How can we use different exercises to train different regions of the hamstrings muscle group?

Chris Beardsley
Nov 27, 2018 · 17 min read

The hamstrings are a key muscle group for athletes, being perhaps the most important muscle in the sprinting movement, and they also make up a large proportion of the posterior thigh, which means that they must also be trained effectively by bodybuilders.

Many strength coaches and bodybuilders treat the hamstrings muscle group as a single entity, which can be trained with a single exercise. However, the hamstrings muscle group is made up of four very different muscles, which are selectively worked by different exercises and types of loading, and each muscle is also made up of two or three different compartments or regions, so they likely also display regional hypertrophy.

The differences between hamstrings muscles (and between regions of each hamstrings muscle) have important practical implications for both athletes and bodybuilders alike.


Why do some muscles within a group grow more than others after certain types of training?

When muscles form groups (such as the hamstrings and quadriceps), we tend to think of them as a single, uniform entity. However, the muscles that make up the group display different anatomical and architectural features, which causes them to respond differently to different strength training exercises.

Muscles within a group can develop differently from one another after a strength training program for two main reasons.

Firstly, this can happen when some muscles within a group are two-joint muscles, and others are single-joint muscles.

Most people are familiar with the fact that the hamstrings and quadriceps include both two-joint and single-joint muscles. The quadriceps include just one two-joint muscle (the rectus femoris) and three single-joint muscles, while the hamstrings include three two-joint muscles and one single-joint muscle (the biceps femoris short head). Consequently, we can train the rectus femoris with hip flexion exercises and the three two-joint hamstrings muscles with hip extension exercises, but we need to use knee flexion exercises to train the biceps femoris short head and we need to use knee extension exercises to train the three single-joint quadriceps muscles.

On the other hand, it is much less well-known that multi-joint hip and knee extension exercises like the back squat are largely unable to train the two-joint rectus femoris. This is likely because the muscle displays only a very small internal moment arm length for knee extension in such movements. Indeed, long-term programs involving the squat tend not to produce gains in rectus femoris size. This actually makes a lot of sense, because we don’t really want the rectus femoris producing a hip flexion turning force during a multi-joint hip and knee extension exercise, as this would reduce our ability to perform the movement successfully.

Secondly, it can happen when either (1) the internal moment arm lengths of the muscles differ from one another (either at all joint angles, or just within one part of the overall joint range of motion), or (2) when other anatomical, architectural, or muscle fiber-related features differ between the muscles of a group, and thereby alter the mechanical loading experienced by each of the muscles in an exercise.

For example, when one muscle functions at a longer sarcomere length for a given muscle fascicle length during a movement compared to the others in the group (as the semitendinosus does during hip extension and knee flexion), then it will be producing force to a greater extent on the descending limb of the length-tension relationship, and therefore may experience more stretch-mediated muscle growth.

Also, when a muscle has a different anatomical attachment at the joint from the others in the group (as the medial and lateral hamstrings do at the knee), then its contribution, and therefore its development following a training program, will be affected by the position of the limb during the exercise.


Why do some regions within a muscle grow more than others after certain types of training?

When some regions of a muscle grow more than others after a strength training program, this is called regional regional hypertrophy.

Regional hypertrophy happens for two main reasons.

Firstly, it happens when muscle fibers are stimulated to grow more in length than in diameter, compared to when they are stimulated to grow more in diameter than in length. Greater increases in fiber length are associated with greater increases in muscle size in the distal region, while the greater increases in fiber diameter are associated with greater increases in muscle size in the middle region.

Eccentric-only strength training, full range of motion strength training, and constant load (weight) strength training all stimulate muscle fibers to grow more in length than in diameter, compared to concentric-only strength training, partial range of motion strength training, and accommodating resistance, respectively.

Secondly, regional hypertrophy can happen when different parts of a muscle are used to carry out different functions, either because the muscle contracts at a different point in the joint range of motion, or because the direction of force application differs.

Just as for individual muscles within a group, the various regions within a muscle can also contribute differently to a given movement because of differences in their internal moment arm lengths, or because of anatomical, architectural or muscle fiber-related reasons. Such differences in the roles of the regions are often mirrored by unique innervation by the central nervous system, allowing them to be controlled separately.


