PDA

View Full Version : replies to : how to define upper limb technical array?



Nancy Black
07-16-1997, 11:17 PM
MY ORIGINAL MESSAGE:
Date: Fri, 4 Jul 1997 08:58:26 -0300

I'm a doctoral candidate, preparing for thesis experiments, and am faced
with the question of whether to use individual markers mounted on definable
"landmarks" on the upper limbs, or whether one is better using
technical arrays of markers on a common base. While individually mounted
markers are generally self-explanatory (except how to define the
limb segments based on the surface markers), I wonder
how to mount, stabilize and place technical array groupings.
I am assuming the arrays would be placed on the upper arm, forearm and
hand.
My questions are:
1) Schmidt et al. (1994) suggest using "foam rubber", and elsewhere I have
read of neoprene. Sounds HOT. Why not a lighter material like tubigrip
(cotton-rubber bandage material)? Would the flexibility of such a
material allow too much individual marker movement?

2) How are the arrays mounted and stabilized on differently sized arms
- are markers on the arrays moved to be proportionally located similarly
on small and large segments?
- does consistency of marker placement on the "base" material matter if a
"calibration" is made with markers over joints?
- what assumptions of joint centres does such a calibrating technique
include?

3) What information would potentially be lost by using technical arrays
rather than individual markers?

I will be using a 3 camera VICON 140 system for recording arm
movements during activities of daily living.
Your suggestions and comments are most appreciated. I will post a
summary of the replies.

SUMMARY OF REPLIES:

Firstly, I apologize for the incomplete reference in my note. Schmidt
et al. (1994) was by Schmidt, R. Disselhorst-Klug, C., Silny, J.,
Rau, G., "Measurement of Cardan Angles and Helical Axis at wrist and elbow
joints during free upper extremity movements", in the Proceedings of the
4th International Symposium on 3D Analysis of Human Movement, Grenoble, France
(4 pages long, but I don't have page number) ....

In direct response to my questions:
1. on what sort of base material may one mount technical arrays
- on just about anything... rigid materials or flexible ones
- Liduin Meershoek mentioned mounting them on velcro strips for the arm
segments and elastic bandage for the hand
- the work out of Oxford University tends to use rigid bases,
although possibly flexible ones may be used. (Leardini)
- markers may be raised to some distance from the skin, or essentially
at skin level. (3cm up wiht 5-8 cm span between each of 4 in a unique
pattern per limb - Williams)
- rigid based systems may be mounted using tape and secured with
elasticated bandage (Williams).

2. one may size arrays according to arm size, but most important is
visibility and calibrating the system to individual joint centres (all..)
- joint centre locations can be determined during arm swing movements
- Williams gives some detail of geometry

3. information potentially lost is particularly in terms of longitudinal
rotation of segments... if arrays are placed near the proximal limb end,
they will not measure it, however the greater intermarker distance the
more one may encounter movement due to being on extreme locations...


Among the key points raised:
* Research has been done on marker arrays by Cappozzo et al. (references
attached) as well as at Oxford University by J. Williams, Alberto
Leardini and their colleagues.
* rigid based arrays may not be the best idea (Leardini)
* technical arrays do allow doing calibration after recording instead of
before (Williams)
* scaling arrays for the arm size is good, and may allow better
visibility of markers
* a calibration motion (like swinging the arm) can help define joint centres

Edited original text of the replies sent (by author) follows...
Thank you all for your assistance!
__|~_
Nancy Black __|~_)_I__)_|~_
Ecole de genie )_ __)_|_)__ __)
Universite de Moncton | )____) | EMAIL: blackn@umoncton.ca
Moncton, New Brunswick \\___|____|____|____// FAX: (506) 858-4082
E1A 3E9 CANADA \\ // PHONE: (506) 858-4079
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~


-------------------------------------------------------------------
THE ORIGINAL REPLIES:

