INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Interlimb coordination during locomotion can be adjusted to meet different environmental demands. Subjects walking on a treadmill with two independent belts adjust a limb's movements based not only on the speed of the belt under that leg, but as a function of the difference between the two belt speeds (Dietz et al. 1994). Subjects pedaling unilaterally in a condition where loading of the pedaling limb is identical to that experienced during bilateral pedaling show increased muscle activity during unilateral, as compared with bilateral, pedaling (Ting et al. 1998). Thus a biomechanically similar task performed unilaterally is not completed using the same muscle activity employed to perform it bilaterally.

Despite clear influences of one limb on the other during locomotion, transfer of adaptation between the limbs is limited in some tasks. Following one-legged hopping on a moving treadmill, subjects inadvertently hop forward when asked to hop on the leg used during hopping on the moving treadmill. This aftereffect is not seen when subjects hop on the opposite leg that was not used during hopping on the treadmill (Anstis 1995). Are there situations where transfer would be possible? If both limbs were active during the task, would transfer between limbs occur? We addressed these questions using podokinetic stimulation. Previous studies have shown that subjects exposed to stepping in place on a rotating disk will inadvertently turn in circles when asked to step in place on a stationary surface with eyes closed (Weber et al. 1998), an adaptive effect called podokinetic after-rotation (PKAR) (Gordon et al. 1995).

We examined PKAR following bilateral stepping in place on a rotating disk and unilateral stepping in place on a rotating disk while the other limb stepped on a stationary surface. We hypothesized that PKAR following unilateral stimulation would be of lower amplitude than that following bilateral stimulation but would result in trunk rotation during stance on both the trained and untrained foot.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects and protocol

Eleven healthy adult volunteers, aged 22-78 yr, participated. All subjects provided informed consent. Each subject completed two sessions separated by at least 24 h. During each session, subjects were exposed to 30 min of PK stimulation, stepping in place with one foot on either side of the axis of a 76-cm rotating disk. In the bilateral session, both feet stepped on the rotating disk so that both limbs turned relative to the trunk. In the unilateral session, the left foot stepped on the rotating disk and the right foot stepped on a stationary surface covering the right half of the disk so that only the left limb turned relative to the trunk.

In both sessions, the disk rotated clockwise (CW) at 45°/s. Subjects stepped at 2 Hz, matching cadence to a metronome affixed to the trunk. Following PK stimulation, PKAR was measured while subjects attempted to step in place in the center of the stationary disk for 30 min while wearing a blindfold and earplugs. The order of bilateral and unilateral sessions was varied; half the subjects were exposed to the unilateral stimulus first and half were exposed to the bilateral stimulus first.

Data collection and analysis

During PKAR, we recorded touchdown and lift-off using foot switches affixed to the plantar surface of each foot under the first metatarsal head. This area was the first to contact and the last to leave the ground, as subjects used a toe-first gait when stepping in-place on the stationary disk. Pelvis position in space was recorded using an electromagnetic coil system (CNC Engineering, Seattle, WA) with the target coil positioned over the right anterior superior iliac spine. Position values were differentiated to obtain pelvis angular velocity in the yaw plane. Velocities in the counterclockwise (CCW) and CW directions were assigned negative and positive values, respectively.

For each subject, plots of velocity versus time for the first 2 min of PKAR from each trial were fitted with exponential rise to maximum curves to obtain 2-min rise maxima and rise time constants. Plots of velocity versus time for the remaining 28 min were fitted with three-parameter exponential decay curves to obtain initial velocity, response decay time constant, and final asymptote values for each trial (Weber et al. 1998). Mean values were calculated by averaging the values obtained for each subject's data (Table 1). A plot of mean PKAR response was generated by averaging velocity data across subjects and plotting average velocity versus time. These average curves were fitted in a manner similar to the individual curves, yielding the equations in Fig. 1B.

View larger version (29K): [in this window] [in a new window]   Fig. 1. Scatter plots of pelvic angular velocity vs. time for podokinetic after-rotation (PKAR) following bilateral () and unilateral () PK stimulation in a single subject (A) and averaged across all subjects (B). Lines drawn through data sets show first-order exponential decay curve fits. Note that the amplitude of the response following unilateral stimulation is less than that following bilateral stimulation, but the direction of turning is counterclockwise for both.

Using foot-switch records, pelvic velocities during left and right single-limb support times for the first 60 strides were determined. Stride was defined as time from one lift-off of the left foot to the next lift-off of the left foot. Thus a stride was composed of two steps and included two periods of single-limb support (1 on the left and 1 on the right). Group means for the bilateral condition were compared with group means for the unilateral condition using paired t-tests (P = 0.05).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All subjects demonstrated PKAR, turning CCW following unilateral or bilateral exposure to the CW-rotating disk. PKAR following unilateral exposure was of lower amplitude than that following bilateral exposure. Figure 1A shows the responses of a single subject whose PKAR initial velocities were 20.24 and 8.99°/s for the bilateral and unilateral conditions, respectively. Figure 1B shows the average across subjects. Equations show the values obtained by fitting the averaged velocity data in Fig. 1B with exponential decay functions. Curve-fit parameters provided in the table are group mean values ± SE, obtained by fitting each subject's data with curves and calculating the mean of these values across subjects. The 2-min maximum, initial velocity, and y-intercept values were significantly lower for the unilateral than for the bilateral condition (P < 0.05, paired t-tests). Time constants and first-order asymptotes were not different between conditions.

