Cross-sectional and longitudinal research approaches: running pace and step characteristics of elite athletes in the 400-m hurdles

The aim of this study was to clarify the running pace and step characteristics among various competitive-level 400-m hurdlers through cross-sectional and longitudinal research approaches. We analysed spatiotemporal data for 13 male world-class and 14 male national-level 400-m hurdlers. We analysed 16.8 ± 4.2 races for each world-class hurdler and 20.0 ± 6.0 races for each national-level hurdler (the total number of analysed runs was 499) using publicly available television and internet broadcasts. Cross-sectional approach showed that positive relationships of finish time were highly obtained between both first- (r = 0.901) and latter-half split times (r = 0.914). In contrast, the first- and latter-half split times were not significantly correlated to SL and SF. A multiple single-subject approach showed that 14/27 hurdlers were identified as being latter-half speed reliant. In contrast, no hurdlers demonstrated first-half reliance. In the latter half of the race, 12/27 hurdlers were identified as being SF reliant; no hurdlers demonstrated SL reliance. In conclusions, important findings regarding high performance in a cross-sectional research approach do not always corresponded with those in a longitudinal research approach. Athletes and coaches should carefully improve performance in the first half of the race based on an individualization principle for training.


35
The 400-m hurdles is a long sprint running event. During the long sprint event, 36 the anaerobic glycolytic system is largely converted to the mechanical energy [1][2][3][4][5][6], 37 which can be externally and kinematically observed using a video camera [5,7]. In 38 particular, researchers can easily analyse spatiotemporal parameters because they can 39 calculate these parameters just by visually counting the number of steps in the fixed 40 moving distance. Therefore, a number of samples can be obtained, for example, through 41 public data analysis [8]. 42 In a 400-m hurdles race for men, a total of ten hurdles at 91.4 cm height are 43 positioned 35 m apart. Hurdlers are required to minimize decelerations into running 44 direction during clearing hurdles and to maintain high running speed during all inter-45 hurdle distances. Split times are often used to evaluate the running speed cross-46 sectionally and longitudinally and these are measured during the hurdles race using the 47 instant of touchdown for the leading leg [9]. Compared to split times determined by 48 touchdown method, it seems to be more difficult during competitive races that split times 49 are determined measured using video camera or photocell, which are the gold standard 50 methods for the experimental measurements [10]. This is because that detecting the 51 instant when the torso frontally passes one thin line is visually difficult using the panned 52 video camera, or the photocell device cannot be directly set close to athletes.

53
Race pace is a key strategy for a good finish time in a long sprint running event 54 [12]. A cross-sectional research reported that the faster the finish times of 400-m 55 hurdlers are, the faster touchdown split-times are during the latter half of the race [11]. 56 However, because "correlation does not imply causation," it is unclear whether the first-57 or latter-half split time is more important for each elite hurdler to improve the finish time.  Fast sprinting speed (short split time) is determined by high step frequency (SF) 61 and/or high step length (SL); therefore, these two step characteristics are often used to 62 evaluate why faster sprinters can run [13,14]  This study was laid out as an observational research design. We obtained the 100 movie data during competitive races from publicly available television and internet  , Table S1). In this study, races in which the individual 107 hurdler was likely to run with maximal effort through the finish line were analysed. We broadcasts used in this study. 114 We used the finish time as the official race time. We determined the split-times 115 for the first and latter halves of the race at the instant of touchdown for the leading leg 116 after clearing the fifth hurdle (hereafter, fifth touchdown).

117
We counted the total number of steps involved in clearing ten hurdles in the race We present all parameters mean ± SD. We checked all datasets for normality and Whitney's U tests. 159 We set statistical significance at P < 0.05.  (Table 1). Greater mean SF contributed significantly to the shorter mean first-166 half split times of the world-class hurdlers compared to those of the national-level athletes.

167
In contrast, no significant differences in the SDs for finish time, split time, and mean SF 168 in all trials were obtained between the two different groups. However, the SD for the 169 latter-half mean SL of the world-class hurdlers was significantly larger than that of 170 national-level hurdlers. As with the other parameters, the personal best times in the 400-171 m race of the world-class hurdlers were significantly shorter than those of the national- during the latter-half phase; SR first = mean step rate during the first-half phase; SL first = 180 mean step length during the first-half phase; SR latter = mean step rate during the latter-half 181 phase; SL latter = mean step length during the latter-half phase. *Significant difference from 182 national-level hurdlers (P < 0.05).

184
Positive relationships of finish time were highly obtained between first-and latter-185 half split times (Fig 1). In contrast, no significant relationships of split-time were 186 obtained between mean SF and mean SL in either the first or latter half of the race. The   during the latter-half phase; SR first = mean step rate during the first-half phase; SL first = 208 mean step length during the first-half phase; SR latter = mean step rate during the latter-half 209 phase; SL latter = mean step length during the latter-half phase. athletes did not favour either characteristic (Fig 2). In the first-half of the race, 19 of the 214 27 athletes were identified as being SF reliant, only one athlete demonstrated SL reliance, 215 and the remaining seven athletes did not favour either characteristic (Fig 3). In the latter 216 half of the race, 12 of the 27 athletes were identified as being SF reliant, no athletes 217 demonstrated SL reliance, and the remaining 15 athletes did not favour either 218 characteristic (Fig 4).   The multiple single-subject approach clarified that, for most hurdlers, the short 286 split-time was not related to high mean SL, but was associated with high mean SF during considered as more sensitive to mean SF rather than mean SL in a 400-m hurdles race.

291
On the other hand, in a 400-m race, the differences in running speed among different level 292 male sprinters are not caused by the differences in mean SF, but rather by the differences 293 in mean SL throughout the whole race [11]. In contrast, our cross-sectional approach 294 results did not support this finding: no relationships were observed between split times 295 and SL during both the first and latter halves of the race. This finding may be due to the 18 296 narrower range of finish times among our subjects compared to those in a previous study 297 [11], which does not have as significant an effect on the relationship. Moreover, this 298 difference might be due to the fact that some hurdlers could not run with an optimal 299 combination between mean SL and mean SF to achieve the highest running speed [19] 300 because they were forced to run between hurdles in a fixed step rhythm [20].

301
In the latter half of the race, although no significant difference in mean SL was 302 observed between the two hurdler groups, the within-subject SD of the SL of the world-303 class hurdlers was significantly greater than that of the national-level hurdlers. As 304 abovementioned, hurdles are not often able to run with an optimal SL for high running 305 speed because of the fixed inter-hurdle distance [20]. The 400-m personal records of 306 the world-class hurdlers were greater than those of the national-level hurdlers, 307 corresponding with the findings in a previous study [28]. Therefore, the SL of the 308 world-class hurdlers might not be more optimal for fast running compared with that of 309 the national-level hurdlers in the latter half of the race. This might lead the SL index of 310 the world-class hurdlers to be higher than that of the national-level hurdlers. Thus, our 311 third hypothesis was accepted.

312
From the results in this study, we can provide the following information as practical in the latter half of the 400-m hurdles rather than in the first half of the race. Second, 316 improving performance in the first half of the race should be considered based on the 317 individualization principle for training. 318 We have a limitation in this study: we could not measure the parameters to assess 319 biochemical energy expenditure. Long sprint running is performed by the glycolytic In conclusion, although the cross-sectional approach showed strong correlations between 329 the finish time and performance in the first and latter halves of the race, the multiple 330 single-subject approach showed latter-half performance is more essential for all hurdlers.

331
Therefore, the important findings regarding high performance in a cross-sectional study 332 approach do not always correspond with those in a longitudinal approach.