Skip navigation

Monthly Archives: October 2011

There has been a debate for a number of years about the effectiveness of the powerlifts (bench press, squat, deadlift) at improving power. The thought has been that while these lifts increase maximal strength, they are performed too slowly to train power adequately. Powerlifting proponents, especially with the advent of training with chains and bands, have argued this is not the case. Swinton et al, in the November issue of the Journal of Strength and Conditioning Research conducted a study looking at the deadlift exercise and came up with some fascinating data on the deadlift that may require that this argument be reevaluated.

The authors had 23 experienced subjects (powerlifters and rugby union athletes) participate in the study. On average, the athletes weighed 107 kilograms and deadlifted 227 kilograms. In other words, these athletes are deadlifting more than twice their body weight.

The athletes performed two testing sessions. During the first session, athletes maxed out on the deadlift. They then performed a single repetition at 30%, 50%, and 70% of their 1-RM at maximal velocity.

During the second session, the athletes performed repetitions at 30%, 50%, and 70% of 1-RM at submaximal velocity. After this, they performed maximal velocity repetitions at 30%, 50%+chains equal to 20% of 1-RM, then 70%+chains equal to 40% of 1-RM. Two repetitions were performed at each weight.

Subjects performed the lift on a force platform and had each lift filmed.

Many of the results are what you’d expect:
• As the weight increased, the velocity of the barbell decreased. The velocity was greatest at the maximal velocity trial with 30%+chains (at 2.2 meters per second).
• For the submaximal velocity trials, peak power increased as the resistance increased. For the maximal velocity trials, it was greatest at the 30%+chain load and decreased as the weights increased. Peak power for the maximal velocity trails was more than double that of the submax trials for every resistance except 70%+chains.
• The acceleration phase for performing the lifts with chains is greater than the submax conditions.

The velocity information is fascinating. These numbers exceed some of the velocities seen in the second pull of the snatch and clean as well as the drive of the jerk exercise. This suggests that the deadlift can be performed in an explosive manner to train for athletic power.

Now, there are some challenges with this study. First, the athletes studied are able to deadlift double bodyweight, which means that they have some skill on the exercise. It’s unclear if the results can be transmitted to other athletes. It is actually likely that there needs to be a strength base present before advanced training tools, like chains, can be effective. Second, the athletes self-selected “submaximal” and “maximal” velocities. Without standardizing the submaximal velocities (i.e. lift at 70% of maximal deadlift velocity, etc.) it makes it difficult to compare the results across lifters and to apply them to a larger pool of athletes. The final limitation that needs to be kept in mind is that this is not a training study. In other words, this study is not looking at the effectiveness of X number of weeks of training the deadlift with chains on power. It’s a snapshot in time and this needs to be kept in perspective when reading about it.

Swinton, P.A., Stewart, A.D., Keogh, J.W.L., Agouris, I., and Lloyd, R. (2011). Kinematic and kinetic analysis of maximal velocity deadlifts performed with and without the inclusion of chain resistance. Journal of Strength and Conditioning Research, 25(11), 3163-3174.


Coaches are always looking for the most effective way to train their athletes. This makes sense as training time and recovery ability are both finite, so if a training approach can be found that provides maximum results with minimum time it would be beneficial. Alcaraz et al in the November issue of the Journal of Strength and Conditioning Research tried to look at this using national-caliber Spanish sprinters.

The authors studied ten sprinters, averaging 22 years of age. The sprinters were able to half squat (down to 90 degrees of knee flexion on a Smith machine) on average 270% of their body weight and ran the 60 meter sprint in a little more than seven seconds.

The authors approach was to have the athletes perform three repetitions at each of 30, 45, 60, 70, and 80% of their 1-RM on the half squat on a force platform. The power production at each of these loads would be analyzed and would also be compared to the athlete’s best 60 meter time to see if there is any relationship between the two.

The results are interesting:
• Peak power increased as the load was increased from 30% to 45% to 60% of 1-RM.
• Peak power was greatest at 60% of 1-RM.
• Peak power decreased as the weight was increased above 60%.
• However, none of these changes were statistically significant. Between 30% and 45% of 1-RM, peak power only increased by 3%. Between 45% and 60% of 1-RM, peak power only increased by 1%. Between 60% and 70% and between 70% and 80%, peak power decreased by 6% each.
• There was no relationship between any of the peak power measures and 60 meter performance.

