Journal Logo


The Timing of Thigh Muscle Activity Is a Factor Limiting Performance in the Deceleration Phase of the 100-m Dash


Author Information
Medicine & Science in Sports & Exercise: June 2022 - Volume 54 - Issue 6 - p 1002-1012
doi: 10.1249/MSS.0000000000002876
  • Open


Although the 100-m dash is the shortest sprint race in track events, the race can be divided into three phases (acceleration (Acc) phase, maximal speed (Max) phase, and deceleration (Dec) phase) (1). The running speed (RS) reaches maximum in the Max phase, the 50- to 70-m section (1–3). Interestingly, the maximum RS cannot be maintained, and RS gradually decreases toward the finish (the Dec phase) (1–3). Indeed, official race analysis of the men’s 100-m dash in the 2009 World Championships reported that all sprinters (i.e., data from 58 sprinters) showed deceleration in the final part (80- to 100-m section) of the race (4). Even when Usain Bolt won the final with the current world record of 9.58 s, he reached the maximal RS of 12.29 m·s−1, and his speed decreased to 11.96 m·s−1 in the final section (5). The main factor determining the 100-m dash record is, of course, the maximal RS (6,7), and many sprint training methods focus on how to improve the maximal RS (8,9). It is, however, also important for running a fast 100 m that the decrease in RS in the Dec phase be made as small as possible.

What happens in the Dec phase to cause the observed decrease in RS? In the world record race, Bolt showed a step frequency (SF) of 4.49 Hz over the 60- to 80-m section (the Max phase), which decreased to 4.23 Hz over the 80- to 100-m section (the Dec phase) (4). On the other hand, step length (SL) became longer in the Dec phase (2.85 m) than in the Max phase (2.77 m) (4). Six other finalists in the same race also showed decreases in SF in the Dec phase (4). Therefore, a decrease in RS in the Dec phase is inevitable even in world class sprinters, because of a decrease in SF. Hence, to lessen the inevitable decrease in RS in the Dec phase, sprinters should maintain the SF as high as possible. Endo et al. (10) investigated changes in the lower extremity movements in the Max phase (at 50 m) and the Dec phase (at 85 m) of maximal effort sprinting, and found that both RS and SF significantly decreased in the Dec phase (RS, 50 m: 10.14 m·s−1 vs 85 m: 9.90 m·s−1; SF, 50 m: 4.58 Hz vs 85 m: 4.40 Hz). Moreover, the hip flexion power also decreased in the Dec phase, accompanying delays in hip flexion torque and onset of energy absorption (10). These results suggest that a difference in the running movement (i.e., muscle activity timing) would cause the decreased SF in the Dec phase.

Various factors influence the SF (11). Previous studies indicate that the timing of thigh muscle activity is associated with SF (12–14). Chumanov et al. (14) reported that the onset of rectus femoris (RF) muscle activity occurred earlier in the running cycle with an intentional 5%–10% greater SF over what they would naturally do than the running with preferred SF. Moreover, the greater the RS and SF, the earlier the onset of RF excitation (13). The RF sustains the impact of the ground contact (15), whereas in the swing phase, the RF has a primary role in hip flexion, to swing the thigh forward, rather than in knee extension (16). On the other hand, the biceps femoris (BF) muscle, one of the antagonist muscles of the RF, shows strong activity with hip joint extension in the late swing and contact phases in association with greater ground reaction forces (15–18). Similar to the RF, the onset of BF activity also occurs earlier in running with higher SF (14). Therefore, the timing of thigh muscle activity seems to be an important determinant of the SF.

We have recently demonstrated that timing of the RF and BF activities influences the SF at the maximal RS (30- to 50-m section) (12). In particular, the higher the SF, the earlier the RF activity occurs in a running cycle. Moreover, the muscle activity of the RF (iRF) in the ipsilateral leg (i.e., swing leg) should occur earlier in relation to the activity of the contralateral leg (i.e., contact leg) BF (cBF) to obtain a higher SF. Furthermore, to obtain a higher SF, it is important to accomplish switching activities of hip flexor (RF) and extensor (BF) in the same leg separately, that is, without their co-contraction in the late swing phase (12).

Because SF is affected by the timing of RF and BF muscle excitation (12–14), the decrease in SF occurring in the Dec phase is related to a change in the timing of thigh muscle activity. Therefore, the purpose of this study was to analyze the changes in the timing of electromyography (EMG) activity of the RF and BF in both legs, together with spatiotemporal variables (RS, SF, and SL) in the Max and the Dec phases of the 100-m dash. No previous study has recorded muscle activity throughout an entire 100-m dash; therefore, our first task was to make a clear recording (Fig. 1). We hypothesized that the decrease in SF in the Dec phase would be related to a change in the timing of muscle activities measured with EMG. In particular, the timing of BF activity in a running cycle would be delayed in the Dec phase, because hip flexion and onset of energy absorption are delayed in the Dec phase (10). Likewise, the timing of RF activity in a running cycle would be delayed in the Dec phase, as was observed in the subjects with lower SFs (12).

