Make your own free website on
"The Effect of Forward and Backward Pedaling on
Leg Muscle Recruitment and Energy Metabolic
Variables In Stationary Bicycling"

The bicycle is a human-powered vehicle. It has evolved as a means, not only for human transportation, but as a racing machine and a toy for children.  The bicycle has become a tool for health and physical fitness development and maintenance. Cycling can be seen in cardiac and neuromuscular rehabilitation centers around the world. Developing finely machined bicycles for various and specific applications is in  increasing demand.

Further bicycle development is a task for ergonomics - the application of biological and engineering data for the mutual adjustment of human and machine. Improved bicycling performance is affected by the interaction of environmental, mechanical, and human variables. Greater bicycle performance and efficiency is achieved through mechanical engineers, while exercise scientists examine cycling performance from the human perspective. To bridge this gap and to develop bicycles more effectively requires continuous communication and cooperation between these two research fields. Variables such as body position, configuration, seat height, the interaction of workload, power output and pedaling rate, alters joint angles, muscle lengths, muscle movement and arm lengths, thus affecting muscle activity and results in measurable changes.

Recently, a bi-directional bicycle gear was developed by which a person was able to pedal both forward and backward. Although there is very little research on backward bicycle pedaling, there are several analogous studies in which the difference of forward and backward walking, running, and wheelchair use was compared. Some researchers (Thorstensson, 1986; Winter et al., 1989) were interested in the difference of muscle excitation patterns between forward and backward walking. Among them, Winter et al. (1989), strongly suggested that muscle activation during backward walking was not similar to that of forward walking.

There were also several reports for rehabilitative merits in backward walking and running. Mackie and Dean (1984), Flynn and Soutas Little (1991) reported that backward running was effective in rehabilitation because knee joint stability was better and compressive forces at patellofemoral joints were less than that of forward running. It also proved that backward running increases quadriceps' power and strength (Mackie and Dean, 1984; Threlkeld et al., 1987).

With respect to cardiopulmonary responses to forward and backward walking and running, Flynn et al. (1994) reported that minute ventilation, oxygen consumption, RER and heart rate, and blood lactate levels were significantly higher during the backward walking and running than their forward counterparts at comparable speed. They also noted shorter stride length and
greater stride frequency during backward walking and running. There have been discussions about the physiological disadvantage of conventional, handrim wheelchair propulsion. Linden et al. (1993) examined the physiological comparisons of forward and reverse wheelchair ergometry, and concluded that the lower V02, YE and FIR responses for reverse exercise at given submaximal power output levels would be due to primary involvement of large back muscles.  While forward propulsion relies predominantly on the relatively small muscle groups of the chest and anterior shoulder .

The purpose of this study was to compare the differences in muscle activity and metabolic responses between forward and backward bicycle pedaling, and to detect the advantages and disadvantages of both pedaling methods.

The subjects for the tests were nine healthy male students, whose majors were physical education. These students voluntarily participated in both forward and backward pedaling. They did not have any medical problems and each subject signed a statement of informed consent. To exclude the effect of bicycle training, we did not select those accustomed to bicycle riding in their daily life.

Experimental Procedure
Subjects practiced adaptive backward pedaling for 10 minutes, and forward pedaling for 5 minutes, pedaling every day for 3 weeks.

During this time we focused not only on the adaptation to backward pedaling but on the rhythmic crank movement, which was to be maintained at 50 rpm. The bicycle used for both the adaptive practice and the main experiment was a stationary ergometer (Monark 818E, Sweden) in which the newly developed MBI bi-directional gear was installed to allow both forward and backward pedaling.

The gear ratio of that device was 52:14, equally in both cases of pedaling. The subjects were randomly divided into 2 groups (group A; n=5, group B; n=4). The first day, members of group A and group B were tested by doing forward and backward respectively, and vice versa the next day. Subjects took a 20 minute supine rest and a 2 minute warm-up of pedaling prior to the test. Initial workload was 50W (lkp, 50 rpm) and increased by 25W every 4 minutes. The test was terminated at the end of the 4th stage with the total 16 minutes of pedaling time.

