Neck Strength and Concussion Risk: What Studies Show

A youth soccer coach texted me last week after her team's third head-to-head collision this season. The athletic trainer had told parents to start their kids on neck exercises before the next matches - for concussion prevention, she said. The coach wanted to know if that advice actually holds up, or if it just sounds responsible without doing much. It is a fair question, and it does not have a clean yes-or-no answer. The research on neck strength and concussion risk is real, but it is also incomplete, mixed in places, and easy to oversell in a sideline pep talk. What follows is what the actual studies measured - the effect sizes, the sample sizes, and where the evidence runs out - along with a training approach built on what we actually know rather than what sounds convincing before kickoff.

The theory behind neck training for concussion prevention rests on basic physics. A concussion results from the brain accelerating and decelerating inside the skull, driven by both linear force (a straight-line jolt) and angular force (a rotational whip). Newton's second law says force equals mass times acceleration - rearranged, acceleration equals force divided by mass. If the neck braces hard enough right before contact, the head effectively couples to the trunk, and the mass resisting that impact force jumps from roughly 4-5 kg (the head alone) to something closer to 50+ kg (head plus trunk). Same impact force, larger effective mass, lower acceleration transmitted to the brain.

That mechanism only works under specific conditions. It requires the athlete to see the hit coming and pre-activate the neck musculature 50-100 milliseconds before contact - what researchers call anticipatory bracing. A blindside hit, a fall where the athlete does not see the ground coming, or contact from an unexpected angle removes that window entirely. This distinction matters more than most concussion-prevention marketing acknowledges: neck strength is a modifier for anticipated impacts, not a shield against every mechanism that produces a concussion.

The most frequently cited paper on this topic is Collins et al. (2014), published in the Journal of Primary Prevention. The research team tested isometric neck strength - flexion, extension, and bilateral lateral flexion - in 6,704 high school athletes across 51 schools before their competitive seasons, then tracked concussions through certified athletic trainers over the following year.

• For every one-pound increase in neck strength, the odds of sustaining a concussion dropped by roughly 5%
• Female athletes scored lower on every neck strength measure than male athletes in comparable sports, and also sustained concussions at a higher rate in those same sports (soccer and basketball had enough female participation for a direct comparison)
• A larger head-to-neck circumference ratio - a bigger head relative to a smaller neck - was independently associated with higher concussion risk

Those numbers sound decisive, and they get repeated that way in coaching clinics. Read the methods section and the picture gets more careful. This was an observational cohort, not a randomized trial - nobody assigned half the athletes to a neck program and left the other half untrained. Strength was measured once, in a static preseason test, not tracked as it changed across the year. And athletes with stronger necks in this sample may simply have had better overall strength and conditioning programs, better coaching, or other protective habits the isometric test was never designed to capture. Collins and colleagues framed their finding as an association, not a proven causal chain, and that framing is worth keeping in mind the next time someone promises a neck program is a guaranteed fix.

Hrysomallis (2016) reviewed the broader neck-strength literature across rugby, gridiron football, and ice hockey in Sports Medicine, and the picture across studies is inconsistent. Some cohorts show the same inverse relationship Collins found - stronger neck, lower injury rate. Others show no relationship at all once playing position and overall body mass are controlled for. The review's own conclusion is that neck strength is a plausible protective factor, not a confirmed one, and it specifically calls for prospective trials that track strength changes over a season rather than a single baseline snapshot.

The most direct test of the biomechanical hypothesis comes from Mihalik et al. (2011), who fitted youth ice hockey players with head-impact sensors for an entire season and separately tested their isometric neck strength. If the bracing theory held up cleanly, stronger-necked players should have shown lower peak head accelerations during real game impacts. They did not - the study found no significant relationship between isometric neck strength and the linear or rotational acceleration recorded during actual collisions. The authors' interpretation is worth sitting with: a static strength test measures peak force production in a controlled position, but a concussion happens during a fast, unanticipated, dynamic event. The muscle qualities that matter in that window - reaction time, activation speed, and coordinated timing across the cervical musculature - are not the same qualities a handheld dynamometer captures on a quiet afternoon in the training room.

Assume for a moment that the association holds and that building neck strength is worth doing regardless of how confidently it prevents concussions - it also protects against neck strain and whiplash-type injuries, which is reason enough on its own. The training literature (Hrysomallis, 2016, and the program descriptions built into Collins-style cohorts) points to a small set of protocols that show up repeatedly:

That last row matters more than it looks. Most neck programs stop at isometric holds and never progress to anything unanticipated, which means they train the exact quality Mihalik's data suggests matters least (peak static force) and skip the one that might matter more (fast, reactive stabilization). If there is time for only one addition to an existing warm-up routine, the reactive perturbation drills are the piece worth prioritizing over another set of isometric holds.

The sex gap in Collins's data deserves more attention than it usually gets. Female athletes in that cohort were not just weaker in absolute neck strength - they also carried a less favorable head-to-neck-circumference ratio on average, and both factors independently predicted higher concussion odds. That is a real, actionable signal: female athletes in collision-adjacent sports may see more benefit per unit of training time invested in cervical strength than male athletes who already sit further up the strength curve. It is also a signal that gets flattened into vague advice about girls doing neck exercises too, without the specificity that would make a program worth an athlete's time.

Here is what nobody can currently claim with a straight face: that a randomized controlled trial has taken two comparable groups of athletes, put one through a structured neck program and left the other alone, and shown a statistically significant drop in diagnosed concussions in the trained group. That trial does not exist yet, largely because concussion incidence is low enough per athlete-season that detecting a difference would require an enormous sample, and because blinding a neck-strength intervention is close to impossible. Until that trial exists, everything here is association, biomechanical plausibility, and one well-designed null result on the direct acceleration measure. That is enough to justify adding neck work to an existing program. It is not enough to promise a parent it will prevent their kid's next concussion.

For a team or individual athlete starting from zero, here is a sequence that matches the dosing described in the research above rather than guessing at volume:

1. Week 0 - baseline test. Use a handheld dynamometer (or a fish scale and strap, if that is what is on hand) to record flexion, extension, and bilateral lateral flexion strength. Flag any side-to-side difference greater than 10-15%, which marks an asymmetry worth addressing before adding load.
2. Weeks 1-2 - isometric foundation. Three sets of 6-10 second holds in all four directions, three times per week, folded into the existing warm-up rather than run as a separate session.
3. Weeks 3-4 - manual resistance. Once holds feel genuinely easy at the 10-second mark, add partner-resisted work through a small range of motion, 2-3 sets of 8-12 reps per direction.
4. Weeks 5-6 - loaded and reactive work. Introduce harness or band-resisted extension for direct strength gains, and layer in unanticipated perturbation drills - a coach applies a light, unpredictable push to the head that the athlete has to resist and stabilize against.
5. Re-test at week 6-8. Repeat the baseline dynamometer test. An athlete who has not improved by at least 10-15% on the weaker side is either under-dosing the program or carrying an asymmetry that needs individual attention beyond the group protocol.

The most common mistake in team settings is compliance, not programming. A four-direction isometric circuit takes under four minutes; teams that fold it into the first four minutes of every warm-up get near-universal compliance, while teams that schedule it as a separate neck day lose most of the room within three weeks. If the choice is a technically superior program nobody actually does or a simpler one that survives contact with a real season, take the one that survives.