Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 May 27.
Published in final edited form as: Ann Biomed Eng. 2026 Feb 12;54(7):2015–2026. doi: 10.1007/s10439-026-04009-9

Intramedullary Strain during Neck Extension is Associated with Microstructural Spinal Cord Injury in Degenerative Cervical Myelopathy

Mahmudur Rahman 1, Karthik Devaraj 1, Balaji Harinathan 1, Omkar Chauhan 1, Matthew Budde 1,2, Narayan Yoganandan 1,2, Aditya Vedantam 1
PMCID: PMC13204404  NIHMSID: NIHMS2173110  PMID: 41678127

Abstract

Purpose

Intramedullary stress and strain during neck motion have been proposed as a cause of spinal cord damage in degenerative cervical myelopathy (DCM). We aimed to determine this relationship using patient-specific modeling and advanced MRI.

Methods

Twenty DCM patients underwent 3D finite element modeling of the cervical spine and spinal cord. We quantified von Mises stress and maximum principal strain at the level of maximum cord compression and across the entire cervical spinal cord during simulated neck flexion and extension. Pre-surgical MRI included T1-weighted, T2-weighted, and diffusion tensor imaging. DTI indices were measured at the maximum compression level and across the entire cervical cord (C2–C7). Associations between FEM metrics and DTI indices were evaluated using stepwise multivariable linear regression.

Results

Twenty DCM patients (13 females, 7 males; median age 64 years [37-76 years]) were enrolled. The median preoperative modified Japanese Orthopedic Association scores was 13 (9–17). The most common cervical spine level of maximum compression was C5-C6 (40%, n=8), followed by C6-C7 (25%, n=5), C4-C5 (20%, n=4), and C3-C4 (15%, n=3). Maximum spinal cord strain during neck extension was associated with lower fractional anisotropy (B=−2.1, R2=0.32, p=0.01), higher mean diffusivity (B=4.8, R2=0.21, p=0.04) and radial diffusivity (B=4.9, R2=0.29, p=0.01) across the whole spinal cord.

Conclusion

Intramedullary strain during neck extension correlates with microstructural spinal cord injury in patients with DCM. Patient-specific biophysical modeling highlights the impact of intramedullary strain during neck motion on the pathophysiology of DCM.

Key Terms: Degenerative Cervical Myelopathy (DCM), Intramedullary Strain, Finite Element Modeling (FEM), Diffusion Tensor Imaging (DTI), Fractional Anisotropy (FA)

1. INTRODUCTION

Degenerative cervical myelopathy (DCM) is the leading cause of spinal cord dysfunction in adults and an important cause of disability and impaired quality of life. The pathophysiology of DCM involves cervical spinal cord compression due to the osteoligamentous spine. Surgical spinal cord decompression remains the primary treatment of DCM. Adverse spinal cord stress/strain during neck motion has been proposed as a contributor to spinal cord injury in DCM [1]. Direct measurement of intramedullary stress or strain is not feasible in humans, and therefore it is not known how neck motion is related to structural spinal cord damage in patients with DCM. Determining the link between neck motion, intramedullary stress/strain and structural spinal cord damage is critical to designing patient-specific surgical treatments that address ongoing spinal cord injury in DCM. Surgical treatments that focus solely on spinal cord decompression without addressing adverse neck motion may limit neurological recovery.

Finite element models (FEMs) of the cervical spine and spinal cord are computational models that can quantify spinal cord biomechanics during neck motion. Validated FEMs of the human cervical spine provide non-invasive assessment of clinical spine biomechanics in relation to injury, degeneration or surgery [2–4]. Intrinsic spinal cord forces measured using spinal cord FEMs vary with the type (median diffuse vs circumferential vs lateral) of cord compression and can predict intramedullary T2-weighted hyperintensity at the level of compression [5–7]. Prior approaches to 3D FEM of the cervical spinal cord, however, have used a generic spinal cord model with material properties from animal spinal cord tissue [5–7]. These models do not incorporate the specific geometries of the individual patient’s cervical spine, which are well known contributors to intramedullary stress and strain. The lack of patient-specific data in these models has limited the clinical translation of biophysical modeling for DCM. The development of MRI-derived patient-specific FEMs of the cervical spine and spinal cord has enabled measurement of intramedullary stress/strain during neck motion for the individual patient [8,9].

Diffusion-weighted MRI has been proposed as an alternative to conventional MRI, and has shown improved detection of injury in DCM [10–14]. Diffusion MRI measures the thermal motion of water molecules to probe tissue microstructure at a resolution far smaller than the actual image resolution [15] and detects subtle microscopic changes in the spinal cord, otherwise undetectable with conventional MRI contrasts [16–18]. Diffusion Tensor Imaging (DTI) is an extension of diffusion-weighted MRI that evaluates water molecule diffusion along three principal axes to assess microstructural integrity and generate indices that provide insights into spinal cord tissue architecture and pathology. Axial diffusivity (AD) measures water diffusion along the primary axonal axis, reflecting axonal integrity, while radial diffusivity (RD) captures diffusion perpendicular to fibers, serving as a marker for demyelination. Fractional anisotropy (FA) reflects the directionality of diffusion, with lower values indicating axonal damage or degeneration. Mean diffusivity (MD) represents the average diffusion magnitude and is elevated in tissue damage or edema. Directional axial diffusivity (filtered apparent diffusion coefficient [fADC]) refines the measurement along the principal eigenvector by filtering diffusion signals that are perpendicular to the principal eigenvector, thereby offering a more specific estimation of axonal injury. DTI has been used in several DCM studies to quantify microstructural spinal cord injury, and DTI metrics, especially FA, have been shown to be a strong diagnostic marker of DCM [19].

