Previously injured limbs exhibited persistent BFlh and BFsh atrophy and subtle muscle-specific mechanical deficits during sprinting, with reduced BFsh force and negative work detectable only with MRI-informed subject-specific musculoskeletal modeling.
Key Findings
Results
Previously injured limbs showed approximately 5% smaller normalized muscle volumes for biceps femoris long head compared to uninjured control limbs.
BFlh normalized volume: 1.67 ± 0.36 vs. 1.76 ± 0.35 mL/kg/m in injured vs. uninjured limbs (p = 0.009)
Study included 79 NCAA Division I male football players with 106 previously injured and 52 uninjured control limbs
Muscle volumetry was conducted using 3-T MRI for whole-hamstring assessment
The ~5% deficit persisted despite athletes being cleared to return to sport and competing at elite collegiate level
Results
Previously injured limbs showed approximately 5% smaller normalized muscle volumes for biceps femoris short head compared to uninjured control limbs.
BFsh normalized volume: 0.92 ± 0.18 vs. 0.97 ± 0.19 mL/kg/m in injured vs. uninjured limbs (p = 0.007)
The deficit magnitude was similar to that observed in BFlh (~5%)
BFsh atrophy was detected via MRI-informed subject-specific modeling, not generic models
These morphological deficits were present bilaterally compared within the same athletes where applicable
Results
Peak muscle-tendon unit lengthening velocity was significantly faster in previously injured limbs for BFlh, semimembranosus, and semitendinosus during maximal-velocity sprinting.
BFlh peak MTU lengthening velocity was 3% faster in injured limbs (p ≤ 0.02)
SM peak MTU lengthening velocity was 5% faster in injured limbs (p ≤ 0.02)
ST peak MTU lengthening velocity was 5% faster in injured limbs (p ≤ 0.02)
Sprint kinematics (joint angles) did not differ between limbs, suggesting the differences arose from muscle morphology rather than movement pattern changes
MTU mechanics were extracted from sprint kinematics collected via inertial measurement units at 100 Hz
Results
MRI-informed subject-specific models detected lower BFsh peak force in previously injured limbs, a difference not detected by scaled-generic models.
BFsh peak force: 1.10 ± 0.21 vs. 1.14 ± 0.22 × body weight in injured vs. uninjured limbs (p = 0.03)
The scaled-generic model did not detect any limb differences in peak force for any hamstring muscle
Both scaled-generic and MRI-informed subject-specific OpenSim models were driven by the same sprint kinematics
This finding demonstrates that personalizing model geometry with MRI data reveals mechanical deficits invisible to standard modeling approaches
Results
MRI-informed subject-specific models detected reduced BFsh negative work in previously injured limbs, a difference not detected by scaled-generic models.
BFsh negative work: -13.4 ± 6.8 vs. -15.3 ± 7.5 J in injured vs. uninjured limbs (p = 0.01)
Negative work represents energy absorbed by the muscle during eccentric loading, relevant to injury risk
The generic model did not detect this limb difference in negative work
No limb differences in peak MTU strain were found for any hamstring muscle in either modeling approach
Results
Sprint kinematics did not differ between previously injured and uninjured limbs despite persistent morphological and mechanical deficits.
No significant limb differences were found in joint kinematics during maximal-velocity sprinting
Athletes performed maximal-velocity over-ground sprints instrumented with inertial measurement units at 100 Hz
The absence of kinematic differences means that observational or standard biomechanical screening would not detect the underlying muscle-level deficits
Peak MTU strain also did not differ between limbs for any of the four hamstring muscles examined
Methods
The study used a within-subject and between-subject design comparing 106 previously injured limbs to 52 uninjured control limbs in 79 NCAA Division I male football players.
Data were drawn from the Hamstring Injury (HAMIR) study (ClinicalTrials.gov: NCT05343052)
All participants were male NCAA Division I football athletes
Four hamstring muscles were analyzed: biceps femoris long head (BFlh), biceps femoris short head (BFsh), semimembranosus (SM), and semitendinosus (ST)
Linear mixed models were used for statistical analysis with α = 0.05
Both scaled-generic and MRI-informed subject-specific OpenSim musculoskeletal models were compared
What This Means
This research suggests that football players who have previously suffered a hamstring muscle strain injury retain measurable deficits in their hamstring muscles even after returning to elite-level competition. Specifically, two parts of the biceps femoris muscle — the long and short heads — were about 5% smaller in volume on the previously injured side compared to uninjured limbs. Despite these differences in muscle size, the athletes moved identically on both sides when sprinting at top speed, meaning a coach or trainer simply watching them run would not notice anything wrong.
The study also found that when researchers used personalized computer models of muscle function — built using each athlete's own MRI scan data — they could detect that the injured-side biceps femoris short head produced less force and absorbed less energy during sprinting. These differences were completely invisible when using standard, one-size-fits-all computer models that are more commonly used in sports science. The muscles in the injured limb also stretched faster during sprinting, which may increase mechanical stress on already-atrophied tissue.
This research suggests that standard return-to-sport assessments, which often rely on how an athlete moves or performs on functional tests, may miss persistent underlying muscle deficits that could contribute to reinjury. Combining MRI-based muscle imaging with wearable sensors and personalized biomechanical modeling could give clinicians and strength coaches a more complete picture of recovery, potentially helping to identify athletes who need additional rehabilitation before being exposed to the high-speed running demands of competitive play.
Check Your Own Numbers
Upload your bloodwork. We'll cross-reference your results against this study and 4,700 others.
Hulm S, Lin Y, Maniar N, Timmins R, Hickey J, Heiderscheit B, et al.. (2026). Impact of Prior Hamstring Strain Injury on Muscle Morphology and Sprinting Biomechanics in Collegiate American Football Athletes.. Scandinavian journal of medicine & science in sports. https://doi.org/10.1111/sms.70317