Table 1 Results table of all 17 studies
Author (year) | Novel Technology | Type of study (multicentric/monocentric) | Population | Participants, n | Sex, M% | Age (y), mean (SD) | Reference technology | Features to characterize RBD | Sensitivity, specificity and diagnostic accuracy (%) | Other results |
|---|---|---|---|---|---|---|---|---|---|---|
Novel Tools | ||||||||||
González et al.26 | Dreem 2 EEG Headband | Multicentric | PD RBD+ | 6 | 70% | 69.5, 7.6 | vPSG | WASO | Sensitivity and specificity: N/A (small sample size) Sleep staging accuracy 80.8% | PD duration and RBD symptoms were associated with more WASO |
PD RBD- | 4 | |||||||||
Cesari et al.25 | Automatic 3D video analysis | Monocentric | RBD | 53 | 84.9% | 65.5, 9.5 | vPSG | Movements | Sensitivity: 83.6% Specificity: 90.2% Diagnostic accuracy 86.6% | N/A |
Non-RBD | 128 | 61.7% | 54.3, 13.4 | |||||||
Oz et al.29 | Soft electrode array | Monocentric | HC | 21 | 61.9% | 56.6, 8.4 | vPSG | RSWA | Sensitivity 85.7% (detection of RSWA) Specificity 58.3% | k = 0.688 k (REM) = 0.723 |
PD | 29 | 65.5% | 65.4, 7.6 | |||||||
Levendowski et al. (A)28 | Sleep ProfilerTM | Multicentric | iRBD | 24 | 79.2% | 63, 12.4 | vPSG | RSWA | Specificity 0.88 chin, 0.93 arm, 0.90 chin or arm Sensitivities 0.81,0.81 and 0.86 | k = 0.68 (chin), k = 0.74 (arms) |
HC | 21 | 34% | 63, 11.7 | |||||||
RBD | 42 | 81% | 61, 17.4 | |||||||
Kataoka et al.27 | Portable two-channel EEG/EOG recording system | Monocentric | PD | 8 | 62.5% | 65.6 + −9.42 | vPSG | RSWA | Calculated sensitivity 75%* ** | Accordance PRS and vPSG in 6 patients (p = 0,686) k = 0.70 (RSWA) |
Actigraphic Tools | ||||||||||
Louter et al.21 | Actiwatch system (AW4, Cambridge Neurotechnology Ltd., Cambridgeshire, United Kingdom) | Monocentric | PD RBD+ | 23 | 66.7% | 61.3, 9.1 | vPSG | Wake bouts | Sensitivity 20.1% Specificity 95.5% | N/A |
PD RBD- | 22 | |||||||||
Raschellà et al.23 | GENEActiv Original wrist actigraph with machine learning classification algorithms | Multicentric | PD RBD+ | 18 | 83.3% | 69.9 ± 8.2 | vPSG | Individual movement episodes and global nocturnal activity | In lab: Sensitivity 94.9 ± 7.4% Specificity 92.7 ± 13.8% Diagnostic accuracy 92.9 ± 8.16% In home recordings: Diagnostic accuracy - in PD: 100% - non-PD: 94.4% | N/A |
PD RBD- | 8 | 50% | 63.8 ± 13.9 | |||||||
Non-PD | 18 | 66.7% | 52.7 ± 15.3 | |||||||
Brink-Kjaer et al. (B)17 | Actigraphy | Multicentric | iRBD | 42 | 78.6% | 67.3, 6.69 | N/A | Abnormal nighttime activity and 24-hour rhythm disruption | Sensitivity 78.9% Specificity 96.4% | N/A |
HC | 42 | 69% | 64.3, 10.3 | |||||||
HC UKBB | 110 | 73.7% | 66.0, 7.0 | |||||||
Brink-Kjaer et al. (A)16 | AX-6 (Axivity, Ltd, Newcastle, UK) Actigraphy and a 9-item questionnaire | Monocentric | iRBD | 42 | 72.2% | 67.3 ± 6.69 | vPSG | Activity count and acceleration | Actigraphy: Sensitivity 95.2% Specificity 90.5% Actigraphy and questionnaire: Sensitivity 88.1% Specificity 100% | N/A |