The intent of the study was to quantify at which intensity a sprinter is developing peak power at the half squat exercise in a Smith machine. Knowing that would allow coaches to really focus the sprinter’s training when they are attempting to develop this quality using this exercise. As the weight on the bar gets heavier, we’re eventually going to move it more slowly. Eventually this results in a decrease in peak power. In this study, while peak power was generated at 60% in this population, this does conflict with other studies using sprinters and with other studies looking at different exercises. It also needs to be pointed out that while peak power occurred at 60%, this was only a 4% increase over the power production at 30% of 1-RM.

The good intention of studies like this, in my opinion, both overly complicates training (i.e. there is a peak range for someone to train at for “x” quality) and overly simplifies it (using the half squat for power production is only one tool among many that have to be balanced in the big picture).

Alcaraz, P.E., Romero-Arenas, S., Vila, H., and Ferragut, C. (2011). Power-load curve in trained sprinters. Journal of Strength and Conditioning Research, 25(11), 3045-3050.

There have been a number of studies over the years looking at how long it takes to achieve hypertrophy as a result of strength training. This is important because it’s one of those things that motivates a lot of people to train. There has been a lot of conflicting information with regards to this, everything from 20 days to eight weeks before anything statistically significant occurs (note that this is different than being able to see it in the mirror).

DeFreitas et al, in the November issue of the European Journal of Applied Physiology, studied this. In their literature review, they noted that many studies that look at hypertrophy suffer from three flaws. First, many studies only measure at the beginning and the end of the study. This makes it difficult to determine the time course of hypertrophy, it only gives you snapshots at the beginning and the end. Second, some studies rely only on anthropometery, in other words using a tape measure around the muscle. Third, many studies just don’t train the subjects hard enough to produce hypertrophy.

The authors of this study looked at 25 untrained men. The subjects trained for eight weeks, using the leg press, leg extension, and bench press. They performed three sets to failure (around 8-12 reps per set), with two minutes rest between sets. Before the study, every fifth day during the study, and at the end of the study the subjects were tested on isometric leg extension strength and the cross-sectional area of the right thigh was measured using quantitative computed tomography (which is considered by the authors to be equivalent to an MRI measurement).

At the end of the study, muscle cross-sectional area had increased by almost 10% and isometric strength had increased by almost 24%.

The time course of these adaptations are very interesting:
• Muscle cross-sectional area had the biggest increases at the end of weeks one, three, five, and six. The gains leveled off in weeks seven and eight.
• Strength had the biggest increases at weeks three, four, seven, and eight.

The fact that muscle hypertrophy gains are leveling off by week seven and eight suggest the importance of variety to continue making gains from exercise. The time course of the strength gains suggests that are increasing their hypertrophy and then translating that to increased strength. It’s not showing a large neural component that results in early strength gains.

DeFreitas, J.M., Beck, T.W., Stock, M.S., Dillon, M.A., and Kasishke II, P.R. (2011). An examination of the time course of training-induced skeletal muscle hypertrophy. European Journal of Applied Physiology, 111: 2785-2790.

I was at the Early Childhood Intervention Advisory Committee meeting yesterday. A lot of the focus for the committee is going to be on writing next year’s annual performance report (APR), which I’ll talk more about later. The rest of the meeting was divided between the changes that are ongoing with ECI in Texas currently and also with a federal survey that was performed.

Changes to ECI:

Changes to ECI have been happening because state funding has been reduced and this is likely to get worse. This will be complicated by a future reduction in Federal funding (i.e. Medicaid) when the debt reduction efforts get serious. As a result, other ways have to be found to financially support ECI (after all, providers aren’t providing ECI services completely out of the goodness of their hearts – it’s a business for them and people want to be paid). The major changes are family cost share and having the ECI providers bill insurance and managed care for the services (see for a summary of these changes). These changes have some interesting and immediate ripple effects. First, ECI providers now have to decide between hiring therapists and hiring billing specialists. In other words, they are sometimes having to enhance their administrative support at the expense of service providers (and this has to happen to get reimbursed for services). Second, this has the potential of making people less likely to take advantage of ECI services which could have long-term implications for their children.