Measurement of Variables
For EMG analysis, vastus lateralis, biceps femoris, gastrocnemius and anterior tibialis were selected, and the electrodes were attached on the site, five-finger-width proximally from patella, the midpoint between ischial tuberosity and fibula head, one-hand-width distally from popliteal, four-finger-width distally from tibial crest, and laterally one-finger-width from that point, respectively. Subjects were prohibited from detaching the electrodes until the test was completed.

The Biopac system and its Acq program (USA) were used for collecting and analyzing EMG and goniometer signals. EMG raw signals from the 4 muscles were digitized at 2000 samples per second, in the frequency range of 100Hz-400Hz, during the last 10 seconds of every stage. For detecting on and off site of the 4 muscles' activity during one crank cycle, the change of knee joint angle measured by goniometer (Penny & Gile, LJK) was prematched with the crank angle. Seven raw EMG signals of each muscle were filtered, rectified, and integrated and then this integrated EMG was divided by 7 to make IEMG/cycle for detecting the amount of muscle activity. Also mean power frequency was measured by the fast Fournier transformation (FFT) method for detecting muscle fatigue.

A combined system of QMC and Q4500 (Quinton, USA) was used for measuring cardiorespiratory variables. Oxygen consumption was automatically calculated every 20 seconds by the mixing chamber method. Blood lactate concentration was calculated by lactate analyzer (1500 sport, YSI, USA). 25 ml of whole blood was sampled from the fingertip at the end of every stage and it was injected into the analyzer.

Work efficiency was calculated by the following equation; mechanical work performed / physiological energy consumption X 100. Physiological energy consumption was calculated from respiratory variables including oxygen consumption, gas exchange ratio, and caloric equivalent. Opto-sensor and reflective marks which were attached on the circumference of the ergometer wheel counted the revolutions and connected with the PC program (SNU extended) to calculate mechanical work.

Statistical Analysis
In this study we hypothesized that muscle recruitment and energy metabolic variables would be different between forward and backward pedaling. Independent variables were the two pedaling types, and muscle activity pattern during one crank cycle.  The amount of muscular activity, muscular fatigue, oxygen consumption, work efficiency, and blood lactate levels were the dependent variables.

Each on and off site of the 4 muscles' activity during one crank cycle of the 9 subjects was averaged, and the difference between the two pedaling types was qualitatively compared.

A paired t-test was done to compare the difference of dependent variables between the two pedaling types at each workload. To test the change of dependent variables with increasing workload in each pedaling type, a one-way repeated ANOV A was performed. The difference of muscle activity and muscle fatigue between the 4 muscles was tested by a one-way ANOV A. For statistical significance the level 0.05 was selected.


Muscle Activity Pattern
In forward pedaling, as is the case shown in Fig .1, vastus lateralis began activation at the point of about 322 degrees, which is where the pedal reached prior to the top of dead center (TPC), and continued to contract until about 124 degrees, the latter part of the power phase of cycling. Biceps femoris acted from the point of 45° to 124°, simultaneously with vastus lateralis, and continued to act to the point of 207 degrees, passing through the bottom of dead space. Gastrocnemius acted from the point of 43° to 204°, and this almost corresponded to the activation range of biceps femoris. Anterior tibialis began to act from the point of 258° to 341°, by passing the activation at point of vastus lateralis.

Fig. 2 shows the muscle activity pattern in backward pedaling. Vastus lateralis began to act at the point of 321° and end at 174°. Biceps femoris acted from the point 101° to 345°. Electromyographical activity was shown from the point of 250° to 136°, and from 84° to 337° in the case of gastrocnemius and anterior tibialis, respectively.

Comparing the two cases, we found that vastus lateralis, gastrocnemius, and biceps femoris had a synergistic function in the power phase during forward pedaling.  On the other hand, during backward pedaling, vastus lateralis mostly alone pushed the crank down. Anterior tibialis acted in the recovery phase in both forward and backward pedaling..