What muscles make up the hamstrings muscle group (and how do they differ)?

There are four muscles that make up the hamstrings muscle group (the biceps femoris (short head), the biceps femoris (long head), the semitendinosus, and the semimembranosus).

These muscles display different (1) innervation, (2) anatomy, (3) muscle architecture, (4) internal moment arm lengths, and (5) sarcomere lengths from one another. For some of these variables, there are also clear regional differences within each of the hamstrings muscles as well.


Each of the hamstrings are innervated separately. Furthermore, each of the hamstrings muscles contain compartments or regions that are also innervated separately.

Among the medial hamstrings, the semimembranosus has two distinct regions (proximal and distal) that are innervated by different primary branches of the same nerve. Similarly, the semitendinosus contains two regions (proximal and distal) that are innervated by separate nerves (and this innervation matches a key anatomical feature of the semitendinosus, which is the separation of the muscle into proximal and distal regions by a tendinous inscription).

Conversely, among the lateral hamstrings, the biceps femoris (long head), contains two main regions that are innervated by primary nerve branches, although these are positioned deep and superficial to one another, instead of proximally and distally. Similarly, the biceps femoris (short head) contains two (or perhaps three) regions that are innervated by different nerves, and these regions are again positioned deep and superficial to one another.

The regional innervation (and corresponding anatomy) seem to be somewhat reflected in the activation of each muscle group during hip extension and knee flexion exercises, as the upper and lower regions of the muscle are activated differently, according to the type of exercise performed. Early research indicated that knee flexion exercise may activate the lower region of both lateral and medial hamstrings to a greater extent, while hip extension may activate the upper regions to a greater extent. More recent research has found that which region is most strongly activated during strength training differs more between muscles, and less between exercises, although there may yet be some small differences.

Moreover, there are differences in the activation of the proximal and distal regions of the biceps femoris (long head) during high-speed running. Specifically, the proximal region is less active, and the difference between proximal and distal regions increases with increasing running speed.


The biceps femoris (short and long heads) comprise the lateral hamstrings, while the semitendinosus and semimembranosus comprise the medial hamstrings. The lateral and medial hamstrings groups have different attachment points at the knee from one another, and this difference in attachment points leads to meaningful effects.

Indeed rotating the foot inwards during leg curls, supine bridges, or single-leg deadlifts increases medial hamstrings involvement, while rotating the foot outwards increases lateral hamstring involvement.

Additionally, not all of the hamstrings share the same anatomical features.

For example, the semitendinosus has a tendinous inscription, which divides the muscle into very discrete proximal and distal anatomical regions, while the other hamstrings muscles do not. Moreover, each of the hamstrings displays different proximal and distal tendon lengths from one another, both when measured in absolute terms and also when measured relative to total muscle-tendon length. Given the differences in elasticity between muscles and tendons, this is likely to have a large impact on the behavior of each of the hamstrings during high-speed movements with light loads, such as during the sprinting gait cycle.


The hamstrings display very different muscle architecture from one another, and there are also differences between regions of each muscle.

The semitendinosus is a fusiform muscle, with the fascicles running longitudinally along its length, while the biceps femoris (long head) and semimembranosus are both heavily pennated. The medial and lateral groups each have one muscle that has a proportionally long normalized fiber length and a small cross-sectional area, and one muscle that has a proportionally short normalized fiber length and a large cross-sectional area.

It has been suggested that the fusiform nature of the semitendinosus (and its long fascicle lengths) might make it more susceptible to being worked during eccentric contractions. Indeed, many studies have found that eccentric leg curl variations, including the Nordic curl, load the semitendinosus more than the other hamstrings. Even so, differences in internal moment arm lengths or sarcomere lengths can more easily explain such observations. Whether differences in muscle architecture can explain why some studies have found that the biceps femoris (long head) is activated more than the semitendinosus in hip extension exercises is less clear, especially as is not always observed.

