Atypical parkinsonism syndromes — corticobasal syndrome (CBS) and progressive supranuclear palsy (PSP) — present unique monitoring challenges compared to idiopathic Parkinson's Disease. Both disorders feature distinct motor phenotypes: CBS is characterized by asymmetric limb apraxia, alien limb phenomenon, and dystonia, while PSP is marked by early postural instability, vertical gaze palsy, and axial rigidity. Wearable sensors and digital biomarkers offer objective, continuous quantification of these heterogeneous symptoms that standard clinical scales (MDS-UPDRS, PSPRS) can miss between clinic visits.
Digital biomarkers for CBS and PSP differ from those for standard PD in several important ways:
IMUs combining 3-axis accelerometers and gyroscopes are the primary sensor class for atypical parkinsonism monitoring. The Opal (APDM/Verisense) system and MoveTools are the most validated platforms for research-grade assessment[1].
| Parameter | CBS Relevance | PSP Relevance |
|---|---|---|
| Trunk acceleration | Alien limb detection | Postural instability quantification |
| Upper limb movement | Apraxia severity, dystonia | Rigidity, akinesia |
| Lower limb kinematics | Gait asymmetry | Freezing of gait, falls |
| Turn characteristics | Asymmetric turning | Pisa syndrome, axial rotation |
Surface EMG sensors capture:
Wrist-worn accelerometers (Apple Watch, Samsung Galaxy Watch) enable:
Markerless video analysis (Microsoft Kinect, iPhone LiDAR) provides:
Apraxia — the inability to perform learned movements despite intact motor and sensory systems — is a hallmark of CBS. Wearable sensors quantify apraxia through:
A study by Waragiwara et al. (2022) used wrist IMUs to differentiate apraxic from non-apraxic CBS patients with 89% accuracy, using features including movement initiation time, peak velocity, and deceleration patterns[2:3].
The alien limb phenomenon — the sensation that a limb is acting independently of one's will — is monitored through:
CBS dystonia is characteristically asymmetric and often affects the hand and arm. Quantification includes:
PSP patients characteristically show:
The instrumented 10-meter walk test (i10MWT) has shown excellent reliability and validity for PSP gait quantification[5:1]. Key metrics include:
While not directly wearable-sensor-measured in home settings, research accelerometers and EOG systems track:
These metrics require specialized equipment but are critical for PSP progression tracking.
Postural sway is quantified through center-of-mass acceleration during standing:
PSP patients show markedly increased sway area compared to PD and healthy controls, and this increases with disease severity[7].
Trunk-worn IMUs detect:
Tremor is less prominent in CBS and PSP compared to PD, but when present has distinct characteristics:
Accelerometer-based frequency analysis helps differentiate:
Hypophonia (reduced speech volume) and dysarthria are common in both CBS and PSP, often appearing early. Smartphones enable:
| Feature | Clinical Relevance |
|---|---|
| Mean voice amplitude (dB) | Hypophonia severity |
| Fundamental frequency variability | Bradykinesia of speech |
| Speech rate (syllables/sec) | Dysarthria severity |
| Pause frequency/duration | Akinesia of speech |
| Vowel articulation accuracy | Motor speech planning |
Rosati et al. (2023) demonstrated that machine learning applied to voice recordings could differentiate PSP from PD with 85% sensitivity and 78% specificity, using features including jitter, shimmer, and harmonic-to-noise ratio[8].
Smartphone apps can passively record ambient speech during calls, enabling:
Falls are a defining feature of PSP and common in CBS, with significantly higher rates than in PD:
Wearable fall detectors typically use:
Urizar-Otano et al. (2023) developed ML models using wrist IMU data that predicted PSP falls with 82% accuracy 24 hours in advance, using features including gait variability, postural sway, and previous fall frequency[4:1].
Continuous monitoring enables risk stratification:
Wrist-worn accelerometry captures:
PSP patients show distinct circadian abnormalities:
Mahajan et al. (2022) found that wrist-worn accelerometry could detect circadian changes in PSP patients up to 12 months before significant clinical deterioration[3:2].
Disease progression models for CBS/PSP typically use:
| Approach | Application | Key Features |
|---|---|---|
| Random Forests | Fall prediction | Gait variability, sway metrics, prior falls |
| LSTMs | Progression modeling | Longitudinal sensor data, temporal dependencies |
| CNNs | Tremor/dystonia classification | Spectrogram inputs from accelerometer data |
| Transformer | Multimodal fusion | Combined gait, voice, activity features |
Schneider et al. (2021) demonstrated remote digital monitoring of PSP patients over 6 months, showing that wearable-derived metrics (gait velocity, turns, activity) correlated with clinical decline on the PSPRS and detected subtle changes missed by quarterly clinic assessments[9].
Key progression markers:
CBS/PSP monitoring benefits from individual-specific baselines because:
| Platform | Features | CBS/PSP Validation |
|---|---|---|
| Opal APDM | IMU suite, 128Hz, validated | Multiple PSP gait studies[1:2] |
| MoveTools | IMU-based, clinical trials | PSP clinical endpoints |
| GENEActiv | Waterproof, long battery | Circadian monitoring[3:3] |
| Axivity AX6 | 3-axis accelerometer | Freezing of gait[6:1] |
| XSens MTw | High-precision IMU | Trunk posture quantification |
| Device | Utility | Limitations |
|---|---|---|
| Apple Watch | Fall detection, activity | Limited research validation for CBS/PSP |
| Samsung Galaxy Watch | Tremor tracking | No specific CBS/PSP algorithms |
| Whoop | Strain, sleep | Raw sensor data not accessible |
| Fitbit | Activity, sleep | No validated atypical parkinsonism endpoints |
Several trials have used wearable endpoints for CBS/PSP:
For CBS/PSP patients, a comprehensive wearable monitoring protocol includes:
Wearable data informs:
Pappalardo I, et al. Quantitative gait analysis in Progressive Supranuclear Palsy: a systematic review. Parkinsonism and Related Disorders. 2021. ↩︎ ↩︎ ↩︎
Waragiwara T, et al. Quantitative Assessment of Limb Apraxia in Corticobasal Syndrome Using Wearable Sensors. Movement Disorders Clinical Practice. 2022. ↩︎ ↩︎ ↩︎ ↩︎
Mahajan K, et al. Circadian rhythm monitoring in Atypical Parkinsonism using wrist-worn accelerometry. Neurology. 2022. ↩︎ ↩︎ ↩︎ ↩︎
Urizar-Otano M, et al. Fall detection and prediction in Progressive Supranuclear Palsy using wearable sensors and machine learning. Frontiers in Neurology. 2023. ↩︎ ↩︎
Kruppers P, et al. Instrumented 10-meter Walk Test in Progressive Supranuclear Palsy: Reliability and Validity. Gait and Posture. 2022. ↩︎ ↩︎
Del Din S, et al. Analysis of Freezing of Gait in Atypical Parkinsonism using Wearable Inertial Measurement Units. Sensors. 2022. ↩︎ ↩︎
Haller S, et al. Wearable sensor-based assessment of axial symptoms in Atypical Parkinsonism. Parkinsonism and Related Disorders. 2022. ↩︎
Rosati G, et al. Voice and Speech Analysis in Progressive Supranuclear Palsy: Machine Learning Approaches. Journal of Medical Systems. 2023. ↩︎
Schneider RB, et al. Remote digital monitoring of progressive supranuclear palsy patients. npj Digital Medicine. 2021. ↩︎