View Full Version : Summary: Estimating movement onset times

Gunter Siegmund
04-01-2003, 03:29 AM
Thanks to those who responded to my posting. Below is the original posting
and a summary of the replies.



The onset of movement is often estimated from biomechanical data using one
of the following techniques:
a) the time at which a signal reaches a threshold (usually based in some way
on the noise level in the pre-movement signal), or
b) the time at which a signal last crosses through zero (or some other
baseline value) before the movement is clearly in progress.
We are attempting to quantify the bias introduced with these techniques for
different transducers, movement rates, etc. and have the following questions
for the group:
1. Are there other techniques that we should also consider?
2. Are there existing articles/references/sources that develop or evaluate
techniques for estimating movement onset?
Our search of the biomechanical literature, the biomch-l archives and the
web has yielded no primary sources for these techniques, only many examples
of their use. A review of a number of signal processing books has also
yielded nothing.
A summary of responses will be posted if there is sufficient interest.


From: "Charles Walter"

You might have a look at the motor control literature. Determining
movement initiation has long been important for reaction time studies, for
example. Here are a few articles dealing with the general issue of
quantifying movement onset:

Teasdale N, Bard C, Fleury M, Young DE, Proteau L (1993). Determining
movement onsets from temporal series. Journal of Motor Behavior, 25 (2):

Corbetta D, Thelen E (1995). A method for identifying the initiation of
reaching movements in natural prehension. Journal of Motor Behavior, 27
(3): 285-293.

Staude GH (2001). Precise onset detection of human motor responses using
a whitening filter and the log-likelihood-ratio test. IEEE Transactions on
biomedical engineering. 48 (11): 1292-1305.

Best Regards
Charles Walter, Ph.D.
Professor and Director of Graduate Studies
School of Kinesiology
University of Illinois at Chicago
Chicago, IL 60608-1516
ph 312-355-1713
fax 312-355-2305

From: "Ted Clancy"

For EMG, Bonato et al. have an IEEE paper: A statistical method for the
measurement of muscle activation intervals from surface myoelectric signal
during gait. IEEE Trans Biomed Eng. 1998 Mar;45(3):287-99.

Good luck,
Ted Clancy

From: "Rachael Seidler"

Here is a reference comparing different onset detection techniques:
Teasdale, N., Bard, C., Fleury, M., Young, D., & Proteau, L. (1993).
Determining movement onsets from temporal series.
Journal of Motor Behavior, 25, 97-106.

best regards,
Rachael Seidler
University of Michigan
Department of Movement Science
Neuroscience Graduate Program Faculty
401 Washtenaw Ave.
Ann Arbor, MI 48109-2214
Phone: 734-615-6224
Lab Phone: 734-764-8186
Fax: 734-936-1925
Email: rseidler@umich.edu
URL: http://www.kines.umich.edu/facstaff/seidler.htm

From: "Gaspar Morey"

You could try the EMG Onset literature, there are several articles
(anyway most don't define the methodology for onset determination). Here
are two hints you might follow.

This one presents a nice methodology that possibly would also fit your
purposes. Further Bonato seemed to me to be quite predisposed to answer
and discuss any questions.

Bonato P, Roy SH, Knaflitz M, De Luca CJ. Time-frequency parameters of
the surface myoelectric signal for assessing muscle fatigue during
cyclic dynamic contractions. 2001 IEEE Transactions on Biomedical

Hodges, P.W.; Bui, B.H. A comparison of computer-based methods for the
determination of onset of muscle contraction using electromyography.
1996 EEG & Clinical Neurophysiology

Best wishes,
Gaspar Morey Klapsing
Department of Biomechanics
German Sport University Cologne
Carl-Diem-Weg, 6
50933 Cologne
Phone: +49 (0)221 4982-776
Fax: +49 (0)221 4971598 Phone: +49 (0)221 4992842
Email: morey@dshs-koeln.de Email: gaspar@loop.de
Snail mail: Kirchweg 7
D-50858 Cologne
Mobile: +49 (0)1794226071

From: Garry T Allison

Onset detection for EMG have similar issues, I summarised some of these
issues and the robustness of certain techniques with increased baseline
Trunk muscle onset detection technique for EMG signals with ECG artefact.
Journal of Electromyography and Kinesiology. (In press - proofs done in
February... not sure when it will actually get to press)


