Table 2.
Summary of the main characteristics of the articles included in the review.
| ID | Paper | Aim | Study Design | Population Characteristics, Age (mean ± SD) and Male/Female Ratio | Population Characteristics | Sensor Type | Sensor Number | Sensor SF | Sensor Range | Raw signal Filter (Cut-Off Frequency) |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Shema-Shiratzky 2019 [37] | To determine which gait features become worse during sustained walking in people with MS | Pilot | 58 relapsing–remitting MS (49.0 ± 10.0, 17/41) | EDSS 2–6 | 3D ACC and 3D GYRO (OPAL, Apdm) | 3 | 128 Hz | ± 16 g; ± 2000 deg/s | / |
| 2 | Retory 2019 [38] | To determine gait parameters in subjects with high or low body mass index | Validation | 10 controls (43.8 ± 12.8, 3/7) 13 non-overweight (42.2 ± 13.6, 4/9) 29 overweight (43.8 ± 12.8, 4/25) |
BMI < 25kg/m2 BMI > 30kg/m2 |
3D ACC (Nox-T3, Polygraph) | 1 | 10 Hz | ± 2 g | LP 5th order Butter. (2.5 Hz) |
| 3 | Zhang 2018 [39] | To propose and evaluate a gait symmetry index | Feasibility | 16 post-stroke (54, range 23–74, 9/7) 9 controls (35, range 25–48, 5/4) |
SIS 190-288 | 3D ACC and 3D GYRO (MTw Awinda, XSens) | 3 | 100 Hz | / | LP 2nd order Butter. (10 Hz) |
| 4 | Byrnes 2018 [40] | To determine characteristics of the attractor for acceleration gait data | Feasibility | 19 sLSS (73.8 ± 5.3, 11/8) 24 controls (59.9 ± 10.5, 9/15) |
ODI 27.9% ± 16.9% | 3D IMU (RehaGait system, Hasomed GmbH) | 7 | 400 Hz | ± 16 g; ± 2000 deg/s; ± 1.3 Gs |
LP 4th order Butter. |
| 5 | Teufl 2018 [41] | To evaluate the performance of an algorithm for the calculation of 3D joint angles | Validation | 28 healthy (24.0 ± 2.7, 13/15) | / | 3D ACC and 3D GYRO (MTw Awinda, XSens) | 7 | 60 Hz | / | / |
| 6 | Proessl 2018 [42] | To investigate agreement between smart device and IMU-based gait parameters during prolonged walking | Validation | 20 healthy (25.0 ± 3.7, 13/7) | / | 3D ACC (Ipod Touch, Apple) | 1 | 100 Hz | / | / |
| 7 | Loske 2018 [43] | To check if gait quality improves postoperatively | Cohort | 20 sLSS 20 controls (60.5 ± 11.4) |
ODI 30.7% ± 16.3% | 3D IMU (RehaGait system, Hasomed GmbH) | 7 | 400 Hz | / | / |
| 8 | Drover 2017 [44] | To validate a novel wearable sensor based faller classification method | Validation | 76 older adults (74.15 ± 7.0) | / | 3D ACC (X16-1C, Gulf Coast Data Concepts) | 3 | 50 Hz | / | / |
| 9 | Brodie 2016 [45] | To validate an adaptive filter designed to improve the quality of accelerometer data | Validation | 5 MS (68 ± 8, 0/5) 13 controls (32 ± 6, 4/9) |
EDSS 4.3 ± 1.0 | 3D IMU (OPAL, Apdm) | 1 | 128 Hz | ± 6 g; ± 2000 deg/s | LP 4th order Butter. |
| 10 | Grimpampi 2015 [46] | To assess the reliability of gait variability assessment in healthy older individuals based on lower trunk accelerations | Validation | 29 older adults (84 ± 5, 5/24) | / | 3D ACC and 3D GYRO (Freesense, Sensorize) | 1 | / | / | / |
| 11 | Brooks 2015 [47] | To develop and validate a self-administered 6MWT mobile application. | Validation | 103 CHF and pHTN | / | 3D ACC (iPhone 4s, Apple) | 1 | / | / | / |
| 12 | Christiansen 2015 [48] | To examine movement symmetry changes over the first 26 weeks following unilateral TKA | Pilot | 24 unilateral TKA (65.2 ± 9.2)19 controls (61.3 ± 9.2) |
/ | 3D ACC (Delsys) | 1 | 1000 Hz | ± 10 g | LP 4th order Butter. (40 Hz) |
| 13 | Juen 2015 [49] | To evaluate six machine learning methods to obtain gait speed during natural walking | Pilot | 28 pulmonary disease (range 50–89, 12/16) 10 controls (age range 18–69, 3/7) |
/ | 3D ACC (S5 and Galaxy Ace, Samsung) | 1 | 60 Hz | / | / |
| 14 | Juen 2014 [50] | To monitor health status using smartphones | Pilot | 30 COPD (53 ± 11, 3/27) | GOLD 1–2 | 3D ACC (Galaxy Ace, Samsung) | 1 | 60 Hz | / | / |
| 15 | Annegarn 2012 [51] | To determine walking patterns during the 6MWT of COPD patients and healthy elderly subjects | Cohort | 79 COPD (64.3 ± 8.9, 47/32) 24 controls (63.7 ± 5.9, 15/9) |