Additionally, the muscle architecture of each of the hamstrings muscles varies between regions. The proximal region of the biceps femoris (long head) has longer fascicles than the distal region. In contrast, the proximal region of the semitendinosus has shorter fascicles than the distal region. There are also variations in pennation angle along the length of both muscles, although the literature in this respect is currently conflicting. Variations within the other hamstrings muscles are unknown.


The internal moment arm length of a muscle is a key determinant of its function, and of its contribution to a given joint action.

Muscles exert turning forces on joints that are equal to the product of their linear tension and their internal moment arm length, so longer internal moment arm lengths lead to greater turning forces. When multiple muscles (or regions of a muscle) act on a joint, those with the longest internal moment arm lengths will contribute to the greatest extent, at any given joint angle.

The hamstrings act on both the hip and on the knee, and therefore have internal moment arm lengths for each joint.

Research has found that these internal moment arm lengths vary between hamstrings, and also vary over the joint range of motion in different ways for each of the hamstrings, but how internal moment arm lengths vary for each region of each hamstrings muscle is still very unclear.


Overall, the semitendinosus has by far the longest internal moment arm length of all the hamstrings acting at the knee joint (albeit to a greater extent in greater degrees of knee flexion) and it is therefore the most active of the hamstrings in knee flexion movements. This may be the real reason why many studies have found that the semitendinosus is the most active hamstrings muscle during leg curl variations.

Additionally, each of the hamstrings displays an internal moment arm length that peaks at a different knee joint angle (shown below in order of the peak at increasing knee flexion).

  • Semimembranosus = full knee extension
  • Biceps femoris (long head) = 50–60 degrees of knee flexion
  • Biceps femoris (short head) = 70–80 degrees of knee flexion
  • Semitendinosus = 90–100 degrees of knee flexion

Consequently, exercises that involve peak forces at different knee joint angles are likely to target different hamstrings muscles, with the semimembranosus being loaded most by exercises involving peak forces closer to full knee extension, and the semitendinosus being loaded most by exercises involving peak forces closer to full knee flexion.


Overall, the semitendinosus again has the longest internal moment arm length of all the hamstrings acting at the hip joint, but this is only apparent at greater degrees of hip flexion. In contrast, closer to full hip extension, the internal moment arm lengths of the hamstrings are actually quite similar to one another.

Consequently, hip extension exercises that involve peak forces at full hip extension (such as supine bridge variations) will likely involve all of the hamstrings to a similar extent (although noting that the gluteus maximus will probably be the main prime mover). Hip extension exercises that involve peak forces in greater degrees of hip flexion (such as most deadlift variations) will likely involve the semitendinosus to a greater extent (although noting that the adductor magnus will probably be the main prime mover).


When muscles produce force, the amount of force they produce is primarily determined by the force-velocity and length-tension relationships of the working muscle fibers.

If different muscles within a group (or different regions within a muscle) are contracting at different velocities or from different starting lengths, then they will produce different amounts of force, and therefore also experience different amounts of mechanical loading. Moreover, when a muscle works predominantly on the descending limb of the length-tension relationship, it is more likely to experience an additive effect of passive and active tension during strength training with a large range of motion, because of greater stretch-mediated signaling.

Little research has measured sarcomere lengths in working human muscles. Estimates have been made for a number of lower body muscles over a range of joint angles, by using a range of assumptions, and actual measurements have been taken for each the hamstrings muscles in cadavers for the anatomical position (however, we have no information about how sarcomere lengths vary across different regions of each of the human hamstrings).

These estimates and measurements are somewhat in agreement, insofar as they show that the medial hamstrings tend to reach longer sarcomere lengths than the lateral hamstrings. Additionally, the estimates suggest that the semitendinosus spends most of its time on the descending limb of the length-tension relationship, which might also help explain why it responds very well to eccentric training (and may also respond well to training at a long muscle length). The semimembranosus appears to function across the plateau region and descending limb of the length-tension relationship, while the biceps femoris appears to function across the ascending limb, plateau region, and descending limb of the length-tension relationship. Therefore, the biceps femoris might be the only muscle to reach full active insufficiency during supine bridging variations.