Date: Mon, 7 Jul 1997 10:16:19 +0200
From: Liduin Meershoek

I have been using some upper limb markes arrays with VICON. We mounted them
with velcro strips (upper arm, fore arm) and elastic bandage (hand). The
elastic bandage was painted black to prevent false reflections. We used
different arrays for large and small arms to get them better stabilized. The
placement of your markers will influence the sensitivety to measurement
errors. The best placement is with equal distances between the markers,
however this might interfere with the recognition and/or the movements.
During calibration your define local coordinate systems through bony
landmarks. The major assumption is that the system (bony landmarks-marker
array) is rigid. You can calculate the rotation axis from the movements (van
den Bogert et al. J.Biomech 27:1477-1488) or assume that they are parallel
with your local coordinate axis. In the latter case you can define a 'zero'
position with all local coordinate systems parallel and calculate
Euler-Cardanic rotations to represent flexion/extension, ab/adduction,
endo/exorotation. This assumes that the rotation axis are mutually
perpendicular and the sequence will influence your results.
You should place your forearm array either as distal as possible (we did) or
as proximal as possible. In this way pro- and suppination will either be
between the upper and forearm arrays or between the forearm and hand and not
somewhere in between.
You will loss information about rigidity of the segment. However using
individual markers you cannot differentiate between skin displacement and
nonrigidity so I don't know whether you really loss something.

Liduin Meershoek
Vakgroep Functionele Morfologie
Faculteit Diergeneeskunde
Universiteit Utrecht
Yalelaan 1
3584 CL Utrecht
the Netherlands
tel. 030-2534324

____________________________
Date: Mon, 7 Jul 1997 10:19:21 GMT0BST
From: Ingram Murray

I read your email query to Biomech-L with great interest. I am
currently in the early stages of a PhD, also looking into motion of
the upper limb. It would seem that we are both at a similar stage of
work as I currently am trying to decide upon the best method of
marker placement and axis definition. I am afraid I haven't yet got
the knowledge to answer any of your questions, but I am very
interested in hearing of the replies you receive. I take it from
your mention of a "calibration" with markers over joints that you are
using a technique similar to that used by Cappozzo (1991), Cappozzo
et al (1995) for measuring movement of the lower limb ? I myself was
considering the use of a modified version of this method for studying
the upper limb.
I think it would be interesting to keep in touch as we seem to be
working in a very similar area. Could you possibly also let me know
the full title for the paper you cite in your email (Schmidt et al
(1994)) as I have not read it and can't seem to find it through my
usual literature search.

Bye for now.....
_____________________________________________
Ingram Murray
CREST
Room M25
Stephenson Building
University of Newcastle upon Tyne
NE1 7RU

Tel: 0191 222 6193
Fax: 0191 222 8600
email: i.a.murray@ncl.ac.uk
------
AND THEN: (after I requested details on the articles Ingram had cited):
Date: Mon, 7 Jul 1997 15:29:23 GMT0BST
From: Ingram Murray

Hi again,
The references for the two papers I mentioned are:

Aurelio Cappozzo :- Three-dimensional analysis of human
walking:Experimental methods and associated artifacts. :- Human
Movement Science 10 (1991) :pp 589-602

A Cappozzo, F Catani, U Della Croce, A leardini:- Position and
orientation in space of bones during movement: anatomical frame
definition and determination. :- Clinical Biomechanics Vol.10, No.4
(1995) pp171-178

Have got my hands on a copy of the Grenoble Symposium Proceedings.
The authors do indeed seem to have taken the idea of "calibrating"
using markers attached for one frame only and then removed, as
suggested by Cappozzo for the lower limb. I am interested however in
measuring motion at the shoulder as well as the elbow and wrist so
would have to take the method a bit further.
Thanks anyway for the reference.
Ingram Murray
_____________________________________________

Date: Fri, 4 Jul 97 21:46:35 UT
From: John Williams

As you will know from Ed. Biden I have done quite a lot of work on upper limb
markers in Oxford and if you are interested would be happy to send you part of
my DM thesis which is relevant to your work. You might perhaps talk to Ed as
well! Please contact me directly if you want this info.

