*3.4.1. Maximum amplitude along the entire cycle*

Computational Intelligence in Electromyography Analysis – 80 A Perspective on Current Applications and Future Challenges

were also activated in the final flight, but after the GN.

muscle electrical activity was present for short periods of time.

**3.3. Profile of muscular electrical activity** 

*3.3.1. Time of onset of maximum intensity* 

in the non-support phase.

when shod can occur earlier.

**3.4. Maximum amplitude** 

We will describe the changes in that order.

jogging cycle.

From the GEAV signal, the order of participation of each muscle can be seen. This order remained constant in all three conditions: tibialis anterior, hamstrings, gastrocnemius, and quadriceps. The AT muscle was the first one to be activated during the stride; it was activated before the flight of the ipsilateral foot, in the last part of unipodal support. The BF and ST were activated after the TA, but in the swing phase, i.e., when the contralateral leg was resting on the floor. The GN was activated in the final flight. Vastus medialis and RF

The profile of the electrical activity is characterized by the length of time of muscle activity, the peak (maximum amplitude), and the time of its appearance. Changes in muscle activity can occur in any of the variables: amplitude, time, and frequency. In all three conditions

The functions of the muscles varied through the cycle of jogging. There is no statistically significant difference between the three conditions with respect to the activity profile of the

Rectus femoris, VM, and GN had a peak at the beginning of the stance phase. Tibialis anterior showed two peaks: one in the swing phase and another in the final flight. Biceps femoris and ST showed two peaks, one at the end of the non-support phase, and one in the initial support.

**Table 4** gives the exact percentages of jogging cycle at which the maximum amplitude occurred. For five muscles there were no differences in the time of peak onset in their GEAV, maintaining their patterns of activity in the three conditions. As shown in Figure 3, the peak of the RF, VM, and GN appeared in the support phase; the peak of the BF, ST, TA,

In both legs, the peak appears in very similar times. Depending on the footwear, the time of occurrence of the peak is slightly modified. In general, muscles that have their peak after the impact when shod may appear a little later; and muscles that have their peak before impact

The peak of the RF, VM, and GN appeared between 6% and 10% in the three conditions and in both legs. The TA peaks between 94-96% in the final flight; the highest peak of the BF appeared between 88% and 90%. Semitendinosus peak appeared about 84%-90% in the shod

The changes of amplitude or intensity of muscle activity can be evaluated considering: (1) the entire cycle, (2) the stance phase after impact, and (3) the non-support, before impact.

condition, ahead respect to the barefoot condition, which appeared between 94-98%.

**3.2. Order of muscle participation** 

**Figure 4** shows the maximum activity displayed during the entire cycle by each of the twelve muscles. In all three conditions the muscles that had the highest amplitude were the VM, GN and TA. The muscles that showed less activity were the BF, ST and RF.

The maximum amplitude of homologous muscles was of the same order of magnitude for both legs, as shown in **Figure 5**.

There is no statistically significant difference between the three conditions with respect to the value of the amplitude peak.

#### *3.4.2. Maximum amplitude during the support phase after impact*

**Figure 6a** shows the maximum activity during the support phase produced after the impact of the heel. The TA, GN, and BF muscles in both legs increased their activity with both types of shoes. The behaviour of RF, VM, and ST varied positively and negatively.

The evolution of electrical activity in buffer muscles (RF and VM) during the loading phase was not the same in both legs, neither with both types of shoes.

**Figure 7** shows the increase (not normalized, in mV) of EMG activity with respect to the barefoot condition. When subjects wore standard shoes with hard soles, the RF of the left leg increased its maximum amplitude after impact, and the left VM decreased their activity.

When subjects wore their own shoes, the RF of the left leg decreased its peak amplitude after impact. The left VM decreased its activity; and the one of the right, increased.

When the shoe was harder, the left RF increased its activity and the VM of both legs decreased. When the shoe was softer, the RF of both legs and the VM of the left leg decreased its activity, and the right VM slightly increased its activity.

Comparison by EMG of Running Barefoot and Running Shod 83

**Figure 5.** Maximum amplitude of the EMG from homologous leg muscles (red: left, blue: right; RF: rectus femoris, VM: vastus medialis, TA: tibialis anterior, GN: lateral gastrocnemius, BF: biceps femoris, ST: semitendinosus) for each locomotion condition (BF: barefoot, SS: standard shoes, and AS: athletic

shoes).

**Figure 4.** Peak of EMG during the whole cycle, for the left (upper graph) and right (bottom) legs muscles for each Locomotion Condition (BF: barefoot, SS: standard shoes, and AS: athletic shoes).

The TA of both legs increased their activity. The right TA increased more than the left one. In both legs, the TA increases more its maximum amplitude in the condition of own jogging shoes; perhaps because this type of shoe offers more cushioning. The GM activity in the condition of standard jogging shoe increased more than in the own shoe condition. When the hardness of the sole was higher, the translational velocity was lower and higher GM activity. Perhaps this trend has to do with balance, because when the speed is less, dynamic stability decreases.

The BF of both legs and both conditions increased their activity. In the standard shoe condition, it increased more. The ST in both conditions decreased its activity.

The response of the agonist muscles from the same muscle group is not the same when the shoe condition changes. Thus, with the standard shoe RF activity increased and decreased the one of VM, the BF increased, and ST decreased.

Computational Intelligence in Electromyography Analysis – 82 A Perspective on Current Applications and Future Challenges

**Figure 4.** Peak of EMG during the whole cycle, for the left (upper graph) and right (bottom) legs muscles for each Locomotion Condition (BF: barefoot, SS: standard shoes, and AS: athletic shoes).

stability decreases.