The goal of this study is to determine how intramedullary spinal cord stress/strain is related to spinal cord damage in DCM. We hypothesize that greater intramedullary spinal cord stress/strain is associated with greater microstructural spinal cord injury in DCM. We aim to use two complementary techniques (patient-specific FEM and DTI) that will be used in combination in the same patient to reveal whether they provide a congruent assessment of spinal cord damage in DCM. Since these measures aim to be clinically accessible, we anticipate the results will support the incorporation of biophysical spinal cord modeling for clinical decision-making in DCM.

2. METHODS

We prospectively recruited 20 DCM patients (7 males and 13 females) who were scheduled for cervical spine surgery at a single academic medical center in the Midwest. DCM patients between the ages of 18 and 90 with one or more signs or symptoms of cervical myelopathy in addition to evidence of cervical spinal cord compression on MRI were included. We excluded patients with a history of trauma, syringomyelia, prior cervical spine surgery, spinal cord hemorrhage, tumor, pregnancy and patients with other neuromuscular diseases that could explain their symptoms. The median age of the final cohort was 64 (range 37-76) with median modified Japanese Orthopedic Association (mJOA) score of 13 (range 9–17). Institutional review board approval was obtained and all participants provided written informed consent.

2.1. Generation of Patient-Specific Surface Model from MRI

Patient-specific FEM of the osteo-ligamentous spine was created from their pre-surgical MRI [8,9]. The sagittal T2-weighted MRI was used to record the spinal cord level of maximum compression and grade intervertebral discs degeneration at all cervical spine levels [20]. The level of maximum spinal cord compression was determined by evaluation of the sagittal and axial T2-weighted MRI images by board-certified and fellowship-trained neurosurgeon. The level with the smallest anteroposterior diameter of the spinal cord (indicating greatest cord compression) was determined to be the level of maximum compression. Vertebral body and spinal cord surface geometries were developed from the DICOM images using the Mimics (Materialise NV, Leuven, Belgium) software package. Surface patches were applied to delineate the endplate and facet regions. Segmentation was performed using intensity-based thresholding and region growing to isolate spinal components. The segmented masks were exported in .STL format. Smoothing was performed in 3-matic to reduce voxel-induced irregularities. NURBS-based surface reconstruction was then applied to convert the STL geometry into .STEP format.

2.2. Finite Element Modeling of Cervical Spine

The 3D cervical spine surface model .STEP was imported into ANSA software (BETA CAE Systems, Farmington Hills, MI, USA) for finite element meshing. The element size was chosen based on the mesh sensitivity analysis using the baseline model and the segmental motion metric [8]. The same element type and size were maintained in the models used in the present study. T2-weighted MRI was used to grade disc degeneration at each level [20], and the corresponding disc properties were incorporated into the model. The spinal cord FEM model consisted of spinal cord parenchyma, pia mater, dura mater, denticulate ligaments, and cerebrospinal fluid [8]. The model used material properties from human spinal cord tissue that have been reported previously [8]. Details of material properties and material models for each component are summarized in Table 1. Prior in-vitro and FEM studies report limiting von Mises stresses of ~4.8 MPa and strain of 9.5% for cancellous bone, and ~113 MPa and 1.8% strain for cortical bone [21]. This patient-specific spinal cord FEM was integrated with the cervical spine model for analysis (Fig. 1). A mesh convergence study was done using five mesh sizes. The average element size was 2.0, 1.5, 1.0, 0.5 and 0.25 mm (161112 elements) for models 1, 2, 3, 4, and 5, respectively. Convergence was set at <3% difference between consecutive refinements, and model 3 was used as the baseline model. The mean differences ranged from 1.3–2.0% for flexion and extension loading. The final mesh shown in Fig. 2 is from the baseline model.

Table 1:

Material properties of the finite element model used for each component of the spinal cord

Element type Constitutive model Parameters
Cortical bone [22] Quadrilateral shell Linear elastic E = 16.8 GPa
ν = 0.3
Trabecular bone [23] Hexahedral solid Linear elastic E = 0.442 GPa
ν = 0.3
Endplates[24] Quadrilateral shell Linear elastic E = 5.6 GPa, ν = 0.3
Facet cartilage [25] Quadrilateral shell Linear elastic E = 0.01 GPa, ν = 0.3
Annulus ground substance [24] Hexahedral solid Hill Foam n=2, C1=0.000115 GPa
C2=0.002101 GPa
C3=−0.000893 GPa
b1= 4, b2= −1, b3= −2
Annulus fibrosus [26] Quadrilateral membrane Orthotropic nonlinear elastic Fiber angle (45-600)
Nucleus pulposus[27] Hexahedral solid Fluid K =1.720 GPa
Ligaments [28] Quadrilateral membrane Non-linear curves Stress-Strain curves
Spinal cord [29] Hexahedral solid Viscoelastic ρ =1.04e6 kg/m3
C1 = 0.5345, C2 = 1.0665, C3 = 1.0113, G1 = 0.8927, G2 = 0.8926, G3 = 0.8917, β1 = −0.0137, β2 = 0.00775, β3 = 0.035
Pia mater [30] Quadrilateral shell Linear Elastic ρ =1.13e6 kg/m3
E=2.3MPa
ν=0.49
Dura mater [31] Quadrilateral shell Linear Elastic ρ =1.130e6 kg/m3
E=80MPa
ν=0.49
Denticulate ligaments [32] Quadrilateral shell Linear Elastic ρ =1.040e6 kg/m3
E=0.0058GPa
ν=0.45
CSF [33] Hexahedral solid Viscoelastic ρ =1.040e-06
K = 2.19 GPa
G0 = 5e-7
GI = 1e-7
Open in a new tab

Fig. 1.

Fig. 1

Open in a new tab

Sagittal T2-weighted MRI of the cervical spine of a DCM patient (a), 3D cervical spine surface model derived from MRI data (b), and finite element mesh of the spine and spinal cord for patient-specific biomechanical analysis (c)

Fig. 2.

Fig. 2

Open in a new tab

Sagittal view of the model with anatomical details and rigid boundary conditions at the inferior end of the spine (a). Additional anatomical details of the facets, discs (b) and spinal cord (c) with boundary conditions are also shown. Graph of the time course is also shown (d).