Example of the four muscle raw EMG signals and spatiotemporal variables during the 100-m dash in the fastest subject (subject B, 10.80 s). Black and white blocks indicate contact and flight phases, respectively. Blue, red, green, and gold indicate iBF, iRF, cBF, and cRF activities, respectively. A portable wireless data logger (H = 12.3 cm, D = 2.0 cm, W = 6.8 cm, mass = 188 g) was carried by each subject on his lower back. We defined the Max phase as the 50- to 70-m section and the Dec phase as the 80- to 100-m section.


Experimental Protocol

Nine well-trained male Japanese track and field athletes (World Athletics (WA) score, 1045.4 ± 58.2 points; height, 1.75 ± 0.06 m; body mass, 67.4 ± 3.0 kg; age, 20.2 ± 1.2 yr) volunteered to participate in the study. Note that 1045 WA score points are equivalent to a record of 10.48 s in the men’s 100 m. The participants specialized in 100-, 200-, 400-, 400-m hurdles.

This study was approved by the Ethics Committees of Waseda University. All subjects were informed of potential risks associated with the experimental procedures. Before the experiments, all subjects gave their written informed consent. All experiments were conducted in accordance with the Declaration of Helsinki.

Experimental Protocol and Data Collection

After a self-selected warm-up activity, subjects ran one 100-m sprint with maximal effort from a crouching start with a starting block (NF155B; NISHI, Tokyo, Japan) in the straight lane of an official 400-m track. Each subject started to run when signaled by the sound of a track and field starting pistol (NG5085B; NISHI, Tokyo, Japan).

Surface muscle EMG data were sampled at 2000 Hz using wireless EMG sensors (Trigno Wireless Sensor; DELSYS, Natick, MA). A portable wireless data logger (H = 12.3 cm, D = 2.0 cm, W = 6.8 cm, mass = 188 g; Trigno Personal Monitor; DELSYS) was attached to the subject’s lower back to record EMG for the entire 100 m (Fig. 1). EMG data were recorded from the RF and BF muscles in both legs. We placed EMG sensors on the RFs at the midpoint on a conceptual line drawn from the anterior spina iliaca superior to the superior part of the patella, and on the BFs at the midpoint on a conceptual line between the ischial tuberosity and the lateral epicondyle of the tibia (19). Before the sensors were attached, the involved area of skin was shaved and treated with alcohol to reduce interelectrode impedance. EMG signals for the four muscles were checked after placing the loggers. To eliminate the influence of motion artifact as much as possible, they were fixed with surgical tape and underwrap tape. Two panning high-speed cameras (LUMIX DMC FZ-300; Panasonic, Osaka, Japan) were used to determine the moments of foot strike (FS) and foot off (FO) from the side of the running track at 240 Hz. Camera 1 was used to record the 0- to 50-m section, and camera 2 was used to record the 50- to 100-m section. Before starting a trial, to synchronize the FS and FO timing with the EMG data, the flash of the wireless all-around light presenter (Synchronizer; DKH, Tokyo, Japan) was recorded with the high-speed cameras. The EMG data and the signal of the synchronizing device were fed into a software analysis program (EMGworks®; DELSYS, Natick, MA).

Data Processing

Spatiotemporal variables

Reference markers were set every 10 m. The length of time it took for a subject’s torso to travel between one reference marker and the next marker 10 m away was defined as the split time (in seconds). Then, RS (in meters per second) was calculated by dividing 10 m by the split time. In addition, the contact time (CT; in milliseconds) and flight time (FT; in milliseconds) for each step were calculated from the number of frames of the high-speed cameras, and the SF (in hertz) was calculated for each step and then averaged. The SL (in meters) was calculated by dividing the RS by the SF. We defined the Max phase as the 50- to 70-m section and the Dec phase as the 80- to 100-m section based on the data from previous studies (1–5), and obtained mean spatiotemporal variables for the two phases (Max and Dec).

EMG data analysis

Based on our previous study (12), the EMG data were imported into the software analysis program and synchronized with the time axis of the camera image based on the time when the optical signal of the synchronizer was confirmed. In the time information of EMG data, the FS and FO times were obtained from the captured video images. In this study, we defined the leg having the first contact with the ground in an analysis section as the “ipsilateral leg” and the leg on the opposite side as the “contralateral leg.”

We defined a running cycle as the time from the moment of the ipsilateral leg FS (iFS) until the next ipsilateral leg FS (iFS2). The ipsilateral leg contact phase was defined from iFS to ipsilateral FO (iFO). The ipsilateral leg swing phase is separated into early, mid, and late as the periods from iFO to the contralateral leg FS (cFS), from cFS to the contralateral leg FO (cFO), and from cFO to iFS2.

To quantify the onset and offset timing of muscle activity, the EMG signal was processed with a Teager–Kaiser Energy Operator (TKEO) filter (20). The TKEO filter has been confirmed to be a reliable method for detecting EMG excitation (20,21). The discrete TKEO Ψ was defined as


where x is the EMG value and n is the sample number. The TKEO was applied after the EMG signal was band-pass filtered (20–450 Hz).

The EMG onset or offset threshold T was defined as


where μ is mean EMG signal during baseline and σ is SD of the EMG signal during baseline. The baseline was defined as the range of 0.05 s in which the SD of each EMG signal was the smallest in the running cycle. The preset variable h defines the level of the threshold, and was set at h = 15 (20). The EMG onset or offset timings were calculated both in the absolute time (Absolute EMG onset/offset timings [ms]) and in the normalized time (Relative EMG onset/offset timings [%]). Relative EMG timings were normalized as the percent of the running cycle (0%–100%). These variables were averaged over 5 cycles in an analyzed section to obtain a representative value for each subject. Subsequently, the band-pass–filtered and rectified EMG was low-pass filtered again at a cutoff frequency of 20 Hz to obtain a linear envelope of EMG activity (22).