GOLD 1–2-3–4 | 3D ACC (Minimod, McRoberts) | 1 | 100 Hz | ± 2 g | LP 4th order Butter. (20 Hz) |
| 16 | Beausoleil 2019 [52] | To quantify the evolution of gait parameters along a 6MWT in LLA population | Pilot | 15 LLA (59 ± 12, 10/5) | / | 3D ACC and 3D GYRO (Physilog 4, GaitUp) | 2 | 200 Hz | ± 3 g; ± 600°/s | / |
| 17 | Galán-Mercant 2019 [53] | To predict physical activity and functional fitness using deep learning | Pilot | 17 older adults (83.26 ± 6.56, 3/14) | / | 3D ACC (iPhone 4, Apple) | 1 | 32 Hz | / | LP 5th order Butter. (16 Hz) |
| 18 | Ameli 2019 [54] | To objectively assess the effects of chemotherapy-induced fatigue on gait characteristics | Pilot | 4 breast cancer (50 ± 2.5) | / | 3D IMU (MTx, Xsens) | 6 | 60 Hz | / | / |
| 19 | Jimenez-Moreno 2018 [55] | To compare accelerometry data between a DM1 cohort and healthy controls | Validation | 30 MD1 (48, range 25–72, 20/10) 14 controls (32, range 23–47, 6/8) |
/ | 3D ACC (GENEActiv, Activinsights) | 4 | / | / | / |
| 20 | Dandu 2018 [56] | To explore the physiological and clinical meaning of four objective measures of walking impairment. | Pilot | 115 MS | Mild (EDSS 0–2.5), moderate (3.0–4.0) and severe ( > 4.0) | 3D ACC (GT3X, ActiGraph)3D ACC and 3D GYRO (in-house) | 6 | 30 Hz, 128 Hz | / | BP (1–3 Hz) |
| 21 | Cheng 2017 [57] | To validate a model for the prediction of pulmonary function, based on motion sensor data from mobile phones | Validation | 25 COPD (76, range 55–95, 15/10) | GOLD 1–2-3 | 3D ACC (Galaxy S5, Samsung and Optimus Zone2, LG) | 2 | / | / | / |
| 22 | Ameli 2017 [58] | To study the effects of fatigue induced by chemotherapy on PPS of cancer patients | Pilot | 4 cancer patients | / | 3D IMU (MTx, Xsens) | 17 | 60 Hz | / | / |
| 23 | Gong 2016 [59] | To propose a causality analysis method that may aid disease diagnosis | Pilot | 28 MS (40.5 ± 9.4, 7/21) 13 controls (39.3 ± 10.3, 6/7) |
Mild (EDSS 0–2.5) and moderate (3.0–4.0) | 3D ACC and 3D GYRO (in-house) | 5 | 128 Hz | ± 16 g; ± 2000°/s | / |
| 24 | Riva 2014 [60] | To evaluate the influence of directional changes and SF on gait variability and stability measures | Validation | 51 healthy (23 ± 3) | / | 3D ACC and 3D GYRO (FreeSense, Sensorize) | 1 | 100 Hz - 200 Hz | / | Signal used unfiltered |
| 25 | Waugh 2019 [61] | To propose an individualized model of gait | Validation | 92 older adults (86 ± 5, 33/53) | / | 3D ACC (X6-2, X6-2mini, X8m-3, X16-2, Gulf Coast Data Concepts) | 3 | 40 Hz - 50 Hz | ± 2, 8, or 16 g | LP 4th order Butter. (10 Hz) |
| 26 | Engelhard 2016 [62] | To discover and validate objective evidence of gait alteration using dynamic time warping | Pilot | 96 MS (46, range 19–61, 13/73) 29 controls (40, range 19–54, 9/20) |
Mild (EDSS 0–2.5), moderate (3.0–4.5) and severe (5.0–6.5) | 3D ACC (ActiGraph GT3X) | 1 | 30 Hz | / | / |
| 27 | Howcroft 2017 [63] | To identify the optimal wearable sensor type, location, and combination for prospective fall-risk prediction | Pilot | 76 older adults (75.2 ± 6.6, 31/44) | Fallers and non fallers | 3D ACC (X16-1C, Gulf Coast Data Concepts) | 4 | 50 Hz | / | LP 5th order Butter. (12.5 Hz) |
| 28 | Cheng 2016 [64] | To propose a gait model to predict saturation categories | Validation | 20 COPD (66.3, range 43–81, 9/11) | GOLD 1–2 | 3D IMU (Droid 4 Mini, Motorola) | 1 | 60 Hz | / | / |
SF: sampling frequency; MS: multiple sclerosis; EDSS: expanded disability status scale; BMI: body mass index; SIS: stroke impact scale; sLSS: symptomatic lumbar spinal stenosis; ODI: Oswestry Disability Index; CHF: congestive heart failure; pHTN: pulmonary hypertension; TKA: total knee arthroplasty; COPD: chronic obstructive pulmonary disease; LLA: lower limb amputee; MD1: myotonic dystrophy type 1; GOLD: Global Initiative for Obstructive Lung Disease Criteria; ACC: accelerometer; GYRO: rate gyroscope; IMU: inertial measurement unit; LP: low-pass; BP: band-pass; butter: Butterworth filter.