Documentation:FIB book/Assessing the Efficacy of Multi-Directional Impact Protection System (MIPS) in Preventing Head Injuries

Overview

Background

Cycling is a common sport and form of transportation. Around 12 million Canadians ages 12 or older (41%) reported cycling, however, 890 cyclists died in Canada from 2006 to 2017, with an average of 74 deaths per year. Collisions with motor vehicles make up 73% of all cycling fatalities. A meta-analysis of 1,809 impacts found that head injuries are the key contributor to fatal and life-changing injuries among cyclists. A study from Baker et al. analyzed 13,684 cyclists and found that 21% of injuries with AIS2+  involved head injuries, and 70% of AIS4+ injuries were head injuries, with skull fractures being the second most common injury.

Head injuries occur in 70-80% of fatal bicycle crashes and contribute one-third of non-fatal injuries faced by cyclists. Head injuries resulting from cycling correspond to a higher risk of getting traumatic brain injury (TBI). TBIs can result from forceful impacts and strikes to the head, or from an object that breaks the skull and enters the brain. Certain types of TBI can cause short-term issues that affect typical brain functions such as thought processes, comprehension, movement, speech and behaviour. More severe cases of TBI on the other hand cause significant and lasting disability and can be fatal. Bicycling has one of the highest sport and recreational-related emergency room (ER) visits for TBI in the United States and the usage of bicycle helmets was found to reduce overall head injury by 48%.

Helmets have repeatedly been found to dramatically reduce focal injuries like skull fractures and subdural hematomas. It also significantly reduces the chance of life-threatening brain injuries. To further improve the protection effect of helmets, the multi-directional Impact Protection System (MIPS) is a popular new helmet design that reduces rotational forces during collisions. It incorporates a low friction layer that is built into the helmet to allow mobility of the original cushioning layer of the helmet. When the cyclist experiences an angled impact, the helmet layers rotate, instead of transferring the rotational forces to the brain. MIPS can reduce peak acceleration and velocity by 30-50% and prevent diffuse axonal injuries (axon damages caused by shearing force), concussions, and brain shear injuries.

Summary of Previous Studies

Experimental Metrics

Significant research has been done on the effectiveness of MIPS, using both physical drop tests and computational models. Common metrics used are peak linear acceleration (PLA), peak angular acceleration (PAA, sometimes called PRA), peak angular velocity change (PAV), and brain injury criterion (BrIC). Generally higher values of PLA are linked to skull fractures and other focal injuries, higher PAA and PAV values are linked to rotational injuries (diffuse axonal injury, and concussions).

Physical Tests

Bonin, DeMarco, and Siegmund (2022)

Bonin et al. evaluated the impact-mitigating effects of the MIPS technology in helmets under oblique impact conditions. Using a customized Echelon II helmet with and without a MIPS slip layer, they conducted 72 controlled drop tests. These involved a 50th percentile Hybrid III headform in three surface conditions: bare, covered with nylon stockings, and fitted with a human hair wig. The headform was released via a cantilevered arm onto a 45° anvil coated with 40-grit sandpaper, simulating real-world crash surfaces. The design allowed for consistent and repeatable impact conditions.

The key metrics analyzed included Peak Linear Acceleration (PLA), Peak Angular Acceleration (PAA), Peak Angular Velocity (PAV), and the Brain Injury Criterion (BrIC). Results showed MIPS significantly reduced angular acceleration and velocity. For Y-axis rotation, PAA decreased by 47% and PAV by 38%. However, the protective effects were diminished in the hair-covered headform condition, with PAA and PAV reductions dropping to 47% and 57% respectively, compared to 60% and 69% for the bare headform. The study also measured friction coefficients, finding that human hair (μ = 0.17) acted like a natural slip layer, potentially reducing the added benefit of MIPS. PLA values showed only minor differences across helmet types, echoing findings from prior computational studies.

Baker et al. (2024)

Figure 1. Sample Testing Apparatus of Baker et al. Paper

Baker et al. further investigated helmet performance across a broader sample by testing the 30 best-selling bicycle helmets in the UK, both with and without MIPS. Testing was conducted at Imperial College London using a drop tower rig. Helmets were fitted to a Cellbond-CEN headform, an improvement over the Hybrid III used by Bonin, as it includes realistic facial geometry and a partial neck segment to better simulate chin strap interaction.