Items calculated to obtain EMG timing

“Switch” and “Scissors” were defined and calculated as follows based on our previous study (12).

“Switch” of the RF and BF activity in the ipsilateral leg (Fig. 2A).

Definitions of the “Switch” (A) and “Scissors” (B).

In this study, to evaluate the timing of switching between ipsilateral leg RF (iRF) and ipsilateral leg BF (iBF) activities, the onset and offset timing of the iRF and iBF activities in the running cycle were determined.

  • iBF-offset: activity offset of ipsilateral leg BF
  • iRF-onset: activity onset of ipsilateral leg RF
  • iRF-offset: activity offset of ipsilateral leg RF
  • iBF-onset: activity onset of ipsilateral leg BF

Based on these values, “Switch” was defined as the length of time it took to switch between the two muscles (iRF to iBF or iBF to iRF) and was calculated as follows:


The calculation produced a negative Switch2 value for some subjects. This indicates co-contraction.

“Scissors” in the bilateral leg (Fig. 2B).

To analyze the coordination of contralateral BF (cBF) and iRF activities in both legs, the activities of the cBF onset and offset timings (%) were calculated as follows:

  • cBF-onset: activity onset of contralateral leg BF
  • cBF-offset: activity offset of contralateral leg BF

Length of “Scissors” is calculated as an index showing the coordination between the cBF as the hip extensor of the contact leg and the iRF as the hip flexor of the swing leg.


Statistical Analysis

A paired-samples t-test was used to determine the difference between the two phases (Max vs Dec). If the results were not normally distributed, the Wilcoxon signed rank test was applied. Because we recruited nine subjects without an a priori power analysis, we instead conducted a sensitivity analysis in G*Power, which indicated that an effect size of 0.91 would be necessary to obtain a power of 80% at an α of 0.05.

Moreover, to analyze which muscle activity timing affects the SF decrease in the Dec phase, we examined the Pearson’s product–moment correlation between the changes in SF (Δ SF) and EMG variables (Δ EMG) from Max phase to Dec phase, as follows:


Statistical analyses were performed using statistical processing software (SPSS ver.25; IBM, Armonk, NY). We set a significance level of 0.05.


Spatiotemporal variables

Measured 100-m time was 11.12 ± 0.27 s. Figure 2 shows spatiotemporal variables of the whole 100-m dash. RS increased sharply after the start, and it reached maximal (10.32 ± 0.23 m·s−1) over the 50- to 60-m segment (Fig. 3A). After that, RS gradually decreased toward the finish (9.99 ± 0.34 m·s−1 over the 90- to 100-m section). SF reached 4.5 Hz over the 0- to 10-m section and peaked over the 20- to 30-m section (4.69 ± 0.20 Hz). SF also gradually decreased toward the finish (Fig. 3B). On the other hand, SL and FT gradually increased toward the finish (Figs. 3B, C), whereas CT decreased sharply after the 10- to 20-m section and was minimized over the 40- to 50-m section (90.6 ± 6.4 ms; Fig. 3C).

Chronological change of the spatiotemporal variables during the 100-m dash. (A) RS, (B) SF, SL, and (C) CT and FT.

Table 1 shows the mean and SD of the spatiotemporal variables and the EMG variables in the Max phase and the Dec phase. RS and SF were significantly decreased in the Dec phase (RS: P < 0.001, dz = 2.26; SF: P < 0.001, dz = 2.55). In contrast, SL and FT were significantly increased in the Dec phase (SL: P = 0.031, dz = 0.60; FT: P < 0.001, dz = 1.67). However, there was no significant difference in CT between the two phases (P = 0.083, dz = 0.34). The relative lengths of the CT (%) and FT (%) were significantly different in the two phases (% CT: P = 0.031, dz = 0.69; % FT: P = 0.031, dz = 0.72). Moreover, the running cycle time was significantly increased in the Dec phase (P < 0.001, dz = 2.18).