The TA of both legs increased their activity. The right TA increased more than the left one. In both legs, the TA increases more its maximum amplitude in the condition of own jogging shoes; perhaps because this type of shoe offers more cushioning. The GM activity in the condition of standard jogging shoe increased more than in the own shoe condition. When the hardness of the sole was higher, the translational velocity was lower and higher GM activity. Perhaps this trend has to do with balance, because when the speed is less, dynamic

The BF of both legs and both conditions increased their activity. In the standard shoe

The response of the agonist muscles from the same muscle group is not the same when the shoe condition changes. Thus, with the standard shoe RF activity increased and decreased

condition, it increased more. The ST in both conditions decreased its activity.

the one of VM, the BF increased, and ST decreased.

Comparison by EMG of Running Barefoot and Running Shod 83

**Figure 5.** Maximum amplitude of the EMG from homologous leg muscles (red: left, blue: right; RF: rectus femoris, VM: vastus medialis, TA: tibialis anterior, GN: lateral gastrocnemius, BF: biceps femoris, ST: semitendinosus) for each locomotion condition (BF: barefoot, SS: standard shoes, and AS: athletic shoes).

Computational Intelligence in Electromyography Analysis – 84 A Perspective on Current Applications and Future Challenges

Comparison by EMG of Running Barefoot and Running Shod 85

**Figure 7** shows that the EMG signal before contact of the heel exhibited greater increase than in the signal after contact. The intensity of the myoelectric activity showed differences between both types of shoes. In both shod conditions the TA muscle activity increased more in absolute terms. The ST, BF and GN increased their activity more than the VM, the activity

**Figure 7.** Absolute GEAV peaks amplitude increments respect to barefoot condition for the other two locomotion conditions (SS: standard shoes and AS: athletic shoes) for both legs muscles after impact

In the left leg, the TA, BF, and ST were more active in the standard shoe condition. The increased activity of the semitendinosus was more important than that of the biceps femoris. In the right leg, the TA, BF, ST, and VM activity increased more in the athletic shoe condition than in the standard shoe condition. The electrical activity of the muscles of both

In order to evaluate the coactivity of agonist-antagonist muscles, we have established an index obtained as the ratio between the peak of the agonist muscle activity and the activity of the oppose muscle or muscles (antagonists). That index can be higher (agonist's peak is bigger than that of the antagonist) or below (opposite case) one (see **Figure 8**). To establish

the role of a muscle as agonist or antagonist, we rely on the movement of joints:

of which increased very little.

(upper graph) and before impact (bottom graph).

**3.5. Activity of antagonist muscles** 

legs showed different evolution depending on the type of shoe.

**Figure 6.** Maximum amplitude of the GEAV post - and pre-impact (after and before initial contact) for the 12 studied muscles for each locomotion condition (BF: barefoot, SS: standard shoes, and AS: athletic shoes).

#### *3.4.3. Maximum amplitude before impact*

As shown in **Figure 6b**, for both legs, five out of the six muscles (TA, GN, BF, ST, and VM) increased their activity in the non-support phase in shod condition. The RF maintained its level of activity; unlike what happened in the stance phase, in which the RF, VM, and ST decreased their activity.

**Figure 7** shows that the EMG signal before contact of the heel exhibited greater increase than in the signal after contact. The intensity of the myoelectric activity showed differences between both types of shoes. In both shod conditions the TA muscle activity increased more in absolute terms. The ST, BF and GN increased their activity more than the VM, the activity of which increased very little.

**Figure 7.** Absolute GEAV peaks amplitude increments respect to barefoot condition for the other two locomotion conditions (SS: standard shoes and AS: athletic shoes) for both legs muscles after impact (upper graph) and before impact (bottom graph).

In the left leg, the TA, BF, and ST were more active in the standard shoe condition. The increased activity of the semitendinosus was more important than that of the biceps femoris. In the right leg, the TA, BF, ST, and VM activity increased more in the athletic shoe condition than in the standard shoe condition. The electrical activity of the muscles of both legs showed different evolution depending on the type of shoe.

#### **3.5. Activity of antagonist muscles**

Computational Intelligence in Electromyography Analysis – 84 A Perspective on Current Applications and Future Challenges

**Figure 6.** Maximum amplitude of the GEAV post - and pre-impact (after and before initial contact) for the 12 studied muscles for each locomotion condition (BF: barefoot, SS: standard shoes, and AS: athletic

As shown in **Figure 6b**, for both legs, five out of the six muscles (TA, GN, BF, ST, and VM) increased their activity in the non-support phase in shod condition. The RF maintained its level of activity; unlike what happened in the stance phase, in which the RF, VM, and ST

shoes).

*3.4.3. Maximum amplitude before impact* 

decreased their activity.

In order to evaluate the coactivity of agonist-antagonist muscles, we have established an index obtained as the ratio between the peak of the agonist muscle activity and the activity of the oppose muscle or muscles (antagonists). That index can be higher (agonist's peak is bigger than that of the antagonist) or below (opposite case) one (see **Figure 8**). To establish the role of a muscle as agonist or antagonist, we rely on the movement of joints:

Computational Intelligence in Electromyography Analysis – 86 A Perspective on Current Applications and Future Challenges

• In the support phase, between 6% and 10% of the jogging cycle, the peaks of the following muscles occur: GN, RF, and VM; there is also a second peak of lesser magnitude in the activity of the BF and ST muscles. Concomitant to the important activity of these muscles, ankle dorsiflexion, knee flexion, and hip extension occur.

Comparison by EMG of Running Barefoot and Running Shod 87

The antagonist muscles showed more activity than agonist ones in the ankle and knee joints, for all the three conditions. In the hip, agonist muscles were more active, mainly in shod condition.

