. 2021 Jan 15;18:8. doi: 10.1186/s12984-020-00797-w

Table 2.

Studies investigating steady-state walking listed by increasing walking speed

Authors	Device [ROI]; method of placing channels	Walking time analyzed, task and speed: m/s (SD)	Results
(Miyai et al., 2003) [43]	fNIRS [PFC, PMC, pre-SMA, SMA, SMC]; relative to Cz. ROI based on MRI of 2 subjects	10–30 s post-start, Treadmill with body weight support: 0.06	1. Baseline: decreased ipsilesional SMC activation compared to contralesional SMC 2. Two-months post inpatient rehab: increased SMC symmetry, greater ipsilesional PMC activity and increased cadence, gait symmetry, Fugl-Meyer LE score 3. Greater pre-SMA and PFC activation in patients with large cortical lesions for both time points. No activation over lesioned cortex 4. Increased SMC symmetry (not PMC, SMA) correlated to increased gait symmetry
(Miyai et al., 2002) [24]	fNIRS [PFC, PMC, pre-SMA, SMA, SMC]; same as Miyai et al. 2003	10–30 s post-start, Treadmill with manual assistance at the leg or facilitation at hip: 0.06	1. Increased bilateral PFC, PMC, pre-SMA, and more prominent contralesional SMC (compared to ipsilesional SMC) activation 2. Greater bilateral pre-SMA and PFC, ipsilesional PMC activation, increased cadence and gait symmetry with facilitation at the hip compared to manual assistance of leg 3. No activation over lesioned cortex
(Contreras-Vidal et al., 2018) [42]	EEG [Whole head]; 10/20 system	10–360 s post-start, Overground, comfortable pace with exoskeleton: (0.14–0.31)*	1. Increased activation over frontal and central regions 2. Consistent pre and post training activation over occipital regions 3. Gait speed doubled after 12 sessions of exoskeleton gait training
(Sangani et al., 2015) [28]	fNIRS [PFC, PMC, SMA, SMC]; relative to Cz	15–25 s post-start, Treadmill, comfortable pace with and without light finger touch: 0.21	1. Increased activation along frontal and sensorimotor cortices with greater activation along contralesional hemisphere 2. Greater overall activation symmetry, localized activation to SMC, and increased gait speed with light finger touch
(Miyai et al., 2006) [25]	fNIRS [PMC, SMA, SMC]; same as Miyai et al. 2003	10–20 s post-start, Treadmill with and without body weight support (BWS): Stroke: 0.30 Healthy: 0.30	1. Overall increased bilateral PMC and SMA, contralesional SMC activity 2. Increased SMC and gait symmetry with BWS 3. Increased bilateral PMC, SMA, SMC for healthy participants and a general increase in activation with BWS 4. For stroke only: increased SMC symmetry correlated to increased gait symmetry and increased SMC (not PMC, SMA) activity correlated to increased cadence
(Mihara et al., 2007) [39]	fNIRS [PFC, SMA, SMC]; same as Miyai et al.2003	24-30 s post-treadmill start, Treadmill: Stroke: Fast, comfortable pace, 0.33 (0.22) Healthy: Comfortable pace, 0.97 (SD not provided)	1. Increased PFC, SMA, and SMC activity 2. Compared to the acceleration phase, PFC and SMA activation decreased in healthy and increased in the stroke group 3. SMC activation remained similar throughout walking for both groups
(Garcia-Cossio et al., 2015) [35]	EEG [Whole head]; 10/20 system	During “constant and stable speed”, Treadmill, comfortable speed with exoskeleton: Stroke: max 0.42 Healthy: 0.42	1. Increased activation at fronto-central and contralesional centro-parietal regions 2. No gait-cycle related modulations in brain activity compared to healthy adults who showed low gamma fluctuations with gait-cycle 3. No correlations between brain activity and walking performance
(Hawkins et al., 2018) [36]	fNIRS [PFC]; high and lateral on the forehead	37–67 s after start command, Overground, preferred speed: Stroke: 0.51 (0.27) Older adults: 1.07 (0.16) Young adults: 1.28 (0.18)	1. Increased PFC activity 2. Greater increase in PFC compared to young adults, no different to older adults
(Hermand et al., 2019) [37]	fNIRS [PFC]; 10/20 system	10–20 s post-start, Overground, comfortable pace: 0.52 (0.23)	1. Increased PFC activity 2. No differences between hemispheres 3. No significant correlations between brain activation and gait performance
(Chatterjee et al., 2019) [33]	fNIRS [PFC]; high and lateral on forehead	7–37 s post-start, Overground, self-selected: 0.6 (0.2)	1. Increase PFC activation during walking 2. Low balance confidence subgroup showed greater PFC, slower walking speed, and shorter stride length (not step width) compared to high balance confidence subgroup 3. No differences between hemispheres
(Chen et al., 2019) [41]	EEG [Whole head]; 10/20 system	0–300 s post-start, Treadmill, Time point 1: 0.58 (0.23) Time point 2: 0.71 (0.31)	1. Increase connectivity between contralesional frontocentral, middle central and ipsilesional centroparietal, greater increase in gait speed and temporal symmetry post turning treadmill 2. Increased connectivity correlated to increased gait symmetry but not gait speed 3. No differences in connectivity after regular treadmill training
(Choi et al., 2016) [34]	EEG [Whole head]; 10/20 system	Entire walking task (10–20 s long), Overground, self-paced: speed not reported	1. Increased activation in contralesional centro-parietal area
(Lee et al., 2018) [30]	fNIRS [PFC, PMC, SMA, SMC]; not reported	5–30 s post-start, Treadmill, self-selected with exoskeleton: speed not reported	1. Decreased SMA, PMC, SMC activity with versus without robotic assistance compared to no assistance
(Caliandro et al., 2020) [23]	fNIRS [PFC]; 10/20 system	5 s to end of walking, Overground, with and without exoskeleton: speed not reported	1. Both groups: greater PFC activation for walking with versus without exoskeleton 2. Greater PFC activation for the stroke compared to healthy group 3. No hemispheric difference
(Saitou et al., 2000) [27]	fNIRS [Ipsilesional forehead]; “forehead of impaired side”	0–300 s post-start, Overground: speed not reported	1. Increased activation for 15 out of 22 participants (cerebral blood volume and cerebral oxygenation volume)

