological interventions group (control). PM rehabilitation was based on the learning of visual imagery techniques.

Outcome Measure: Test Ecologique de Memoire Prospective (TEMP), Visual Discrimination Task, Semantic Association Task, Letter Visualization Task, Digit Symbol, Cancellation Task, Trail Making Test A & B, Brown-Peterson Task, Digit Span, Sullivan Logical Memory, Rey Auditory Verbal Learning Test, Brief Visuospatial Memory Test, Semantic Verbal Fluency, Mazes, Stroop Interference and Flexibility, CAPM (relative and participant versions).

1.        The experimental group performed significantly better on the TEMP post PM training than the control group (p<0.05).

2.        During the learning phase, cued recall improved for those in the experimental group, although this improvement was not found to be significant.

3.        Participants who took part in the rehabilitation program improved their performance on the PM experimental task (p<0.05).

4.        No significant group effects were found for any neuropsychological tests, except with the digit symbol test (p<0.05).

5.        Self-evaluated PM failures was significantly lower post-test in the rehabilitation group (p<0.05) but not the control group.

Grilli & Glisky

(2013)

USA

Pre-Post

N=30

Population: Patient Group: TBI=13, ABI=2; Mean Age=51.3yr; Gender: Male=7, Female=8. Healthy Control (n=15): Mean Age=50.7yr; Gender: Male=7, Female=8.

Intervention: Participants were exposed to five intentional coding conditions over two days. Controls did all five in one day. For each trial a word was on the screen for 10sec. A sentence specifying the task (condition) would appear above the target word. The conditions were: baseline, semantic elaboration, semantic self-referential processing, episodic self-referential processing, and self-imagining.

Outcome Measure: Immediate free-recall test.

*only results for the TBI group are reported

1.        For the patient group, self-imagining showed better free recall than baseline (p<0.001), semantic elaboration (p<0.001), episodic self-referential processing (p<0.001), and semantic self-referential processing (p<0.05).

2.        Self-referential processing enhanced free recall more than episodic self-referential processing (p<0.05).

3.        Semantic elaboration and episodic self-referential processing showed better free recall than scores attained at baseline (p<0.05, p<0.01, respectively).

4.        Self-descriptive trait adjectives were recalled more than non-self-descriptive trait adjectives among only those in the self-imagining (p<0.05) and semantic self-referential processing conditions (p<0.05).

Sumowski et al.

(2010)

USA

Case-Control

N=28

Population: Mean Age of TBI=38.4yr; Etiology of injury: motor vehicle accidents (n=9), falls (n=2), sports injuries (n=2), and assault (n=1). Condition: TBI=14, Control=14.

Intervention: Examining the effects of retrieval practice in delayed memory recall than simple restudy. Using a verbal paired associate paradigm examined recall abilities between controls and TBI patients.

Outcome Measures: Delayed cue recall test.

1.        A significant learning condition by group interaction was discovered (p<0.001).

2.        Healthy controls benefited from spaced restudy over massed restudy (p<0.001).

3.        Both groups greatly benefited from retrieval practice over massed and spaced restudy (p<0.001, p=0.23).

Schefft et al.

(2008)

USA

PCT

N=20

Population: Mean Age: 31.8yr; Gender: male=13, female=7; Condition: TBI

Intervention:

Study 1: Read condition: words were presented in pairs-1 pair per card, which participants were asked to read aloud. Generate condition: participants were shown one word on the card with the first letter of second word and asked to read aloud the words as soon as they knew the second word. The first recall test was given immediately after the presentation of the 50-word pairs, followed by the recognition memory test. Free recall test had patients write down as many of the second words from each pair that could be remembered. Recognition Test: 50 items corresponding to the appropriate input list and each item was composed of 2 previously unseen distractor words and 1 target word from the learning task. Word pairs were presented in the same order at testing as they had been presented during the learning trials.

Population: Mean Age: 34.3yr; Gender: male=18, female=2, Condition: TBI.

Study 2: Both the read and generate conditions were identical to study 1; however, here there was no recognition test.  Patients were given a cued recall trail, where each word pair association rule was provided as a cue for memory and a cued recall trail where the first word in the pair was presented. Free recall test had participants write down as many of the second words from the pair they could remember. For the cued recall with rules test they were given a sheet of paper with the title on it and one example of each rule. They were then asked to write down as many of the second words they could remember.