2.3. Loading and boundary conditions

The patient-specific FEMs were constrained at the inferior surface of the T1 vertebra in all degrees-of-freedom, and the sagittal loadings were applied at the superior vertebra (Fig. 2a–c). The pure moment load level was 2 Nm each in flexion and extension modes over 4000 ms (Fig. 2d). An additional follower force of 75N to simulate the head mass and muscle force was applied during flexion and extension using the method described by Patwardhan et al [34]. The loading condition represents quasi-static physiological cervical spine loading in adult males and females [23,33]. The applied follower load accounts for the stabilizing action of the human cervical musculature [33].

The 75N loading was maintained using LS-PrePost v4.3 and LS-Dyna v9.0.1 (Livermore, USA). Motion-related biomechanical output including range of motion (ROM), von Mises stress (VMS) and maximum principal strain (MPS) from each disc-level regions (e.g., C2/C3, C3/C4) of the spinal cord was calculated (Fig. 3). MPS and vMS at the disc-level of maximum spinal cord compression as well as the maximum MPS and vMS across the entire cervical spinal cord were recorded.

Fig. 3.

Fig. 3

Open in a new tab

Patient-specific finite element modeling of the cervical spine showing intramedullary von Mises stress (a) and maximum principal strain (b) during neck flexion and extension, alongside sagittal T2-weighted MRI and axial fractional anisotropy (FA) maps from diffusion tensor imaging with the average FA voxels demarcated in yellow (c). The transparent white rectangles represent the spinal levels and imaging slices that were used to quantify stress/strain measures and DTI metrics respectively.

2.4. Quantitative MRI

MRI was obtained for all patients using a head and neck spine coil using a 3T Premier system (GE Medical Systems, Milwaukee, Wisconsin). T2-weighted imaging used a 3D fast spin-echo sequence with isotropic resolution of 0.8 x 0.8 x 0.8 mm3, a 25.6 cm2 field-of-view (FOV), and TR/TE of 2500/125 ms. Diffusion weighted imaging (DWI) used a single-shot, spin-echo EPI sequence with reduced field of view excitation, 11 cm2 FOV, yielding an acquired resolution of 1.14 mm2, 16 slices with a slice thickness of 5 mm, and TR/TE of 4000/52 ms. For diffusion tensor imaging (DTI), images were acquired with 6 directions at a b-value of 100 s/mm2, 30 unique directions at a b-value of 800 s/mm2, and two images without diffusion weighting (0 s/mm2). Additionally, a filtered-diffusion acquisition was obtained with the same imaging parameters but an alternative encoding strategy with diffusion weighting (b-value=2000 s/mm2) along the left-right axis and 5 diffusion directions uniformly distributed along the other two axes (anterior-posterior and inferior-superior) at a b-value of 500 s/mm2 [37].

2.5. Image processing

Image analysis used off-line processing and correction with pipelines developed in nipype in combination with the Spinal Cord Toolbox [38]. Original images were denoised using a MPPCA algorithm and motion corrected [39]. Maps of fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) were estimated using a tensor model, and maps of filtered ADC (fADC‖) were estimated using a 2D tensor model, both in dipy. All images were aligned to a common template (PAM50) for quantification. The spinal cord was segmented using a machine learning algorithm (sct_deepseg) and spatially registered to the template. All DWI maps used the warping field as initialization with a final registration using slicewise warping. At each vertebral level including at the level of maximum spinal cord compression, several metrics were derived, including the cross-sectional spinal cord area (cm2), anterior-posterior diameter (cm), and mean DWI metrics from the whole-cord including both white and gray matter. Units for all measures of diffusivity are expressed as μm2/ms, and FA is a unitless parameter. This technique has been used in multiple prior studies with DCM patients [40] and DTI metrics have been shown to correlate with baseline neurological function in DCM [11].

2.6. Statistical analysis

Descriptive statistics were used for FEM outputs and spinal cord morphometry. Stepwise linear regression models were created to identify the relationship between spinal cord stress/strain and DTI metrics, while controlling for degree of spinal cord compression as measured by the antero-posterior spinal cord diameter. Statistical analysis was performed using statsmodel, seaborn and matplotlib statistical packages, SPSSv23 (IBM, Armonk, NY) and statistical significance was set at p<0.05.

3. RESULTS

3.1. Pre-surgical intramedullary stress and strain

The stress and strain during flexion and extension were calculated at the level of maximum spinal cord compression as well as for the entire cervical spinal cord (Table 2). The most common cervical spine level of maximum compression was C5-C6 (40%, n=8), followed by C6-C7 (25%, n=5), C4-C5 (20%, n=4), and C3-C4 (15%, n=3). In 40% of patients (n=8), the level of maximum compression coincided with the level of maximum von Mises stress during flexion. Six patients (30%) had intramedullary T2-weighted hyperintensity on MRI. There was no significant difference in the maximum intramedullary stress/strain between patients with and without intramedullary T2-weighted hyperintensity (maximum stress in flexion [p=0.35], maximum stress in extension [p=0.72], maximum strain in flexion [p=0.66], maximum strain in extension [p=0.84]). Similarly, there was no significant difference in the intramedullary stress/strain at the level of maximum cord compression (maximum stress in flexion [p=0.9], maximum stress in extension [p=0.97], maximum strain in flexion [p=1.0], maximum strain in extension [p=0.97]).

Table 2.