In the ankle and knee joints, agonist muscles showed a slightly higher activity than the

The peaks of TA, BF, and ST muscles occur in the non-support phase (see Figure 3), concomitant with the dorsiflexion, the extension of the knee, and the hip extension. In all three conditions, the agonist muscles show greater activity than the antagonist muscles, as shown by the coactivity index being higher than one (see Figure 8). The results about the muscular coactivity during jogging barefoot and shod are just preliminary, but indicate that

**Figure 8.** Coactivity of ago- and antagonist muscles during support (upper graph) and non-support

The aim of our study was to evaluate the influence of footwear on the electrical activity of muscles of both legs when jogging barefoot and shod with two types of shoes, with different

antagonists. In the hip and knee, agonist muscles activity was higher.

*3.5.1.3. Coactivation during non-support phase* 

it is worth a study in greater depth on the subject.

(bottom graph) phases.

**4. Discussion** 

• In the phase of non support, the peak of the TA, BF, and ST are concomitant with an ankle dorsiflexion, knee extension, and hip extension.

There is no statistically significant difference between the three conditions with respect to the level of coactivation.

### *3.5.1. Coactivation during support phase*

#### *3.5.1.1. Barefoot*


The peaks of RF and VM were not taken into account because they occur while acting as antagonists.

#### *3.5.1.2. Shod*


The antagonist muscles showed more activity than agonist ones in the ankle and knee joints, for all the three conditions. In the hip, agonist muscles were more active, mainly in shod condition.

#### *3.5.1.3. Coactivation during non-support phase*

Computational Intelligence in Electromyography Analysis – 86 A Perspective on Current Applications and Future Challenges

the level of coactivation.

*3.5.1.1. Barefoot* 

dorsiflexion.

activity than the BF.

much higher.

antagonists. *3.5.1.2. Shod* 

*3.5.1. Coactivation during support phase* 

ankle dorsiflexion, knee extension, and hip extension.

The BF (hip extensor) shows increased activity.

increase their activity, so the index remained similar.

decreases because the ST activity decreases.

decreases because the RF activity increases.

because the activity of VM is much higher.

• In the support phase, between 6% and 10% of the jogging cycle, the peaks of the following muscles occur: GN, RF, and VM; there is also a second peak of lesser magnitude in the activity of the BF and ST muscles. Concomitant to the important activity of these muscles, ankle dorsiflexion, knee flexion, and hip extension occur. • In the phase of non support, the peak of the TA, BF, and ST are concomitant with an

There is no statistically significant difference between the three conditions with respect to

• Coactivation of tibialis anterior (agonist) / gastrocnemius (antagonist): the peak activity of GN occurs between 8% and 10% of the jogging cycle. The TA shows no peak in the stance phase. The index is below unity. The LG shows greater activity during

• Coactivation of hamstring (agonist) / rectus femoris (antagonist): index above unity.

• Coactivation of hamstring (agonist) BF / vastus medialis (antagonist): the index is less than unity and therefore smaller than that showed by the antagonist muscles BF and RF

• Coactivation of semitendinosus (agonist) / rectus femoris (antagonist): index greater than one, but less than that found between the BF and RF because the ST has less

• Coactivation of semitendinosus (agonist) / vastus medialis (antagonist): the index is less than unity and therefore is less than that presented by the antagonist muscles BF and RF because the ST activity was smaller than that of the BF, and the activity of the VM is

The peaks of RF and VM were not taken into account because they occur while acting as

• Coactivation of tibialis anterior (agonist) / gastrocnemius (antagonist): both muscles

• Coactivation of semitendinosus (agonist) / vastus medialis (antagonist): the index is

• Coactivation of hamstring (agonist) / vastus medialis (antagonist): the index increases

• Coactivation of semitendinosus (agonist) / rectus femoris (antagonist): the index

• Coactivation of biceps femoris (agonist) / rectus femoris (antagonist): the index

maintained because the two muscles decrease their respective activity.

slightly because the BF activity increases and that of the VM decreases.

In the ankle and knee joints, agonist muscles showed a slightly higher activity than the antagonists. In the hip and knee, agonist muscles activity was higher.

The peaks of TA, BF, and ST muscles occur in the non-support phase (see Figure 3), concomitant with the dorsiflexion, the extension of the knee, and the hip extension. In all three conditions, the agonist muscles show greater activity than the antagonist muscles, as shown by the coactivity index being higher than one (see Figure 8). The results about the muscular coactivity during jogging barefoot and shod are just preliminary, but indicate that it is worth a study in greater depth on the subject.

**Figure 8.** Coactivity of ago- and antagonist muscles during support (upper graph) and non-support (bottom graph) phases.

#### **4. Discussion**

The aim of our study was to evaluate the influence of footwear on the electrical activity of muscles of both legs when jogging barefoot and shod with two types of shoes, with different geometry and damping features. Our findings show that shoes increase the intensity of muscle activity, especially in the non-support phase. These results are consistent with those of Nigg et al. (2003) that showed specific changes in the intensity of muscle activation before heel strike.

Comparison by EMG of Running Barefoot and Running Shod 89

absolute values, the time of activity would be different because the stance phase was shorter in the condition of athletic shoes and the phase of non support was longer. Thus, the time of muscle activity would be greater in the stance phase and lower in the non-support phase.