It is likely that in the future state funding is going to continue to be reduced to ECI programs. It is also likely that federal funds will dry up as well as part of debt reduction. This is going to have a number of effects. First, state and federal dollars will probably end up being prioritized to those populations least able to afford the services. In other words, ECI will become an indigent program. Second, ECI providers will see declining profit margins resulting in fewer (and larger) programs that can take advantage of economies of scale. This has pros and cons. Third, ECI providers will increasingly need to rely on managed care and family cost share to service those populations not covered by state/federal funds. This has the potential to have more people focus on private therapy or to have more people simply not use the services.

Federal Survey:

The Feds had a third party put out an online survey about ECI to parents. Part of yesterday’s meeting was devoted to covering the results of the survey. The survey was very positive and painted a picture of general satisfaction with ECI. But there was a problem. About 1100 people submitted a survey. Of those, about 8% had kids currently in ECI programs (this is 88 surveys). This is a problem because 28,000 children receive ECI services each month (i.e. this represents 0.3% of the population). A response rate this low completely invalidates the results of the surveys, in other words the information is useless.

I asked about this and found out that the Feds did not want any ECI providers or DARS associated with distributing the survey (they might bias the results), so this means that it did not go out to very many parents.


I posted about ECI’s annual performance report earlier (see as there are some challenges with it. I will have a lengthy phone conversation with DARS’ staff about the items relating to being a statistics geek (I did not want to drag out the meeting with things that only bother me), but the long and the short is that the DARS staff realize many of the shortcomings behind the indicators and measures being used in the report. The reality is that the federal Office of Special Education Programs is mandating many of these things even though they are flawed and don’t allow for meaningful conclusions to be drawn.

Some of these useless reports and surveys represent a great example of how we can save taxpayer dollars. People are paid to develop the reports and surveys, analyze them, write summaries, write follow ups, continually track the data, etc. Eliminating a lot of these would result in real savings to taxpayers and might alleviate the need to cut services to needy people.

There has been a small debate in the strength and conditioning literature in recent years concerning how a field sport athlete should start a sprint. Generally options include starting them track style (i.e. one foot back), starting them from the athletic ready position with the feet parallel to each other, or starting from the parallel position but taking a quick step backwards prior to sprinting forward. In theory all positions have some advantages. The track style is how sprinters begin and (since sprinters do it) should result in a faster sprint. The parallel position is probably more realistic given that many athletes are in the ready position during the sport and may not have time to set up for an artificial starting position. The step backwards, in theory, would result in an enhanced stretch reflex resulting in a faster sprint.

Frost and Cronin looked at the kinetics of each of those starting positions in the October issue of the Journal of Strength and Conditioning Research. They had 27 men with an athletic background (had participated in sports, but no one was national caliber) perform three 5 meter sprints, one using each starting style. Each athlete received an audio command on the start and (to standardize) each moved the right foot first for each of the starts.

With regards to results, the authors subdivide their subjects into fast and slow subjects. The results listed below are for the faster subjects:
• The split start results in the fastest 2.5 meter and 5 meter sprints, the false start is the second fastest group, and parallel is slowest. Now, the differences aren’t great (for example, in the fast group they made it to 2.5 meters in 0.56 seconds, the false group made it in 0.6 seconds, and the parallel group made it in 0.67 seconds). At 5 meters, the differences is a hundredth of a second between the split and false groups (i.e. split starts are a hundredth of a second faster than false starts).
• The parallel start has the greatest horizontal impulse (meters/second), followed by split starts, followed by false starts.
• The parallel start results in the greatest horizontal peak force, followed by the split starts, followed by the false starts.
• The parallel start had the greatest ground contact time, followed by the split start, followed by the false start.
• The horizontal time to peak force was greatest in the parallel start, followed by the split start and then the false start.

The results are interesting and seem to indicate that in terms of kinetic variables, the false start is superior to the parallel and split styles over extremely short distances (five meters). Even though there was not an advantage in terms of time to 2.5 meters and 5 meters, the kinetic variables were superior suggesting that with training this might result in faster times over shorter distances.

Now, this needs to be applied with care. From a static starting position over very short distances these results may hold true. It needs to be kept in mind that highly trained athletes may perform completely differently on this study. It also needs to be kept in mind that in most field sports, the situation is too fluid to have an artificial static starting position like what has been described here. Usually the sprint will occur from another type of movement. For example, an athlete is jogging then must break into a sudden sprint. It’s unclear if these results would apply in that kind of situation.