TABLE 1 - Spatiotemporal and EMG variables (mean ± SD): difference between Max and Dec phases.
Variables Max (50–70 m), Mean SD Dec (80–100 m), Mean SD Paired t, P Effect Size, Cohen’s dz
Spatiotemporal variables
 RS (m·s−1) 10.30 ± 0.26 10.03 ± 0.33 <0.001 a 2.26 b
 SF (Hz) 4.65 ± 0.20 4.48 ± 0.22 <0.001 a 2.55 b
 SL (m) 2.22 ± 0.11 2.24 ± 0.13 0.031 a 0.60
 CT (ms) 91.9 ± 6.2 93.4 ± 7.4 0.083 0.34
 FT (ms) 123.8 ± 7.9 130.2 ± 7.9 <0.001 a 1.67 b
 Relative length of the CT (%) 42.6 ± 2.4 41.8 ± 2.2 0.031 a 0.69
 Relative length of the FT (%) 57.4 ± 2.4 58.2 ± 2.2 0.031 a 0.72
 Running cycle time (ms) 431.3 ± 18.9 447.2 ± 23.6 <0.001 a 2.18 b
Relative EMG onset/offset timings
 iRF-onset (%) 36.8 ± 2.2 38.9 ± 3.5 0.002 a 1.39 b
 iRF-offset (%) 72.8 ± 3.7 74.7 ± 3.8 0.008 a 0.99 b
 iBF-offset (%) 28.5 ± 4.7 31.8 ± 5.4 0.049 a 0.63
 iBF-onset (%) 73.8 ± 1.7 74.4 ± 3.7 0.304 0.16
 cBF-onset (%) 22.7 ± 1.3 24.0 ± 2.0 0.067 0.56
 cBF-offset (%) 79.2 ± 4.5 82.3 ± 4.3 0.017 a 0.84
 “Switch1” (%) 8.3 ± 5.1 7.1 ± 4.3 0.228 0.27
 “Switch2” (%) 1.0 ± 4.3 −0.3 ± 5.0 0.104 0.45
 “Scissors1” (%) 14.1 ± 2.8 15.0 ± 2.1 0.041 a 0.73
 “Scissors2” (%) 6.4 ± 4.3 7.6 ± 3.7 0.236 0.26
Absolute EMG onset/offset timings
 iRF-onset (ms) 159 ± 14 174.3 ± 21 <0.001 a 1.68 b
 iRF-offset (ms) 314 ± 23 334.2 ± 26 <0.001 a 1.95 b
 iBF-offset (ms) 123 ± 23 142.9 ± 30 0.011 a 1.19 b
 iBF-onset (ms) 318 ± 16 332.7 ± 21 0.019 a 1.16 b
 cBF-onset (ms) 98 ± 6 107.3 ± 11 0.009 a 0.98 b
 cBF-offset (ms) 341 ± 22 368.2 ± 28 <0.001 a 1.53 b
 “Switch1” (ms) 36 ± 22 31 ± 18 0.234 0.32
 “Switch2” (ms) 4 ± 19 −2 ± 23 0.082 0.62
 “Scissors1” (ms) 61 ± 13 67 ± 12 0.002 a 1.19 b
 “Scissors2” (ms) 27 ± 18 34 ± 17 0.157 0.25
aSignificant difference was observed (P < 0.05).
bDetected effect size (dz > 0.91).
Numbers in bold indicate significant P values.

EMG variables

Raw EMG signals from four muscles (RF and BF in both legs) and split times every 10 m as well as chronological change of the RS, SF, and SL during the entire 100-m dash in the fastest subject (subject B) are shown in Figure 3. Relative excitation timing (%) and actual averaged EMG waveforms (% maximal EMG amplitude) of the ipsilateral leg (i.e., iRF and iBF) and both legs (i.e., iRF and cBF) of all individual data and mean ± SD are presented in Figures 4 and 5, respectively. As for the iBF and iRF activity of the ipsilateral leg (“Switch,” Fig. 4), the iBF continued to be active until 28.5% ± 4.7% in the Max phase and 31.8% ± 5.4% in the Dec phase. A period followed in which neither iBF nor iRF was active in all subjects; that is, iRF began to be active after iBF completely terminated its activity. iRF was active from the early swing phase (Max: 36.8% ± 2.2% vs Dec: 38.9% ± 3.5%) to the late swing phase (Max: 72.8% ± 3.7% vs Dec: 74.4% ± 3.8%). Then, approaching the iFS2 (100%), iBF began to become active again in the late swing phase (Max: 73.8% ± 1.7% vs Dec: 74.4% ± 3.7%). The iRF and iBF of four subjects (A, B, F, G) showed co-contraction (negative “Switch2” values) in Max, whereas six subjects (A, B, F, G, H) showed co-contraction in Dec. This mean that the iBF became active before the iRF finished its activity. The remaining subjects showed clear switching from the iRF to the iBF (positive “Switch2” values).

Difference of onset and offset timing of iRF and iBF between Max (A) and Dec (B) phases of all individual data (bottom figure) and the mean ± SD (top figure). Note that these results correspond with Figure 6A. Blue and red indicate iBF and iRF activities, respectively. White indicates the time when both muscles are inactive. Purple indicates co-contraction.
Difference in onset/offset timings of iRF and cBF between Max (A) and Dec (B) phases of all individual data (bottom figure) and mean ± SD (top figure). Note that these results correspond with Figure 6B. Green and red indicate cBF and iRF activities, respectively. White indicates the time when both muscles are inactive.

Regarding iRF and cBF activity in both legs (“Scissors,” Fig. 5), the contralateral contacting leg BF (cBF) was active for a longer period than the ipsilateral swinging iRF both in the Max phase (22.7% ± 1.3% to 79.2% ± 4.5%) and in the Dec phase (24.0% ± 2.0% to 82.3% ± 4.3%) in most subjects. Subject D, however, showed an iRF-offset that occurred after the cBF-offset. The cBF was always activated before the iRF.

