Table 1.
Summary of the included studies.
| First author + YOP | Journal | n. of patients | n. of controls | Longitudinal | Type of sensor | Wearing position | Duration of use | Location of monitoring | Measured disease characteristics | Main results |
|---|---|---|---|---|---|---|---|---|---|---|
| Myers 1979 [36] | Biol Psychiatry | 10 mHD, 15 at risk HD | 0 | no | Accelerometer | Not specified | Not specified | Clinic | Tremor | Accelerometer measures can detect and characterize tremor in manifest and pre-manifest HD |
| Folstein 1983 [35] | Neurobehav Toxicology and teratology | 17 mHD, 27 at risk HD | 10 | no | Three-axial piezoelectric accelerometer (Wilcoxon Model no.139) | Dorsal surface of subject's hands | 4 tasks, 5 10-s trials for each task | Clinic | Involuntary movements; some voluntary movements (simple reaction time, finger tapping, movement time) | Motor abnormalities can be detected in manifest and at risk HD; screening of motor abnormalities in the population |
| van Vugt 1996 [5] | Movement Disorders | 14 mHD | 14 | no | Wrist-worn activity monitor (accelerometer) (Gaehwiler Electronic, Switzerland) | Non-dominant wrist | 5 successive days and nights | Home | General daytime motor activity | Higher hypokinesia in HD patients |
| van Vugt 2001 [46] | Movement Disorders | 64 mHD | 67 | yes | Wrist-worn activity monitor (accelerometer) (Gaehwiler Electronic, Switzerland) | Non-dominant wrist | 5 successive days and nights | Home | General daytime motor activity | Higher hypokinesia in HD patients; correlation with impaired voluntary movements, disturbed posture and gait, and reduced functional capacity; progresses with functional disability |
| Hurelbrink 2005 [45] | J Neurol | 8 mHD | 8 | no | Actiwatch-Neurologica (Cambridge Neurotechnology) | Preferred wrist | 48 h | Home | Day- and night-time involuntary movements; Sleep-wake activity | Greater activity levels in HD while awake and during sleep; HD sleep longer than controls |
| Grimbergen 2008 [49] | Movement Disorders | 45 mHD | 27 | no | Digitally-based angular velocity transducer (SwayStar) | Lower back | Time to walk on the GaitRite carpet) | Clinic | Trunk movements | Trunk displacement significantly greater in patients than controls; increased trunk sway in fallers compared to non-fallers; clinical chorea scores positive correlated to the range of angular trunk motion |
| Khalil 2010 [29] | JNNP 2010-EHDN suppl | 10 mHD 5 pHD | 6 | no | AD_BRC sensor with a three-axial accelerometer | Sternum | Time of TUG performance | Clinic | Performance of Timed Up and Go Test | Accelerometer objective measures can be useful to catch disease specific features and so to differentiate between groups |
| Dalton 2013 [38] | Gait and Posture | 14 mHD 10 pHD | 10 | no | AD_BRC sensor with a three-axial accelerometer | Chest | Unspecified (duration of the examination in clinic) | Clinic | Balance; gait | An accelerometer based sensor may be an effective means of differentiating between premanifest and manifest Huntington's disease subjects |
| Rudzinska 2013 [43] | Neurologia I Neurochirurgia Polska | 43 DA 28 mHD 23 tic disorders |
51 | no | Three-axial accelerometer (BIOPAC) | Proximal phalanx of the third finger | 1.5 min (accelerometer registration) | Clinic | Tremor | Postural and essential type tremor found in 10% of HD; prevalence of tremor is considerably higher among patients with degenerative ataxias compared with HD, tic disorder and the control group. The most common type of tremor accompanying ataxias, HD and tic disorders is essential tremor type |
| Norberg 2013 [26] | AFMR 2013 CA | 15 PD or mHD | 0 | no | Wireless three-axial accelerometers (UCLAWireless Health Institute) | Both ankles | 4 50-foot timed training walks + 3 days of monitoring | Clinic and home | Gait | Wireless sensors can obtain multiple measures of gait and other physical activities in an inexpensive and unobtrusive manner |