What does this mean in practice?

In practice, we can see that there are clear regions within each individual hamstrings muscle, as shown by the patterns of innervation and by some differences in muscle architecture. However, we lack information about the regional internal moment arm lengths and sarcomere lengths of each region, which makes it difficult to infer how the different regions might contribute to different sporting movements or exercises.

Some early research indicated that the upper regions of the hamstrings might contribute more to hip extension movements, while the lower regions might contribute more to knee flexion movements, but recent research suggests that which region is most strongly activated during strength training differs more between muscles, and less between exercises.

Additionally, we can see that the hamstrings muscles display different innervation, anatomy, muscle architecture, internal moment arm lengths, and sarcomere lengths from one another, and this means that they will contribute differently to different movements, and will be trained most effectively by different exercises and approaches.

  • Anatomy — rotating the foot inwards in either hip extension or knee flexion exercises increases medial hamstrings involvement, while rotating the foot outwards increases lateral hamstring involvement.
  • Knee internal moment arm lengths — altering the point at which peak force is produced in a leg curl alters which of the hamstrings is loaded, due to differences in internal moment arm lengths (the semimembranosus can be targeted by peak forces at full knee extension, the biceps femoris can be targeted by peak forces in the middle of the exercise range of motion, and the semitendinosus will be targeted by peak forces in full knee flexion).
  • Hip internal moment arm lengths — altering the point at which peak force is produced in a hip extension exercise also alters which of the hamstrings is loaded, due to differences in internal moment arm lengths (the semitendinosus will be targeted by peak forces in full hip flexion, while the hamstrings are loaded to largely the same extent when closer to hip extension).
  • Sarcomere lengths — the semitendinosus might be loaded the most effectively by long ranges of motion, while the medial hamstrings might be loaded more during supine bridges compared to the lateral hamstrings, due to greater active insufficiency in the biceps femoris long head than in the medial hamstrings.

Overall, we can see that it is very easy to train the semitendinosus selectively, but trickier to train the semimembranosus and biceps femoris selectively. How we can train different regions of each muscle is currently unclear, although hip extension exercises might work proximal regions to a greater extent, and knee flexion exercises might load distal regions to a greater extent.

Overall, the largest differences between exercises occur between the seated leg curl, which will preferentially load the semitendinosus very markedly, and the horizontal back extension, which will load all of the two-joint hamstrings fairly equally. For high-level athletes and bodybuilders, a complete selection of exercises can be obtained by using a leg curl with outwardly turned feet to shift the load towards the biceps femoris (short and long heads), and a lying leg curl with the strength curve altered to require peak force at full knee extension to load the semimembranosus selectively.


How are the hamstrings used during sprinting?

During sprinting, the hamstrings primarily work to produce high forces while the muscle is either contracting isometrically or eccentrically, during the terminal swing phase, while the hip is somewhat flexed but moving rapidly towards full hip extension, and while the knee is extending. Their main role is to perform large amounts of negative work done at the knee, and the amount of negative work done is closely linked to running speed.

As the hip and knee approach full extension, the internal moment arm lengths of each of the hamstrings at both joints are at their most similar, and therefore contribute most equally. At greater degrees of hip and knee flexion, the semitendinosus becomes progressively more important. When the knee gets within approximately 40 degrees of full knee extension, the contributions of the semitendinosus and semimembranosus are roughly equal. Consequently, it seems likely that each of the hamstrings contribute fairly equally to force production in terminal swing (and certainly at the point of ground contact), at least based on an analysis of internal moment arm lengths (whether each region contributes similarly is a different matter, however).

Early research identified that the medial hamstrings may be more active during sprint running. However, more recent research has shown that the activity of the medial and lateral hamstrings differs only in the middle of the gait cycle, and not in the important terminal phase, which is the phase in which work done determines running speed.

