Relationship between swimming velocity and trunk twist motion in short-distance crawl swimming

This study aimed to estimate the trunk twist angle from the shoulder and hip rotation angles in short-distance crawl swimming and to elucidate the twist motion of the relationship between the trunk and the rotation angular velocity in response to changes in swimming speed. Swimming speed during the experimental trials was computed from the subject’s best times in the 50 and 100m crawl swims. Wireless self-luminous LED markers were attached to seven locations on the body. The actual coordinate values of the LED markers were obtained using 18 cameras for underwater movements and 4 on the water for above-water movements. A comparison of the rate of change between trials revealed a high correlation (r = 0.722, p < 0.01) between the twist angle and shoulder rotation angular velocity in the push phase. In the same phase, a high correlation (r =0.748, p < 0.01) was also found between the twist angle and the angular velocity of hip rotation. These results suggest that swimmers increase the twist angle of their trunks to obtain a higher swimming speed. Moreover, the trunk muscle group increases its activity before starting the main motion of the twist back motion, and stretch-shortening cycle (SSC) motion may be performed in the trunk.


Introduction 4
A swimming race is a competition in which swimmers travel a defined distance from the 5 start to the finish line in a defined discipline and within the predefined time required to 6 reach the finish line. The race is divided into the start phase (i.e., 15 m from the start), the 7 turn phase (i.e., 5 m before the turn), the finishing phase (i.e., 5 m before the finish), and 8 the stroke phase (i.e., phases other than start, turn, and finish). The stroke phase is the 9 most common, and it is important to obtain a high swimming speed by upper limb strokes 10 to improve performance (Wakayoshi, 1992). There are four types of swimming strokes in 11 competitive swimming: butterfly, backstroke, breaststroke, and crawl. Crawl swimming 12 is considered the swimming stroke that can attain the highest swimming speed, and 13 requires the swimmers to maintain a horizontal position in the water and exert propulsive 14 force in an environment without a support point (Hollander et al., 1986). Relationship between swimming velocity and trunk twist motion in short-distance crawl swimming 5 a race. Focusing and by focusing on the changes in a swimming motion is expected to 1 discover findings that improve maximum swimming speed in short-distance events. In 2 crawl swimming, it is known that the propulsive force obtained by upper limb motion is 3 greater than that of other swimming strokes. It has been reported that upper limb motion 4 accounts for approximately 90% of the total whole-body propulsive force (Watkins et al., Furthermore, crawl swimming involves a rotational motion of the trunk around the long 7 axis in conjunction with upper limb motion, and this motion is defined as a trunk rotation 8 (Yanai, 2003). In crawl swimming, it has been reported that trunk rotation affects the 9 direction and speed of internal and external hand motions during the pull phase, which is 10 the phase before the hand enters the water and moves just below the shoulder (Grimston 11 & Hay, 1986;Payton et al., 2002). In a previous study that measured the fluid force 12 generated at hand using pressure distribution measurements at the swimmer's hand, it was 13 also reported that an increase in hand velocity contributes to an increase in the propulsive  angular velocity during a volleyball spike action. The same study also reported that the 10 angular velocity of shoulder rotation during the roll back phase of the spike motion affects 11 hand velocity. In crawl swimming, hand speed also affects performance. It is possible that 12 hand velocity, which is important for improving swimming speed, is affected by twisting 13 motion. However, no studies focus on the relationship between the angular velocity of 14 trunk rotation and trunk twisting motion in crawling swimmers. Furthermore, the 15 Relationship between swimming velocity and trunk twist motion in short-distance crawl swimming 8 parameters in these studies were only between subjects, and no previous studies have 1 focused on changes in swimming motion at different speeds. 2 The purpose of this study was to analyze three-dimensional motion during crawl 3 swimming using a motion capture system, to calculate the trunk twist angle from the 4 shoulder and hip rotation angles during short-distance crawl swimming, and to clarify the 5 relationship between the trunk twisting motion and the rotation angular velocity in 6 response to changes in swimming speed.  (Table 1) participated in this study (age: 21.0 ± 2.5 years, height: 179.8 ± 6.7 11 cm, weight: 77.3 ± 5.9 kg). All subjects trained six days a week and all had participated 12 in national competitions.  If the subjects were minors, their guardians were informed prior to the experiment that 4 they could withdraw their consent to participate in this study of their own free will at any 5 time. Even after informed consent, they would not be treated unfavorably because of such 6 withdrawal. In addition, we informed them that personal information obtained during the Measurement method 1 The experiment was conducted in a reflux water tank (Igarashi Industries, Japan), where 2 swimming speed could be controlled by artificially flowing water inside the pool. The 100-m crawl average swimming speed test ("V100 m") and a 50-m crawl average 10 swimming speed test ("V50 m") for 10 s each in the first and second sessions, respectively.