| Trojaniello 2014 [33] | IEEE Conference 2014 | 10 mHD | 10 | no | MIMU (Opal, APDM, Inc) | Ankle | 1 min walking | Clinic | Gait | The MIMU has about 30% of errors associated to the best estimates of gait direction changes for patients, compared to gold standard (GAITRite Math) |
| Collett 2014 [37] | Gait & Posture | 7 pHD 28 mHD | 22 | no | IMU (Pi-node Philips, Netherlands) | Taped over the fourth lumbar vertebra | 8.8 or 10 m walking | Clinic | Gait | More variability in gait parameters in mHD compared to controls; no differences between pHD and HC, except for 1 parameter of the phase plot analysis, which also correlated with UHDRS-TMS and DBS. Phase plot analysis as a sensitive method to detect gait changes in HD |
| Trojaniello 2015 [42] | Gait & Posture | 10 stroke 10 PD 10 mHD |
10 | no | MIMU (Opal, APDM, Inc) | Over the subject lumbar spine, between L4 and S2 | 1 min walking | Clinic | Gait | Comparison of 3 different methods to detect gait events. None of the tested methods outperformed the others in terms of gait parameter determination accuracy. Missed or extra gait events were found for all methods where pathological populations were analysed |
| Hogarth 2015 [28] | ICPDMD 2015 | 5 mHD | 5 | no | Shoe-worn inertial sensor (APDM Inc) | Both shoes | walking hours for 7 days | Home | Gait | Gait parameters correctly identified subjects. Significant differences between HD and HC in gait parameters |
| Townhill 2016 [40] | J Neurosci Meth | 9 mHD 4 pHD |
9 | no | Actiwatch-Neurologica (Cambridge Neurotechnology) + ambulatory EEG | Non-dominant wrist | 24 h (EEG); 7 days continuously (Actiwatch) | Home | Circadian Rhythm | Actiwatch is not a reliable tool for measuring awake/sleep periods in patients with movement disorders; no differences in circadian rhythmicity between groups |
| Andrzejewski 2016 [48] | J of HD | 15 mHD | 4 | no | Accelerometer-based wearable PAMSys-X (BioSensics, Cambridge, MA) | Both ankles, both wrists, and chest | 7 days | Clinic and home | General daily motor activity; gait | Same level of physical activity; differences in gait measures between HD and controls; feasible use of wearable sensors |
| Mannini 2016 [44] | Sensors | 17 mHD 15 post-stroke |
10 | no | MIMU (Opal, APDM, Inc) | Both ankles, and over the subject's lumbar spine between L4 and S2 | Unspecified (duration of the examination in clinic) | Clinic | Gait | Propose and validation of a new machine learning framework for gait classification (normal vs pathological) |
| Dinesh 2016 [25] | IEEE Xplore Digital Library | 16 PD 10 mHD |
15 | no | Accelerometer-based BioStampRC wearable sensors, MC10 Inc (Lexington, MA) | Both anterior thighs, both proximal anterior forearms, and medial chest | 2 days | Clinic and home | Gait | Signal analysis of light-weight body-affixed sensors can detect motor symptoms associated with PD and HD |
| Bennassar 2016 [27] | Procedia Computer Science (20th International Conference on Knowledge Based and Intelligent Information andEngineering Systems) | 15 mHD | 7 | no | GENEActiv three-axial accelerometer (Activinsights Ltd, Cambridgeshire, UK) | Both wrists, and chest | Few minutes (time of completing the Moneybox-Test tasks) | Clinic | Movements of the upper limbs during the execution of the Money Box Test | Introduction of a new approach to automatically classify HD and controls (upper-limb movements) |
| Kegelmeyer 2017 [50] | J Neurol Sci | 41 mHD | 36 | no | iPod with the Level Belt Pro software installed | Back at the level of L5 and of the lower border of scapulae | Unspecified (duration of the examination in clinic) | Clinic | Trunk control | Significant greater amplitude of thoracic and pelvic movements in HD vs controls (++ in static than in dynamic tasks) |
| Maskevich 2017 [39] | J of HD | 4 pHD 3 mHD |