Each helmet was dropped onto a 45° abrasive-coated anvil at 6.5 m/s across four impact locations: front, rear, side, and front-side. Twelve tests were performed per helmet (three per site across multiple samples). Metrics such as PLA, PAA, and BrIC were again evaluated.

The study found that the top eleven helmets ranked by overall risk all featured MIPS technology. Quantitatively, MIPS reduced Peak Resultant Velocity (PRV) by 19%, PRA by 27.6%, and BrIC by 18.5%. Like Bonin, Baker found no consistent reduction in PLA, reinforcing the idea that MIPS primarily improves angular kinematics rather than linear.

Abderezaei et al. (2021)

Figure 2. Effect of the presence or absence of the mitigation system on bicycle helmet performance in impact tests

Abderezaei et al. conducted a meta-analysis of published drop test data comparing conventional helmets and those with advanced injury mitigation technologies like MIPS. The study reviewed 148 drop tests and applied standardized injury metrics such as PLA, PAA, and GAMBIT. GAMBIT is a generalized model that integrates translational and rotational acceleration into a single injury prediction score, where a value of 1 represents a 50% chance of sustaining an AIS > 3 (serious brain injury).

To ensure fair comparison, impact velocities were normalized across studies. Tests were grouped based on helmet type and whether a neck surrogate was used. The analysis revealed that helmets with additional mitigation features had consistently lower values for all injury metrics. For MIPS-equipped helmets at lower velocities (4.2 ± 0.4 m/s), PAA was reduced by 16.8% and GAMBIT by 49.3% compared to conventional designs. At higher velocities (5.9 ± 0.6 m/s), MIPS helmets continued to outperform in PLA, PAA, and GAMBIT, confirming their benefit across a range of impact severities.

Computational Tests

Han et al. (2024)

A study by Han et al. used finite element analysis (FEA) to assess the performance of MIPS. The MIPS equipped helmet was 3D scanned and turned into a point cloud, where an FEA mesh was applied; this mesh was able to include detailed representations of the rotational shells. The shell was modeled using plastic (MAT_24) and the foam lined was modeled using (MAT_63) foam. The gel and helmet linear were modeled with a face to face contact with a coefficient of 0.17. For comparison a non MIPS helmet was also simulated.  The FEA model and headform was validated by simulating a Working Group 11 (WG11) rotational impact test with Economic Commission for Europe (ECE) 22.05 requirements.  In the simulation the helmets were dropped on a 45° inclined anvil, across 4 specified positions; 30°, 45°, and 60°, with speeds of 5m/s and 8m/s - these results were compared against a physical test with impact positions at 0°, 135°, 180°, and 270°. The key metrics obtained were PLA and PAA.

Figure 3. Verification model for the MIPS helmet-headform. (A) The model construction of the MIPS helmet coupled to the headform. (B) The surface-to-surface contact coupling between the helmet strap and the head model

Figure 4. The results against validation of MIPS bicycle helmet of FE model. (A) Impact_0°, (B) Impact_135°, (C) Impact_180°, and (D) Impact_270°

From Figure 4, it's clear that the kinetic curves closely align. The peak metrics are also very close; for example at 0° impact location the PLA was 97g and PAA was 1810 rad/s/s in the physical test, while the simulation produced 103g PLA and 189 rad/s/s PAA.  Comparing the results across different locations and speeds, the models were found to be generally more accurate with uniform impact surfaces and less accurate with complex contact conditions. Across all models however FE models were found to reliably predict head dynamics with MIPS (see Table 1), with a max error of 6.19%.

Table 1. Helmets rotation experiments compared with simulation

Test points	Parametric	Experiment values	Simulation values	Error ratio (%)
Impact_0°	PLA (g)	97	103	6.19
	PAA (rad/s2)	1810	1890	4.42
Impact_135°	PLA (g)	80	82	2.50
	PAA (rad/s2)	791	840	6.19
Impact_180°	PLA (g)	60	57	5.00
	PAA (rad/s2)	1054	1051	0.28
Impact_270°	PLA (g)	142	139	2.11
	PAA (rad/s2)	1192	1236	3.69

