Table 4.
Study characteristics for studies that apply ML techniques.
| Ref. | Year | Pathology | Technology | Study Goals |
|---|---|---|---|---|
| [95] | 2021 | Ataxia | Opal (APDM Wearable Technologies, Portland, OR, USA) | ataxia severity assessment |
| [96] | 2020 | Stroke | XSens MTw Awinda (XSens Technologies, Enschede, The Netherlands) | upper-limb impairment estimation |
| [97] | 2022 | Stroke (hemiparesis) | XSens MTw Awinda (XSens Technologies, Enschede, The Netherlands) | impairment severity estimation |
| [98] | 2021 | PD | Opal (APDM Wearable Technologies, Portland, OR, USA) | comparison between upper- and lower-limb inertial data to diagnose PD and determine the stage of the disease |
| [99] | 2019 | PD | Custom fingers/wrist-worn composed by Cortex-M3 CPU (STMicroelectronics, Geneva, Switzerland) and LSM9DS (STMicroelectronics, Geneva, Switzerland) | PD stage quantification |
| [100] | 2022 | PD | BWT901CL (Wit-MotionCompany (Wit-MotionCompany, ShenZhen, China) | PD diagnosis |
| [101] | 2023 | PD | Custom fingers/wrist-worn composed by MPU9250 (TDK InvenSense, Tokyo, Japan) and nRF52832 (Nordic Semiconductor, Trondheim, Norway) | Distinguish Parkinson’s patients |
| [102] | 2022 | Not specified | Shimmer3 (Shimmer, Dublin, Ireland) | Classification of the load exerted on the shoulders |
| [103] | 2023 | Not specified | APDM Wearable model not specified (APDM Inc., Portland, OR, USA) | Predict pain |
| [104] | 2023 | Not specified | STEVAL-STLKT01V1 (ST Microelectronics, Geneva, Switzerland) | Joint prediction |
| [105] | 2022 | |||
| [106] | 2021 | Not specified | STEVAL-STLKT01V1 (ST Microelectronics, Geneva, Switzerland) | Detect elbow movement intention |
| [107] | 2023 | Stroke | STEVAL-STLKT01V1 (ST Microelectronics, Geneva, Switzerland) | Motion intention detection |
| [108] | 2019 | Stroke | Shimmer model not specified (Shimmer, Dublin, Ireland) | Monitor physiotherapeutic evolution |
| [109] | 2020 | Stroke | Shimmer2 (Shimmer Sensing, Dublin, Ireland) | Clinical scores estimation |
| [110] | 2020 | Stroke | Smartwatch W270 (LG, Seoul, South Korea) | Identify and log the variety and frequency of rehabilitation exercises performed by the individual |
| [111] | 2021 | Stroke | Actigraph GT3X-BT or GT9X-Link (ActiGraph, Pensacola, FL, USA) | Categorization of upper-limb performance |
| [112] | 2019 | Stroke | Finger and wrist-worn (Arcus, ArcSecond Inc., San Diego, CA USA) | Hand use during ADLs estimation |
| [113] | 2022 | Not specified | Wrist-worn [64] | Comparison of different measures for assessing upper limb |
| [114] | 2022 | Stroke (hemiparesis) and Traumatic brain injury | Shimmer2r (Shimmer Sensing, Dublin, Ireland) | Estimation of clinical scores reflecting upper-limb movement quality |
| [115] | 2022 | Stroke | ZurichMOVE (ZurichMOVE, Zurich, Switzerland) | Estimation task scores |
| [116] | 2021 | Stroke (hemiparesis) and traumatic brain injury | Shimmer2 (Shimmer Sensing, Dublin, Ireland) | Predict and monitor the recovery process and assess the responsiveness to treatment on an individual basis |
| [117] | 2019 | Stroke (hemiplegia) | wrist-worn (not specified) | Motion recognition |
| [118] | 2021 | Stroke | Custom Arm band composed by MPU9250 (TDK InvenSense, Tokyo, Japan), Arduino Nano V3 (Arduino Srl, Turin, Italy) and Wireless Transmitter HC-06 | Motion recognition of upper-limb exercises and improvement of recognition performance |
| [119] | 2019 | Not specified | Smartwatch W270 (LG, Seoul, South Korea) | Optimal size estimation of the time window for classifying real-time motions |