Study 1:

1.        Self-generation encoding procedures improved recognition memory test performance, but not free recall, compared with the didactic presentation.

Study 2:

1.        Self-generation strategy improved cued recall, but not free recall compared with the didactic condition.

2.        Study results also indicated that cued recall was also important as it was found to be effective when presented with the first word of the word pair.

Hillary et al.

(2003)

USA

Case Series

N=20

Population: Age Range: 18-55yr; Gender: male=16, female=4; Mean Time Post-Injury: 4.1yr; Condition: moderate to severe TBI.

Intervention: Examining if learning in TBI patients can be improved using spaced repetitions of a procedure compared to consecutive presentations of a procedure. A list of 115 words were chosen for recall, words were presented either once (single condition), twice consecutively (massed condition), or twice with 11 words between presentations (spaced condition).

Outcome Measures: Immediate and Delay Recall; Delay Recognition Trials, neuropsychological tasks.

1.        Spaced words were more likely to be recalled during the immediate recall than massed words (p=.018).

2.        On the delayed recall spaced words were more likely to be correctly recalled than massed words or once presented words during delayed recall performance (p<0.001).

3.        On the recognition performance test, individuals were able to correctly identify spaced words over massed (p=0.001) or once presented words (p=0.017).

4.        Significant main effect for study condition on immediate recall in the neuropsychological tasks (p<0.001).

Milders et al.

(1998)

Netherlands

Prospective Control Trial

N=26

Population: Closed Head Injury (CHI)=13; Healthy Controls=13. CHI Group: Mean Age=39yr; Mean Time Post Injury=5yr.

Intervention: Individuals with TBI completed exercises with standardized instructions that help make the new name more meaningful to the learner (8, 60-90min sessions over 4mo). Participants were assessed at baseline (3x) and 1wk and 6mo after training.

Outcome Measures: Name Learning Test, Name-Occupation-Town Learning Test, Famous Faces Naming Test, Digit Span Forwards, Auditory Verbal Learning Task.

*only results for the TBI group are reported

1.        A main effect for the patient’s group was found for the Name-Occupation Town Test (p<0.001).

2.        Performance on the name learning test for the patient’s group from pre-to post training (meaningful names= 12.8±4.6 to 14.0±3.6; meaningless names=11.6±3.9 to 11.7±3.2).

3.        There were improvements on the Name-Occupation-Town Learning Test in the patient group (names= 16.8±7.7 to 21.6±7.2; Occupations + town= 22.4±9.4 to 23.5±8.2).

Thoene & Glisky

(1994)

Germany

PCT

N=12

Population: Mean age=45.58yr; Gender: Male=6, Female=6; Mean time post-injury=7.38yr.

Intervention: Individuals attempted to learn the names associated with 4 faces in 3 conditions (mnemonic, vanishing cues, and video). Mnemonic trials consisted of associating a face with an elaborate verbal association. The video condition consisted of the ‘face subject’ introducing themselves via video to the participant. The vanishing cues condition consisted of cueing the individual to remember the name during training sessions by cueing them with letters from the target name.

Outcome Measures: Naming errors: Omission errors, other-set intrusions (information from another condition), same-set intrusions, other errors in naming, reaching criterion threshold, incidental recall (information not related to names).

1.        There as a significant effect of condition where the only condition to reach the criterion threshold was the mnemonic condition (p=0.001). Post hoc tests confirmed that individuals required fewer trials in the mnemonic condition to reach criterion (p=0.017).

2.        While participating in the vanishing cues condition, individuals required less cues to remember target names over time.

3.        There were significant differences between conditions for omission made, with the mnemonic group making significantly less (p=0.000).

4.        There were significantly fewer other-set intrusions in the mnemonic group, compared to the other groups (p=0.04).

5.        There were significantly fewer same-set intrusions in the mnemonic condition than other conditions (p=0.01).

6.        The incidental recall of the target’s professions was significantly higher in the video condition compared to other conditions (p=0.04).