It is interesting to find that during the jogging, as happens in walking, there are phases in which no muscle activity is taking place, although there are movements that would make one expect some muscular activity. This result is consistent with other studies that have shown that the main goal of locomotion is to transport the body from one position to another using lower limb locomotor coordination and effective mechanisms. The nervous system integrates the movement of different body segments and controls the activity of the

This lack of muscle activity is not surprising. In jogging, the decrease of kinetic and potential energies in the first part of the support is concomitant with the accumulation of elastic energy in the tendons of the muscles; this energy is later released during the remainder of the stance phase to facilitate the progression and stability dynamics. Margaria and Cavagna (1965) found that during jogging, elasticity accounts for 50% of the work. During the unipodal support, the contralateral leg swing provides a second pushing force generated by the acceleration of the leg. Once the centre of gravity has shifted over the foot on the ground, the weight of the body becomes a driving force. When the contralateral leg is slowed by the hamstring muscles, the ipsilateral leg acquires kinetic energy, which is used to start the double flight. Inman (1968) argues that the slowdown of the swing leg during walking can contribute more to the forward movement of the body that the thrust of the ipsilateral leg. Once the leg is in motion, as the first law of Newton states, a force is needed to stop it, and that is the role of the hamstrings when they become active at the end of the swing phase. While the hamstrings are active, the

RF and VM become activated to improve the accuracy of the landing of the foot.

The amplitude of the myoelectric activity, expressed in absolute values, varied when using the two different types of footwear; these results are consistent with those found by Gollhofer & Komi (1987), who found differences in the electrical activity of muscles when subjects were at first barefoot and then shod. The effects of athletic shoes on the amplitude of muscle activity were significant. Changes in hardness and shape of the shoe did not affect

Different considerations have led us to not normalize the EMG signal amplitude: (1) to assess absolute muscle activity when a sole is interposed between the foot and the ground, and (2) to maintain the electrodes in place when the geometry of the footwear and the

The subjects ran faster in the athletic shoe condition and muscle activity was similar in the two footwear conditions: higher in the non-support phase. There is no statistically significant difference between the three conditions with respect to the speed of locomotion.

There was similar muscle activity in both shod conditions, which was different from the barefoot condition. In this paper we have classified the hardness of footwear subjectively. Although the interaction subject-shoe-ground occurs only during the stance phase, the intensity of the EMG activity was modified both during the stance phase and during the phase of non support. At first, one might think that the presence of the sole alters the physical

muscles to decrease energy expenditure (Inman 1968).

the muscle activity.

elasticity of the sole varied.

The speed increase is concomitant with decreased activity of the VM and ST in the stance phase. The increase in the intensity of TA in the stance phase is less than in the non-support phase, a result consistent with Komi et al. (1987).

The speed of jogging at which the EMG activity was recorded in the three conditions was freely chosen by the subjects because previous research (e.g., Kadaba et al. 1985) have shown that EMG activity is more reproducible when the speed is chosen freely than when it is imposed.

We note that the footwear influences the speed of locomotion, increasing it in the athletic shoe condition. This result suggests that the speed has to be one more parameter to consider when assessing the influence of footwear on the interaction between man-shoe-surface. The speed in similar condition of hardness (barefoot and standard shoe) was similar, but muscle activity was different in both conditions, being more similar that of the standard shoe and athletic shoe conditions.

In the different conditions under study, the pattern of muscle activity kept its profile. The starting and ending times of the activity were constant, but the amplitude of the EMG signal changed when using footwear. Also, the duration of the phases in the shod condition was reversed with respect to the bare condition, decreasing the stance phase and increasing the non-support one.

A limitation of this work was that has considered only 6 muscles out of the 57 having each of the lower extremities. The advantage over previous studies is that it has analyzed those six muscles from both legs simultaneously. The fact that the homologous muscles of both legs have similar amplitude allows us to use their absolute values.

The timing of muscle activity was assessed using visual inspection of the actual records, which is the detection method used to interpret the real surface EMG signal. Usually, visual inspection provides a high level of accuracy because all the details of the signal can be measured. Detecting the beginning and end of the activity by using different types of algorithms is based on establishing a threshold, usually based on intuitive criteria. In the end, both visual interpretation and detection using algorithms are based on subjective criteria. The advantage of the subjectivity of the experimenter is based on the acquisition of the ability to interpret the EMG signal and experience that allows the evaluator to use the capabilities of the human brain to perform a more sophisticated interpretation of highly variable data (Staude & Wolf 1999).

The order of participation of the muscles studied did not vary with the type of shoes. First, the TA muscles activated, afterwards the muscles of the posterior part of the thigh (BF and ST), followed by the activation of the GN of the back of the leg, and finally the activation of the RF and VM of the anterior part of the thigh.

The length of time of muscle activity in the three conditions did not vary with respect to the normalized jogging cycle. In all conditions, muscle activity was present in the first half of the stance phase, disappearing in the second, when the body weight fell on the forefoot. In absolute values, the time of activity would be different because the stance phase was shorter in the condition of athletic shoes and the phase of non support was longer. Thus, the time of muscle activity would be greater in the stance phase and lower in the non-support phase.

Computational Intelligence in Electromyography Analysis – 88 A Perspective on Current Applications and Future Challenges

phase, a result consistent with Komi et al. (1987).

athletic shoe conditions.

variable data (Staude & Wolf 1999).

the RF and VM of the anterior part of the thigh.

non-support one.

geometry and damping features. Our findings show that shoes increase the intensity of muscle activity, especially in the non-support phase. These results are consistent with those of Nigg et al. (2003) that showed specific changes in the intensity of muscle activation before heel strike. The speed increase is concomitant with decreased activity of the VM and ST in the stance phase. The increase in the intensity of TA in the stance phase is less than in the non-support

The speed of jogging at which the EMG activity was recorded in the three conditions was freely chosen by the subjects because previous research (e.g., Kadaba et al. 1985) have shown that EMG activity is more reproducible when the speed is chosen freely than when it is imposed.