Frost, D.M. and Cronin, J.B. (2011). Stepping back to improve sprint performance: A kinetic analysis of the first step forwards. Journal of Strength and Conditioning Research, 25(10), 2721-2728.

As part of my role as a parent member on the ECI Advisory Committee, I was sent the link to ECI’s annual performance report (APR) and state performance plan (SPP). The link to both can be found at:

After receiving the email about it, I have read through both the APR and the SPP. The APR is the report to the Federal government about how ECI is meeting the targets in the SPP from Sep 2009 to August 2010. What follows are the comments and questions that I submitted to DARS about the information contained in the APR. It would be most helpful to open the APR to follow these, the APR is pretty long and this would be a very long post to repeat everything in it.

Stakeholder Involvement:
Throughout the documents, it mentions many areas where the ECI Advisory Committee was heavily involved in reviewing and providing input. Is this still being done in an ongoing manner?

Under Indicator 1:
• The APR mentions most children receiving services within 28 days of the signing of the Individualized Family Service Plan (IFSP). I understand that in the real world it takes time to mobilize staff and resources, but it also represents almost a month of lost opportunities. Is 28 days the mean for the delivery of services? If not, do we know what that mean is?
• The data that is reported for this indicator (and several others as well) is from March to May 2010 (i.e. a 3-month period). The assumption is that this 3-month period is representative of the other nine months in that year, though clearly they may not be. Is this the standard for other states (i.e. is it best practices)?
• With regards to untimely delivery of services:
o Extrapolating data from 20% of a sample size of 277 and attempting to apply it to almost 9600 has some flaws.
o With the table (page 5), are we saying that for 138/139 untimely delivery of services is the result of family-related reasons?
• Do we have any numbers on CAPTA referrals? Is this a burden on the ECI system?
• I appreciate the attempt to tie this back to inadequate funding.
• There’s mention (page 7) of a newly formed committee of program directors and supervisors to streamline paperwork and processes. Is this still around?

Under Indicator 2:
• Is a one-day child count a representative sample? What if it were a weird day (it’s Halloween after all)?
• Has the guidance for implementing group services in ECI been issued (page 9)?

Under Indicator 3:
• On page 13 (progress data table), this seems to say that 38-45% of children are not functioning at the level of their age-matched peers by the time they leave the program. This would seem to argue for the severity of what ECI sees, the need for more funding, the fact that this is going to be transferred to the school districts, and the need for more specialists…

Under Indicator 4:
• The approach to come up with the numbers (i.e. the number who had a mean score of four or higher on the Likert scale items) is interesting. Normally you see a breakdown of how many people marked each of 1, 2, 3, to 5 and then an average of the responses. By only focusing on four plus you are ignoring some of the sample.

Under Indicator 5:
• See above about the one-day sample.
• Texas results in 1.07% vs. the national 1.03%. Is this because Texas has different standards or because Texas has different demographics?

Under Indicator 6:
• Texas has 2.29% with IFSPs from birth to 3, from birth to 1 it’s 1.07%. Why do you think Texas is greater than the national average from birth to 1 but lower from birth to 3?

Are indicators 5 and 6 things that DARS can do anything about? If not, are they appropriate to measure?

Under Indicator 7:
• See above about the 3 month time sample.
• On page 23 it discusses DARS staff developing processes and procedures to identify noncompliance and verify the correction. What’s the status of this?

Under Indicator 8:
• 3 month sample period.
• The comment (page26) about a single person (part B 619 coordinator) holding up the entire process is concerning.
• How is the TEA restructure impacting things?
• What is the status of the three revisions on page 27?
• There’s a lot of cutting and pasting of verbiage in this document…

Under Indicator 9:
• Am I reading this correctly that 11 programs were the source of all 63 findings of noncompliance? Are there any penalties or anything like that?
• On page 31 there’s reference to changing the contract structure and new monitoring/oversight. Has this been done?

Sprinting is thought to be made up of several qualities; acceleration (the ability to increase velocity), maximum velocity (the fastest an athlete can run), and speed endurance (the ability to maintain velocity). Not all of these qualities are thought to be equally important for every athlete. Outside of sprinters, most athletes don’t have to worry about speed endurance. For most sports, acceleration and maximum velocity will be the most important qualities.

