228
K. Kodono and H. Kimura
to the transition of leg loading from LF to RF even not so sufficient. Since
similar results about durations of B1 and B2 and duty ratios at early adaptation
with those at early adaptation in Fig. 1 are obtained in experiments using the
decerebrate cat model, we can consider that early adaptation is induced at the
spinal cord of a decerebrate cat [6].
While keeping split-belt walking, the STPD of LF is adjusted by learning
functions in Eq. (4). Consequently, adjustment of the relative phase between
LF and RF is carried out by learning functions at the cerebellum, leg loading
between LF and RF is exchanged smoothly at [H] in Fig. 8-(c), and the stable
gait in split-belt walking different from the gait in tied-belt walking is induced
as a result of late adaptation at [F] in Fig. 8-(c). In the late adaptation of the
decerebrate cat model, the duration of B1 is prolonged at (iii) in Fig. 9-(a) and
at [E] in Fig. 9-(c). But the duration of B2 is also prolonged at (iii) in Fig. 9-(a).
Such difference from the result of a decerebrate cat, including the duration of
the swing/stance phases should be fixed in our next study.
5 Conclusion
We constructed the spinal cat model integrating our previous leg controller with
Frigon’s spinal cord model, and the decerebrate cat model adding motor learning
function at the cerebellum. We showed by experiments that early adaptation
was obtained in the spinal cat model, and late adaptation was obtained by
motor learning as the step distance adjustment in the decerebrate cat model.
The validity of those models was evaluated by comparing durations of forelegs
bisupport phases and duty ratios between decerebrate cats and a quadruped
robot with the decerebrate cat model. It does not necessarily mean that this is
how the control in the biological system works, but we confirmed that the basis
of proposal model concepts could explain the adaptation of split-belt walking.
References
1. Yanagihara, D., Udo, M., Kondo, I., Yoshida, T.: A new learning paradigm: adaptive changes in interlimb coordination during perturbed locomotion in decerebrate
cats. Neurosci. Res. 18, 241–244 (1993)
2. Yanagihara, D., et al.: Climbing fiber responses in cerebellar vermal Purkinje cells
during perturbed locomotion in decerebrate cats. Neurosci. Res. 19, 245–248 (1994)
3. Yanagihara, D., Kondo, I.: Nitric oxide plays a key role in adaptive control of
locomotion in cat. Natl. Acad. Sci. 93, 13292–13297 (1996)
4. Frigon, A., et al.: Left-right coordination from simple to extreme conditions during
split-belt locomotion in the chronic spinal adult cat. J. Physiol. 595(1), 341–361
(2017)
5. Ito, S., Yuasa, H., Luo, Z., Ito, M., Yanagihara, D.: A mathematical model of
adaptive behavior in quadruped locomotion. Biol. Cybern. 78, 337–347 (1998)
6. Fujiki, S., et al.: Adaptation mechanism of interlimb coordination in human splitbelt treadmill walking through learning of foot contact timing: a robotics study.
J. Roy. Soc. Interface 12(110), 20150542 (2015). https://doi.org/10.1098/rsif.2015.
0542
K. Kodono and H. Kimura
to the transition of leg loading from LF to RF even not so sufficient. Since
similar results about durations of B1 and B2 and duty ratios at early adaptation
with those at early adaptation in Fig. 1 are obtained in experiments using the
decerebrate cat model, we can consider that early adaptation is induced at the
spinal cord of a decerebrate cat [6].
While keeping split-belt walking, the STPD of LF is adjusted by learning
functions in Eq. (4). Consequently, adjustment of the relative phase between
LF and RF is carried out by learning functions at the cerebellum, leg loading
between LF and RF is exchanged smoothly at [H] in Fig. 8-(c), and the stable
gait in split-belt walking different from the gait in tied-belt walking is induced
as a result of late adaptation at [F] in Fig. 8-(c). In the late adaptation of the
decerebrate cat model, the duration of B1 is prolonged at (iii) in Fig. 9-(a) and
at [E] in Fig. 9-(c). But the duration of B2 is also prolonged at (iii) in Fig. 9-(a).
Such difference from the result of a decerebrate cat, including the duration of
the swing/stance phases should be fixed in our next study.
5 Conclusion
We constructed the spinal cat model integrating our previous leg controller with
Frigon’s spinal cord model, and the decerebrate cat model adding motor learning
function at the cerebellum. We showed by experiments that early adaptation
was obtained in the spinal cat model, and late adaptation was obtained by
motor learning as the step distance adjustment in the decerebrate cat model.
The validity of those models was evaluated by comparing durations of forelegs
bisupport phases and duty ratios between decerebrate cats and a quadruped
robot with the decerebrate cat model. It does not necessarily mean that this is
how the control in the biological system works, but we confirmed that the basis
of proposal model concepts could explain the adaptation of split-belt walking.
References
1. Yanagihara, D., Udo, M., Kondo, I., Yoshida, T.: A new learning paradigm: adaptive changes in interlimb coordination during perturbed locomotion in decerebrate
cats. Neurosci. Res. 18, 241–244 (1993)
2. Yanagihara, D., et al.: Climbing fiber responses in cerebellar vermal Purkinje cells
during perturbed locomotion in decerebrate cats. Neurosci. Res. 19, 245–248 (1994)
3. Yanagihara, D., Kondo, I.: Nitric oxide plays a key role in adaptive control of
locomotion in cat. Natl. Acad. Sci. 93, 13292–13297 (1996)
4. Frigon, A., et al.: Left-right coordination from simple to extreme conditions during
split-belt locomotion in the chronic spinal adult cat. J. Physiol. 595(1), 341–361
(2017)
5. Ito, S., Yuasa, H., Luo, Z., Ito, M., Yanagihara, D.: A mathematical model of
adaptive behavior in quadruped locomotion. Biol. Cybern. 78, 337–347 (1998)
6. Fujiki, S., et al.: Adaptation mechanism of interlimb coordination in human splitbelt treadmill walking through learning of foot contact timing: a robotics study.
J. Roy. Soc. Interface 12(110), 20150542 (2015). https://doi.org/10.1098/rsif.2015.
0542