Tweet
Herman's invitation to comment on the use of net joint moments
and powers has prompted me to broaden the present debate somewhat
to include the (f)utility of relating mechanical energy measures
to metabolic energy costs, and more generally still, the choice
of performance criteria to determine `good' gait.
There have been many reports in the literature of investigations
into the energy requirements and `efficiency' (although for most
cyclic, level movements true efficiency must be zero) of various
activities, typically walking and running, by analysis of
mechanical work (See Kaneko [J. Biomech 1990, pp 57-63]).
Burdett et al. [J. Orthop. Res. 1983, pp 63-72] compared various
mechanical energy measures with metabolic energy cost and
achieved a close correlation with work done per second on the
centre of gravity. Aleshinsky [J. Biomech. 1986 a..e, pp 287-
315] suggested energy measures based on work done at the joints
rather than segmental energies.
Although these investigations may be interesting, it is surely
misguided to attempt to relate whole body mechanical energy to
metabolic energy, they are measuring very different things.
Mechanical energy measures ignore the effects of co-contraction,
the metabolic costs of an isometric contraction, the extra
metabolic requirements related to increased cardio-vascular
workload, changes in the contractile efficiency caused by
chemical changes in the muscle, elastic energy storage e.t.c.. In
short they represent an ideal energy cost figure, which bears a
complicated, time varying relationship to metabolic energy costs.
If one wishes to measure metabolic energy costs then one should
surely use a more direct measure such as oxygen consumption or
increase in heart-rate.
However, this brings me to the area of interest for me, which is
the establishment of appropriate performance indicators for
assessing the `quality' of an activity (whether the activity is
actually measured from a subject or is the result of a computer
simulation where the quality measure forms the cost function for
an optimisation routine). These performance indicators should be
both RELEVANT to the constraints on the activity being studied
and also straightforward to measure and interpret.
Net metabolic energy consumption is relatively easy to measure
and interpret, but has the following disadvantages when assessing
pathological gait:
1. It requires a period of time (minutes) for a steady state
to be reached, this may be longer than some subjects can
remain walking. (are studies that get around this by
measuring oxygen debt valid?)
2. It can not isolate the metabolic costs of movements at
individual joints.
3. It can not isolate the metabolic costs of different
phases of the movement.
4. It often requires the use of cumbersome equipment which
may disturb the gait (can a child or elderly subject be
expected to walk normally on a treadmill whilst breathing
into a Douglas Bag?).
As an example, I wish to assess swing through crutch aided gait
in paraplegics, the limiting factor in the use of the gait is
often not the overall energy cost, but specific fatigue of the
shoulders and arms due to high loads during the body-swing phase.
Various researchers have examined net moments at the appropriate
joints, raised to various powers and/or their time integrals
[Opila, J. Biomch. Eng. 1987 285-290],[Crowninshield and Brand, J
Biomch. 1981 793-801], others have looked at mechanical work
[Wells, J. Biomch. 1979, pp 579-585].
The first approach ignores the fact that a moment applied during
an eccentric contraction have a lower metabolic cost than one for
an isometric contraction, which is lower still than a concentric
contraction (Of course the relationship depends on
shortening/lengthening speed rather than just direction of
movement). The second ignores the metabolic cost of sustaining
an isometric force. They are probably both limited predictors of
joint fatigue.
If the actual muscular force distribution around a joint were
reliably known, together with the speed of lengthening/shortening
of the muscle fibres (do all the fibres in a muscle change length
at the same rate along all their length?), then one could
approximate the energetics of each muscle from anatomical
information and experiments performed on isolated muscles that
have been reported in the physiology literature.
However it is more likely that we will only have net joint
torques, so my question is (finally):
Can a relationship be established between the net joint
torque, angle, and angular velocity and the metabolic energy
cost at that joint (or better still the tendency to
fatigue)? The relationship should be valid for positive,
negative and zero values of angular velocity. An exact
relationship would not be needed, a monotonically increasing
one would suffice to show if a particular change in the gait
pattern was beneficial or not.
My thought is to isolate the movement of interest in a
dynamometer that is capable of driving a limb eccentrically (such
as a KIN/COM). Then to calibrate the increase in metabolic cost
(measured t hrough oxyg en consumpt ion) agains t
shortening/lengthening speed and load. These values would be
used in a look-up table to approximate the metabolic cost at the
same conditions in the actual movement. I realise that there are
many assumptions and shortcomings in this technique, can anyone
suggest a better measure?
My apologies for straying away from the high ground of
theoretical mechanics to the boggy depths of rehabilitation
engineering and exercise physiology!
