What the Cervical Spine Can Tolerate

Why this data exists

Nobody was ever going to load a living neck to failure to find out what it takes. The numbers below exist because the automotive industry needed them to certify restraint systems, and they were obtained through cadaveric testing, volunteer sled experiments at sub-injurious levels, animal models, and reconstruction of real collisions.

This is by a wide margin the most rigorous body of work ever done on what the cervical spine can withstand. It is freely available. It is almost never cited in any discussion of neck training, which is a strange gap given that everyone training a neck is applying force to one.

A caution before the tables: these are failure thresholds under acute impact loading, not training guidelines. Nothing you do in a gym approaches them. Their value is calibration — understanding the actual scale of what the structure tolerates, and what the safety margin in ordinary training really is.

Nij — the standard criterion

Nij is the normalised neck injury criterion, introduced by the US National Highway Traffic Safety Administration in 1999 and referenced by essentially all subsequent work. It combines axial force and bending moment into a single dimensionless number:

Nij = F_z / F_int + M_y / M_int

where F_int and M_int are size-specific intercept values. Those intercepts:

Nij = 1.0 corresponds to approximately a 22% risk of AIS≥3 neck injury — that is, serious injury — derived by logistic regression on porcine data.

Eppinger R, Sun E, Bandak F, Haffner M, Khaewpong N, Maltese M, Kuppa S, Nguyen T, Takhounts E, Tannous R, Zhang A, Saul R. Development of Improved Injury Criteria for the Assessment of Advanced Automotive Restraint Systems – II. NHTSA, November 1999.

Two things are worth noticing in that table. Extension tolerance is roughly 40% of flexion tolerance across every body size — the neck is far less able to withstand being bent backward than forward under load. And the child values are dramatically lower, which is relevant to anyone thinking about youth athletes.

Where the numbers came from

The intercepts are not arbitrary. Each traces to specific experimental work:

Compression, 4,000 N — Mertz and colleagues, 1978, from tackle-block impacts on a Hybrid III 50th-percentile male dummy. Pintar’s later dynamic cadaveric testing gave roughly 4,500 N compressive failure in males aged 30 to 35.

Tension, 3,300 N — Nyquist and colleagues, 1980, from real-world collision reconstruction. Yoganandan’s 1996 cadaveric tension tests failed intact specimens at a mean of 3,373 N — a striking convergence between reconstruction and direct testing.

Bending moments — Mertz, 1971, combining volunteer sled tests to establish pain thresholds with cadaveric tests to establish injury. A small-stature cadaver sustained ligamentous damage at 47.5 N·m of extension, scaled to 57 N·m for the mid-size male. Maximum measured flexion reached 190 N·m.

Combined loading — Prasad and Daniel, 1984, using matched juvenile porcine and three-year-old dummy airbag pairs, established that axial tension and extension moment combine linearly, and proposed 2,000 N and 34 N·m as paired limits.

NIC — the rear-impact criterion, with a usable graded scale

Nij handles frontal loading well and rear impact poorly. The Neck Injury Criterion was developed for the whiplash case:

NIC(t) = 0.2 · a_rel(t) + v_rel(t)²

where a_rel and v_rel are the relative acceleration and velocity between head and torso. The commonly used threshold is 15. More useful than the single threshold is the graded evidence behind it:

Li F, Liu N, Li H, Zhang B, Tian S, Tan M, Sandoz B. A review of neck injury and protection in vehicle accidents. Transportation Safety and Environment, 2019;1(2):89–105. DOI 10.1093/tse/tdz012.

That progression — nothing, then pain, then structural damage — is the most intuitively useful injury scale in this literature, because it maps a continuous physical quantity onto what a person would actually experience.

Tissue-level thresholds

Below the whole-neck level, individual structures have been tested directly.

Ligament

Intervertebral disc

Clinical instability markers

Transverse atlantal ligament lateral displacement beyond 7 mm, and an atlas–dens interval beyond 3 mm, are the figures used to identify instability at the upper cervical spine. These are the structures whose integrity conditions like rheumatoid arthritis can compromise — which is why that diagnosis appears on the safety page as requiring clearance before training.

Spinal cord

Stress above 8.1 kPa and strain above 0.0117 in finite element models; fragment impact velocity above 4.5 m/s in bovine models.

Li F et al., 2019, as above.

Lateral loading, which Nij does not cover

Side impact has its own thresholds, established for aerospace seat certification:

Meijer R, Philippens M, van Hoof J. Side Impact Neck Injury Criteria and Tolerances in Aerospace Safety. TNO Automotive, Proceedings of the Thirtieth International Workshop on Injury Biomechanics Research, 2002.

Note how much lower the lateral bending moment tolerance is than the flexion figures in the Nij table. The neck is weakest sideways, and lateral flexion is the direction most often skipped in training.

What this does and does not tell you

It does not tell you how much weight to use. These are acute impact failure thresholds. Training loads are orders of magnitude below them, and the relationship between muscular strength and structural tolerance is not established by any of this work.

It does tell you the shape of the risk. Extension tolerance is consistently lower than flexion. Lateral tolerance is lower still. The upper cervical ligaments fail at forces in the hundreds of newtons, not thousands. Children’s tolerances are dramatically lower than adults’.

And it tells you what the injury conversation is actually about. When people discuss whether neck strength protects against injury, the mechanism being proposed is that a stronger, stiffer neck changes how much of an external force reaches these structures. Whether it does so meaningfully is a separate question, addressed in head impact and concussion. The thresholds themselves are not in dispute.