The Activation of the Mirror Neuron System during Action Observation and Action Execution with Mirror Visual Feedback in Stroke: A Systematic Review

Table 1

Characteristics of studies regarding the effects of long-term training on brain activation in stroke patients.

Study	Characteristics of stroke participants	Severity of hemiplegia	Intervention protocol	Intervention protocol	Imaging modality	Main results
						Neuroimaging findings	Behavioral improvements
Ertelt et al. [22]	15 chronic stroke patients (>six months after stroke), aged 38–69.	Moderate arm paralysis (FAT: 0 to 5; WMFT (time): 2.41 to 41.29 seconds).	EG (): AOT (videos of upper limb actions followed by practice of observed actions, using the paretic upper limb); CG (): control (nonbiological videos followed by practice of the same actions, using the paretic upper limb).	Four-week intervention (90 minutes/session, 18 sessions in total).	fMRI (EG = 7 and CG = 6).	More evident activation of PMv, SMA, insula, and STG over the nonaffected hemisphere and PMv, SMG, and STG over the affected hemisphere in EG than CG.	FAT posttreatment (EG > CG); WMFT-posttreatment (−); SIS-posttreatment (EG > CG).; No significant difference between posttreatment and follow-up in three assessments.
Michielsen et al. [52]	40 chronic stroke patients (>one year after stroke), aged 55.3 ± 12.0 (MT)/58.7 ± 13.5 (CG).	Brunnstrom stage for the upper extremity III-V.	MT (): bimanual exercise with MVF of the unaffected hand; CG (): bimanual exercise with direct view of both hands.	Six-week intervention (one hour/session, five sessions/week).	fMRI (MT = 9 and CG = 7).	A shift of activation toward the M1 over the affected hemisphere in MT.	FMA posttreatment (MT > CG); FMA 6-month follow-up (−); Jamar handheld dynamometer (−); Tardieu scale (−); ARAT (−); Stroke-ULAM (−); EQ-5D (−).
Bhasin et al. [46]	20 stroke patients (three 14 months after stroke), aged 45.45 ± 6.6.	Brunnstrom stage of the hand II-IV.	Bilateral hand exercise with virtual MVF of the nonparetic hand.	Eight-week intervention (five days/week, 60–90 minutes/session).	fMRI.	LI of BA 4 and BA 6 was increased at eight weeks.	FM scale (posttreatment and 24-week follow-up > pretreatment); MBI (posttreatment and 24-week follow-up > pretreatment); MRC grade scale (−); Brunnstrom stage (−).
Bae et al. [45]	20 stroke patients (<six months after stroke), aged 55.2 ± 8.5 (MT)/52.6 ± 11.2 (CG).	Brunnstrom stage of the hand II–IV.	MT (): bilateral upper limb exercise with MVF; CG (): paretic arm exercise only.	Four-week intervention (30 minutes/session, five times/week).	EEG.	Mu suppression at C3, Cz, and C4 was higher in MT than CG.	MFT posttreatment (MT > CG).
Sun et al. [28]	10 stroke patients (<two months after stroke), aged 59.4 ± 4.94.	Severe arm paralysis (FMA: 10 to 25).	On top of the CR, patients received the training: EG (): AO (paretic upper limb actions) with MI; CG (): MI alone.	Four-week intervention.	EEG.	AO with MI increased to higher mu suppression over C3 than MI at week 2, 3 and 4.	FMA week 2, 3 and 4; (EG > CG); PST week 3 and 4 (EG > CG).
Brunetti et al. [59]	11 stroke patients (15–92 days after stroke), aged 49–74.	Severe hand paralysis (a wrist extension of less than 20 degrees and metacarpophalangeal joint extension of less than 10 degrees).	On top of the CR, patients performed a bilateral exercise with MVF of nonparetic side.	Four-week intervention (30 minutes/session, five sessions/week).	fNIRS.	The activation pattern of M1 and precuneus was stable over time.	Six of eleven patients showed improvement (gain scores from one to eight) in FMA-finger.

FAT: Frenchay arm test; WMFT: Wolf motor function test; EG: experimental group; CG: control group; AOT: action observation training; fMRI: functional magnetic resonance image; PMv: ventral premotor cortex; SMA: supplementary motor area; STG: superior temporal gyrus; SMG: supramarginal gyrus; SIS: stroke impact scale; MVF: mirror visual feedback; MT: mirror therapy; M1: primary motor cortex; FMA: Fugl-Meyer assessment; ARAT: action research arm test; Stroke-ULAM: stroke upper limb activity monitor; EQ-5D: EuroQOL five-dimension questionnaire; LI: laterality index; BA: Broadman area; MBI: modified Barthel index; MRC: Medical Research Council; EEG: electroencephalography; BA: Broadman area; MFT: manual function test; PT: physical training; CR: conventional rehabilitation; AO: action observation; MI: motor imagery; ERD: event-related desynchronization; PST: pinch strength test; fNIRS: functional near-infrared spectroscopy.