For a 100 meter sprinter, maximum velocity isn’t reached until 60-80 meters. If this holds true for other types of athletes, it suggests that most athletes are in an acceleration situation when sprinting in sports. There is a lot of mixed research on acceleration in the sense that some research shows that maximum strength is important for acceleration, some shows that it is not. Lockie et al in the October issue of the Journal of Strength and Conditioning Research investigated the variables that differentiate field sport athletes with better acceleration from those with slower acceleration.

The athletes were evaluated on 4×10 meter sprints, 3 trials each of five alternate bounds for distance, counter-movement jump, 40cm drop jump, leg stiffness was evaluated, 5 meter sprints over a force platform, and a 3-RM back squat (in a Smith machine) was used to evaluate strength.

The results are interesting:
• The faster group had a higher velocity from 0-5 meters and 0-10 meters.
• Stride length was similar between the faster and slower groups.
• Stride frequency had some differences, though they were not statistically significant. From 0-5 meters, the stride frequency for the faster group was almost 10% faster than the slow group. From 0-10 meters it was almost 7.5% faster than the slower group.
• Both groups exerted a similar amount of force against the ground, but the faster group had a much shorter ground contact time than the slower group.
• There were differences between the faster and slower groups in terms of the strength and power tests. The table below shows how much higher (or lower) in terms of percentage the faster group was when compared to the slower:

5 Bound Test


Counter-Movement Jump


Drop Jump Height


3-RM Squat


3-RM Squat/Body Weight


This study suggests that those field sport athletes who have higher velocities during acceleration may be achieving this due to a number of factors. First, they have a greater stride frequency (i.e. they are moving their limbs more quickly). Second, they are spending less time on the ground. Third, they have greater strength and power. All of these are qualities that are trainable.

This is an interesting study for a number of reasons. First, it uses field sport athletes. Unfortunately, we are not given any information on what level of athlete or what sport, which would be useful. But it is significant because it is likely that studies on elite sprinters probably won’t transfer to a soccer player. Second, it suggests that strength, power, and moving the limbs more quickly is important for acceleration. This is significant because the old school of thought says that we get faster by moving our limbs more quickly (i.e. technique, A drills, etc.). The other school of thought says that we get faster by exerting more force against the ground. This study suggests that both schools of through are correct.

Lockie, R.G., Murphy, A.J., Knight, T.J., and De Jonge, X.A.K.J. (2011). Factors that differentiate acceleration ability in field sport athletes. Journal of Strength and Conditioning Research, 25(10), 2704-2714.

Daniel Robbins conducted an analysis of the physical characteristics of football players drafted into the NFL between 2005 and 2009 that also participated in the combine. The analysis is broken down by position and by performance test.

Centers, defensive tackles, offensive guards, and offensive tackles were the slowest at the 9.1, 18.3, and 36.6 meter sprints as well as the agility drills (18.3m shuttle and L drill). These positions also had the shortest jumps (vertical jump and standing broad jump). However, they had the most repetitions at the bench press and the greatest predicted 1-RM bench press.

Wide receivers, corner backs, free safeties, outside linebackers, strong safeties, and running backs were the fastest on the sprints and agility drills. These positions also had the greatest jumps (horizontal and vertical). In terms of bench press repetitions and predicted bench press 1-RM, these positions performed the fewest number of repetitions and had the lowest predicted 1-RM.

Not surprisingly, the heaviest players are the ones that are slowest, have the lowest jumps, and are the strongest. Also interesting, offensive and defensive players that oppose each other have similar physical characteristics.

Based upon the physical characteristics reported in this article, the cornerback position seems to be the most athletic (fastest, most agile, best jumper) where the offensive guard seems to be the least (slowest, least agile, worst jumper).

It needs to be recognized that this is studying athletes that were invited to the combine and were drafted by the NFL. There is no analysis relating the performance on these tests to performance in the NFL (i.e. to their eventual success). There is also no indication on whether these measures were changing from 2005 to 2009, which would also be interesting to know.

Having said that, if performance on these tests does relate to performance in the NFL, then they can be used to both help select athletes and to help drive the athlete’s strength and conditioning.

Robbins, D.W. (2011). Positional physical characteristics of players drafted into the National Football League. Journal of Strength and Conditioning Research, 25(10), 2661-2667.

Depending upon the sport, female athletes are between two and eight times more likely to suffer an ACL injury than their male counterparts. There are a lot of theories about why this is so, one relates the hamstrings. The theory is that female athletes may have weaker hamstrings, which causes them to use a quadriceps-driven landing strategy during jumps. When this is done, it forces the knees forward and in (valgus) which is believed to increase the strain on the ACL.