The relative EMG onset/offset timings (%) of RF and BF for Switch and Scissors are compared between the Max and Dec phases in Figure 6A. Note that they are based on the identical data as Figures 4 and 5. Significant differences were observed in iRF-onset (P = 0.002, dz = 1.39), iRF-offset (P = 0.008, dz = 0.99), iBF-offset (P = 0.049, dz = 0.63), and cBF-offset (P = 0.017, dz = 0.84). Moreover, a significant difference was observed in “Scissors1” (P = 0.041, dz = 0.73), whereas there was no significant difference between “Switch1” (P = 0.228, dz = 0.27), “Switch2” (P = 0.104, dz = 0.45), and “Scissors2” (P = 0.236, dz = 0.26; Table 1).

Difference in mean and SD of relative onset and offset timings of iRF, iBF, and cBF between Max and Dec (A). Blue and red indicate iBF and iRF activities, respectively. White indicates the time when both muscles are inactive. Purple indicates co-contraction. Green indicates cBF activity. White indicates the time when both muscles are inactive. Note that iRF and cBF excitation overlap. Similarly, absolute EMG timings also presented (B). *A significant difference was observed between Max and Dec. These results correspond with Table 1.

The absolute EMG onset/offset timings (in milliseconds) are also summarized in Table 1 and illustrated in Figure 6B. Naturally, we found that all muscle onset/offset timings significantly delayed in the Dec phase (iRF-onset: P < 0.001, dz = 1.68; iRF-offset: P < 0.001, dz = 1.95; iBF-offset: P = 0.011, dz = 1.19; iBF-offset: P = 0.019, dz = 1.16; cBF-onset: P = 0.009, dz = 0.98; cBF-offset: P < 0.001, dz = 1.53). Moreover, “Scissors1” had significantly prolonged in the Dec phase (P = 0.002, dz = 1.19).

In addition, we analyzed the correlations between ΔSF and ΔEMGs to elucidate which muscle activity timings were associated with the decrease in SF. As a result, significant negative correlations were found between ΔiRF-onset and ΔSF (P = 0.045, r = −0.597), and between ΔiBF-offset and ΔSF (P = 0.036, r = −0.626; Fig. 7). No significant correlation with ΔSF was obtained for other ΔEMGs (Fig. 7).

Scatter plot of the Δ SF and Δ EMG onset/offset timings (A), and Δ values of “Switch” and “Scissors” (B). There were significant negative correlations between ΔSF and ΔiRF-onset, and between ΔSF and ΔiBF-offset.


The primary purpose of this study was to clarify the differences in the timing of RF and BF muscle activities in both legs between the Max (50–70 m) and Dec (80–100 m) phases. To the best of our knowledge, this is the first attempt to record EMG activity throughout an entire 100-m dash; we were successful, thanks to the recent development of new recording techniques (Fig. 1).

Both RS and SF were significantly decreased in the Dec phase. On the other hand, FT and SL were significantly increased in the Dec phase. These results were consistent with previous studies (1,4,5). Therefore, decreased RS in the Dec phase may have been caused by the change in SF, not in SL. Furthermore, the relative lengths (%) of CT and FT significantly differed between the Max and Dec phases. These results mean that the spatiotemporal variables of the Dec phase are not just a linear scaling of the Max phase.

In the present study, it was hypothesized that the timing of thigh EMG activity in the Dec phase would change from the Max phase together with a decrease in SF, because SF was shown to be associated with changes in thigh EMG activity timing in the previous study (12). It is natural that absolute EMG onset/offset timings delayed in the Dec phase (Fig. 6B), because decreasing SF means increasing running cycle time. Interestingly, present result showed that relative EMG timings (%), that is, iRF-onset, iRF-offset, iBF-offset, and cBF-offset, also significantly delayed despite the normalized time by running cycle (Fig. 6A). Moreover, “Scissors1” was also significantly prolonged in the Dec phase. These results reveal that muscular coordination patterns between the Dec phase are not just a linear scaling of the Max phase. Furthermore, change in EMG timings (ΔiRF-onset and ΔiBF-offset) has significant negative correlations between change in SF (Fig. 7). In other words, a greater change in EMGs has a greater change in SF. We think these results also mean that muscular coordination patterns differ between two phases and influence the decreasing SF in the Dec phase.

RF activity

Especially, both relative iRF-onset and iRF-offset timings showed larger than detected effect size (dz > 0.91). The RF acts as a hip flexor during the swing phase to swing the thigh forward (16). However, both relative iRF-onset and iRF-offset timings were seen later in the running cycle in the Dec phase, which would produce a delay in the leg recovery in the swing phase. Previous research has shown that thigh angle decreased in repeated sprinting (i.e., in a condition of fatigue) (23). Moreover, Endo et al. (10) pointed out that one deceleration factor in the 100-m dash is delay of the hip flexion movement (i.e., recovery mechanics) in the swing phase. One possible explanation for these findings is that RF activity in the swing leg occurred later in the Dec phase (Fig. 6 and Table 1). If the movement of the swing leg was delayed, swing time (i.e., FT) would be prolonged, and thus, SF would decrease. Indeed, significant negative correlation was found between ΔiRF-onset and ΔSF (Fig. 7). Hence, the delayed timing of RF activity would be a factor in the decreased SF.