Intramedullary Stress and Strain from Patient-Specific Finite Element Models (n=20) during flexion and extension

FEM Measurement (Units) Neck Motion Parameter Median (Range) T2W intramedullary hyperintensity + Median (Range) T2W intramedullary hyperintensity − Median (Range)
Intramedullary Stress (MPa) Flexion Max Stress 10.21 (4.03–23.91) 12.18 (4.87–23.91) 7.63 (4.03–17.32)
Flexion Max Compression 7.475 (2.12–17.32) – –
Extension Max Stress 7.14 (3.34–17.06) 7.36 (5.01–9.26) 7.11 (3.34–17.06)
Extension Max Compression 5.6 (1.60 – 10.97) – –
Intramedullary Strain (%) Flexion Max Strain 0.05 (0.03–0.10) 0.06 (0.03–0.09) 0.05 (0.03–0.10)
Flexion Max Compression 0.04 (0.2-0.01) – –
Extension Max Strain 0.02 (0.01–0.06) 0.02 (0.02–0.04) 0.02 (0.01–0.06)
Extension Max Compression 0.02 (0.01–0.05) – –
Open in a new tab

3.2. DTI metrics

Average spinal cord DTI metrics were calculated for the whole cervical spinal cord (C2-C7) as well as at the level of maximum compression. There was no significant difference between DTI metrics across the whole cord and the level of maximum compression for FA (p=0.15), MD (p=0.18), AD (p=0.38), RD (p=0.16) and fADC (p=0.09) (Fig. 4).

Fig. 4.

Fig. 4

Open in a new tab

Diffusion tensor imaging (DTI) metrics of the cervical spinal cord, comparing fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), radial diffusivity (RD), and filtered apparent diffusion coefficient (fADC) at the level of maximum compression (Orange) and across the entire cord (Green)

3.3. Relationship Between DTI metrics and Intramedullary Stress/Strain Across the Cervical Spinal Cord

Stepwise multivariable regression, controlling for the degree of spinal cord compression, demonstrated that maximum spinal cord strain during neck extension correlated with lower FA (B=−2.1, R2=0.32, p=0.01), higher MD (B=4.8, R2=0.21, p=0.04) and RD (B=4.9, R2=0.29, p=0.01) (Table 3). Statistically significant relationships between maximum spinal cord strain and these diffusion metrics are illustrated in Fig. 5. There was no significant association between intramedullary stress/strain and fADC.

Table 3.

Linear stepwise regression controlling for degree of cord compression

Whole cord diffusion indices Intramedullary biomechanical property across the whole cord R2 p-value B 95% CI low 95% CI high
Fractional anisotropy (FA) Maximum strain during neck extension 0.32 0.01* −2.1 −.36 −0.57
Mean diffusivity (MD) Maximum strain during neck extension 0.21 0.04* 4.8 0.18 9.4
Axial diffusivity (AD) - - - - - -
Radial diffusivity (RD) Maximum strain during neck extension 0.29 0.01* 4.9 1.14 8.6
filtered apparent diffusion coefficient (fADC) - - - - - -
Open in a new tab
*

Statistically significant with p<0.05

Fig. 5.

Fig. 5

Open in a new tab

Association between maximum spinal cord strain in extension and DTI metrics across the entire cervical cord: (a) fractional anisotropy, (b) mean diffusivity, and (c) radial diffusivity. Red lines indicate linear regression fits with 95% confidence intervals

3.4. Relationship Between DTI metrics and Intramedullary Stress/Strain at the Level of Maximum Compression

At the level of maximum compression, microstructural injury as measured by DTI indices was not associated with intramedullary stress/strain. Multivariate stepwise regression showed statistically significant relationship between MD (B=−0.17, R2=0.23, p=0.035), RD (B=0.14, R2=0.22, p=0.038) and mean anteroposterior spinal cord diameter but not intramedullary stress/strain. No significant association was noted for FA and fADC at the level of maximum compression and intramedullary stress/strain.

4. DISCUSSION

Our study demonstrates that intramedullary strain during neck extension is associated with microstructural spinal cord injury across the whole cervical spinal cord in patients with DCM (R2 ≤ 0.32). Intramedullary strain during neck extension was linked to reduced spinal cord fractional anisotropy (FA), increased mean diffusivity (MD) and radial diffusivity (RD) in the whole cord reflecting axonal injury, edema, and demyelination. These results suggest a potential association between spinal cord strain during neck extension and microstructural spinal cord injury in patients with DCM.

The observed relationship between increased intramedullary strain and microstructural injury in DCM aligns with prior biomechanical studies of neural tissue deformation. Harrison et al. [41,42] highlighted that prolonged spinal cord tension from postural loads, particularly in flexion, elevates intramedullary pressure, disrupts blood flow, and impairs neural transport, leading to oxidative metabolic dysfunction and progressive neural damage. These findings parallel our observations that increased strain during neck extension is associated with axonal injury, demyelination, and white matter edema. The mechanisms proposed by Harrison et al. [41,42] including hypoperfusion and axoplasmic flow disruption, may explain the microstructural changes observed in our study. Additionally, their identification of a positive feedback loop, where increased intraneural pressure exacerbates neural dysfunction, provides a framework for understanding the progression of spinal cord damage in DCM.

In this study, intramedullary strain was associated with DTI metrics across the entire cervical spinal cord but not at the level of maximum compression. At the level of maximum cord compression, DTI metrics were most closely linked to the degree of spinal cord compression and not intramedullary stress/strain. This finding is supported by prior data showing that the DTI metrics are altered significantly at the level of maximum cord compression compared to other levels in the spinal cord [18,43]. While our model predicted higher stress/strain during flexion, the lack of correlation with DTI metrics may be due to extension-specific geometric effects in the compressed spinal cord. Dynamic MRI studies consistently show that neck extension increases cord compression, worsens multilevel stenosis, and is associated with intramedullary T2-hyperintensity—features linked to microstructural injury on DTI [44–46]. During cervical extension, the spinal cord shortens and the spinal canal narrows, promoting posterior buckling of the ligamentum flavum and reducing space available for the cord [47]. Finite element analyses show that this configuration produces a pincer-type mechanism, in which posterior ligamentous infolding and anterior compression localize intramedullary stress and strain [48]. In contrast, flexion lengthens the cord and increases canal dimensions, resulting in more distributed deformation despite higher global stress and strain magnitudes. These extension-specific geometric effects provide a mechanistic explanation for the association between intramedullary strain during extension and DTI-derived microstructural injury. Additionally, DTI metrics may be preferentially sensitive to specific mechanical states potentially reflecting altered axonal tension, microvascular perfusion, or extracellular diffusion environments rather than peak compressive loading during neck motion [49,50]. Thus, diffusion abnormalities may reflect cumulative or position-dependent microstructural vulnerability rather than instantaneous high-stress conditions.