As shown in Table 1, integral EMG of tibialis anterior was higher in backward pedaling than in forward pedaling at all workloads, but only at the load of 100W and 125W were there significant differences between the two types. IEMG was significantly changed with increasing workload only in backward pedaling. There were no significant differences of IEMG between the two pedaling types in gastrocnemius, and only during backward pedaling there were significant changes of IEMG with increasing workload. As in the case of gastrocnemius, BF did not show any differences in IEMG between the pedaling types at all workloads except 125W. Significant change with increasing workloads was shown only in backward pedaling. In the case of vastus lateralis, significant differences of IEMG between the two pedaling types were most predominant among the 4 muscles. At the workload of 75W, 100W, and 125W, testing showed significant difference of IEMG between the two pedaling types, and significant change was shown in both forward pedaling and backward pedaling. Table 1 also shows that ZIEMG was significantly different between the two pedaling types at the workload of 100W and 125W, and that ZIEMG of both types significantly changed with increasing workload.

Table 1. The comparison of IEMG between Forward and Backward pedaling and the change of IEMG in each pedaling type.

14.39 +/- 3.38
16.96 +/- 8.86
13.64 +/- 2.01
15.94 +/- 5.57
13.48 +/- 1.69
17.04 +/- 4.86 (+)
14.66 +/- 2.15
22.07 +/- 6.43 (+*)
11.13 +/- 2.45
9.94 +/- 1.39
10.91 +/- 2.40
9.63 +/- 1.51
19,92 +/- 3.38
10.67 +/- 2.74
11.20 +/- 2.79
12.01 +/- 2.53 (*)
5.72 +/- 1.76
5.66 +/- 1.88
5.63 +/- 1.15
6.19 +/- 1.83
6.09 +/- 1.46
6.65 +/- 1.86
6.20 +/- 1.46
9.35 +/- 4.48 (+*)
9.61 +/- 0.81
10.91 +/- 1.76
10.81 +/- 1.36
14.07 +/- 3.84 (+)
12.48 +/- 2.84
15.37 +/- 3.49 (+)
14.35 +/- 3.18 (*)
19.13 +/- 5.89 (+*)
40.85 +/- 6.52
43.47 +/- 6.52
40.99 +/- 5.30
45.83 +/- 8.95
44.05 +/- 5.30
49.83 +/- 7.11 (+)
46.40 +/- 5.30 (*)
62.56 +/- 11.86 (+*)
+ : significant difference between two pedaling types
*: significant change with increasing workload

However, as shown in Table 2, MPF, referring to the level of muscle fatigue, there was no significant difference between pedaling types and there was no significant change with increasing workload in all muscles except vastus lateralis. Vastus lateralis showed lower MPF in backward pedaling at all workloads, but only the difference at 125W was significant.

Table 2. The comparison of MPF between Forward and Backward pedaling and the Change of MPF in each pedaling   Type.

348.95 +/- 17.6
358.45 +/- 16.6
254.17 +/- 15.5
355.70 +/- 15.8
352.21 +/- 16.9
350.96 +/- 16.9
358.40 +/- 17.3
350.97 +.- 15.7
376.07 +/- 13.1
348.07 +/- 17.0
374.89 +/- 17.9
373.70 +/- 18.7
377.31 +/- 17.5
360.92 +/- 22.4
350.97 +/- 15.7
377.90 +/- 16.0
346.55 +/- 17.6
348.68 +/- 30.4
345.55 +/- 19.7
342.83 +/- 33.1
338.38 +/- 16.7
336.71 +/- 30.7
366.86 +/- 29.6
342.50 +/- 21.8
336.11 +/- 25.2
328.27 +/- 16.5
327.08 +/- 19.4
316.55 +/- 18.9
315.29 +/- 21.8
304.33 +/- 9.3
336.93 +/- 23.6 (*)
296.35 +/- 14.1 (+*)
+ : significant difference between two pedaling types
*: significant change with increasing workload

As shown in Table 3, among the 4 muscles, the vastus lateralis showed the highest change in the ratio between the 4th stage and the 1st stage in IEMG, and the lowest change in MPF ratio. But the statistical significance in IEMG did not show in backward pedaling.

