|
North American Congress on Biomechanics Canadian Society for Biomechanics - American Society of Biomechanics University of Waterloo Waterloo, Ontario, Canada August 14-18, 1998 |
|
The analysis of human movement often assumes the body segments are rigid. In measuring the motion of these segments problems arise due to the motion of skin mounted markers. Part of the reason for the motion of the skin is due to the motion of the underlying muscular masses. This study examines intra-segmental soft tissue motion and its implications in defining the motion of human body segments and in separating signal from noise.
Biomechanical data are usually low pass filtered, with cut-off frequencies between 4 and 10 Hz, before derivatives are taken and used in the estimation of forces in a rigid body model. When these data are used to estimate ground reaction forces the higher frequency components of the signal cannot be reproduced (e.g. Bobbert et al., 1991; Sanders et al., 1991).
Part of the noise that is commonly filtered out is the skin marker movement. Studies which have looked at the comparison of surface mounted markers versus bone mounted markers have found soft tissue artifacts of up to 30 mm (Fuller et al., 1997; Cappozzo et al., 1996). Fuller et al (1997) also found the frequency content of the bone markers and the skin markers to be very similar to each other during a gait cycle. The relative motion of the tibia and femur showed a heel strike transient at 14 Hz. They suggested that soft tissue motion could introduce higher frequency components from relative motion between bone and soft tissue. The skin markers did not produce the 14 Hz transient that they predicted when they examined the unsmoothed vertical displacement data and from this they inferred a well damped response.
This study used an array of markers on the lateral surface of the thigh to investigate the soft tissue motion of the thigh during a drop landing. If the marker movement is random noise the array will divulge no more information than the net motion of the segment as a rigid body. However if the marker movement is a result of soft tissue motion the array information can be used to evaluate and quantify aspects of this soft tissue motion during an impact.
Two subjects, one male: 1.75 m, 84 kg, and age 26, and one female: 1.65 m, 59 kg, and age 22, performed five one legged drops with a stiff leg onto a force platform, sampling at 1000 Hz, from a height of 25 cm. The left lateral portion of the thigh of each subject was covered with an array of 28 retro-reflective markers, forming a 7 by 4 grid. The marker movement was recorded in two-dimensions using an automatic motion analysis system (Qualisys Pro Reflex). Data were recorded for one second at 240 Hz, with recording initiated during the drop. This allowed the subject time to land and stand still. The marker positions were used to estimate areas of the thigh. This gave a 6 by 3 grid of area sectors. These sectors were used to evaluate the redistribution of soft tissue during the impact within this single body segment.
The frequency content of the raw marker data and of the area data were examined. The spectral power for the force plate data, the sum of the marker data and the sum of the sectors data were calculated and the relative distribution of power among the frequencies were evaluated.
The marker data were used to examine the change in area of the sectors throughout the trial. Each sector's area was calculated assuming each sector was a trapezoid and stored in an array that mimicked the topography of the thigh. Contour plots of these data were then created to examine the extent to which parts of the thigh deformed with time.
The total area of the thigh varied from 94% to 106% of the initial area, suggesting little area was lost due to rolling out of the 2-D plane. Changes in the area of sectors ranged from around ± 10% to ± 20%, this corresponds to 6 cm 2 in the larger sectors. The change in area of a sector relative to its resting area was dependent on the location on the thigh and time. Figure 1 shows a contour plot of the thigh for one frame just before impact. The dark colors represent the areas with the highest decrease in area and the white represents the highest increase in area, relative to the rest area. The non-uniform motion of the soft tissue can be seen with the posterior portion decreasing in area and the central to lower-anterior portion increasing in area.
Figure 1: A contour plot of the thigh of the male subject showing change in area relative to normal standing. (Dark gray = large negative change. White = large positive change. Anterior is to the left.)
Table 1 shows the observable peaks in the frequency spectrum for the markers, the areas and the force plate for the female subject. It also indicates the percentage occurrence of these peaks as they are not present for all areas or markers. The presence or absence of peaks was dependent on the position of the data set on the thigh. Sectors which underwent greater changes in area had the more distinct peaks. Markers whose 1st derivatives at impact were greater also had the more distinct peaks, these tended to be towards the anterior portion of the leg. The frequency below which 96% of the spectral power was contained was 16.5 Hz for the area data, and 21 Hz for the marker data.
| Markers | Areas | Force Plate | |
|---|---|---|---|
| 1st Peak | 2 to 3 | 2 to 2.5 | 3 |
| % | (100) | (100) | (100) |
| 2nd Peak | 8 to 9 | 6 | 6 |
| % | (100) | (95) | (100) |
| 3rd Peak | 13 | 12 to 13 | 10 to 16 |
| % | (82) | (89) | |
| 4th Peak | 19 | 18 to 20 | |
| % | (7) | (66) | (100) |
Table 1: Peak frequencies in Hz and their percentage occurrence in the three data sets.
A 28 marker set on the thigh can provide information on the soft tissue motion during an impact. The deformation of the tissue can be tracked and quantified and the frequency of the motions calculated. The thigh can clearly be seen to act as a non-uniform non-rigid body with distinct regions of differing visco-elastic behavior. Although the marker movement had a higher frequency for the 96th percentile of power than the area movement, the area movement had distinct higher frequency peaks and higher power in frequency ranges as seen in the force plate data.
It is hoped that this information on changes in area can be extended to changes in volume and mass and used in conjunction with the frequency information to fill in the missing higher frequency components when reproducing ground reaction force. This would be an end in itself but would also help to validate quantification of intra-segment motion.
Bobbert, M. F. et al. J. Biomech., 12, 1095-1105, 1991.
Cappozzo, A. et al. Clin. Biomech., 2, 90-100, 1996.
Fuller, J. et al. Hum. Mvmt. Sci., 16, 219-242, 1997.
Sanders, R. et al. Int. J. Sports Biomech., 7, 330-343, 1991.