Although we obtained the MPS and VMS at each disc level from the inferior margin of the superior vertebral body and the superior margin of the inferior vertebral body, our model did not examine the transmission of these forces into the adjacent levels since impingement occurs clinically at the disc levels. As described earlier, the transmission of stresses and strains over time may extend inferiorly and/or superiorly affecting spinal cord function, and this temporal response is not included the present finite element model. Evaluation of the association between local variables (stress/strain) and tissue characteristics are not possible in patients due to the lack of feasibility of spinal cord tissue sampling. However, multiple animal studies have been performed to evaluate this relationship. In 2D plain-strain FEM guinea-pig white matter, combined mechanical injury (both stress/strain) correlated strongly with horseradish peroxidase uptake (R2 = 0.917), indicating membrane disruption in high-strain regions [51]. In a primate cervical contusion model, peak principal strains and von Mises stresses at the injury epicenter predicted (R2 up to 0.76) the chronic lesion area [52]. Static compression simulations further showed that physiologic variations in both gray and white-matter stiffness can alter local stress and strain by as much as 50% [53]. Together, these studies suggest that regions of mechanical stress/strain coincide with histological damage in both grey and white matter of the spinal cord.

Intramedullary stress/strain is linked to spinal cord damage as measured by T2-weighted intramedullary hyperintensity on MRI. T2-weighted intramedullary hyperintensity is seen in 58-85% of DCM patients and these signal changes represent a variety of histopathological features including spinal cord edema, Wallerian degeneration, demyelination, gliosis and/or cavitation [54]. Takahashi et al. [5], used FEM and clinical MRI data, reported a threshold of intramedullary stress required to induce T2-weighted intramedullary hyperintensity on MRI. The correlation between stress magnitude, T2-weighted intramedullary hyperintensity, and clinical symptoms provides further evidence that mechanical strain contributes directly to neurological impairments in DCM. Although we found higher intramedullary stress/strain in patients with T2-weighted hyperintensity compared to those without, the differences were not statistically significant. This result may be because T2-weighted spinal cord hyperintensity may not represent permanent spinal cord injury since it can result from both reversible edema and irreversible spinal cord damage [54,55]. Additionally, T2-weighted hyperintensity in DCM develops due to repetitive spinal cord damage during neck motion, which was not simulated in our study and may have affected our ability to detect correlations. These findings suggest that intramedullary strain during neck extension plays a role in determining the pattern and severity of spinal cord injury, reinforcing the biomechanical basis of DCM pathophysiology observed in this study. This mechanistic understanding bridges structural stress distributions with clinical manifestations, offering potential avenues for tailored therapeutic interventions.

This study highlights the potential clinical use of FEM to quantify intramedullary stress/strain in DCM patients. In DCM, dynamic MRIs have shown that neck extension can worsen existing cervical stenosis and cord compression [44–46]. Intramedullary stress and strain during neck movement can lead to cumulative microstructural injury that exacerbates neurological deficits over time, highlighting a common discrepancy between radiological findings and clinical symptoms [56]. Recognizing the impact of these dynamic forces is essential not only for early diagnosis but also for more accurate prognostication and surgical planning. DCM patients experiencing high intramedullary stress during neck motion may benefit more from stabilization procedures like fusion surgery rather than motion-preserving approaches such as laminoplasty or laminectomy.

Our patient-specific FEM did not incorporate patient-specific neck range of motion data, which could affect the accuracy of stress and strain estimations during neck movements. Additionally, the relatively small sample size may limit the generalizability of the findings to a broader DCM population. The reliance on assumptions regarding tissue material properties in FEM modeling introduces potential variability, as these properties can differ among individuals. Meshing anatomical structures of different materials may produce local stress concentrations at bone soft tissue interfaces, and this is a limitation of these types of FEMs. While the present study applied the loading based on accepted method and magnitudes, it is only a first step. A more detailed study should be conducted that include patient-related factors such as changes in the external loading due to limitations in physiological motions of flexion-extension, lateral bending and axial rotation due to the differing pain status of each patient, changes in material properties of discs and other components that can also vary with patient anatomy, including changes in bone mineral density (due to age, sex, chronic use of certain medications like steroids). Detailed sensitivity analyses can be conducted with the acknowledgment that validation is not always possible with patient-specific modeling without additional experimental studies on each patient. The cross-sectional study design also precludes establishing causal relationships between intramedullary stress, strain, and microstructural injury. Longitudinal studies, including postoperative follow-up, are required to establish temporal relationships. Despite these limitations, this study is one of the first to provide in vivo evidence that adverse biomechanical tension is associated with increased spinal cord damage in DCM.

Our study demonstrates that intramedullary strain during neck extension is significantly associated with microstructural spinal cord injury in patients with degenerative cervical myelopathy. These findings highlight the importance of biomechanical forces in the pathophysiology of DCM and underscore the potential of patient-specific finite element modeling and advanced MRI as tools for personalized diagnosis and treatment planning.

Statements and Disclosures:

  • The research leading to these results received funding from North American Spine Society, AO Spine North America and NIH/NINDS (R01NS129724).

  • The authors have no relevant financial or non-financial interests to disclose.

