The participant walks with the phone in a front trouser pocket, once normally and once while counting backwards in threes; the phone's motion sensors measure stride time and its dual-task cost.
Configure parameters and run an interactive preview exactly as participants will experience it. No data is recorded.
Adjust parameters below, then start the preview on the right.
Include practice trials
Shown with feedback before the main task
Task parameters
BetaBeing checked on iOS and Android devices.
Phones only: the task records the phone's motion sensors. The phone vibrates at the start and end of each recording, and the screen stays on. The web preview shows the screens without measuring.
The phone goes in a front trouser pocket; participants walk in a quiet room or hallway, turning when needed. Validated in healthy adults aged 18–35 (Manor et al., 2018). The counting is not recorded or scored.
20–120. Default 45 (Manor et al., 2018).
1–3, each a normal and a counting walk. Default 1, as at home; 3 per laboratory visit (Manor et al., 2018).
Default 10 (no published value).
2 walks × 45 s (about 2 min). Counting backwards in threes from a random number between 200 and 999 (Manor et al., 2018), shown before the walk.
Fixed: the vertical acceleration at 100 Hz (Manor et al., 2018), filtered with a 2nd-order 25 Hz zero-phase Butterworth (Bourke et al., 2020; Manor et al. do not state theirs). Heel strike = the trough after each relatively high peak (Manor et al., 2018), one per stride period from the autocorrelation (Moe-Nilssen & Helbostad, 2004); strides during turns (area × duration of the turning rate above 2 rad·s, Manor et al., 2018) or outside walking (Bourke et al., 2020) are left out, and stride times more than 50% from the walk's median are dropped (no published value). Outputs: mean stride time per condition and the dual-task cost, 100 × (dual − normal) / normal (Manor et al., 2018); stride-time CV and cadence, exploratory.
This is a researcher preview. No participant data is recorded.
Simulated participant view
The participant walks with the phone in a front trouser pocket, once normally and once while counting backwards in threes; the phone's motion sensors measure stride time and its dual-task cost.
No data is recorded
Phones only. The participant puts the phone in a front trouser pocket and walks for 45 seconds at a normal pace, and for 45 seconds at a normal pace while counting backwards out loud in threes from a number between 200 and 999 shown before the walk, in random order. A vibration marks the start and the end of each walk. Walking in a quiet room or hallway, turning when needed. Stride time comes from the vertical acceleration; turns are detected from the turning rate and left out. The web preview shows the screens only.
Useful in ageing, neurology and intervention studies that want the cost of thinking while walking (cognitive-motor interference) measured at home and repeatedly.
Not for participants who cannot walk safely on their own, or where there is no safe space to walk for 45 s. The counting answers are not recorded or scored. The validation was in healthy adults aged 18–35 (Manor et al., 2018); results in older or clinical groups are not yet published. Beta: being checked on iOS and Android devices.
The defaults follow the smartphone assessment of Manor et al. (2018): the phone in the participant's preferred front trouser pocket, one 45 s normal walk and one 45 s walk while counting backwards aloud in threes from a random number between 200 and 999, the order randomized within the pair, in a quiet room or hallway, turning when needed. Manor's app spoke the start number; here it is shown before the walk. Manor et al. ran three pairs per laboratory visit and one at home.
Longer stride times while counting mean that walking slowed under the dual task. Against a GAITRite mat, the app's stride times (r² = .98) and dual-task costs (r² = .95) agreed closely, with a mean absolute error of 16.9 ms; ICCs were .82–.94 (Manor et al., 2018). Manor et al. did not publish their peak-picking rules; the heel strikes here follow their description (the trough after each relatively high peak of the vertical acceleration), with the stride period from the signal's autocorrelation (Moe-Nilssen & Helbostad, 2004) and stride times more than 50% from a walk's median dropped. Stride-time variability was not validated by Manor et al. and is exploratory.
Add Dual-Task Walking to your ESM study on SMAAT — free to get started.