Oxygen consumption and blood lactate concentration was higher in backward pedaling, and efficiency was lower in backward pedaling at all workloads. However, the significant difference was shown at workloads of 75W, 100W, and 125W in oxygen consumption, efficiency, and in blood lactate concentration.  The difference at workloads of l00W and 125W was significant. All the metabolic variables changed significantly with increasing workloads in both pedaling types.

Table 3. The comparison of change in IEMG and MPF (stage 4 - stage 1) among the 4 muscles.

4.74 (+)
48.85 (+)
(% ratio)
-27.14 (+)
-7.97 (+)
(% ratio)
-9.64 (+)
+ : significant difference among the 4 muscles

Table 4. The comparison of Metabolic Variable between Forward and Backward pedaling and the change of Metabolic variable in each pedaling type.

17.57 +/- 1.26
21.98 +/- 1.33
27.51 +/- 1.51
33.06 +/- 1.96 (*)
18.66 +/- 1.45
24.34 +/- 1.55 (+)
32.03 +/- 2.85 (+)
40.76 +/- 4.22 (+*)
13.05 +/+ 0.98
15.33 +/- 0.91
16.58 +/- 0.89
17.15 +/- 0.95 (*)
12.48 +/- 1.38
14.04 +/- 0.90 (+)
14.37 +/- 0.93 (+)
13.43 +/- 1.4 (+*)
1.91 +/- 0.32
2.31 +/- 0.66
2.21 +/- 0.79
3.13 +/- 0.89 (*)
2.01 +/- 0.50
3.47 +/- 1.73
4.19 +/- 1.22 (+)
6.45 +/- 1.51 (+*)
+ : significant difference between two pedaling types
* : significant change with increasing workload

Muscle Activity Pattern: During forward pedaling, vastus lateralis, biceps femoris, and gastrocnemius co-worked as synergists in the power phase of pedaling.  This indicates that these muscles played an important role of pushing down the crank. Knee extension is necessary for a pushing down action and this supports vastus lateralis activity as a knee extensor to be undoubtful. By the same token, activation of gastrocnemius told us that ankle plantar flexion occurred during the power phase of pedaling. Considering that biceps femoris is not only a knee plexor but also a hip extensor, it is not unusual for biceps femoris to act as a synergist of vastus lateralis (which is one of the typical knee extensors). As shown in Fig. 2, because of this, the dual function biceps femoris can be recruited when the leg is straightened, like the pushdown motion of pedaling. This was supported from Ericson (1986) and Faria & Cavanagh (1978). However, tibialis anterior was activated through the opposite range to that of gastrocnemius, in the recovery phase of pedaling, to indicate that dorsiflexion occurred in the latter part of pedaling near TDC. All of the above results have been evidently supported by many researchers (Houtz & Fischer, 1959; Despires, 1986; Faria & Cavanagh, 1978; Gregor et al. 1982; Jorge & Hull, 1986).

As shown in Fig. 1 and Fig. 2, the predominant difference of muscle activity between the two types of pedaling could be found in the activation of the biceps femoris. Biceps femoris worked in the power phase during forward pedaling, but its role was changed during backward pedaling to work in the recovery phase. Even though there was little information about muscle activity during backward pedaling, we inferred that in this type of pedaling, biceps femoris worked more seriously as a knee flexor in the recovery phase, than as a hip extensor in the power phase.

Fig. 2 also shows vastus lateralis acted mainly in the power phase of backward pedaling to illustrate how the vastus lateralis plays a major role in bicycling. While gastrocnemius was activated at the latter part of the power phase, it ended its activity in the early phase of recovery, passing by the BDC. It was not enough with this result to support that gastrocnemius was the prime mover in the power phase during backward pedaling, because vastus lateralis had already activated, to generate muscle power and pedaling acceleration before gastrocnemius activated.

Tibialis anterior's activation pattern during backward pedaling was similar to that of forward pedaling. In both cases this muscle acted in the latter part of the recovery phase.  This may also be due to ankle dorsiflexion occurrence as said above.