From: Jean Blouin

You may be interested in the following paper:
Teasdale et al. (1993) Determining movement onsets from temporal series.
Journal of Motor Behavior 25: 97-102

Jean Blouin
UMR Mouvement et Perception
CNRS et Universite de la Mediterranee
163, avenue de Luminy
13288 Marseille

From: "K. Han Kim"

We normally place a micro button switch in the "home position" for the
right hand and asked the subject to place the right hand on top of the
button until the onset of the movements. With the onset of the right
hand toward the target, the push button switch is disengaged, and
eventually engaged again when the right hand returns back to that "home

It seems that the "cutting points" obtained from the switch activities
work very well in general, even for the torso and arm movements
which might involve some "preparatory" movements before the onset of
the actual hand movements. The problem is, however, that the head
gaze movements, which normally advance the hand movements, are not
adequately captured with this method. We had to determine the onset
of the head movements manually in most cases, accordingly.

K. Han Kim
Ph.D Candidate
The Center for Ergonomics
The University of Michigan

From: "Deric Wisleder"

I was considering using one method described in:

Teasdale N, Bard C, Fleury M, Young D, Proteau L (1993) Determining
movement onsets from temporal series. J Motor Behav 25(2) p 97-106.
to determine movement onset. It takes into account max velocity and
SD of the static period leading up to a threshold. I believe the SD,
however, was referenced to the threshold and not the mean. That does not
seem appropriate, as they are not directly related (Hence, my question in
the attached figure). In addition, the recursive routine repeats the
process and selects the lesser of the start points, but this seems subject
to random fluctuation of the baseline, and not necessarily indicative of
the true start of motion.

It seems that one could locate the points (.09*Vmax, .05*Vmax, or similar)
sequentially backward from Vmax and select the start of motion from that
set (usually the first point backward from Vmax). In addition, one might
reference the SD and mean of the initial baseline period, though I have not
yet worked out the method completely. Please let me know what you find. I
will determine movement onset and offset before doing sub-movement analysis
of target pointing data sets.

Deric Wisleder, Ph.D.
Research Associate
Motor Control Laboratory
PEBE 158
Department of Kinesiology
Arizona State University
Tempe, AZ 85287
Phone: 480 965 9081
Department Fax: 480 965 8108

From: "Sian Jenkins"

The source I've seen most often quoted for biomechanical techniques in gait
initiation studies is:

Polcyn AF, Lipsitz LA, Kerrigan C, Collins JJ. Age-related changes in the
intiation of gait: Degredation of central mechanisms for momentum
generation. Arch Phys Med Rehabil 1998; 79:1582-1589

I don't know if that'll be of any help to you.
Good Luck,
Sian E. M. Jenkins
Lady Youde Clinical Research Fellow in Applied Biomechanics
Centre for Rehabilitation and Ageing, Division of Geriatric Medicine
St George's Hospital Medical School, University of London
tel: +44 (0)208 725 5319

From: "christian peham"

We developed a method to determine the stance phase in horses via the
speed distribution of the hoofmotion.
It was published in the Journal of Biomechanics 32 (1999) 1119-1124
Limb locomotion - speed distribution analysis as new method for stance
phase detection. C Peham, M Scheidl, T Licka

I hope this information is useful.

A. Univ. Prof. Dr. Christian Peham
Clinic of Orthopaedics in Ungulates
University of Veterinary Medicine Vienna
Veterinaerplatz 1
A-1210 Vienna
Phone : ++43/1/250 77/5506
Fax: ++43/1/250 77/5590
Email: christian.peham@vu-wien.ac.at
Homepage: http://www.vu-wien.ac.at/i111

__________________________________________________ _____

Gunter Siegmund, Ph.D., P.Eng.

Adjunct Professor
School of Human Kinetics
University of British Columbia, Vancouver, BC, Canada

MacInnis Engineering Associates
11-11151 Horseshoe Way, Richmond, BC, Canada, V7A 4S5

office: 604 277 3040
fax: 604 277 3020
tollfree: 800 565 3040
__________________________________________________ _____

To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l