Skip to main content

Table 2 Summary of studies that compared ER to training without error modification

From: The effects of error-augmentation versus error-reduction paradigms in robotic therapy to enhance upper extremity performance and recovery post-stroke: a systematic review

Article

Study design

Number of participants: experimental group (E) and control group (C)

Participants characteristics

Equipment

Experimental Protocol

Outcomes and assessment tools

Main results and interpretation (means ± standard deviation)

Effect size (Cohen’s d)

Quality of study (PEDro score)

Kahn et al. [55]

Randomized controlled trial.

Trial not registered

19 participants with stroke in total

E = 10

C = 9

E: age (mean) = 55.6 ± 12.2, four males and six females, months post-stroke (mean) = 75.8 ± 45.5, CMc (mean) = 3.5 ± 0.9

C: age (mean) = 55.9 ± 12.3, seven males and two females, months after stroke (mean) = 103.1 ± 48.2,

CM (mean) = 3.2 ± 1.0

ARM Guide (the Assisted Rehabilita-tion and Measure-ment Guide), a robotic device that drives user’s hand along a linear rail

Experimental task: reaching

ERa group: the subjects received robotic active assistance

Control group: they were not attached to the device

Training parameters:

A single session consisted of 50 trials

Kinematic:

Supported fraction of range (FRs), supported fraction of speed (FSs), unsupported fraction of speed (FRu).

Clinical:

CM

FRs: ER group showed more improvement than control group (0.14 ± 0.11 vs 0.12 ± 0.09), but not significantly (p = 0.844)

FSs: ER group showed more improvement than control group (0.22 ± 0.09 vs 0.21 ± 0.13), but not significantly (p = 0.898)

FRu: control group showed more improvement than ER group (0.02 ± 0.07 vs 0.01 ± 0.09), but not significantly (p = 0.687)

CM: control group showed more improvement than ER group (0.3 ± 0.5 vs 0.2 ± 0.4), but not significantly (p = 0.414)

FRs: 0.2 (small effect)

FSs: 0.09 (very small effect)

FRu: 0.13 (very small effect)

CM: 0.22 (small effect)

4/10, fair quality

Takahashi et al. [56]

Randomized controlled trial

Trial registered

NCT01244243

13 participants with stroke in total

Group A-A: 7

Group ANA-A: 6

A-A group: age (mean) = 58.6 ± 16, months since stroke (mean) = 14.4 ± 13.2 AMFMb = 40.4 ± 10.5

ANA-A group: age (mean) = 67.3 ± 15, months since stroke (mean) = 57.8 ± 87.6, AMFM = 49.5 ± 8.6

HWARD (The Hand Wrist Assistive Rehabilita-tion Device) is 3 degrees of freedom that assists the hand in grasp and in release movement

Experimental task: Grasp and release exercises

A-A (Active Assist) group: the subjects received robotic assistance throughout the training sessions. ANA-A (Active Non-Assist to Active Assist) group: the subjects were attached to the robotic device but did not receive assistance for the first half of the treatment phase, then received robotic assistance in the second half of the phase.

Training parameters:

15 sessions over three weeks. Each session lasted around one hour and half.

Clinical: ARATd, AMFM,

Box and Blocks test

Other:

EMGf, fMRIg

ARAT: by the end of session, the A-A group showed more improvement than ANA-A group (5.3 ± 2.1 vs 2.8 ± 1.8), the difference was significant [F(2,10) = 5.2, p < 0.03].

AMFM:: by the end of session, the A-A group showed more improvement than ANA-A group (9.1 ± 2.1 vs 5.8 ± 1.6), the difference was significant [F(2,10) = 4.8, p < 0.04]

Box and Blocks: no significant difference found between the two groups

EMG: no significant change within subjects in the recruitment pattern of muscle

fMRI: the percent signal change in the left primary sensorimotor cortex remained stable (p > 0.3)

ARAT: 1.28 (very large effect)

AMFM:1.78 (very large effect)

5/10, fair quality

Timmer-mans et al. [57]

Randomized controlled trial.

Trial registered on ISRCTN registry

ISRCTN82787126

22 participants with stroke in total

E = 11

C = 11

E: age (mean) = 61.8 ± 6.8, eight males and three females, months post-stroke (mean) = 44.4 ± 36,

AMFM (median) = 53

C: age (mean) = 56.8 ± 6.4, eight males and three females, years after stroke (mean) = 33.6 ± 34.8,

AMFM (median) = 50

The robotic system Haptic Master, with six degrees of freedom

Experimental task: functional tasks (‘drinking from a cup,’ ‘eating with a knife and fork,’ ‘taking money from a purse’ or ‘using a tray.’

ER group: subjects received trajectory guidance from the robot.

Control group: the subjects were not supported by the robot.

Training parameters:

Two sessions of 30 min (break of 0.5 to 1 h) per day, four days per week for eight weeks.

Clinical

AMFM, ARAT, MAL AU and QUe

AMFM: by the end of 8 weeks training, the control group showed more improvement than ER group (3.5 ± 5.9 vs 1.6 ± 17.9), but the result was not significant (p = 0.51).

ARAT: the control group showed more improvement than ER group (16.1 ± 26.5 vs 9.0 ± 11.0), but the result was not significant (p = 0.79).

MAL AU: the control group showed more improvement than ER group (33.1 ± 64.0 vs 9.0 ± 37.5), but the result was not significant (p = 0.33).

MAL QU: the control group showed more improvement than ER group (46.5 ± 40.9 vs 41.3 ± 33.1), but the result was not significant (p = 0.4).

AMFM: 0.16 (very small effect)

ARAT: 0.38 (small effect)

MAL AU: 0.47 (small effect)

MAL QU: 0.14 (very small effect)

8/10, high quality

  1. aEA/ER: error augmentation/error reduction
  2. bFM/AMFM: Fugl-Meyer assessment/Arm Motor Fugl-Meyer
  3. cCM: Chedoke-McMaster scale score
  4. dARAT: Action Research Arm Test
  5. eMAL: Motor Activity Log, AU: Amount of use, QU: Quality of use
  6. fEMG: Electromyography
  7. gfMRI: Functional Magnetic Resonance Imaging