We note that the footwear influences the speed of locomotion, increasing it in the athletic shoe condition. This result suggests that the speed has to be one more parameter to consider when assessing the influence of footwear on the interaction between man-shoe-surface. The speed in similar condition of hardness (barefoot and standard shoe) was similar, but muscle activity was different in both conditions, being more similar that of the standard shoe and

In the different conditions under study, the pattern of muscle activity kept its profile. The starting and ending times of the activity were constant, but the amplitude of the EMG signal changed when using footwear. Also, the duration of the phases in the shod condition was reversed with respect to the bare condition, decreasing the stance phase and increasing the

A limitation of this work was that has considered only 6 muscles out of the 57 having each of the lower extremities. The advantage over previous studies is that it has analyzed those six muscles from both legs simultaneously. The fact that the homologous muscles of both

The timing of muscle activity was assessed using visual inspection of the actual records, which is the detection method used to interpret the real surface EMG signal. Usually, visual inspection provides a high level of accuracy because all the details of the signal can be measured. Detecting the beginning and end of the activity by using different types of algorithms is based on establishing a threshold, usually based on intuitive criteria. In the end, both visual interpretation and detection using algorithms are based on subjective criteria. The advantage of the subjectivity of the experimenter is based on the acquisition of the ability to interpret the EMG signal and experience that allows the evaluator to use the capabilities of the human brain to perform a more sophisticated interpretation of highly

The order of participation of the muscles studied did not vary with the type of shoes. First, the TA muscles activated, afterwards the muscles of the posterior part of the thigh (BF and ST), followed by the activation of the GN of the back of the leg, and finally the activation of

The length of time of muscle activity in the three conditions did not vary with respect to the normalized jogging cycle. In all conditions, muscle activity was present in the first half of the stance phase, disappearing in the second, when the body weight fell on the forefoot. In

legs have similar amplitude allows us to use their absolute values.

It is interesting to find that during the jogging, as happens in walking, there are phases in which no muscle activity is taking place, although there are movements that would make one expect some muscular activity. This result is consistent with other studies that have shown that the main goal of locomotion is to transport the body from one position to another using lower limb locomotor coordination and effective mechanisms. The nervous system integrates the movement of different body segments and controls the activity of the muscles to decrease energy expenditure (Inman 1968).

This lack of muscle activity is not surprising. In jogging, the decrease of kinetic and potential energies in the first part of the support is concomitant with the accumulation of elastic energy in the tendons of the muscles; this energy is later released during the remainder of the stance phase to facilitate the progression and stability dynamics. Margaria and Cavagna (1965) found that during jogging, elasticity accounts for 50% of the work. During the unipodal support, the contralateral leg swing provides a second pushing force generated by the acceleration of the leg. Once the centre of gravity has shifted over the foot on the ground, the weight of the body becomes a driving force. When the contralateral leg is slowed by the hamstring muscles, the ipsilateral leg acquires kinetic energy, which is used to start the double flight. Inman (1968) argues that the slowdown of the swing leg during walking can contribute more to the forward movement of the body that the thrust of the ipsilateral leg. Once the leg is in motion, as the first law of Newton states, a force is needed to stop it, and that is the role of the hamstrings when they become active at the end of the swing phase. While the hamstrings are active, the RF and VM become activated to improve the accuracy of the landing of the foot.

The amplitude of the myoelectric activity, expressed in absolute values, varied when using the two different types of footwear; these results are consistent with those found by Gollhofer & Komi (1987), who found differences in the electrical activity of muscles when subjects were at first barefoot and then shod. The effects of athletic shoes on the amplitude of muscle activity were significant. Changes in hardness and shape of the shoe did not affect the muscle activity.

Different considerations have led us to not normalize the EMG signal amplitude: (1) to assess absolute muscle activity when a sole is interposed between the foot and the ground, and (2) to maintain the electrodes in place when the geometry of the footwear and the elasticity of the sole varied.

The subjects ran faster in the athletic shoe condition and muscle activity was similar in the two footwear conditions: higher in the non-support phase. There is no statistically significant difference between the three conditions with respect to the speed of locomotion.

There was similar muscle activity in both shod conditions, which was different from the barefoot condition. In this paper we have classified the hardness of footwear subjectively. Although the interaction subject-shoe-ground occurs only during the stance phase, the intensity of the EMG activity was modified both during the stance phase and during the phase of non support. At first, one might think that the presence of the sole alters the physical characteristics of the force of action and reaction and influences the feedback mechanisms of muscle activity. The results show that although the interaction subject-shoe-ground occurs only during the stance phase, the amplitude of the EMG activity was modified both in the stance and in the swing phases, the major changes occurring in the non-support phase.

Comparison by EMG of Running Barefoot and Running Shod 91

When wearing footwear, the right and left BF increased their activity at the end of the swing phase, in the final flight, and in the onset of the stance phase. At the end of the swing phase, those muscles act as antagonists of the extension of the knee and of the hip flexion in the swing phase; from the final flight phase, they act as agonists of the extension of the hip.

The increased activity of BF in the shod condition at the end of the swing phase could be justified by the higher speed and greater participation of these muscles to slow down the flexion of the thigh. In principle, its action would facilitate the advancement of the leg by inertia, and then its action would stop the leg being extended to facilitate the support of the foot on the floor. Perhaps, the increased activity in the phase of non-support has an effect on increasing the speed of the contralateral leg, as suggested by the studies of Inman (1968). At the beginning of the stance phase, the BF and ST intervene as agonists of the extension of the hip and knee flexion, cooperating synergistically with the torque that has the lower extremity and maintained as a result of inertia when the foot has been slowed by friction with the ground. When the sole absorbs more impact, shock muscles (RF and VM) decreased its activity. When the shoe was harder, the left RF increased its activity and that of

The fact that the muscles of both legs do not modify its maximum amplitude in the same percentage is understood in the context that the activity of a leg is adapted to the activity of the other. However, in the left leg there are significant differences in the activity of some of its muscles when wearing different types of footwear (see for example O'Connor 2005).