While Fig. 1 focuses on overall PKAR velocity across the entire 30-min period of PKAR, Fig. 2 shows pelvic angular velocity for left and right single-limb support periods over the first 60 strides, i.e., the first 60 s of the response. Thus the 60-s period shown in Fig. 2 corresponds to the first six data points of each curve in Fig. 1, as each point in Fig. 1 represents average velocity calculated over a 10-s bin. Following bilateral stimulation (Fig. 2A), the pelvis rotated CCW with the same velocity during both right and left single-limb support. Following unilateral stimulation (Fig. 2B), however, pelvic angular velocity was significantly different during right versus left single-limb support across the first 13 strides (P < 0.05). The pelvis rotated CCW during left single-limb support () but rotated CW or very little during right single-limb support () following left unilateral stimulation. Across the first 13 strides, pelvic angular velocities during right and left single-limb support migrated toward each other and by the 14th stride were not significantly different. From stride 14 on, pelvic angular velocity was generally in the CCW direction and was similar during left and right single-limb support on the majority of strides.

View larger version (34K): [in this window] [in a new window]   Fig. 2. Average pelvic angular velocity ± SE of all subjects during right () and left () single-limb support periods plotted vs. stride number. Data for bilateral PK stimulation (A) and unilateral left stimulation (B) are shown. There were no significant differences between left and right sides for the bilateral condition. For the unilateral condition, differences between left and right sides were significant (P < 0.05) for strides 1-13 (to the left of the - - -), after which the left and right sides were similar for most strides.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Interlimb influences

The difference in pelvic rotation during left and right single-limb support over the initial strides following unilateral PK stimulation demonstrates that adaptation of the left limb, which stepped on the rotating disk, was different from that of the right limb, which stepped on a stationary surface. The left side adapted to the CW rotation of the disk. The right side did not show evidence of this adaptation during the initial strides of PKAR. In fact, during the initial strides following unilateral PK stimulation, pelvic rotation during right single-limb support was in the CW direction. We hypothesize that touchdown of the unadapted right foot occurs while the pelvis is rotating CCW over the adapted left foot. This rotation is not "expected," as the unadapted right limb is programmed not to turn relative to the trunk. To compensate for this unexpected rotation, the pelvis may rotate CW during right single-limb support in an effort to achieve alignment of the pelvis over the foot and obtain zero relative rotation between the trunk and feet.

Although the left and right sides adapted differently, differences between the sides were quickly resolved. Integration of signals between the left and right sides shows clear influences of the PK state of one limb on that of the other limb. Differences in the overall velocity of PKAR following unilateral and bilateral stimulation suggest that neither the adapted nor the unadapted limb totally dominated the response. Rather the two limbs influenced each other, suggesting transfer of information between the two sides.

Comparison to interlimb interactions in other locomotor tasks

The interlimb influences observed here are in keeping with the reports of Ting et al. (1998), who showed that the sensorimotor state of one limb influences the output of the other limb. However, integration of inputs between the two sides in the PK system may seem to contradict reports that aftereffects observed following hopping on a treadmill do not transfer between limbs (Anstis 1995). A major difference between the present study and that of Anstis is that both limbs are active during stepping. In the hopping paradigm, only one limb at a time was active. The active engagement of both limbs may allow for integration and transfer between the limbs. Had Anstis asked subjects to hop on both feet following unilateral hopping on the treadmill, we think he would have observed an aftereffect. Conversely, had we tested unilateral hopping on the left and right sides following unilateral left PK stimulation, we think we would have observed PKAR during left-footed hopping but not during right-footed hopping.

We acknowledge that interlimb influence in the stepping task may result in part from the indirect mechanical linkage between the limbs via their connection to the pelvis. However, our reports are quite similar to those obtained with split-belt treadmill studies where adaptation to a differential in belt speeds occurred in 12-15 strides. This similarity in time required for coordination of the bilateral response in the present study and the split-belt study, despite clear differences in the mechanics of the two situations, supports the notion that this is a neural process and not merely a consequence of biomechanical linkages.

In summary, our results suggest that local somatosensory information regarding the rotation of the trunk with respect to each limb can be used to drive each leg independently. During bilateral PK stimulation, the local somatosensory information for the two limbs is similar and both limbs adapt similarly. With unilateral PK stimulation, local somatosensory information differs for the two limbs, and they adapt independently. Nevertheless, when walking on a stationary surface, the locomotor networks integrate information from the two sides.