This doctoral dissertation study developed three-dimensional biomechanical neck models to investigate the internal loads on the structures of the human neck that result from isometric external loads on the head. In biomechanical modeling procedure, there are two basic approaches to estimate muscle forces and spinal loads, given the indeterminacy of system equations:​ optimization method and EMG (electromyography) based method. Each method has its own advantages and weak points. Moroney et al. [56] applied the double optimization method to a human neck model to calculate the muscle forces and spinal loads. But, their model had the limitations that the general optimization method has. Namely, their model did not predict antagonistic muscle forces, therefore the predicted spinal loads are possibly underestimated due to not including the antagonistic muscle forces. On the other hand, Cholewicki and McGill [20] introduced new hybrid method called EMG assisted optimization (EMGAO) method. This new approach provided physiologically based muscle recruitment patterns, as did the EMGbased method, together with the exact fulfillment of moment constraints about three joint axes [20], Especially, in investigating the cervical spine case, no one has tried to apply the EMG-baesed method and EMGAO method to solve the statically indeterminant problem.

The first goal of the study was to investigate the EMG behavior of neck musculature during isometric exertions from EMG experiments. The second was to apply double-optimization (DOPT) method, EMG-based (EMG) method, and EMG assisted optimization (EMGAO) method to biomechanical modeling of the human neck to estimate muscle forces and C4/5 cervical joint loads. The third was to compare the results of the different methods, and the fourth was to select the method that best describes the human neck muscle forces and spinal loads.

To evaluate the muscle forces and spinal loads in the human neck model, a double optimization model, an EMG-based model, and an EMG assisted optimization model were formulated. The effects of co-contraction of neck musculature were assessed by comparing the three methods since the methods predicted varying amounts of co-contraction. To formulate the EMG-based model, electromyographic signals were collected from ten male subjects. EMG signals were obtained from eight sites around the C4/5 level of the neck by Ag-AgCl surface electrodes while the subject performed near maximum, isometric, ramp efforts in flexion, extension, left lateral bending and right lateral bending. During maximum voluntary contraction (MVC) efforts, most of the muscles showed relatively high levels of EMG signal expressed as % MVC in twisting efforts. The greatest mean (± SD) EMG (% MVC) activity of each neck muscle during peak isometric twisting ranges up to 91.0 (± 12.0) % MVC in trapezius, 77.6 (± 18.5) % MVC in sternocleidomastoid, 77.3 (± 19.9) % MVC in platysma, and 66.4 (±21.2) % MVC in levator scapulae. Posterior part of the neck (trapezius) was the only electrode site in which maximum activity consistently occurred during the same loading mode (twisting) in all subjects. However, it proved difficult to find a loading mode that consistently produced maximum EMG signals at a certain location of neck musculature for all subjects.

Muscle forces and spinal loads of 4 point (25 %, 50 %, 75 %, and 100 % of peak efforts) from the ramp trials were calculated by the three different methods. The mean values (± SD) of calculated neck muscle forces during peak exertions were 302 N (± 89 N) by EMG, 308 N (±97 N) by EMGAO, and 250 N (±52 N) by DOPT. The mean values (±SD) of calculated C4/5 joint compressive forces during peak exertions were 1654 N (±308 N) by EMG method, 1674 N (±319 N) by EMGAO method, and 1208 N (± 123 N) by DOPT method. The muscle forces and spinal loads predicted by the DOPT model were roughly equivalent to those reported by Moroney et al. [56]. In contrast to the EMG and EMGAO methods which showed activation of all the muscles including the antagonists, the DOPT approach nullified forces in a majority of flexors during the extension, extensors during flexion and right neck side muscles during left lateral bending. The root mean square (RMS) differences in the muscle forces obtained from a DOPT approach compared with an EMG-based approach were 60.7 % in extension, 81.1 % in flexion, 134.4 % in left lateral bending and 99.6 % in right lateral bending. EMGAO predictions gave RMS differences of 3.5 % in extension, 19.1 % in flexion, 52.8 % in left lateral bending and 81.5 % in right lateral bending compared with an EMG approach. EMG and EMGAO approaches showed various load sharing patterns among the agonist muscles during generation of the same magnitude of moments, especially in lateral bending. The EMGAO approach is an attempt to incorporate the biological sensitivity of the EMG approach which accommodate co-contraction of antagonists and variation of load sharing patterns of agonists while including the characteristic of the DOPT approach to satisfy moment equilibrium constraints. For this reason, EMGAO approach is considered an improvement on both the EMG and DOPT methods.

This study suggests that higher levels of spine loads are likely at maximum effort than has been previously reported for the C4/5 level by a double optimization model [4]. EMG and EMGAO methods provide data that will require a clinically more conservative approach to treatment of the cervical spine.