The lower activity of the RF and of the VM observed at the beginning of the stance phase may facilitate hip extension and knee flexion caused by increased activity of the hamstrings. This finding is consistent with the results of De Wit et al. (2000) who found that the knee is more flexed in the medium supporting phase in the shod condition. Also in that condition, the TA increased its activity in the final flight and at the beginning of the stance phase, being its activity lower during part of the swing. The role of the TA in the final flight and at the beginning of the stance phase is to promote dorsiflexion. The increased activity in the final flight can be explained based on segmental geometry: the hip extends and the foot flexes dorsally to avoid bumping into the ground until the proper stride length has been obtained. When jogging without shoes, De Wit et al. (2000) reported increased plantar flexion at the beginning of the support phase. Our results are consistent with their findings: the lower activity of the TA muscle in this condition would favour plantar flexion, apparently intended to reduce the pressure, which is higher in the barefoot condition because the

One explanation for the decreased activity of ST in the stance phase in the shod condition could be the following: this muscle is extensor and adductor of the hip, and also flexor and rotator of the knee. Shod, the movement from supine-pronation can be affected by the change in the distance between the point of application of the GRF and the subtalar or calcaneus-talus joint. The shoes would favour a faster decrease of the inversion with a faster rotation of the knee in a lower limb that is externally rotated, favouring a twist in the knee. If the ST does not decrease its activity, it would increase that twisting torque affecting the knee. A more active GN in the stance phase could influence the stability of the ankle. In the

the VM of both legs decreased.

impact force is increased and occurs in a shorter time.

Before the foot contacts the ground, there is no information available from the ground; therefore, no feedback mechanism is in use and muscle activity would be controlled by forward mechanisms. The changes in the amplitude of muscle activity in the two footwear conditions were more homogeneous in both legs during the post-impact than during the pre-impact. This finding is in agreement with those of Nigg (2003), who suggests that the EMG activity before and after impact corresponds to two different events. Before the foot contact with the ground, there is no information available from the ground, then no feedback mechanism is in use. It is speculated that this would be some pre-programmed muscle activity based on the expected impact and would be related to the "muscle tuning" and controlled by a forward mechanism. Thus, the configuration of the leg when the heel hits the floor is prepared beforehand during the non-support phase, following the strategy adopted (De Wit et al. 2000). In the other hand, the activity that appears after heel contact is due to a reflex effect. In this context, one can understand the variability of electrical activity in the pre-impact and its consistency in the post-impact.

The changes in the impact force described under the influence of footwear may be a consequence of changes in muscle activity in the non-support phase to reduce soft tissue vibrations during locomotion. Wakeling et al. (2002) explains how the frequencies of the impact force and the soft tissues may overlap during walking and running. It can be expected that the impact forces will cause vibrations within the range of vibration of the soft tissues. However, observations show that such vibrations are minimal. Wakeling et al. (2002) proposed that muscle activity is adjusted before impact in order to reduce possible vibrations resulting from the impact.

The reflex effect has to be understood in the context that the heel contact with the ground is a stimulus that triggers feedback mechanisms. This stimulus does not trigger the onset of muscle activity, which had already begun in all muscles during the swing phase or during the final flight phase (or even during the initial flight phase), but could just change it. The reflex time is around 110 ms in the normal population (Kroll & Clarson 1978), and is divided into a latency time of 20 ms and a motor time of 90 ms. A peak muscle activity for most of the muscles studied appears on the 8% of 700ms-long jogging cycle, i.e. at 56 ms, which correspond well with the reflex motor time.

The maximum activity peak appears in the three conditions in the same phase. The increase in mean activity increased recruitment of motor units and increased force production. The decrease means the opposite. The observation of decreased activity of the VM in the postimpact is consistent with previous findings (De Wit et al. 2000) about the influence of shoes on the ground reaction force. In the shod condition, impact peak appears later perhaps due to decreased activity of the damper (shock absorber) muscles. The loading speed is influenced by muscle activity, which affects the acceleration of the segments.

When wearing footwear, the right and left BF increased their activity at the end of the swing phase, in the final flight, and in the onset of the stance phase. At the end of the swing phase, those muscles act as antagonists of the extension of the knee and of the hip flexion in the swing phase; from the final flight phase, they act as agonists of the extension of the hip.

Computational Intelligence in Electromyography Analysis – 90 A Perspective on Current Applications and Future Challenges

in the pre-impact and its consistency in the post-impact.

vibrations resulting from the impact.

correspond well with the reflex motor time.

characteristics of the force of action and reaction and influences the feedback mechanisms of muscle activity. The results show that although the interaction subject-shoe-ground occurs only during the stance phase, the amplitude of the EMG activity was modified both in the

Before the foot contacts the ground, there is no information available from the ground; therefore, no feedback mechanism is in use and muscle activity would be controlled by forward mechanisms. The changes in the amplitude of muscle activity in the two footwear conditions were more homogeneous in both legs during the post-impact than during the pre-impact. This finding is in agreement with those of Nigg (2003), who suggests that the EMG activity before and after impact corresponds to two different events. Before the foot contact with the ground, there is no information available from the ground, then no feedback mechanism is in use. It is speculated that this would be some pre-programmed muscle activity based on the expected impact and would be related to the "muscle tuning" and controlled by a forward mechanism. Thus, the configuration of the leg when the heel hits the floor is prepared beforehand during the non-support phase, following the strategy adopted (De Wit et al. 2000). In the other hand, the activity that appears after heel contact is due to a reflex effect. In this context, one can understand the variability of electrical activity

The changes in the impact force described under the influence of footwear may be a consequence of changes in muscle activity in the non-support phase to reduce soft tissue vibrations during locomotion. Wakeling et al. (2002) explains how the frequencies of the impact force and the soft tissues may overlap during walking and running. It can be expected that the impact forces will cause vibrations within the range of vibration of the soft tissues. However, observations show that such vibrations are minimal. Wakeling et al. (2002) proposed that muscle activity is adjusted before impact in order to reduce possible

