it is known that keeping a neutral lower back under load is crucial to avoid injuries and/or low back pain. This because both an overextended back or a flat/flexed back squeeze the spinal discs, as explained for instance in each stronglift guide.

Also a rounded upper back is known to be bad under load. I mean something like that.

enter image description here

But, I don't understand why keeping an arched upper back (instead of a neutral one) isn't a problem in many exercises. I mean all the exercises where, for any reasons, retracting the shoulder-blades is required.


1) SQUAT: The shoulder-blades are retracted (and elevated) to keep the barbell stable throughout the exercise. Hence, the upper back is arched. It is explicitly written also here:

Arch your upper-back to create support for the bar. Squeeze your shoulder-blades and raise your chest.

enter image description here

2) BENCH PRESS: as explicitly written here (and everywhere), the middle-upper back must be arched in order to be stable on the bench and prevent front delt from activating too much.

enter image description here

3) PLENTY OF OTHER EXERCISES: dumbbell lateral raises, shrugs, the top phase of barbell row etc etc

Why doesn't an arched upper back squeeze the spinal discs?

1 Answer 1


Shoulder blade retraction is not the same thing as either upper back rounding/thoracic flexion or bridging/lumbar and thoracic extension.

The shoulder blades move independently of the thoracic spine, so you can have either scapular retraction or protraction occurring in any combination with thoracic flexion or extension.

The Stronglifts description of the upper back being "arched" in the squat is a misnomer. It's not clear from the article, but maybe he's referring to either the natural curvature of the thoracic in the neutral position, or to the retraction of the shoulder blades. Normally "arching" is only used to describe the spine in the bench press, where only the hips and shoulders are in contact with the bench, while the chest is raised between them to form an arch under the lifter's torso.

In the bench press, the spine is not subject to any external loading, so people don't care that it's in a non-neutral position.

The other examples you give (lateral raises, shrugs, rows) are all performed with a neutral thoracic, and either retracted shoulder blades (lateral raises) or shoulder blades moving throughout the lift (shrugs, rows).

  • Perfectly said, on all accounts. I was just about to write precisely the same thing.
    – POD
    Jul 1, 2021 at 4:41
  • Thank you very much. I've noticed I'm not able to clearly understand if my thorax is extended or not. Are there some cues to avoid excessive extension whilst retracting the shoulder-blades (for instance of the top of a row)? For instance, is the thorax in the following picture overextended? Is the arch I've been highlighted excessive? i.ibb.co/sgcjzrB/arcj.jpg
    – Kinka-Byo
    Jul 1, 2021 at 5:18
  • 1
    I think you guys mean thoracic flexion. thoracic extension is the opposite of rounding
    – user32213
    Jul 1, 2021 at 15:33
  • 1
    @Kinka-Byo I just wouldn't worry about it. Yes, that image appears to show thoracic extension, but there's no reason to think that's harmful. In the genuinely heavily axially loaded lifts (squat, deadlift), gravity is trying to pull you out of extension anyway, so over extending your spine is pretty much impossible there. Jul 2, 2021 at 2:36
  • 1
    @AceCabbie Thanks! I was getting mixed up with arching like in a bench press. I've edited it to correct/clarify it now. Jul 2, 2021 at 2:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.