Ben Heller, Bioengineering Unit, Strathclyde University, Glasgow,Scotland.
and powers has prompted me to broaden the present debate somewhat
to include the (f)utility of relating mechanical energy measures
to metabolic energy costs, and more generally still, the choice
of performance criteria to determine `good' gait.
There have been many reports in the literature of investigations
into the energy requirements and `efficiency' (although for most
cyclic, level movements true efficiency must be zero) of various
activities, typically walking and running, by analysis of
mechanical work (See Kaneko [J. Biomech 1990, pp 57-63]).
Burdett et al. [J. Orthop. Res. 1983, pp 63-72] compared various
mechanical energy measures with metabolic energy cost and
achieved a close correlation with work done per second on the
centre of gravity. Aleshinsky [J. Biomech. 1986 a..e, pp 287-
315] suggested energy measures based on work done at the joints
rather than segmental energies.
Although these investigations may be interesting, it is surely
misguided to attempt to relate whole body mechanical energy to
metabolic energy, they are measuring very different things.
Mechanical energy measures ignore the effects of co-contraction,
the metabolic costs of an isometric contraction, the extra
metabolic requirements related to increased cardio-vascular
workload, changes in the contractile efficiency caused by
chemical changes in the muscle, elastic energy storage e.t.c.. In
short they represent an ideal energy cost figure, which bears a
complicated, time varying relationship to metabolic energy costs.
If one wishes to measure metabolic energy costs then one should
surely use a more direct measure such as oxygen consumption or
increase in heart-rate.
However, this brings me to the area of interest for me, which is
the establishment of appropriate performance indicators for
assessing the `quality' of an activity (whether the activity is
actually measured from a subject or is the result of a computer
simulation where the quality measure forms the cost function for
an optimisation routine). These performance indicators should be
both RELEVANT to the constraints on the activity being studied
and also straightforward to measure and interpret.
Net metabolic energy consumption is relatively easy to measure
and interpret, but has the following disadvantages when assessing
pathological gait:
1. It requires a period of time (minutes) for a steady state
to be reached, this may be longer than some subjects can
remain walking. (are studies that get around this by
measuring oxygen debt valid?)
2. It can not isolate the metabolic costs of movements at
individual joints.
3. It can not isolate the metabolic costs of different
phases of the movement.
4. It often requires the use of cumbersome equipment which
may disturb the gait (can a child or elderly subject be
expected to walk normally on a treadmill whilst breathing
into a Douglas Bag?).
As an example, I wish to assess swing through crutch aided gait
in paraplegics, the limiting factor in the use of the gait is
often not the overall energy cost, but specific fatigue of the
shoulders and arms due to high loads during the body-swing phase.
Various researchers have examined net moments at the appropriate
joints, raised to various powers and/or their time integrals
[Opila, J. Biomch. Eng. 1987 285-290],[Crowninshield and Brand, J
Biomch. 1981 793-801], others have looked at mechanical work
[Wells, J. Biomch. 1979, pp 579-585].
The first approach ignores the fact that a moment applied during
an eccentric contraction have a lower metabolic cost than one for
an isometric contraction, which is lower still than a concentric
contraction (Of course the relationship depends on
shortening/lengthening speed rather than just direction of
movement). The second ignores the metabolic cost of sustaining
an isometric force. They are probably both limited predictors of
joint fatigue.
If the actual muscular force distribution around a joint were
reliably known, together with the speed of lengthening/shortening
of the muscle fibres (do all the fibres in a muscle change length
at the same rate along all their length?), then one could
approximate the energetics of each muscle from anatomical
information and experiments performed on isolated muscles that
have been reported in the physiology literature.
However it is more likely that we will only have net joint
torques, so my question is (finally):
Can a relationship be established between the net joint
torque, angle, and angular velocity and the metabolic energy
cost at that joint (or better still the tendency to
fatigue)? The relationship should be valid for positive,
negative and zero values of angular velocity. An exact
relationship would not be needed, a monotonically increasing
one would suffice to show if a particular change in the gait
pattern was beneficial or not.
My thought is to isolate the movement of interest in a
dynamometer that is capable of driving a limb eccentrically (such
as a KIN/COM). Then to calibrate the increase in metabolic cost
(measured t hrough oxyg en consumpt ion) agains t
shortening/lengthening speed and load. These values would be
used in a look-up table to approximate the metabolic cost at the
same conditions in the actual movement. I realise that there are
many assumptions and shortcomings in this technique, can anyone
suggest a better measure?
My apologies for straying away from the high ground of
theoretical mechanics to the boggy depths of rehabilitation
engineering and exercise physiology!
Ben Heller, Bioengineering Unit, Strathclyde University, Glasgow,Scotland.