The reflex effect has to be understood in the context that the heel contact with the ground is a stimulus that triggers feedback mechanisms. This stimulus does not trigger the onset of muscle activity, which had already begun in all muscles during the swing phase or during the final flight phase (or even during the initial flight phase), but could just change it. The reflex time is around 110 ms in the normal population (Kroll & Clarson 1978), and is divided into a latency time of 20 ms and a motor time of 90 ms. A peak muscle activity for most of the muscles studied appears on the 8% of 700ms-long jogging cycle, i.e. at 56 ms, which

The maximum activity peak appears in the three conditions in the same phase. The increase in mean activity increased recruitment of motor units and increased force production. The decrease means the opposite. The observation of decreased activity of the VM in the postimpact is consistent with previous findings (De Wit et al. 2000) about the influence of shoes on the ground reaction force. In the shod condition, impact peak appears later perhaps due to decreased activity of the damper (shock absorber) muscles. The loading speed is

influenced by muscle activity, which affects the acceleration of the segments.

stance and in the swing phases, the major changes occurring in the non-support phase.

The increased activity of BF in the shod condition at the end of the swing phase could be justified by the higher speed and greater participation of these muscles to slow down the flexion of the thigh. In principle, its action would facilitate the advancement of the leg by inertia, and then its action would stop the leg being extended to facilitate the support of the foot on the floor. Perhaps, the increased activity in the phase of non-support has an effect on increasing the speed of the contralateral leg, as suggested by the studies of Inman (1968). At the beginning of the stance phase, the BF and ST intervene as agonists of the extension of the hip and knee flexion, cooperating synergistically with the torque that has the lower extremity and maintained as a result of inertia when the foot has been slowed by friction with the ground. When the sole absorbs more impact, shock muscles (RF and VM) decreased its activity. When the shoe was harder, the left RF increased its activity and that of the VM of both legs decreased.

The fact that the muscles of both legs do not modify its maximum amplitude in the same percentage is understood in the context that the activity of a leg is adapted to the activity of the other. However, in the left leg there are significant differences in the activity of some of its muscles when wearing different types of footwear (see for example O'Connor 2005).

The lower activity of the RF and of the VM observed at the beginning of the stance phase may facilitate hip extension and knee flexion caused by increased activity of the hamstrings. This finding is consistent with the results of De Wit et al. (2000) who found that the knee is more flexed in the medium supporting phase in the shod condition. Also in that condition, the TA increased its activity in the final flight and at the beginning of the stance phase, being its activity lower during part of the swing. The role of the TA in the final flight and at the beginning of the stance phase is to promote dorsiflexion. The increased activity in the final flight can be explained based on segmental geometry: the hip extends and the foot flexes dorsally to avoid bumping into the ground until the proper stride length has been obtained. When jogging without shoes, De Wit et al. (2000) reported increased plantar flexion at the beginning of the support phase. Our results are consistent with their findings: the lower activity of the TA muscle in this condition would favour plantar flexion, apparently intended to reduce the pressure, which is higher in the barefoot condition because the impact force is increased and occurs in a shorter time.

One explanation for the decreased activity of ST in the stance phase in the shod condition could be the following: this muscle is extensor and adductor of the hip, and also flexor and rotator of the knee. Shod, the movement from supine-pronation can be affected by the change in the distance between the point of application of the GRF and the subtalar or calcaneus-talus joint. The shoes would favour a faster decrease of the inversion with a faster rotation of the knee in a lower limb that is externally rotated, favouring a twist in the knee. If the ST does not decrease its activity, it would increase that twisting torque affecting the knee. A more active GN in the stance phase could influence the stability of the ankle. In the shod condition, there is more damping of the impact force. Cushioning and stability are conflicting (Perry 1992): more cushioning brings less stability.

Comparison by EMG of Running Barefoot and Running Shod 93

changing the subject-shoe-ground interface. According to Ferris et al. (1999), persons adjust the degree of stiffness of the legs when they run on different surfaces. The stiffness is given by the coactivation of the ago-antagonists muscles that cross the joints. In the support phase, the coactivation in the shod condition was greater than in the barefoot condition. At the stage of non support, the coactivation in the shod condition was not homogeneous, neither for both

A major function of footwear is to cushion both the strength of action the subject performs on the ground and absorb its reaction force, in order to protect the musculoskeletal system. Athletic shoes diminish –or even nullify- the impact peak of the vertical component and are involved in delaying the onset of the vertical support force, by changing the load gradient

The impact force acts as an input signal in the body and influences the vibration of the soft tissues (Nigg et al. 2000). The vertical component is biphasic and has two peaks: the impact peak appears after 20 ms of the impact, representing 140-160% -200% of BW, in those runners that support the heel first in the ground. A second peak, in the stance phase, appears at 80 ms and can almost triple the BW. Previous works (De Wit et al. 2000, Nigg 1983) have shown that in jogging, the magnitude of the passive and active peaks of the vertical component of the GRF does not vary with the shoes. The load gradient was lower in the shod condition due to the later occurrence of the impact force (33 ms against 11 ms).

In our study, peak muscle activity appeared in the following cycle time: 2-6% for the ST, 6- 8% for the VM and RF, 8-10% for GN. In absolute values, the 2% amounts to 14 ms, 6% to 42 ms, and 10% to 70 ms. The activity peak of four muscles occur between the peaks of the vertical component of the reaction force. Despite finding a second peak of greater magnitude in the shod condition, muscle activity was higher in only three of the six muscles studied. In two of the other muscles (VM and ST) in the shod condition the activity was

The increased activity of the TA in the shod condition could lead to the appearance of the anterior compartment syndrome of the lower leg. The results here presented about the increased EMG signal amplitude suggest further studies to corroborate or refute the argument used by athletes and shoe manufacturers on increased performance with the use

*Faculty of Physical Activity and Sport Science, University of Basque Country, Vitoria, Spain* 

lower. In the RF, the activity response was not constant.