0 | no | Actiwatch Spectrum Pro (Philips/Respironics), Fitbit One and Jawbone UP2 | Non-dominant wrist | Overnight | Clinic | Sleep characteristics | Three monitors less accurate of polysonnography to estimate sleep parameters in HD. Can't be a good replacement, but sufficient for overall estimations of sleep-wake patterns, and/or to assess gross level changes over time |
| Adams 2017 [24] | Digit Biomark | 15 mHD 5 pHD 16 PD | 20 | no | Accelerometer-based BioStampRC wearable sensors, MC10 Inc (Lexington, MA) | Both anterior thighs, both proximal anterior forearms, and medial chest | 2 days | Clinic and home | General daytime motor activity | Patients with HD spent more time lying down; participants happy with the sensors |
| Saadeh 2017 [32] | IEEE Conferences 2017 | 13 ALS, 20 mHD, 15 PD | 16 | no | Flexi-force sensing resistor (A201 Tekscan) | Shoe sole | Unspecified (used of an existing database?) | Clinic | Gait | The system classified the different groups with high sensitivity and specificity and a high classification accuracy |
| Youdan 2018 [30] | HSG 2018 | 37 mHD | 15 | no | MIMU (Opal, APDM, Inc) | Medial chest, medial lower back, both ankles and both wrists | Time of task performing | Clinic | Gait; cognition | Dual-task impairment in HD compared to HC, as showed by increased total sway area, decreased gait speed and decreased correct response to cognitive tasks |
| Waddel 2018 [34] | HSG 2018 | 14 subjects | ? | yes | Android smartphone app (GEORGE) | Smartphone | 1 month | Clinic and home | Gait, involuntary movements, voice, balance, dexterity, mobility, socialization | Feasibility of the app |
| Lipsmeier 2018 [31] | JNNP 2018-EHDN suppl | 46 mHD | 0 | yes | Smartphone and Smartwatch (ROCHE platform) | Preferred wrist (smartwatch) and belt or trouser pocket (smartphone) | 8- week preliminary results | Home | General daytime motor activity; motor tasks; chorea; balance; cognition; mood; quality of life | Good adherence; feasibility |
| Lauraitis 2018 [47] | IEEE j of Biomedical and Health Informatics | 11 mHD | 11 | no | Android tablet app | Tablet | Once or twice a week for an unspecified period | Home | Motor and cognitive abilities trough three tasks | High classification accuracy of the app and useful support for automated medical examination |
| Acosta-Escalante 2018 [52] | IEEE Special edition on trends, perspectives and prospects of machine learning applied to biomedical systems in internet of medical things | 7 mHD | 7 | no | Movement sensors on two smartphones iPhone 5S | Both ankles | Walking on a 20-m math during visits of 7 consecutive days | Clinic | Gait | Meta-classifier algorithms useful for improving accuracy in classification and reducing the number of sensor devices needed. Best performance of Logitboost & RandomForest combination |
| Bennasar 2018 [51] | IEEE transactions on neural systems and rehabilitation engineering | 44 mHD | 48 | no | Three-axial accelerometer GENEactiv | Both wrists, and chest | Few minutes (time of completing the Moneybox-Test tasks) | Clinic | Movements of the upper limbs during the execution of the Money Box Test | Presentation of a system for an objective and continuous assessment of motor impairment during a novel upper limb task for HD patients |
| Bartlett 2019 [41] | Neurobiol of Sleep and Circadian Rhythms | 32 pHD | 29 | no | Wrist-worn actigraphy GT3X + ActiGraph monitor | Non-dominant wrist | 7 nights | Home | Circadian rhythm and habitual sleep characteristics | Decreased habitual sleep efficiency and increased awakenings in pHD compared with HC. No association between hypothalamic volume and circadian rhythm or habitual sleep outcomes in pre-HD |
YOP: year of publication; HD: Huntington's disease; mHD: manifest Huntington's disease; pHD: pre-manifest Huntington's disease; PD: Parkinson's disease; DA: degenerative ataxia; ALS: amyotrophic lateral sclerosis; HC: healthy controls; IMU: inertial measurement unit; MIMU: magnetic inertial measurement unit; UHDRS-TMS: unified Huntington's disease rating scale – total motor score; DBS: disease burden score.