Stuart McGill on Abdominal Training

The science of spine stability: Effective spine stabilization approaches must begin with a solid understanding of what stability is. From a spine perspective it has little to do with the ability to balance on a gym ball. This is simply the ability to maintain the body in balance which is important but does not address the unstable spine. In fact, in many instances the unstable spine is also flexion intolerant and with associated intolerance to compression. Sitting on an exercise ball performing movement exercises increases spine compression to a flexed spine. This retards progress – it is generally a poor choice of back exercise until quite late in a therapeutic progression. True spine stability is achieved with a “balanced” stiffening from the entire musculature including the rectus abdominis and the abdominal wall, quadratus lumborum, latissimus dorsi and the back extensors of longissimus, ilioicostalis and multifidus. Focusing on a single muscle generally does not enhance stability but creates patterns that when quantified result in less stability. It is impossible to train muscles such as transverse abdominis or multifidus in isolation – people cannot activate just these muscles. Do not perform abdominal hollowing techniques as it reduces the potential energy of the column causing it to fail at lower applied loads (McGill, 2009). Interestingly a recent clinical trial (Koumantakis et al, 2005) compared the efficacy of many of the exercises that I quantified and published in Physical Therapy (McGill 1998), with the same exercises combined with specific transverse abdominis isolation (hollowing etc.). Adding the specific transverse abdominis training reduced efficacy! Instead, the abdominal brace (contracting all abdominal muscles) enhances stability. Target contraction levels for bracing and training techniques are described in McGill (2006). Finally, some provocative tests, such as a shear test, will help reveal which classification of patient is best suited for a stabilization approach (Hicks et al, 2005).

Linking Anatomy with Function: Consider the usual and popular approach to train the abdominal wall muscles by performing situps or curl-ups over a gym ball for example. But consider the rectus abdominis where the contractile components are interrupted with transverse tendons giving the “six pack” look. The muscle is not designed for optimal length change but rather to function as a spring. Why have these transverse tendons in rectus abdominis? The reason is that when the abdominals contract, “hoop stresses” are formed by the oblique muscles that would split the rectus apart. In addition to the spring-like architecture of the muscle consider how it is used. People rarely flex the rib cage to the pelvis shortening the rectus in sport or everyday activity. Rather they stiffen the wall and load the hips or shoulders – if this is performed rapidly such as in a throw or movement direction change, the rectus functions as an elastic storage and recovery device. When lifting weights it stiffens to efficiently transmit the power generated at the hips through the torso. Those individuals who do actively flex the torso (think of cricket bowlers and gymnasts) are the ones who suffer with high rates of disc damage and pain. Now revisit the common training approach of curling the torso over a gym ball that replicates the injury mechanics while not creating the athleticism that enhances performance. This is a rather poor choice of exercise for most situations. Yet many clients will expect that a gymball be used. Play a trick on these clients and retain the gymball but change the exercise from a spine breaking curlup to a plank where the elbows are placed on the ball. Now “stir the pot” to enhance the spring and spare the spine – this is a much superior exercise for most people.


gluteal muscle activation retraining based primarily on the original work of Professor Janda has been honed in our own lab (see figure 4). This cannot be accomplished with traditional squat training (McGill, 2007). Chronic back pain tends to cause hip extension using the hamstrings and subsequent back extension using the spine extensors creating unnecessary crushing loads. Gluteal muscle reintegration helps to unload the back. 

Finally consider exercises such as the squat. Interestingly when we measure world class strongmen carrying weight, NFL footballers running planting the foot and cutting – neither of these are trained by the squat. This is because these exercises do not train the quadratus lumborum and abdominal obliques which are so necessary for these tasks. In contrast, spending less time under a bar squatting and redirecting some of this activity with asymmetric carries such as the farmers walk (or bottoms-up kettlebell carry – see figure 7) builds the athleticism needed for higher performance in these activities in a much more “spine friendly” way. The core is never a power generator as measuring the great athletes always shows that the power is generated in the hips and transmitted through the stiffened core. They use the torso muscles as anti-motion controllers, rarely motion generators (of course there are exceptions for throwers etc but the ones who create force pulses with larger deviations in spine posture are the ones who injure first). Many more progressions to enhance athleticism in a spine sparing way are provided in my text “Ultimate back fitness and performance”