BF activity

Although the RF acts as a hip flexor in the swing phase, the BF acts as a hip extensor during the late swing phase and the contact phase to propel the body (15,17,24–26). Hip extensor muscles (e.g., BF) have been regarded as important for swinging the leg backward to obtain large ground reaction forces (18). We observed longer activity (later timing of the offsets) in both relative iBF and cBF timings during the contact phase of the Dec phase than in the Max phase; this longer activity would produce greater hip extension to push longer against the ground. Although no significant difference in CT was observed between the Max and Dec phases, FT was significantly increased in the Dec phase. Athletes should swing a leg forward as soon as it finishes pushing against the ground (27). However, longer BF activity would cause greater thigh extension, placing the lower limb far behind the total body center of gravity. This would require a greater range of lower-limb motion for the next ground contact, which inevitably increases FT in the Dec phase. To sum up, in the Dec phase, SF decreases with the increase in FT that is caused by the delayed timing of BF activity. Thus, the elongation in the BF activity period would also be a factor for the decreased SF. Moreover, in the individual level analysis, we found significant negative correlations between ΔiBF-offset and ΔSF. This suggests that the change in these EMG activities were certainly related to the change in SF, thus supporting the result obtained in the group data.

Interleg muscular coordination: “Scissors”

As for interleg muscular coordination, the time difference between the onsets of iRF activity in the swing leg and cBF activity in the contact leg was negatively correlated with SF in the maximal speed phase (“Scissors1”) (12). On the other hand, the time difference between offsets of iRF and cBF (“Scissors2”) had a positive relationship with SF in the maximal speed phase (12). In other words, swing leg activity (iRF) should be activated earlier relative to contralateral leg activity (cBF) to obtain higher SF. Therefore, we focused on “Scissors1” and “Scissors2” as the indicators of interleg muscle coordination, and hypothesized that “Scissors1” would be prolonged and “Scissors2” would be shortened in the Dec phase. Indeed, in parallel with the decrease in SF in the Dec phase, relative length of the “Scissors1” became significantly longer in the Dec phase than in the Max phase; however, it was not beyond the detected effect size (Cohen’s dz; “Scissors1” [0.73]). However, “Scissors2” did not change between the two phases. Indeed, we did not observe significant correlations between ΔSF and Δ“Scissors1,” and between ΔSF and Δ“Scissors2.” This result imply that onset timing of recovery muscle activity (iRF) is delayed relative to the contact leg muscle activity (cBF) in the Dec phase in group-level analysis. Hence, our second hypothesis was partially supported; however, whether interleg muscle coordination will associate decreasing SF was not accepted in the individual level.

Samozino et al. (28) pointed out that the limiting factors of cycling power and speed outputs were related to muscle coordination (EMG onset and offset). They concluded that suboptimal muscle coordination deteriorates sprint cycling performances. The present study showed a similar phenomenon; interleg muscle coordination changed in the Dec phase. A previous study demonstrated that sprinters had higher SF because their specific running technique minimizes the distance between the positions of both knees at the time of ground contact compared with distance runners (29). Several studies have pointed out the importance of recovery mechanics (swinging the thigh forward) in obtaining a higher SF (8,11,27). If the recovery movement is delayed, the duration of swing time will be prolonged, inducing a decrease in SF. Our results suggest that the disturbed coordination of the iRF and cBF would influence the decreased SF in the Dec phase of the 100-m dash.

Switching agonist–antagonist muscle activity: “Switch”

Moreover, it is important to accomplish the switch between hip flexion (iRF) and hip extension (iBF) of the same leg at the late swing phase (“Switch2”) smoothly (without minimal co-contraction) to obtain higher SF (12). Although we observed six subjects (A, B, F, G, H) who exhibited co-contraction (negative “Switch2” values) in the Dec phase (Fig. 4B), neither “Switch1” nor “Switch2” was significantly different between two phases. Moreover, there were not significant correlations between ΔSF and Δ“Switch1,” and between ΔSF and Δ“Switch2.” Therefore, the question of whether co-contraction is more likely to occur and how it might be associated with the decreased SF in the Dec phase remains for future studies to explore.


Finally, the SF depends not only on neuromuscular factors but also on physiological factors (e.g., the energy system) (11). Further study is needed to further elucidate the mechanism of deceleration during the 100-m dash. Present study showed that iRF-onset, iRF-offset, iBF-offset, and cBF-offset relative timings shifted to later in the running cycle, and the interleg muscle coordination, that is, “Scissors1,” significantly prolonged in the Dec phase. Moreover, change in EMG timings (ΔiRF-onset and ΔiBF-offset) has significant negative correlations with change in SF. These differences of muscle timings and coordination estimated that some of them have an effect on the decrease in the SF in the Dec phase. In the Dec phase, the interleg coordination changes, which may lead to a decrease in the SF. Therefore, we suggest that it is important to train so that the interleg coordination is as little impaired as possible in the Dec phase. Training menu such as the dash longer than 100 m would be beneficial for building up the ability to maintain speed in the final part of the 100-m race.

The present result will provide coaches and athletes with a new insight; interleg muscular coordination could be one of the key factors in maintaining higher SF during the Dec phase of the 100-m dash. This has practical implication for both coaches and athletes who should consider the concept of interleg muscle activity training.