Amount of Muscle Activity: The higher amount of IEMG refers to higher amounts of muscle activity (Kond & Vitasalo, 1976; Komi et al. 1987; Hakkinen et al, 1985; Tesch et al. 1983). So, as shown in Table 1, it was believed that the amount of activity of the tibialis anterior was higher in backward pedaling than in forward pedaling. This indicates that recovering the crank during forward pedaling is easier than in that of its backward counterpart, and it means that this muscle is likely to generate more force for pulling the crank upward. This is evident when you see that IEMG of that muscle changed with increasing workload only in backward pedaling.

Ericson (1986) reported that IEMG of gastrocnemius did not change with increasing workload, while other muscles did. Also Duchateau et al. (1986) agreed with him that soleus activity increased with progressive overload, but gastrocnemius did not change. Marsh et al. (1995) supported the above results by reporting that gastrocnemius, could change its activity level only when pedaling speed was increased. In this study there were no differences of IEMG in gastrocnemius between forward pedaling and backward pedaling, and IEMG did not change during forward pedaling. In spite of those results, it is interesting, however, that there were meaningful increases of IEMG during backward pedaling. It may also be due to the difference of activity range as shown Fig. 1 and Fig. 2.

Biceps femoris showed higher IEMG in backward pedaling as in the case of tibialis anterior. It means that both biceps femoris and tibialis anterior worked more in backward pedaling. As shown in Fig. 1 and Fig. 2, 3 muscles, including biceps femoris worked for pushing down the crank in forward pedaling, while in backward pedaling, 2 muscles, including biceps femoris was likely to do the upward pulling of the crank. This indicates that biceps femoris worked less during the power phase in forward pedaling, and worked more during the recovery phase in backward pedaling. Ericson (1986) has reported that knee extensors worked more than knee flexors and hip extensors during bicycling. Also the fact that significant changes in IEMG occurred in biceps femoris during backward pedaling, strongly supported the above results.

It is undoubtful that vastus lateralis was one of the prime movers for bicycle pedaling. Many researchers (Kond & Vitasalo, 1976; Kond et al. 1987; Hakkinen et al, 1985) have already reported that the IEMG of vastus lateralis predominantly increased during bicycle pedaling. Weineck (1993), Jorge and Hull (1986) also regarded quadriceps femoris as an agonist in pedaling. Moreover, Ericson (1986) reported that among muscles, vastus lateralis showed the largest amount of IEMG. However, the IEMG of this muscle was higher in backward pedaling than that of forward pedaling.  This is likely to be because vastus lateralis' activation was more serious in backward pedaling. The reason for this may be due to the fact that vastus lateralis co-worked with 2 muscles for pushing the crank down in forward pedaling, while in backward pedaling, crank pushdown depends mainly on vastus lateralis. In both forward and backward pedaling, the IEMG changed significantly with increasing workload to indicate that this muscle was certainly a prime mover in pedaling.

The sum of the 4 muscles' IEMG was changed with increasing workload in both forward and backward pedaling. Though IEMG of biceps femoris, tibialis anterior, and gastrocnemius was not significantly increased, vastus lateralis played a major role of increasing total muscle activity during forward pedaling. However, it was no wonder that total muscle activity in backward pedaling increased, for all the muscles changed their amount of activation with increasing workload. It was inferred that differences in EIEMG between the two pedaling types might make a difference in energy metabolic variables. Table 3 shows the change of muscle activity, and its ratio between the 1st work stage and the 4th work stage. It confirms again that vastus lateralis was the prime worker among the 4 muscles, and each muscle was activated much more during backward pedaling than during forward pedaling.