This same research has identified that the activity of the medial and lateral hamstrings differs between accelerating and maximum speed running. This could be taken to imply that these two phases of sprint running require a proportionally different involvement from the medial and lateral hamstrings. Even so, given that athletes probably cannot easily alter accelerating running ability without also altering maximum speed, such differences in hamstrings muscle activation in the middle of the gait cycle may not be relevant. The similar activation of the lateral and medial hamstrings in terminal swing may be why the qualities of acceleration and maximum speed are in fact quite closely linked (and are underpinned by a common limiting factor).


Why are the hamstrings strained during sprinting?

The main role of the hamstrings muscle group during the sprinting movement is to absorb energy in the terminal swing phase of the gait cycle, repeatedly over time.

Since this activity involves producing high forces while the muscle group is being forcibly lengthened and fatigued, this therefore puts the whole muscle group at high risk of being strained.

Researchers have identified that the (proximal region of the) biceps femoris long head is the most commonly strained hamstrings muscle during sprinting, but the exact reasons for this are unclear. This may be related to anatomical or morphological features that affect the extent to which this muscle is required to absorb elastic energy in the terminal swing phase.

The common location of the strain in the proximal region may be linked to differences in activation of the proximal and distal regions of the biceps femoris (long head) during high-speed running, or to the differences in muscle architecture muscle between regions, or to any number of other factors that have not yet been identified.


Can we reduce the risk of hamstrings strains by selectively training muscles or regions?

Given that the biceps femoris long head has been identified as the hamstrings muscle that is most frequently injured during high-speed running, it can be tempting to identify exercises that can be used to train this muscle selectively.

However, whether selectively training one muscle out of the whole hamstrings muscle group would have a beneficial effect on hamstrings strain injury risk seems doubtful, since it is likely that all of them are approximately equally important in the terminal swing phase, and a reduced capacity to produce force in one muscle can probably be compensated for by an increased capacity to produce force in another muscle.

In contrast, whether different regions of each hamstrings muscle might be trained selectively to reduce the risk of hamstrings strain injury is more difficult to identify, because of the lack of information about the various different regions within each of the muscles. Some new research suggests that there may be unique (architectural) features of the proximal region of the biceps femoris (long head) that make it more susceptible to injury, but this requires further work to clarify.


How can we help reduce the risk of the hamstrings being strained during sprinting?

Since it is unlikely that selectively training different hamstrings muscles might be able to reduce the risk of a strain, what other methods are available for reducing the risk of a hamstrings strain?

To understand the best methods requires an understanding of what causes muscle strains. While many aspects of muscle strains remain unclear, the following key points form the basis of our understanding.

Essentially, the research tells us that eccentric strength (force production while lengthening), muscle fascicle length (maximum distance lengthened), and fatigue resistance can each affect how likely it is that a muscle will be strained in a given movement. Greater force while lengthening, greater fascicle lengthening, and greater fatigue all increase the risk of a muscle strain occurring.

There are therefore two ways in which we might reduce the chances of a strain injury: (1) to increase resistance to fatigue, or (2) to increase the ability of the muscle to produce force while lengthening.

Increasing fatigue resistance requires increasing the oxidative capacity of the muscle, thereby increasing muscular endurance, and this has the potential downside of fiber type shifts, as well as decreases in muscle size. In addition, fatigue resistance can simply lead to the muscle strain occurring later, because the athlete just runs faster for longer before the strain occurs. Consequently, while increasing fatigue resistance can seem logical at first glance, it is neither the best course of action when high performance is a key concern, nor is it as effective as the alternative.

In contrast, eccentric hamstrings training reduces the risk of strain injury by increasing the active stiffness of the muscle, partly by increasing the ability to produce voluntary force with the central nervous system, and partly by increasing the tensile strength of the passive tissues. Moreover, an increase in the eccentric strength of the muscle relative to the concentric strength means that the hamstrings muscles will almost always be able to absorb more elastic energy than the amount of kinetic energy they can produce in the sprinting gait cycle, given that the kinetic energy is produced metabolically and the elastic energy is absorbed partly metabolically and partly passively.


What is the takeaway?

The hamstrings are a very varied muscle group, and each muscle within the group can be targeted with a different exercises or loading type. There are also different regions within each muscle, as identified by their separate innervation, anatomy, and architecture, but how these regions can be targeted is currently unclear.