Ford et al, in the October issue of the Journal of Applied Biomechanics, investigated the quadriceps and hamstring recruitment strategies during drop jumps of different heights in female athletes. The authros studied sixteen high school volleyball players and had them perform drop jumps from heights of 15cm, 30cm, and 45cm. Each drop jump was done onto a force platform.

The authors found that:
• In the preparatory phase (i.e. prior to ground contact) the hamstring to quadriceps activity increased as the drop height increased. It went from being a hamstring-dominated activity to being a quadriceps-dominated activity.
• This is interesting because the hamstring activity did not change, but the quadriceps activity increased by almost 50% as the drop height increased from 15cm to 45cm.
• During the reactive phase, quadriceps activity increased by 20-33% (depending upon the muscle) as drop height increased. Hamstring activity in the biceps femoris essentially did not change with increased drop height, but the semitendinosus increased by almost 23%.
• Ground reaction force increased by almost 50% as drop height increased.
• Hip flexion angles on landing decreased by almost 21% and knee flexion angles at landing decreased by almost 10%.

There are a number of interesting things from these results. First, as the drop height increased the female athletes had a shallower landing (this can be seen from the hip and knee flexion angles at landing). This tied in with the increased quadriceps activity that was seen as the drop height increased. The fact that hamstring activity did not increase in proportion to quadriceps activity may suggest that these athletes are more predisposed to an ACL injury from landing, certainly this would be true if the hamstrings are one of the contributors to ACL injuries as many people think today. There is a school of thought that in order to help prevent noncontact ACL injuries, work needs to be done enhancing hamstring strength and landing mechanics and this research would seem to support that need.

It needs to be pointed out that the results of this study may be limited to these sixteen athletes. In other words, just because these athletes react to different drops heights in this manner does not mean that all female athletes will do the same. It is possible that different training backgrounds, techniques, and strength levels could impact landing mechanics. It is also likely that these athletes will function differently in a game situation than in an artificially induced laboratory situation.

Ford, K.R., Myer, G.D., Schmitt, L.C., Uhl, T.L., and Hewett, T.E. (2011). Preferential quadriceps activation in female athletes with incremental increases in landing intensity. Journal of Applied Biomechanics, 27, 215-222.

Stokes et al in the October issue of Clinical Biomechanics are looking at the ability of various abdominal muscle groups to contribute to spinal stability. The idea behind spinal stability is that some sort of buckling at the spine leads to lower back injuries. However, the authors point out that this has never been documented in vivo. The thinking is that not only will the abdominal wall muscles maintain lumbar stability, but they will also increase the intra-abdominal pressure which will also help to maintain stability.

This study was meant to look at the effects of activating different abdominal muscles on spinal stability. It needs to be emphasized that this study is using a mathematical model of the five lumbar vertebrae with the muscles that act on them. The model was loaded in flexion, extension, lateral bending, and axial rotation with moments and intra-abdominal pressure gradually increasing. Different patterns of muscle activation (rectus abdominus, internal/external obliques, and transversus abdominus) at 10% and 20% of maximum) were studied during this loading.

The authors found that increasing intra-abdominal pressure increased the stability of the spine. However, preferentially recruiting abdominal muscles had no impact on stability.

These results need to be considered carefully and in context. First, it’s a mathematical model which means that there are major limitations to its applicability. Mathematical models cannot take into account the infinite variables such as training history, training adaptations, injury history/adaptations, fluid in structures, the fact that muscles work together, etc. Second, by examining the contribution of individual muscle activation, it ignores the fact that in reality all the muscles of the abdominal wall would be working together which could easily change the results. Third, a mathematical model has to make assumptions about movement parameters and the shape of structures which may or may not influence the applicability of the results.

All mathematical models of human motion have limitations, this is why this type of research needs to be considered extremely carefully before it is applied. As far as this research goes, it suggests that preferential recruitment of individual abdominal wall muscles may not have an impact on spinal stability, but this study does have some limitations.

Stokes, I.A.F., Gardner-Morse, M.G., and Henry, S.M. (2011). Abdominal muscle activation increases lumbar spinal stability: Analysis of contributions of different muscle groups. Clinical Biomechanics, 26, 797-803.