*Faculty of Medicine, University of Basque Country, Bilbao, Spain* 

*Biorobotics Department, TECNALIA, Bilbao, Spain* 

legs nor for all muscles, as it may decrease or increase respect to the barefoot condition.

(Nigg 1983, De Wit et al. 2000).

of athletic footwear.

**Author details** 

Gonzalo A. Garcia\*

Corresponding Author

 \*

Begoña Gavilanes-Miranda

Juan J. Goiriena De Gandarias

Nigg et al. (2003) speculated that the body reacts to changes in input signal to adapt the muscle activity to reduce the vibrations of the soft tissues. In principle, the proximal muscles would be less required than the distal ones. These authors found that the TA activity increased more than that of the proximal ones (BF and ST). Our results are consistent with those findings respect to the TA and the ST; but not for the BF.

According to Mundermann et al. (2003), changes in the intensity of leg muscles activity predict better the differences in comfort than changes in the impact force. As the muscle activity decreased in the stance phase, the kinematic differences between the conditions disappeared (De Wit et al. 2000). This result is consistent with the absence of muscle activity. The signal amplitude indicates the level of activation of the muscle: there is a relationship between activity and strength, implying that increased activity is the source of increased muscle strength. The change in muscle electrical activity may be due to the role of the muscles to adapt to the characteristics of elasticity and friction of the shoes used.

According to Arsenault et al. (1987), it seems clear that the kinematics of locomotion is not very variable. From the scientific literature one can obtain the parameters describing the movement and explain muscle activity during jogging in different conditions. When the peak of the RF, VM, GN, BF, and ST muscles appears, the hip stretches, the knee flexes, and the ankle flexes dorsally. Rectus femoris and VM undergo an eccentric contraction, BF and ST concentric, and GN concentric in the knee and eccentric in the ankle. Biceps femoris and GN promote the flexing of the knee and the GN transfers energy from the ankle to the knee.

The activation of the muscles acting as antagonists (i.e., contracting eccentrically) promote the absorption of energy recovered in the immediate concentric contraction. In the condition barefoot, the BF presented electrical activity at the end of the swing phase, in the final flight, and at the onset of stance phase, acting first as antagonist and then as agonist. With shoes, the activity increased in both phases, increasing the energy absorbed in the non-support phase. During the concentric contraction at the beginning of the stance phase, the increased activity means more power generation. The increase of the absorbed energy and the energy produced could be interpreted as an effort at the beginning of the stance phase in order to do hip extension, perhaps to counterbalance the increased frictional force that opposes the advance in the shod condition. At the beginning of the support phase, the GN, RF, and VM are active in eccentric contraction. The reduction of the duration of the support phase in the shod condition may improve the efficacy of the stretching-shortening cycle.

The potential problem of crosstalk was reduced using a double differential technique, which is based on a single amplifier fed with three electrodes (De Luca & Merletti, 1988; Winter, 1990; Winter et al., 1994). It is already widely accepted that double-differential technique reduces the level of cross-talk (see, *e.g.*, De Luca & Merletti, 1988; and Meinecke et al., 2004). In any case, cross-talk cannot ever be fully cancelled.

As the human body is a biological system that has many possibilities of action and reaction, it would be advisable to evaluate the electrical activity of other additional muscles when changing the subject-shoe-ground interface. According to Ferris et al. (1999), persons adjust the degree of stiffness of the legs when they run on different surfaces. The stiffness is given by the coactivation of the ago-antagonists muscles that cross the joints. In the support phase, the coactivation in the shod condition was greater than in the barefoot condition. At the stage of non support, the coactivation in the shod condition was not homogeneous, neither for both legs nor for all muscles, as it may decrease or increase respect to the barefoot condition.

A major function of footwear is to cushion both the strength of action the subject performs on the ground and absorb its reaction force, in order to protect the musculoskeletal system. Athletic shoes diminish –or even nullify- the impact peak of the vertical component and are involved in delaying the onset of the vertical support force, by changing the load gradient (Nigg 1983, De Wit et al. 2000).

The impact force acts as an input signal in the body and influences the vibration of the soft tissues (Nigg et al. 2000). The vertical component is biphasic and has two peaks: the impact peak appears after 20 ms of the impact, representing 140-160% -200% of BW, in those runners that support the heel first in the ground. A second peak, in the stance phase, appears at 80 ms and can almost triple the BW. Previous works (De Wit et al. 2000, Nigg 1983) have shown that in jogging, the magnitude of the passive and active peaks of the vertical component of the GRF does not vary with the shoes. The load gradient was lower in the shod condition due to the later occurrence of the impact force (33 ms against 11 ms).

In our study, peak muscle activity appeared in the following cycle time: 2-6% for the ST, 6- 8% for the VM and RF, 8-10% for GN. In absolute values, the 2% amounts to 14 ms, 6% to 42 ms, and 10% to 70 ms. The activity peak of four muscles occur between the peaks of the vertical component of the reaction force. Despite finding a second peak of greater magnitude in the shod condition, muscle activity was higher in only three of the six muscles studied. In two of the other muscles (VM and ST) in the shod condition the activity was lower. In the RF, the activity response was not constant.

The increased activity of the TA in the shod condition could lead to the appearance of the anterior compartment syndrome of the lower leg. The results here presented about the increased EMG signal amplitude suggest further studies to corroborate or refute the argument used by athletes and shoe manufacturers on increased performance with the use of athletic footwear.