Footnotes

Ethics approval:

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Medical College of Wisconsin (protocol code: PRO00043133)

References:

  • 1.Henderson FC, Geddes JF, Vaccaro AR, Woodard E, Berry KJ, Benzel EC. Stretch-associated injury in cervical spondylotic myelopathy: new concept and review. Neurosurgery. 2005;56:1101–13; discussion 1101-1113. [PubMed] [Google Scholar]
  • 2.Wheeldon JA, Stemper BD, Yoganandan N, Pintar FA. Validation of a finite element model of the young normal lower cervical spine. Ann Biomed Eng. 2008;36:1458–69. 10.1007/s10439-008-9534-8 [DOI] [PubMed] [Google Scholar]
  • 3.John JD, Saravana Kumar G, Yoganandan N. Cervical spine morphology and ligament property variations: A finite element study of their influence on sagittal bending characteristics. J Biomech. 2019;85:18–26. 10.1016/j.jbiomech.2018.12.044 [DOI] [PubMed] [Google Scholar]
  • 4.Yoganandan N, Choi H, Purushothaman Y, Jebaseelan D, Baisden J, Kurpad S. Effects of different severities of disc degeneration on the range of motion of cervical spine. J Craniovertebral Junction Spine. 2020;11:269–75. 10.4103/jcvjs.JCVJS_158_20 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Takahashi K, Ozawa H, Sakamoto N, Minegishi Y, Sato M, Itoi E. Influence of intramedullary stress on cervical spondylotic myelopathy. Spinal Cord. 2013;51:761–4. 10.1038/sc.2013.94 [DOI] [PubMed] [Google Scholar]
  • 6.Nishida N, Kanchiku T, Imajo Y, Suzuki H, Yoshida Y, Kato Y, et al. Stress analysis of the cervical spinal cord: Impact of the morphology of spinal cord segments on stress. J Spinal Cord Med. 2016;39:327–34. 10.1179/2045772315Y.0000000012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Lévy S, Baucher G, Roche P-H, Evin M, Callot V, Arnoux P-J. Biomechanical comparison of spinal cord compression types occurring in Degenerative Cervical Myelopathy. Clin Biomech Bristol Avon. 2021;81:105174. 10.1016/j.clinbiomech.2020.105174 [DOI] [PubMed] [Google Scholar]
  • 8.Vedantam A, Purushothaman Y, Harinathan B, Scripp S, Budde MD, Yoganandan N. Spinal Cord Stress After Anterior Cervical Diskectomy and Fusion: Results from a Patient-Specific Finite Element Model. Ann Biomed Eng. 2022; 10.1007/s10439-022-03118-5 [DOI] [PubMed] [Google Scholar]
  • 9.Rahman M, Banurekha Devaraj K, Chauhan O, Harinathan B, Yoganandan N, Vedantam A. Intramedullary Stress and Strain Correlate with Neurological Dysfunction in Degenerative Cervical Myelopathy. Appl Sci Basel Switz [Internet]. 2025. [cited 2025 Mar 6];15:886. 10.3390/app15020886 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Vedantam A, Rao A, Kurpad SN, Jirjis MB, Eckardt G, Schmit BD, et al. Diffusion Tensor Imaging Correlates with Short-Term Myelopathy Outcome in Patients with Cervical Spondylotic Myelopathy. World Neurosurg. 2017;97:489–94. 10.1016/j.wneu.2016.03.075 [DOI] [PubMed] [Google Scholar]
  • 11.Rao A, Soliman H, Kaushal M, Motovylak O, Vedantam A, Budde MD, et al. Diffusion Tensor Imaging in a Large Longitudinal Series of Patients With Cervical Spondylotic Myelopathy Correlated With Long-Term Functional Outcome. Neurosurgery. 2018;83:753–60. 10.1093/neuros/nyx558 [DOI] [PubMed] [Google Scholar]
  • 12.Martin AR, Aleksanderek I, Cohen-Adad J, Tarmohamed Z, Tetreault L, Smith N, et al. Translating state-of-the-art spinal cord MRI techniques to clinical use: A systematic review of clinical studies utilizing DTI, MT, MWF, MRS, and fMRI. NeuroImage Clin. 2016;10:192–238. 10.1016/j.nicl.2015.11.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Murphy RKJ, Sun P, Xu J, Wang Y, Sullivan S, Gamble P, et al. Magnetic Resonance Imaging Biomarker of Axon Loss Reflects Cervical Spondylotic Myelopathy Severity. Spine. 2016;41:751–6. 10.1097/BRS.0000000000001337 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Ying J, Zhou X, Zhu M, Zhou Y, Huang K, Zhou B, et al. The Contribution of Diffusion Tensor Imaging to Quantitative Assessment on Multilevel Cervical Spondylotic Myelopathy. Eur Neurol. 2016;75:67–74. 10.1159/000443270 [DOI] [PubMed] [Google Scholar]
  • 15.Vedantam A, Jirjis MB, Schmit BD, Wang MC, Ulmer JL, Kurpad SN. Diffusion tensor imaging of the spinal cord: insights from animal and human studies. Neurosurgery. 2014;74:1–8; discussion 8; quiz 8. 10.1227/NEU.0000000000000171 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Ellingson BM, Ulmer JL, Kurpad SN, Schmit BD. Diffusion tensor MR imaging in chronic spinal cord injury. AJNR Am J Neuroradiol. 2008;29:1976–82. 10.3174/ajnr.A1272 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Budzik J-F, Balbi V, Le Thuc V, Duhamel A, Assaker R, Cotten A. Diffusion tensor imaging and fibre tracking in cervical spondylotic myelopathy. Eur Radiol. 2011;21:426–33. 10.1007/s00330-010-1927-z [DOI] [PubMed] [Google Scholar]
  • 18.Kara B, Celik A, Karadereler S, Ulusoy L, Ganiyusufoglu K, Onat L, et al. The role of DTI in early detection of cervical spondylotic myelopathy: a preliminary study with 3-T MRI. Neuroradiology. 2011;53:609–16. 10.1007/s00234-011-0844-4 [DOI] [PubMed] [Google Scholar]