JohnW
John Williams MA DM FRCS(Orth)
9, Dobson Close,
Hall Lane,
Appley Bridge,
WIGAN.
WN6 9ES

Tel: 01257 256379 (Home)
Tel: 01257 252211 (Work)
E-mail: williams_john@msn.com
--------------
Date: Tue, 8 Jul 1997 14:20:51 -0300 (ADT)
From: Nancy Black
To: John Williams

I have read a copy of your DM thesis, and have thus far based my marker
placement (individual markers) on those you suggested, given your
extensive checks on skin movement etc. However: since for my ADL
activities, the medial epicondyle of the humerus is very
rarely visible, I thought about using the lateral epicondyle which is
consistently visible, but upon further checking, is almost colinear with
the origin of the brachioradialis and the insertion of the deltoid markers.
Hence one of the reasons I thought about considering technical arrays.
The other reason was / is your suggestion that likely such arrays may be
more exact than individually mounted markers. This is particularly
interesting and important to me since I am not an orthopaedic surgeon and
likely have a relatively large error in placement associated with each
marker as a result.
Unfortunately, I have not yet been able to find a full definition of what
one is looking for in a technical array. Ex. Should markers be placed
consistently as much as possible between subjects (ie. mounting markers
onto a sleeve which is already on the subject, trying to place markers
relative to bony landmarks (or other landmarks) .. .or should markers be
permenantly anchored on the sleeve,... or does it make a difference? I'm
assuming that indeed a calibration of the technical array must precede
recording sessions... I have a copy of the abstract of your poster from
CORS in Quebec City, 1996, but this doesn't give me much information. to
go on (A. Leardini suggested it...). Are there other documents I should be
consulting? I recently
ordered a copy of Cappozzo et al. (1995 - Clinical Biomechanics), but it
will take a while for it to reach me here...
---------- his reply:
Date: Wed, 9 Jul 97 21:47:09 UT
From: John Williams

With regard to technical markers the nice thing about them is that the
calibration is done at the end so you can play around with the position of
them until you get good visualisation of the markers in the array, then
collect your data and then dot he calibration. Make the technical markers
with four markers on the array and make sure that it has a unique pattern i.e.
it is not a mirror image about one axis. Make then not too big or else they
wobble and not too small or else the markers merge. The ones I played with in
Oxford were about 3cm off the base plate and about 5-8 cm in span. (ask
Alberto or one of the Oxford people to measure them as they are still in the
lab somewhere -Danielle used them last). I attached then to people with a
combination of double sided tape on the 4x3 cm bases and a circumferential
band of elasticated bandage (approx 8cm wide).

I suggest you contact Alberto Leardini about his/Bologna's CAST routine for
further details.

JohnW
----------
From: Nancy Black
Sent: 09 July 1997 23:06

What I had been considering were basically
arrays against the skin (rather than at a distance away from the
segment). Ed (Biden) is encouraging me to instead move
one of my upper limb markers to say the body of the biceps muscle (to
make the 3 non-colinear). We will see the results of that and then I'll
make a final decision.
--------- his reply:
Date: Fri, 11 Jul 97 01:33:08 UT
From: John Williams

I would be wary of placing markers on muscles that may change shape and size
during motion.

johnW

_________________________________________
Date: Tue, 8 Jul 1997 07:19:03 -0300 (ADT)
From: "Movement Analysis Lab."

I've been doing some work in Oxford in order to apply the 'anatomical
landmark calibration' technique (first introduced for lower limb: Cappozzo
et al. Clin Biomech 1995) to the kinematics of the upper limb. We designed
and manufactered technical arrays in order to be able to collect upper limb
segment positions during several reaching tasks. It would not difficult for
you to find the abstract of the preliminary results presented at the last
year Canadian Orthopaedic Research Society (Quebec, May 1996).
Good luck
Alberto Leardini