Sleep clinic patients | 21 | 63.1 ± 11.8 | ||||||||
HC | 21 | 65.5 ± 8.6 | ||||||||
Filardi et al.18 | MicroMini-Motionlogger Watch Actigraphy | Monocentric | HC | 16 | 56.2% | 43.63 + −15.66 | vPSG | i > o index (98.32) | Sensitivity 63.2% Specificity 89.1% | N/A |
RBD | 19 | 84.2% | 71.68 + −7.85 | |||||||
SAS | 19 | 94.7% | 50.53 + −11.29 | |||||||
RLS | 20 | 50% | 47.50 + −14.19 | |||||||
Ko et al.19 | ASUS VivoWatch BP (HC-A04) (ASUSTeK Computer Inc. Taipei, Taiwan) | Monocentric | PD | 20 | 51.9% | 62.3 + −9.51 | N/A | REM stage detection (K-means clustering algorithm | N/A (no reference technology)** | Abnormal REM % p: 0.007 |
HC | 18 | 50% | 61.7 + −9.2 | |||||||
Stefani et al.24 | MicroMini-Motionlogger Actigraphy | Monocentric | iRBD | 20 | 74% | 54 (range 22–84) IQR 40–68 | vPSG | Activity, through a piezoelectric tri-axial accelerometer | Sensitivity 85%-95% Specificity79%-91% Diagnostic accuracy 81%-91% | N/A |
RLS | 20 | |||||||||
RLS + SA | 10 | |||||||||
SA | 20 | |||||||||
HC | 20 | |||||||||
Sandala et al.20 | MotionWatch, CamNtech Itd. with RBDSQ | Monocentric | RBD | 45 | 86.9% | 66.8, 9.7 | vPSG | Short immobile bouts (SIB) with 31% cut-off | Actigraphy: Sensitivity: 86.6 Specificity: 66.0 Actigraphy and questionnaire: Sensitivity 100% Specificity 83.3% Diagnostic accuracy 87.6% | N/A |
SRMD | 30 | 83.3% | 51.5, 12 | |||||||
HC | 20 | 90% | 70.4, 6.3 | |||||||
Naismith et al.22 | Actiwatch Firmware Version 01.01.0007 (Minimitter – Respironics Inc., Bend, Or) and questionnaires | Monocentric | iPD RBD- | 9 | 50% | 63.4, 7.5 | N/A | Wake bouts | N/A (no reference technology)** | Actigraphy and questionnaire: p = 0,011 RBD+ & RBD- |
iPD RBD+ | 13 | |||||||||
Novel Modalities | ||||||||||
Levendowski et al. (B)31 | Machine learning algorithm on Sleep Profiler | Multicentric | AD | 37 | 73% | 73.0,7.8 | vPSG | RSWA | Sleep biomarkers for NDD conditions 1 Supine sleep: - Sensitivity 63% - Specificity 59% 2 Sleep efficiency - Sensitivity 41% - Specificity 89% | 27% iRBD misclassified as normal sleep |
LBD | 18 | 89% | 70.0, 6.2 | |||||||
iRBD | 33 | 73% | 66.0, 10.0 | |||||||
PD | 29 | 76% | 65.0, 9.6 | |||||||
MCI | 78 | 55% | 70.0, 9.9 | |||||||
HC | 186 | 54% | 62.0, 9.5 | |||||||
Abdelfattah et al.30 | Optical flow computer vision algorithm | Monocentric | iRBD | 81 | 65% | 65.0, 8,6 | vPSG | Movements (features of rate, ratio, magnitude, velocity of movement and ratio of immobility) | Sensitivity 84.0% Specificity 99.0% Diagnostic accuracy 91.9% | N/A |
HC | 91 | 67% | 64.1, 9,3 | |||||||
Possti et al.32 | Wearable system/mask with automated algorithm for quantifying RSWA | Monocentric | HC | 24 | 45.5% among all groups | 66 years among all groups, | vPSG | RSWA | Sensitivity 83% Specificity 79% Balanced accuracy 81% | RSWA showed 60% correlation between home and lab |
PD RBD- | 13 | |||||||||
PD RBD+ | 15 | |||||||||
iRBD | 3 | |||||||||
