Biomedical Engineering Reference
In-Depth Information
quires an underlying associative process, and there is some evidence that the
cerebellum is involved comparably, as in classical conditioning of eyeblink (22)
and lower limb withdrawal reflexes (42,43).
We studied classical conditioning of postural reflexes in a group of young
and healthy subjects (42) and in a group of 8 cerebellar patients suffering from
diffuse degeneration disorders who were compared with a group of sex- and
age-matched controls (43). Subjects were standing on a tilting platform allowing
rotational movements, representing the US, with the rotation axis through the
ankle. In paired trials a preceding auditory signal was given as CS. Subjects
were tested in 70 US-alone trials, in 80 trials with paired stimuli, followed by a
brief section of 20 US-alone trials. EMG signals were recorded from the main
muscle groups of the leg (tibial anterior muscle (TA), GA, rectus femoris mus-
cle, and biceps femoris muscle). In a study on a group of young and healthy sub-
jects (42), it turned out that 22% of the subjects established CR in all main
groups of the leg muscles tested and thus responded to the unexpected perturba-
tions similarly as in experiments on classical conditioning of the eyeblink or
withdrawal reflex (73). The preponderant proportion of the subjects, however,
established CR in the GA only but developed a substantial decay of the UR am-
plitude in the TA, which was much larger compared with that observed during
US alone trials (Figure 5B). Subjects with CR in
all
muscles (strategy-I sub-
jects) were assumed to apply a different strategy than those with CR in the GA
only (strategy-II subjects). In cerebellar patients, however, no such strategy was
observed. Stacks of typical TA responses recorded during US-alone and paired
trials obtained from a healthy strategy-I subject, from a strategy-II subject and
from a cerebellar patient are shown in Figure 5. Patients hardly established any
CR, and if they did show responses within the CS-US window, the responses
were not time-locked. Consequently, the resulting CR-incidence did not follow a
characteristic exponential learning curve.
The platform was equipped with strain gauges to record vertical forces ex-
erted by the subjects. From these forces the center of vertical pressure (CVP)
was calculated (43). The CVP represents the final outcome of all muscles in-
volved during a given trial with excursions primarily in the sagittal plane. Figure
6 shows the deviations of the CVP in the sagittal and frontal planes for the con-
trols and the group of cerebellar patients. Consequently, if a subject established
a muscle-related CR there was also a deviation in the CVP prior to the US,
which has to be attributed to a CVP-related CR. This deviation was to the rear
and was small compared with the deviation to the front due to the perturbation
representing the UR. In contrast to the generally accepted rules for classical
conditioning (25), the deviation of this CVP-based CR is in the opposite direc-
tion to the UR and can be interpreted as a preceding, preparatory, and compen-
satory shift of the CVP. As mentioned above, in cerebellar patients no muscle
activity-based-CR has been observed. Although we recorded from all main mus-
cle groups of the leg, there is the general problem of missing muscles involved