The FEA simulation predicted that at 8m/s impact speed and 30° inclined anvil, the MIPS helmet reduced the PAA from 8243 rad/s to 3225 rad/s. Similarly the PLA was reduced from 292gs to 281gs. Across all simulations MIPS reduce PAA by an average of 54.8%, and PLA  by an average of 15.6% (see Figure 5)

Figure 5. PLA and PAA with and without MIPS helmet group under different oblique impact tests. (A) PAA and PLA on a 30° anvil, (B) PAA and PLA on a 45° anvil, and (C) PAA and PLA on a 60° anvil

Discussion

Overall, MIPS helmets are effective in reducing PLA and PAA values, and preventing injury to the head caused by angular rotation during a collision or accident. The PAA values were lower in all computational simulations and physical tests with MIPS helmets compared to normal helmets. Studies found that MIPS performance was dependent on the impacting angles; at lower impact angles the reduction of PAA is more effective. As the alteration in impact angle could affect load distribution, hence influencing the sliding kinematics of the MIPS helmets, overestimation of helmet protective effect might be introduced at steeper impact angle. However, impact velocity has no significant impact on the effectiveness of MIPS helmets. Hair was also found to reduce the effectiveness of MIPS significantly, since the hair layer reduces the coefficient of friction, and allows the helmet to slide relative to the head. Computational models were proven to be effective in modeling MIPS helmets and impacts; similar results in terms of PAA and PLA values were found with both physical and computational models.

Some limitations of the discussed work includes the absence of real-world accident data and a lack of consideration for parameter differences among helmets. For instance, a MIPS-equipped helmet might outperform a helmet without MIPS simply due to increased helmet thickness rather than the MIPS technology itself. Such variables have to be controlled in experiment so that only the presence of MIPS could affect experimental results. Most studies rely on a 50th percentile Hybrid III headform or slight variations of it, representing a limited population sample. In addition to representing a limited population sample, the Hybrid III doesn’t have realistic interactions between the chin strap and the neck, further contributing to possible inaccuracies. As the angular rotation in the head is closely related to the neck, especially for cervical extension and flexion, a more biofidelic subject such as human cadaver can be used in tests for more accurate results. Many studies also only used a single helmet to test the MIPS helmet, meaning there is limited research on helmet to helmet comparisons.

Controversies

One computational analysis found that MIPS helmets were less effective in side impacts, with only minimal reductions in PRA and PVA. In contrast, another study demonstrated that MIPS helmets performed better with side impacts when verified through physical testing. This discrepancy could be attributed to factors such as a larger rotation radius in the midsagittal plane and fewer constraints on rotational motion, allowing greater helmet movement and thus maximizing protective effectiveness. From a real-world perspective, this inconsistency raises concerns about whether current helmet designs provide reliable protection across different impact scenarios, potentially leading to gaps in safety for cyclists. From a design standpoint, it underscores the need for improved helmet testing methodologies and refinements in MIPS technology to ensure consistent protection across all impact angles.

Similarly, Abderezaei et al. found that MIPS effectively reduced PLA, while Baker et al. found no significant reduction . This could be due to using different helmets to test the MIPS helmet against. The thickness, shape and material of the helmet could result in different responses during impact. In general, no concrete and specific conclusions can be reached on how effectively MIPS reduce PLA compared to regular helmets.

There is also some debate about whether current helmet standards do not adequately address rotational injuries. Current helmet testing standards, such as CPSC and EN1078, mainly regulate the performance of helmets from linear impacts, by measuring PLA. Only little regulatory tests include oblique tests for rotational energy management. This leads to a gap in regulation for MIPS or similar systems which aim at reducing rotation impact, which implies that injuries resulting from rotational acceleration remain inadequately regulated.

Future Work

Although the works presented in this review draw interesting conclusions about MIPS helmets, there is still progress to be made. In order to provide a more direct comparison between conventional helmets and helmets with injury mitigation features, a series of drop tests could be performed amongst the various helmet types with consistent parameters, such as liner thickness, helmet mass, same neck surrogate, and headform model. Further work in drop testing could also be done to capture the effects of MIPS for a range of helmet sizes and styles rather than just one size, as well as how MIPS respond to different impact directions since current results are incoherent. Additionally, both computational models and physical testing need to be conducted to cover a broader range of anthropometric characteristics and impact scenarios, such as vehicle-to-head collisions rather than just impacts on kerbs or stationary surfaces, including the study of real accident cases.