Static Balance

Phones only. The participant puts the phone in a belt bag or running belt at the front of the waist (or, if chosen, a trouser pocket), taps Start, and gets 10 seconds to get into position. The phone vibrates, the participant stands still without support, feet apart, eyes open, arms relaxed at the sides, for 30 seconds, and the phone vibrates again. The phone's acceleration in the horizontal plane gives the sway. The participant is asked to stand next to a wall or a sturdy chair. The web preview shows the screens only.

Useful as a frequent, self-administered balance marker in ageing, multiple sclerosis, Parkinson's disease and intervention studies, where change over repeated tests matters more than a single score. The Floodlight and Roche studies aggregated daily tests over 2 weeks.

Not a clinical balance or fall-risk assessment and not a substitute for a force plate: sway path did not correlate with the Berg Balance Scale in people with MS (r = −.20; Montalban et al., 2022), and no comparison of the phone test with a force plate has been published. Not for participants who cannot stand safely without support. Beta: being checked on iOS and Android devices.

The defaults follow the smartphone Static Balance Test of the Floodlight MS and Roche PD apps: one 30 s test, standing still and unsupported, feet apart, eyes open, arms relaxed at the sides, the phone in a running belt at the front of the waist, with a vibration at the start and end (Montalban et al., 2022; Lipsmeier et al., 2018; Rinderknecht et al., 2023). Sway path follows Floodlight's definition (Montalban et al., 2022, Table S2); jerk and the preprocessing (the horizontal plane, a 3.5 Hz zero-phase low-pass filter) follow ISway (Mancini et al., 2012), which the Roche papers cite for jerk, because the Roche papers do not publish their filter. Plan repeated tests: reliability was reported for medians over 2 weeks.

• Test length (default 30 s; Montalban et al., 2022; Lipsmeier et al., 2018)
• Tests per session (default 1; one test per day in Montalban et al., 2022)
• Phone position: running belt at the front of the waist (default; Rinderknecht et al., 2023; Lipsmeier et al., 2022) or a trouser pocket (Lipsmeier et al., 2018; a later study discarded pocket recordings, Lipsmeier et al., 2022, JMIR)
• Seconds to get into position before the start vibration (default 10; no published value)
• Store the horizontal acceleration of each test, about 20 KB (default on; no published value)
• Fixed: resampled to 50 Hz (Rinderknecht et al., 2023); a 3.5 Hz zero-phase Butterworth low-pass (Mancini et al., 2012; order not stated there, 2nd order here); the recording starts 0.5 s after the start vibration (no published value)

• Sway path: the sum of the horizontal acceleration's sample-to-sample changes, m/s² (Montalban et al., 2022)
• Sway jerk, m²/s⁵, and its natural log (Mancini et al., 2012; log sway jerk: Lipsmeier et al., 2022)
• RMS of the horizontal acceleration
• Flags: a phone lying on a table (Montalban et al., 2022, Table S1), walking during the test, very unequal thirds of the test (Mancini et al., 2012), a sampling rate of 33 Hz or less (Kriara et al., 2025)
• The horizontal acceleration per test (optional)

Larger sway path and jerk mean more postural sway. The values depend on where the phone is worn, so keep the position constant within a study. Test-retest ICC was .71 in people with MS and .40 in healthy controls, computed on 2-week medians (Montalban et al., 2022); log sway jerk had an ICC of .82 in early Parkinson's disease but correlated only weakly with clinical postural instability (ρ = .14; Lipsmeier et al., 2022). The phone's sign conventions do not matter: both features use only the horizontal plane, whatever the phone's orientation.

• In the Floodlight study in multiple sclerosis, sway path from a daily 30 s phone test had a test-retest ICC of .71 and correlated with disability (EDSS) and self-reported physical impact (MSIS-29), but not with the Berg Balance Scale (Montalban et al., 2022).
• In the Roche PD app, the balance test differed between people with Parkinson's disease and controls (Lipsmeier et al., 2018), and log sway jerk was reliable (ICC .82) with a weak link to postural instability (Lipsmeier et al., 2022).
• ISway, the same features from a sensor at the lower back, was reliable (jerk ICC .86–.87) and related to postural instability in Parkinson's disease (Mancini et al., 2012).

• Montalban, Graves, Midaglia et al. (2022), A smartphone sensor-based digital outcome assessment of multiple sclerosis, Multiple Sclerosis Journal 28:654–664 (sway path)
• Rinderknecht, Zanon, Boonstra et al. (2023), An observational study to assess validity and reliability of smartphone sensor-based gait and balance assessments in multiple sclerosis, Digital Health 9 (the protocol)
• Mancini, Salarian, Carlson-Kuhta et al. (2012), ISway: a sensitive, valid and reliable measure of postural control, J NeuroEngineering Rehabil 9:59 (jerk and preprocessing)
• Lipsmeier, Taylor, Kilchenmann et al. (2018), Evaluation of smartphone-based testing to generate exploratory outcome measures in a phase 1 Parkinson's disease clinical trial, Movement Disorders 33:1287–1297
• Lipsmeier, Taylor, Postuma et al. (2022), Reliability and validity of the Roche PD Mobile Application for remote monitoring of early Parkinson's disease, Scientific Reports 12:12081
• Kriara, Dondelinger, Capezzuto et al. (2025), Investigating measurement equivalence of smartphone sensor-based assessments, JMIR 27:e63090 (sampling-rate exclusion)

• Rest and Postural Tremor
• Dual-Task Walking
• Finger Tapping Variability