11
To prevent changes in swimming speed during the test, markers were placed at the bottom 12 of the pool so that the subject did not move from the measurement section. The subject's 13 kicking motion was the same as that in a race, i.e., a six-beat kick with six kicking motions 14 in one stroke. The test was performed without breathing to eliminate the effect of 15 Relationship between swimming velocity and trunk twist motion in short-distance crawl swimming 11 breathing on the stroke motion.  (for a total of 18 cameras used to capture underwater motion. Four cameras were set up 10 above the turning basin for the above-water motion to acquire real coordinate values for 11 LED markers in the calibration volume. The cameras were positioned at a sampling 12 frequency of 100 Hz. The angle of view was adjusted so that the markers attached to the 13 body could be photographed from two or more cameras (Fig 3). To avoid the 14 misidentification of the LED markers due to reflections from the water and glass surfaces 15 Relationship between swimming velocity and trunk twist motion in short-distance crawl swimming 13 during the test and calibration, the room was kept dark with curtains to block sunlight 1 from the outdoors. 2 For the analysis, image analysis software (VENUS 3D R, Nobby Tech Inc., Japan) was 3 used to define a virtual perpendicular line to the space from the coordinate values (X, Y) 4 on the two-dimensional plane using the angle and distance of the two-dimensional plane 5 obtained from each camera with an epipolar matching algorithm. The 3D space was 6 constructed from the 3D coordinate values of the body as follows. A fixed right-handed  (Fig 4). The obtained real coordinates were 11 smoothed using a low-pass filter to reduce noise by omitting frequencies above the 12 outliers.

13
Before the experiment, more than 1,500 samples were acquired with each camera by 14 waving the wand within the angle of view of each camera to perform dynamic calibration using a dedicated calibration wand to construct a three-dimensional coordinate space for 1 the measurement range. Dynamic calibration was performed separately underwater and 2 above water to eliminate light refraction from the water surface. After calibration, the 3 standard error of motion capture was less than 0.5 mm in and on the water. Then, a 4 calibration LED marker plate with three points arranged in an L-shape was floated on the 5 water's surface for a 30-s calibration period to combine the underwater and above-water 6 three-dimensional coordinate spaces. The water surface was identified by photographing 7 the marker positions from underwater and above water, and the origin of the space was 8 determined by aligning the coordinate axes of underwater and above water. Because the 9 plate has a thickness and there is a gap between the markers placed in the water and on 10 the water surface, the length of the gap was measured and reflected to fill in the difference 11 when aligning the coordinate axes. In addition, the calculation error was calculated by  The timing of the maximum shoulder and hip rotation angles was analyzed, abbreviated 3 as "peak time." The relative time (%) from the beginning of the phase to the maximum 4 rotation angle is shown for the phases in which the maximum rotation angle was observed.