Movement Analysis Lab.
Department of Orthopaedic Surgery - Istituto Ortopedico Rizzoli
Via Di Barbiano, 1/10 40136 Bologna, ITALY
tel: ++39 51 6366520 (secretary)
++39 51 6366571 (direct)
fax: ++39 51 583789
email: VI6BOQ71@ICINECA.CINECA.IT
-------
TO WHICH I REPLIED:
> I had seen the abstract form the CORS meeting in Quebec 1996, but what I
> have is simply a 1 paragraph summary (Poster #71, isn't it, p. 61 of the
> ringed document...? it was a photocopy from a colleague, so I can't be
> sure this was the proceedings as such). If you have a more detailed
> description of the array you manufacture, I would be most interested in
> receiving it. I wonder if it would be appropriate for activities of
> daily living where the forearms tend to be rotated toward the centre of
> the body, as well as for reaching activities?
> Since posting my message to biomch-l I have ordered a copy of Cappozzo et
> al.'s Clinical Biomechanics 1995 article.
> Thank you again for your response, and your assistance.
> Nancy
-------
Date: Wed, 9 Jul 1997 05:34:33 -0300 (ADT)
From: "Movement Analysis Lab."

yes, the arrays were designed to be seen during many actvities: reaching
objects above or in front the subject, touching the head or the back, etc.
Anyway, we just pointed out that 'technical' arrays instead of direct
landmark location of the markers would allow for a better visibility of the
markers, particularly in large range of motion such those necessary for the
upper limb. At that time we manufactered rigid arrays (constant distances
between markers). Now I would be more for deformable arrayes, as those
obtained just stucking markers on the skin, of course in positions suitable
to be seen in every body segment position during the activity under
analysis. So that the only criteria of the design would be marker visibility
and easy-to-track. I would suggest to define your own arrays according to that.
No much more of that was included in the abstract.
Please do not hesitate to contact me for further suggestions.
Alberto Leardini

_____________________________
Date: Wed, 09 Jul 1997 08:46:05 +1000
From: Richard Smith

I think the main question you have to answer is "What do you want to
measure with the markers?". I you place an array on the limb (either
forearm or upper arm) using a base it will move with the "average" skin
surface at that point, probably in the middle because the array must not
interfere with joint movement. If you look at the skin movement at this
point during rotation about a longitudinal axis you will see that it moves
only about half the distance that the humerus moves (as indicated by
rotation at the elbow joint). My solution (for the upper arm) has been to
use a shoulder marker and two markers on the lateral and medial
epicondyles. This is not perfect by any means but I think it is more
representative of upper arm movement than an array in the middle of the
arm. For the forearm the same: two markers at the wrist joint.
To define the axes you could video a "neutral" position and make all your
angles relative to that position. Place the limb in the anatomical position
relative to the lab axes, video a few frames, track and calculate the
rotations required to bring a temporary segment axis system to being
parallel with the lab axes. This transformation can then be used to
calculate a segment fixed axis system. The two segment fixed systems can
then be used to calculate the relative angle between the two segments. If
you only want relative angles then you don't have to worry about joint
centres and translations.
If you have only three cameras and you are doing extreme rotations you may
have to resort to your own suggestion to avoid to many lost markers.
Good luck,

Richard

__________________________

Date: Wed, 9 Jul 1997 13:19:38 -0300 (ADT)
From: "Ed Biden (BioMed)"

Before the issue of technical markers goes too far, I have a real problem
with them in terms of begin able to use them for routine clinical
assessments. Having a technical array generally means a fairly bulky and
to some extent, "in the way" arrangement which can be a problem especially
with smaller children, or others who are readily distractable. I
personally view a technical array as a poor second choice if there is any
sensible way to do the tests with anatomical markers.

Ed

_________________________________
And on a related note:
Date: Sun, 13 Jul 1997 21:52:51 +0200
From: (Carolyn Anglin) Eric Nodwell
To: 'Nancy Black'

One of the most difficult problems with markers on the arm is simply being
able to see them in at least 2 cameras at all times. I would therefore
check this out before you finalise anything.

I'm very attracted to technical arrays since it's somewhat of a pain to attach
individual markers. However, I can foresee two problems: for the sake of
accuracy you want the markers to be widely spaced, but that may make it more
difficult to attach or to hold steady. The main goal, whether you're you're
using individual markers or technical arrays is to relate the markers to
joint centres. Van der Helm et al. (several papers) suggest relating the
markers to known bony landmarks such that data from different centres can
be easily compared and swapped. You also have to decide ahead of time
whether you care about scapular movement. Some programs don