  • 19.Guan X, Fan G, Wu X, Gu G, Gu X, Zhang H, et al. Diffusion tensor imaging studies of cervical spondylotic myelopathy: a systemic review and meta-analysis. PloS One. 2015;10:e0117707. 10.1371/journal.pone.0117707 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Suzuki A, Daubs MD, Hayashi T, Ruangchainikom M, Xiong C, Phan K, et al. Magnetic Resonance Classification System of Cervical Intervertebral Disk Degeneration: Its Validity and Meaning. Clin Spine Surg. 2017;30:E547–53. 10.1097/BSD.0000000000000172 [DOI] [PubMed] [Google Scholar]
  • 21.Yoganandan N, Khandelwal P, Porwal V, Humm J, Banerjee A. Human Thoracolumbar Spine Tolerance to Injury and Mechanisms From Caudo-Cephalad Loading: A Parametric Modeling Study. J Eng Sci Med Diagn Ther [Internet]. 2021. [cited 2026 Jan 12];4:011007. 10.1115/1.4049523 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Reilly DT, Burstein AH. The elastic and ultimate properties of compact bone tissue. J Biomech. 1975;8:393–405. 10.1016/0021-9290(75)90075-5 [DOI] [PubMed] [Google Scholar]
  • 23.Kopperdahl DL, Keaveny TM. Yield strain behavior of trabecular bone. J Biomech. 1998;31:601–8. 10.1016/s0021-9290(98)00057-8 [DOI] [PubMed] [Google Scholar]
  • 24.Panzer MB, Cronin DS. C4-C5 segment finite element model development, validation, and load-sharing investigation. J Biomech. 2009;42:480–90. 10.1016/j.jbiomech.2008.11.036 [DOI] [PubMed] [Google Scholar]
  • 25.Yamada H, Evans FG. Strength of biological materials [Internet]. Williams & Wilkins; 1970. [cited 2022 Jun 22]. https://scholar.google.com/scholar_lookup?title=Strength+of+biological+materials&author=Yamada%2C+Hiroshi&publication_year=1970. Accessed 22 Jun 2022 [Google Scholar]
  • 26.Cassidy JJ, Hiltner A, Baer E. Hierarchical structure of the intervertebral disc. Connect Tissue Res. 1989;23:75–88. 10.3109/03008208909103905 [DOI] [PubMed] [Google Scholar]
  • 27.Iatridis JC, Setton LA, Weidenbaum M, Mow VC. Alterations in the mechanical behavior of the human lumbar nucleus pulposus with degeneration and aging. J Orthop Res Off Publ Orthop Res Soc. 1997;15:318–22. 10.1002/jor.1100150224 [DOI] [PubMed] [Google Scholar]
  • 28.Mattucci SFE, Moulton JA, Chandrashekar N, Cronin DS. Strain rate dependent properties of younger human cervical spine ligaments. J Mech Behav Biomed Mater. 2012;10:216–26. 10.1016/j.jmbbm.2012.02.004 [DOI] [PubMed] [Google Scholar]
  • 29.Bilston LE, Thibault LE. The mechanical properties of the human cervical spinal cord in vitro. Ann Biomed Eng. 1996;24:67–74. 10.1007/BF02770996 [DOI] [PubMed] [Google Scholar]
  • 30.Fradet L, Arnoux P-J, Callot V, Petit Y. Geometrical variations in white and gray matter affect the biomechanics of spinal cord injuries more than the arachnoid space. Adv Mech Eng. SAGE Publications; 2016;8:1687814016664703. 10.1177/1687814016664703 [DOI] [Google Scholar]
  • 31.Persson C, Evans S, Marsh R, Summers JL, Hall RM. Poisson’s ratio and strain rate dependency of the constitutive behavior of spinal dura mater. Ann Biomed Eng. 2010;38:975–83. 10.1007/s10439-010-9924-6 [DOI] [PubMed] [Google Scholar]
  • 32.Greaves CY, Gadala MS, Oxland TR. A three-dimensional finite element model of the cervical spine with spinal cord: an investigation of three injury mechanisms. Ann Biomed Eng. 2008;36:396–405. 10.1007/s10439-008-9440-0 [DOI] [PubMed] [Google Scholar]
  • 33.Mao H, Zhang L, Jiang B, Genthikatti VV, Jin X, Zhu F, et al. Development of a finite element human head model partially validated with thirty five experimental cases. J Biomech Eng. 2013;135:111002. 10.1115/1.4025101 [DOI] [PubMed] [Google Scholar]
  • 34.Patwardhan AG, Havey RM, Ghanayem AJ, Diener H, Meade KP, Dunlap B, et al. Load-carrying capacity of the human cervical spine in compression is increased under a follower load. Spine. 2000;25:1548–54. 10.1097/00007632-200006150-00015 [DOI] [PubMed] [Google Scholar]
  • 35.Panjabi MM. Hybrid multidirectional test method to evaluate spinal adjacent-level effects. Clin Biomech Bristol Avon. 2007;22:257–65. 10.1016/j.clinbiomech.2006.08.006 [DOI] [PubMed] [Google Scholar]
  • 36.Barrey C, Rousseau M-A, Persohn S, Campana S, Perrin G, Skalli W. Relevance of using a compressive preload in the cervical spine: an experimental and numerical simulating investigation. Eur J Orthop Surg Traumatol. 2015;25 Suppl 1:S155–165. 10.1007/s00590-015-1625-2 [DOI] [PubMed] [Google Scholar]
  • 37.Lee S-Y, Meyer BP, Kurpad SN, Budde MD. Diffusion-prepared fast spin echo for artifact-free spinal cord imaging. Magn Reson Med. 2021;86:984–94. 10.1002/mrm.28751 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.De Leener B, Lévy S, Dupont SM, Fonov VS, Stikov N, Louis Collins D, et al. SCT: Spinal Cord Toolbox, an open-source software for processing spinal cord MRI data. NeuroImage. 2017;145:24–43. 10.1016/j.neuroimage.2016.10.009 [DOI] [PubMed] [Google Scholar]