Authors

Device
[ROI]; method of placing channels

Walking time analyzed, task and speed: m/s (SD)

Results

(Miyai et al., 2003) [43]

fNIRS

[PFC, PMC, pre-SMA, SMA, SMC]; relative to Cz. ROI based on MRI of 2 subjects

10–30 s post-start,

Treadmill with body weight support: 0.06

1. Baseline: decreased ipsilesional SMC activation compared to contralesional SMC

2. Two-months post inpatient rehab: increased SMC symmetry, greater ipsilesional PMC activity and increased cadence, gait symmetry, Fugl-Meyer LE score

3. Greater pre-SMA and PFC activation in patients with large cortical lesions for both time points. No activation over lesioned cortex

4. Increased SMC symmetry (not PMC, SMA) correlated to increased gait symmetry

(Miyai et al., 2002) [24]

fNIRS

[PFC, PMC, pre-SMA, SMA, SMC]; same as Miyai et al. 2003

10–30 s post-start,

Treadmill with manual assistance at the leg or facilitation at hip: 0.06

1. Increased bilateral PFC, PMC, pre-SMA, and more prominent contralesional SMC (compared to ipsilesional SMC) activation

2. Greater bilateral pre-SMA and PFC, ipsilesional PMC activation, increased cadence and gait symmetry with facilitation at the hip compared to manual assistance of leg

3. No activation over lesioned cortex

(Contreras-Vidal et al., 2018) [42]

EEG

[Whole head]; 10/20 system

10–360 s post-start, Overground, comfortable pace with exoskeleton: (0.14–0.31)*

1. Increased activation over frontal and central regions

2. Consistent pre and post training activation over occipital regions

3. Gait speed doubled after 12 sessions of exoskeleton gait training

(Sangani et al., 2015) [28]

fNIRS

[PFC, PMC, SMA, SMC]; relative to Cz

15–25 s post-start,

Treadmill, comfortable pace with and without light finger touch: 0.21

1. Increased activation along frontal and sensorimotor cortices with greater activation along contralesional hemisphere

2. Greater overall activation symmetry, localized activation to SMC, and increased gait speed with light finger touch

(Miyai et al., 2006) [25]

fNIRS

[PMC, SMA, SMC]; same as Miyai et al. 2003

10–20 s post-start,

Treadmill

with and without body weight support (BWS):

