117 volunteers · ages 20–94
Added inertia moved the mean from 7.7 to 5.2 Hz
Raethjen and colleagues found a reduction of about 32% after loading the hand. Frequency changed with mechanical properties.
7.75.2 Hz
Mechanical physiology · postural tremor
Ten oscillations a second
Measurements of outstretched hands in healthy volunteers often find acceleration energy between 8 and 12 Hz. Resonance in the hand–muscle–tendon system shapes that rhythm, and the phone used to measure it changes the mechanics.
In a human centrifuge experiment, resonant mechanics predicted 99% of peak tremor frequency. Across the classic band, one cycle lasts 83–125 milliseconds, giving 96–144 oscillations in 12 seconds.
Measure your hand
Simulated trace · 1.5 s
- One cycle
- 125 ms
- In 12 seconds
- 96 oscillations
An experiment on your body
Separate table from palm
Two 12-second runs place the sensor noise from a table beside the acceleration of a phone held on your palm. The calculation stays on your device and disappears when the page reloads.
Start with the control run. Your phone may ask for motion-sensor access.
| Run | 8–12 Hz peak | Concentration | Sample rate |
|---|---|---|---|
| Table | … | … | … |
| Palm | … | … | … |
After both runs, this line will show the ratio between 8–12 Hz band energy on your palm and on the table.
The mechanism
A living spring with inertia
Holding a hand in the air requires continuous muscle force. Small fluctuations drive the hand mass against the elasticity of muscles, tendons and ligaments, while resonance amplifies a narrow portion of the spectrum.
A simplified model links natural frequency to effective stiffness k and inertia I. Greater inertia lowers frequency when the other conditions remain comparable.
f0 ∝ √(k/I)
The 2015 experiment separated weight from inertia with a human centrifuge. Changing gravitational force increased muscle effort without adding mass to the hand. Under the tested conditions, resonant mechanics predicted 99% of peak frequency.
Neural activity and motor-unit recruitment supply some of the fluctuations that drive the system. Hand mechanics shapes the peak seen by the accelerometer. Experiments also find neural components and several dominant frequencies.
- FForce fluctuatesPostural contraction produces fluctuations across a range of frequencies.
- kTissues supply elasticityMuscle, tendon and ligaments form the effective spring.
- IThe hand supplies inertiaMass and its distribution set the mechanical response of the segment.
- ƒResonance selectsAcceleration grows near natural frequency, giving the spectrum a peak.
28 healthy, right-handed volunteers · ages 18–40
The instrument enters the measurement
The 169 g phone had almost 30 times the mass of the 5.7 g sensor. For the dominant hand in a postural test, the mean peak was 18% higher and peak spectral amplitude in the 5–12 Hz band was 72% greater in the phone recording.
5.7 g169 g
Method
What the run can tell you
The instrument looks for periodicity in acceleration. It describes phone motion during the measured 12 seconds and keeps the control run separate.
- Three axesThe accelerometer supplies motion along x, y and z in metres per second squared.
- Uniform rhythmSamples are placed on equal intervals, then the mean and linear drift are removed from each axis.
- Time windowA Hann window reduces the abrupt edge of the 12-second run.
- Combined spectrumPower from all three axes is added on a 0.1 Hz grid between 2 and 15 Hz. Analysis stops if the sensor cannot supply the rate needed for the 8–12 Hz band.
Primary sources
Where the figures come from
The claims about frequency, resonance and sensor mass come from experiments that measured acceleration, muscle activity or response to loading.
- Lakie et al., 2015, The Journal of PhysiologyThe human centrifuge experiment and mechanical prediction of peak frequency.
- Lakie et al., 2012, The Journal of PhysiologyThe relationship between muscle activity, acceleration and hand resonance.
- Raethjen et al., 2000, Clinical NeurophysiologyAccelerometry and electromyography in a sample of 117 people.
- Santos et al., 2022, Scientific ReportsA comparison between a 169 g phone and a 5.7 g sensor.
- Veluvolu and Ang, 2011, SensorsTime-frequency analysis of multiple components in physiological tremor.