The present study analyzed the data on the premise that the RF acts for hip flexion and the BF acts for hip extension. Anatomically, because both RF and BF are biarticular muscles, their activity could be modified by other joint and segment interactions. Furthermore, synergist muscles other than the BF, such as the gluteus maximus and semitendinosus muscles, are also involved in the hip extension movement (25). Indeed, studies on multichannel surface EMG have suggested that the same muscle may act differently along proximal to distal sites (30). Therefore, it is not possible to conclusively explain the thigh movements by only the activity of RF and BF. In addition, the present study has the limitation of not being able to elucidate the interaction between EMG activity and the running movement because the running kinematic data were not obtained. Recently, the results of the study by Clark et al. (31) revealed that faster RS was achieved with higher SF and greater total amplitudes of thigh angular motion, resulting in greater thigh angular velocities. Moreover, at the top speed, the slope of the thigh angular position versus time curve was steeper for the sprinter than for the recreational athletes. Therefore, further investigation of the relationship between “Scissors” and “Switch” obtained from EMG activity and actual thigh movements will be necessary.

Furthermore, there is an electromechanical delay (EMD) between the onset of EMG activity and the exertion of force (32). EMG does not always correspond to actual movements, so we need to be cautious in interpreting the present results. We utilized telemetric sensors with an interelectrode distance of 10 mm, which avoided the problem of crosstalk. As for the problem of “representativeness,” it cannot be solved with the recording system used in the present study. Of course, it is possible to put several sensors on the RF and BF to monitor the activities in different portions of the muscles. However, it is impractical for a subject to carry, say, 3 sensors on each muscle, in total 12 on both legs, because the bulkiness would interfere with the normal movement of the thigh. In addition, in such a scenario, it could not be guaranteed that each sensor does not pick up the activity of the nearby muscle (crosstalk). Therefore, with the currently available apparatus, one sensor on each muscle was the point of compromise, because the first priority of this study was to record EMG activity throughout an entire 100-m dash in as natural a condition as possible. We always tried to locate a sensor in the middle of a muscle. We believe that the data thus obtained will be able to be compared with future data recorded using more advanced technology. However, we believe that understanding the timing of thigh muscle activity in both legs during the 100-m dash will provide meaningful insights for both athletes and coaches, leading to improvements in training programs.

Unfortunately, we could recruit only a limited number of 100-m specialists. Therefore, this study focused on athletes who specialize in short sprints and hurdles who train at least 5 d·wk−1. Interestingly, however, the decrease in RS in the final part of the 100 m was observed not only in the 100-m specialists but also in the other subjects, even in the 400- or 400-m hurdles athletes. Therefore, we believe that the existence of the Dec phase is a more generalized phenomenon that limits human running ability.

Finally, the main finding that EMG variables and sprint spatiotemporal variables significantly differed between the Max and Dec phases was reached based on a group-level analysis. In the individual analyses, ΔEMG (ΔiRF-onset and ΔiBF-offset) and ΔSF had significant negative correlations. However, because correlation does not necessarily indicate causation, further studies with more and a greater variety of subjects will be needed to conclude whether the change in EMG activities obtained in this study really produced the change in SF/RS and whether this effect is specific to sprinters.


In the present study, we mainly focused on EMG activity during the 100-m dash and succeeded in making clear recordings throughout the entire period. We analyzed the differences in timing of RF and BF activities in both legs, together with spatiotemporal variables, in the Max and Dec phases. RS decreased in the Dec phase with decreased SF. Relative timing of iRF-onset, iRF-offset, iBF-offset, and cBF-offset shifted to later in the running cycle in the Dec phase than in the Max phase. Moreover, the time difference between EMGs of the swing leg (iRF-onset) and the contact leg (cBF-onset; “Scissors1”) was greater in the Dec phase than in the Max phase. Furthermore, change in EMG timings (ΔiRF-onset and ΔiBF-offset) has significant negative correlations between change in SF. In other words, a greater change in EMGs have a greater change in SF. We concluded that the timings of thigh muscle activity are a factor limiting performance in the deceleration phase of the 100-m dash.

We thank Dr. Candace O’Connor for editing the English in the manuscript.

This study was supported by JSPS KAKENHI (grant numbers (JP19K19957, JP19K22822) and Waseda University Grant for Special Research Projects (2020C-415).

There are no conflicts of interest. The results of the present study do not constitute endorsement by the American College of Sports Medicine. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.