Stroke: 0.30

Healthy: 0.30

1. Overall increased bilateral PMC and SMA, contralesional SMC activity

2. Increased SMC and gait symmetry with BWS

3. Increased bilateral PMC, SMA, SMC for healthy participants and a general increase in activation with BWS

4. For stroke only: increased SMC symmetry correlated to increased gait symmetry and increased SMC (not PMC, SMA) activity correlated to increased cadence

(Mihara et al., 2007) [39]

fNIRS

[PFC, SMA, SMC]; same as Miyai et al.2003

24-30 s post-treadmill start, Treadmill:

Stroke: Fast, comfortable pace, 0.33 (0.22)

Healthy: Comfortable pace, 0.97 (SD not provided)

1. Increased PFC, SMA, and SMC activity

2. Compared to the acceleration phase, PFC and SMA activation decreased in healthy and increased in the stroke group

3. SMC activation remained similar throughout walking for both groups

(Garcia-Cossio et al., 2015) [35]

EEG

[Whole head]; 10/20 system

During “constant and stable speed”, Treadmill, comfortable speed with exoskeleton:

Stroke: max 0.42

Healthy: 0.42

1. Increased activation at fronto-central and contralesional centro-parietal regions

2. No gait-cycle related modulations in brain activity compared to healthy adults who showed low gamma fluctuations with gait-cycle

3. No correlations between brain activity and walking performance

(Hawkins et al., 2018) [36]

fNIRS

[PFC];

high and lateral on the forehead

37–67 s after start command, Overground, preferred speed:

Stroke: 0.51 (0.27)

Older adults: 1.07 (0.16)

Young adults: 1.28 (0.18)

1. Increased PFC activity

2. Greater increase in PFC compared to young adults, no different to older adults

(Hermand et al., 2019) [37]

fNIRS

[PFC];

10/20 system

10–20 s post-start, Overground, comfortable pace: 0.52 (0.23)

1. Increased PFC activity

2. No differences between hemispheres

3. No significant correlations between brain activation and gait performance

(Chatterjee et al., 2019) [33]

fNIRS

[PFC];

high and lateral on forehead

7–37 s post-start, Overground, self-selected: 0.6 (0.2)

1. Increase PFC activation during walking

2. Low balance confidence subgroup showed greater PFC, slower walking speed, and shorter stride length (not step width) compared to high balance confidence subgroup

3. No differences between hemispheres

(Chen et al., 2019) [41]

EEG

[Whole head];

10/20 system

0–300 s post-start,

Treadmill,

Time point 1: 0.58 (0.23)

Time point 2: 0.71 (0.31)

1. Increase connectivity between contralesional frontocentral, middle central and ipsilesional centroparietal, greater increase in gait speed and temporal symmetry post turning treadmill

2. Increased connectivity correlated to increased gait symmetry but not gait speed

3. No differences in connectivity after regular treadmill training

(Choi et al., 2016) [34]

EEG

[Whole head]; 10/20 system

Entire walking task (10–20 s long), Overground, self-paced: speed not reported

1. Increased activation in contralesional centro-parietal area

(Lee et al., 2018) [30]

fNIRS

[PFC, PMC, SMA, SMC]; not reported

5–30 s post-start,

Treadmill, self-selected with exoskeleton: speed not reported

1. Decreased SMA, PMC, SMC activity with versus without robotic assistance compared to no assistance

(Caliandro et al., 2020) [23]

fNIRS [PFC]; 10/20 system

5 s to end of walking, Overground, with and without exoskeleton: speed not reported

1. Both groups: greater PFC activation for walking with versus without exoskeleton

2. Greater PFC activation for the stroke compared to healthy group

3. No hemispheric difference

(Saitou et al., 2000) [27]

fNIRS

[Ipsilesional forehead]; “forehead of impaired side”

0–300 s post-start, Overground: speed not reported

1. Increased activation for 15 out of 22 participants (cerebral blood volume and cerebral oxygenation volume)

All results are reported in comparison to baseline (unless otherwise stated)

ROI region of interest, PFC prefrontal cortex, PMC premotor cortex, SMA supplementary motor area, SMC sensorimotor cortex, M1 primary motor cortex, fNIRS functional near-infrared spectroscopy, EEG electroencephalography

*Interpreted from graph