Muscle Fatigue: The vivid information about the electrical power of muscle contraction can be acquired from frequency spectrum analysis, including mean power frequency and median power frequency (Lindstiorm et al, 1974). Mean power frequency was considered to be the parameters of muscular fatigue when it transfers from high frequency to low frequency (Petrofsky & Lind, 1980; Basmajian and DeLuca, 1985). MPF of tibialis anterior at a workload of 100W, and 125W didn't show the difference between forward and backward pedaling, despite the fact that IEMG at these workloads was different between the 2 types of pedaling. This indicates that the muscle's activity during the recovery phase was not so serious as to lead to muscular fatigue. Referring to Taylor et al (1997), muscle fatigue is closely related to the amount of muscle activation, which disagrees with the above result. From Table 1 and Table 2, it was reasonable to assume that the MPF of gastrocnemius did not change in both forward and backward pedaling, because the IEMG of this muscle appeared almost to be similar in both pedaling types. This could be supported from Taylor et al. (1997).  MPF of biceps femoris was not different between forward and backward pedaling, and the change of MPF in both types of pedaling was not significant. In forward pedaling, no change in MPF seemed to be related to no change in IEMG, and it agreed with the results of Petrofsky (1979), Garnet (1993), and Moritani (1993). On the other hand, biceps femoris did not show a meaningful change in MPF, in spite of a change in IEMG with increasing workload. Even though it did not agree with those researchers above, it can be partly explained, as in the case of tibialis anterior, that muscles working in the recovery phase were less tired than the muscles in the power phase. To see Fig. 2, it is certain that biceps femoris worked in the latter part of the recovery phase. From Table 1 and Table 2, we found that vastus lateralis had a tendency to decrease in MPF nearly linearly according to the increase of IEMG in both pedaling types.  This agrees with Petrofsky (1979), Garnet et al. (1993), Oda & Moritani (1995), Hori et al. (1995). The linear relationship between the change of MPF and IEMG in vastus lateralis indicates this muscle to be the most active and fatiguable among the 4 muscles tested in this study. Table 3 additionally supported this by showing that the highest negative change of MPF occurred in vastus lateralis.

Metabolic Variables: Oxygen consumption increased linearly with the change of workload in both forward and backward pedaling.  This was certainly related to the increase in total amount of muscle activity (Table 1, Table 3). Many researchers had already discussed the relationship between IEMG and oxygen consumption to concur that these two variables positively related (Bigland & Woods, 1976; Kunstlinger et al., 1985; Mateika & Duffin, 1994; Moritani et al., 1993; Takaishi et al., 1996; Glass et al., 1997).

The difference in work efficiency between forward and backward pedaling, matched with that of oxygen consumption to confirm that more oxygen consumed at the same workload, generated lower efficiency. Moreover, it shows that the efficiency in forward pedaling was made better with increasing workload, while the efficiency in backward pedaling suddenly dropped after a 100W workload. It may be related to the sudden increase of V02, or to the sudden increase of muscle activation at that workload in backward pedaling. Komi (1987) has reported that the more IEMG is increased, the lower the work efficiency, and vice versa.

There has been controversy about the relationship between blood lactate concentration and the amount of muscle activity. Some researchers suggested that an increase of muscle activity was highly related with an increase of blood lactate concentration (Nagata et al, 1981 ; Airaksinen et al., 1992), while others opposed this opinion (Sebum et al., 1992; Taylor et al., 1994). In our study, the workload where the VO2 and IEMG increased accorded with the workload where the blood lactate increased. This was supported from the results of Duffin et al. (1994), Glass et al. (1997), and Nagata et al. (1981 ). It is also noticeable that lactate threshold was likely to occur during backward pedaling, while during forward pedaling it showed a blunt increase of blood lactate.

In conclusion, backward pedaling appeared to be more intense to perform than its forward counterpart, showing higher IEMG, lower MPF, and accompanying higher VO2, higher blood lactate, and lower work efficiency. It is not desirable to use this system for the purpose of efficient bicycling.  However, it was considered to be excellent for strengthening the muscles, especially for the competitive cyclist as several researchers (Mackie et al., 1984: Thorstensson, 1986; Flynn et al. 1991) have suggested. Unfortunately, we did not focus on the availability of this pedaling system for rehabilitation. For further studies it would be recommended to verify the effect on that issue.


All information on this page is correct to the best of our knowledge.

Back To Health Studies Page