5
The start of the phase was set at 0%, and the end of the phase was set at 100%. The timing 6 at which the shoulder rotation angle reached its maximum value (ShR peak time) and that 7 at which the hip rotation angle reached its maximum value (HiR peak time) were 8 calculated. To make comparisons at different swimming speeds, the time taken for the 9 target phase of each trial was normalized to 100%. The time from the beginning of the 10 phase to the maximum value of the rotation angle was expressed as phase percentage (%).     2.75, p < 0.05). There were no significant differences between trials in all phases for the   Table 3. Mean Values of ShR, HiR and TA Variables at V50 m and V100 m.  Table 4 shows the correlation coefficients between the percent change in TA at V50 m 1 relative to V100 m and the percent change in ShRAV and HiRAV for each phase. The    ShR for each trial was significantly lower for V50 m than for V100 m. No significant 10 differences in HiR were observed between the trials. In a previous study, Yanai (2003) 11 reported an association between ShR and SF, with ShR showing a decrease with 12 increasing SF. Although the set swimming speed in the present study was higher than that 13 in the previous studies because the set swimming speed was tested in the short-distance 14 events, SF was significantly higher in the V50 m and the ShR was lower, indicating that 15 Relationship between swimming velocity and trunk twist motion in short-distance crawl swimming 23 the results were similar in the short-distance events as in the previous studies of the set 1 swimming speed in the middle and long-distance events.   However, the test was conducted in a reflux water tank to control the swimming speed.
3 Therefore, the propulsive force may increase the shoulder rotation angular velocity in the 4 static water environment and a reflux water tank in the push phase. The results showed that the rotation 8 of the hips was greatly influenced by the kicking motion, which was performed with the 9 lower limbs, and showed that the direct relationship between the rotation of the hips and 10 the hand motion was small. Furthermore, in this study, the waist rotation was indirectly 11 related to ShRAV by changing the timing in the pull phase as swimming speed increased, 12 contributing to an increase in TA. These results inferred that hip rotation motion did not 13 directly contribute to hand propulsion but might have affected SF and hand propulsion by 14 influencing ShRAV. Therefore, in the future, a more detailed investigation is needed on the relationship between hip rotation and kicking motion, which is considered highly 1 related to hip rotation motion. the pull-roll-back phase and the rate of change in the push phase (Fig 6). In HiRAV, a 10 significant correlation was found between the rate of change in TA and the rate of change 11 in the push phase (Fig 7). Takahashi et al. (2018) reported that for the trunk twisting 12 motion, increasing the activity of the external obliques and the latissimus dorsi muscles, 13 which are the primary muscles of the trunk twisting motion, before the start of the main 14 motion, increases the moment of the shoulder exerted. A previous study by Hiruma et al. 1 of the trunk muscle groups, which contributed to an increase in ShRAV of the twist back 2 after ShR reached its maximum value. This study also showed a high correlation between 3 the change in TA and the rate of change in the pull roll back phase and the push phase, 4 which are the twist back phase in ShRAV, similar to previous studies that investigated the 5 trunk twisting motion in other sports. Therefore, it was inferred that an increase in TA 6 during short-distance crawling induced the SSC motion of the trunk muscle groups, which 7 affected ShRAV during the twist back phase (i.e., pull-roll-back phase and push phase) 8 after ShR reached its maximum value. 9 Furthermore, these results were related to hand velocity during the push phase, which is 10 considered to have a significant effect on the propulsive force during short-distance crawl reported that the rotation motion during crawling might be caused by the buoyancy 14 moment generated by the shoulder sinking into the water, suggesting that not only the 15 Relationship between swimming velocity and trunk twist motion in short-distance crawl swimming 28 SSC motion in the trunk muscle group but also the buoyancy peculiar to underwater 1 exercise may affect the rotation motion. Because only the propulsive phase was 2 considered in this study, the rotational moment generated around the long axis of the trunk, 3 which includes the recovery phase, was not taken into account. In the future, it is 4 necessary to investigate the relationship between the factors that cause trunk rotation, 5 including upper limb recovery motion, and the trunk muscle groups. In short-distance crawl swimming, it was suggested that to obtain a higher swimming 9 speed, the swimmer increases the trunk twist angle and increases the activity of the main 10 muscles, such as the external obliques and the latissimus dorsi muscles, before the start 11 of the main motion of the twist back motion, resulting in the SSC motion of the trunk 12 muscle groups. This may increase the rotation angular velocity by increasing the moment 13 exerted at the shoulder. In crawl swimming, it is important to reduce the total projected 14 area to reduce propulsive drag, and isometric training is widely used to improve trunk 15 Relationship between swimming velocity and trunk twist motion in short-distance crawl swimming 29 muscles. However, in short-distance crawl swimming, an increase in ShRAV has been 1 reported to be important because increased SF is a determinant of performance (Kudo et  The instructor should introduce training involving relaxation and contraction in the 6 muscle groups mobilized in the trunk twisting motion to increase not only the contraction 7 of the muscles but also the angular velocity of the twist back, which may be important in 8 achieving higher swimming speed in short-distance crawling. This study suggests that an increase in TA during short-distance crawl swimming induces 12 the motion of SSC in the trunk muscle groups, which affects ShRAV in the pull-roll-back 13 and push phases after ShR has reached its maximum value. Thus, to obtain a higher 14 swimming speed, swimmers increase their trunk angle and increase the activity of their 15 Relationship between swimming velocity and trunk twist motion in short-distance crawl swimming 30 primary muscles, such as the external obliques and the latissimus dorsi muscles, before 1 the start of the main motion of the twist back motion, which may induce the SSC motion 2 in the trunk muscle group. However, we did not examine the actual muscle activity of the 3 trunk muscles in this study, and it is necessary to investigate muscle activity during the 4 trunk rotation motion using electromyography in the future.