1. Ae M, Ito A, Suzuki M. The men’s 100 metres. New Stud Athletics. 1992;7(1):47–52.
2. Bae YS, Park YJ, Park JJ, Lee JS, Chae WS, Park SB. Biomechanics Research Project in the IAAF World Championships Daegu 2011. Biomechanics Research Project at IAAF World Championships in Athletics—Daegu. Daegu (Korea): Korean Society of Sports Biomechanics; 2011:13–23.
3. Bissas A, Walker J, Tucker C, Paradisis DG. Biomechanical Report for the IAAF World Championships London 2017: 100 m Men’s. Monte Carlo (Monaco): IAAF; 2017. pp. 1–58.
4. Hommel H, Badura M, Böttcher J, et al. Biomechanical analyses of selected events at the 12th IAAF World Championships in Athletics, A Project by the German Athletics Federation, Berlin. Darmstadt (Germany): Deutscher Leichtathletik-Verband; 2012. pp. 1–15.
5. Graubner R, Nixdorf E. Biomechanical analysis of the Sprint and hurdles events at the 2009 IAAF World Championships in Athletics. New Stud Athletics. 2011;26(1/2):19–53.
6. Slawinski J, Termoz N, Rabita G, et al. How 100-m event analyses improve our understanding of world-class men’s and women’s sprint performance. Scand J Med Sci Sports. 2017;27(1):45–54.
7. Healy R, Kenny IC, Harrison AJ. Profiling elite male 100-m sprint performance: the role of maximum velocity and relative acceleration. J Sport Health Sci. 2022;11(1):75–84.
8. IAAF. RUN! JUMP! THROW! The Official IAAF Guide to Teaching Athletics. IAAF Coaches Education and Certification System. Lincolnshire (United Kingdom): Warners Midlands plc; 2009. pp. 5–11.
9. Haugen T, Seiler S, Sandbakk O, Tonnessen E. The training and development of elite sprint performance: an integration of scientific and best practice literature. Sports Med Open. 2019;5(1):44.
10. Endo T, Miyashita K, Ogata M. Kinetics factors of the lower limb joints decelerating running velocity in the last phase of 100 m race. Jpn J Phys Educ Health Sport Sci. 2008;53(2):477–90.
11. Ross A, Leveritt M, Riek S. Neural influences on sprint running. Sports Med. 2001;31(6):409–25.
12. Kakehata G, Goto Y, Iso S, Kanosue K. Timing of rectus femoris and biceps femoris muscle activities in both legs at maximal running speed. Med Sci Sports Exerc. 2021;53(3):643–52.
13. Andersson EA, Nilsson J, Thorstensson A. Intramuscular EMG from the hip flexor muscles during human locomotion. Acta Physiol Scand. 1997;161(3):361–70.
14. Chumanov ES, Wille CM, Michalski MP, Heiderscheit BC. Changes in muscle activation patterns when running step rate is increased. Gait Posture. 2012;36(2):231–5.
15. Nummela A, Rusko H, Mero A. EMG activities and ground reaction forces during fatigued and nonfatigued sprinting. Med Sci Sports Exerc. 1994;26(5):605–9.
16. Mero A, Komi PV. Electromyographic activity in sprinting at speeds ranging from sub-maximal to supra-maximal. Med Sci Sports Exerc. 1987;19(3):266–74.
17. Mero A, Komi PV, Gregor RJ. Biomechanics of sprint running. A review. Sports Med. 1992;13(6):376–92.
18. Morin JB, Gimenez P, Edouard P, et al. Sprint acceleration mechanics: the major role of hamstrings in horizontal force production. Front Physiol. 2015;6:404.
19. Hermens H, Freriks B, Merletti R, et al., editors. European recommendations for surface electromyography: results of the SENIAM Project. 1999.
20. Solnik S, Rider P, Steinweg K, DeVita P, Hortobagyi T. Teager–Kaiser energy operator signal conditioning improves EMG onset detection. Eur J Appl Physiol. 2010;110(3):489–98.
21. Li X, Zhou P, Aruin AS. Teager-Kaiser energy operation of surface EMG improves muscle activity onset detection. Ann Biomed Eng. 2007;35(9):1532–8.
22. Yu B, Queen RM, Abbey AN, Liu Y, Moorman CT, Garrett WE. Hamstring muscle kinematics and activation during overground sprinting. J Biomech. 2008;41(15):3121–6.
23. Pinniger GJ, Steele JR, Groeller H. Does fatigue induced by repeated dynamic efforts affect hamstring muscle function?Med Sci Sports Exerc. 2000;32(3):647–53.
24. Schache AG, Dorn TW, Blanch PD, Brown NA, Pandy MG. Mechanics of the human hamstring muscles during sprinting. Med Sci Sports Exerc. 2012;44(4):647–58.
25. Howard RM, Conway R, Harrison AJ. Muscle activity in sprinting: a review. Sports Biomech. 2018;17(1):1–17.
26. Dorn TW, Schache AG, Pandy MG. Muscular strategy shift in human running: dependence of running speed on hip and ankle muscle performance. J Exp Biol. 2012;215(Pt 11):1944–56.
27. Mann R. The Mechanics of Sprinting and Hurdling. Part 5: Chapter 12. Front Side and Back Side Sprint Mechanics. Scotts valley (CA): CreateSpace Independent Publishing Platform; 2011. pp. 133–41.
28. Samozino P, Horvais N, Hintzy F. Why does power output decrease at high pedaling rates during sprint cycling?Med Sci Sports Exerc. 2007;39(4):680–7.
29. Bushnell T, Hunter I. Differences in technique between sprinters and distance runners at equal and maximal speeds. Sports Biomech. 2007;6(3):261–8.
30. Watanabe K, Vieira TM, Gallina A, Kouzaki M, Moritani T. Novel insights into Biarticular muscle actions gained from high-density electromyogram. Exerc Sport Sci Rev. 2021;49(3):179–87.
31. Clark KP, Meng CR, Stearne DJ. ‘Whip from the hip’: thigh angular motion, ground contact mechanics, and running speed. Biol Open. 2020;9(10):bio053546.
32. Vigotsky AD, Halperin I, Lehman GJ, Trajano GS, Vieira TM. Interpreting signal amplitudes in surface electromyography studies in sport and rehabilitation sciences. Front Physiol. 2017;8:985.


Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American College of Sports Medicine.