Although magnetic resonance imaging (MRI) has been increasingly used as a diagnostic tool for cervical spine injuries in canines, a comprehensive normal MRI anatomy of the canine cervical spine muscles is lacking. Therefore, the purpose of this study was to build a magnetic resonance imaging atlas of the normal cross sectional anatomy of the muscles of the canine cervical spine. MRI scans were performed on a canine cadaver using a combination of T1 and T2-weighted images in the transverse, sagittal and dorsal planes acquired at a slice thickness of 1mm. Muscle contours were traced manually in each slice, using local osseous structures as reference points for muscle identification. Twenty-two muscles were traced in 401 slices in the cervical region. A three dimensional surface model of all the contoured muscles was created to illustrate the complex geometrical arrangement of canine neck muscles. The cross-sectional area of the muscles was measured at the mid-level of each vertebra. The accuracy of the location of the mapped muscles was verified by comparing the sagittal view of the 3D model of muscles with still photographs obtained from anatomic canine cadaver dissection. We believe that this information will provide a unique and valuable resource for veterinary researchers, clinicians and surgeons who wish to evaluate MRI images of the cervical spine. It will also serve as the foundation for ongoing work to develop a computational model of the canine cervical spine in which anatomical information is combined with electromyographic, kinematic and kinetic data.

Due to the frequency of cervical spine injuries in canines, the purpose of this effort was to develop an EMG-driven dynamic model of the canine cervical spine to assess a biomechanical understanding that enables one to investigate the risk of neck disorders. A canine subject was recruited in this investigation in order to collect subject specific data. Reflective markers and a motion capture system were used for kinematic measurement; surface electrodes were used to record electromyography signals, and with the aid of force plate kinetics were recorded. A 3D model of the canine subject was reconstructed from an MRI dataset. Muscles lines of action were defined through a new technique with the aid of 3D white light scanner. The model performed well with a 0.73 weighted R value in all three planes. The weighted average absolute error of the predicted moment was less than 10% of the external moment. The proposed model is a canine specific forward-dynamics model that precisely tracks the canine subject head and neck motion, calculates the muscle force generated from the twelve major moment producing muscles, and estimates resulting loads on specific spinal tissues.