  • 39.Veraart J, Fieremans E, Novikov DS. Diffusion MRI noise mapping using random matrix theory. Magn Reson Med. 2016;76:1582–93. 10.1002/mrm.26059 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Chernysh AA, Loftus DH, Zheng B, Arditi J, Leary OP, Fridley JS. Utility of Diffusion Tensor Imaging for Prognosis and Management of Cervical Spondylotic Myelopathy: A PRISMA Review. World Neurosurg. 2024;190:88–98. 10.1016/j.wneu.2024.07.032 [DOI] [PubMed] [Google Scholar]
  • 41.Harrison DE, Cailliet R, Harrison DD, Troyanovich SJ, Harrison SO. A review of biomechanics of the central nervous system--Part III: spinal cord stresses from postural loads and their neurologic effects. J Manipulative Physiol Ther. 1999;22:399–410. 10.1016/s0161-4754(99)70086-2 [DOI] [PubMed] [Google Scholar]
  • 42.Harrison DE, Cailliet R, Harrison DD, Troyanovich SJ, Harrison SO. A review of biomechanics of the central nervous system--part II: spinal cord strains from postural loads. J Manipulative Physiol Ther. 1999;22:322–32. 10.1016/s0161-4754(99)70065-5 [DOI] [PubMed] [Google Scholar]
  • 43.Lindberg PG, Sanchez K, Ozcan F, Rannou F, Poiraudeau S, Feydy A, et al. Correlation of force control with regional spinal DTI in patients with cervical spondylosis without signs of spinal cord injury on conventional MRI. Eur Radiol. 2016;26:733–42. 10.1007/s00330-015-3876-z [DOI] [PubMed] [Google Scholar]
  • 44.Narvekar M, Dave BR, Krishnan A, Degulmadi D, Mayi S, Rai RR, et al. Utility of cervical dynamic magnetic resonance imaging for evaluating patients with cervical myelopathy: a retrospective study. Asian Spine J. 2024;18:647–53. 10.31616/asj.2024.0176 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Muhle C, Metzner J, Weinert D, Falliner A, Brinkmann G, Mehdorn MH, et al. Classification system based on kinematic MR imaging in cervical spondylitic myelopathy. AJNR Am J Neuroradiol. 1998;19:1763–71. [PMC free article] [PubMed] [Google Scholar]
  • 46.Shin JJ, Yoo SJ, Kim TW, So J-Y, Jeong WJ, Lee MH, et al. Radiological and Clinical Significance of Cervical Dynamic Magnetic Resonance Imaging for Cervical Spondylotic Myelopathy. Neurospine. 2024;21:443–54. 10.14245/ns.2448166.083 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Joaquim AF, Baum GR, Tan LA, Riew KD. Dynamic Cord Compression Causing Cervical Myelopathy. Neurospine. 2019;16:448–53. 10.14245/ns.1938020.101 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Nishida N, Kato Y, Imajo Y, Kawano S, Taguchi T. Biomechanical analysis of cervical spondylotic myelopathy: The influence of dynamic factors and morphometry of the spinal cord. J Spinal Cord Med [Internet]. 2012. [cited 2026 Jan 12];35:256–61. 10.1179/2045772312Y.0000000024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Yu Z, Cheng X, Chen J, Huang Z, He S, Hu H, et al. Spinal Cord Parenchyma Vascular Redistribution Underlies Hemodynamic and Neurophysiological Changes at Dynamic Neck Positions in Cervical Spondylotic Myelopathy. Front Neuroanat. 2021;15:729482. 10.3389/fnana.2021.729482 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Tykocki T, English P, Minks D, Krishnakumar A, Wynne-Jones G. Predictive value of flexion and extension diffusion tensor imaging in the early stage of cervical myelopathy. Neuroradiology [Internet]. 2018. [cited 2026 Jan 25];60:1181–91. 10.1007/s00234-018-2097-y [DOI] [PubMed] [Google Scholar]
  • 51.Galle B, Ouyang H, Shi R, Nauman E. Correlations between tissue-level stresses and strains and cellular damage within the guinea pig spinal cord white matter. J Biomech. 2007;40:3029–33. 10.1016/j.jbiomech.2007.03.014 [DOI] [PubMed] [Google Scholar]
  • 52.Jannesar S, Salegio EA, Beattie MS, Bresnahan JC, Sparrey CJ. Correlating Tissue Mechanics and Spinal Cord Injury: Patient-Specific Finite Element Models of Unilateral Cervical Contusion Spinal Cord Injury in Non-Human Primates. J Neurotrauma. 2021;38:698–717. 10.1089/neu.2019.6840 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Sparrey CJ, Manley GT, Keaveny TM. Effects of white, grey, and pia mater properties on tissue level stresses and strains in the compressed spinal cord. J Neurotrauma. 2009;26:585–95. 10.1089/neu.2008.0654 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Nouri A, Martin AR, Mikulis D, Fehlings MG. Magnetic resonance imaging assessment of degenerative cervical myelopathy: a review of structural changes and measurement techniques. Neurosurg Focus. 2016;40:E5. 10.3171/2016.3.FOCUS1667 [DOI] [PubMed] [Google Scholar]
  • 55.Kadanka Z, Adamova B, Kerkovsky M, Kadanka Z, Dusek L, Jurova B, et al. Predictors of symptomatic myelopathy in degenerative cervical spinal cord compression. Brain Behav. 2017;7:e00797. 10.1002/brb3.797 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Witiw CD, Mathieu F, Nouri A, Fehlings MG. Clinico-Radiographic Discordance: An Evidence-Based Commentary on the Management of Degenerative Cervical Spinal Cord Compression in the Absence of Symptoms or With Only Mild Symptoms of Myelopathy. Glob Spine J. 2018;8:527–34. 10.1177/2192568217745519 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES