Motor Score Patterns Complement the AIS Grade in the Analysis of Segmental Motor Recovery after Spinal Cord Injury

Neurological recovery following spinal cord injury (SCI) is commonly studied through changes in high-level descriptors of injury severity, such as the American Spinal Injury Association Impairment Scale (AIS) grade, or total upper and/or lower extremity motor scores. More recently, the analysis of segmental motor scores has attracted interest as it provides a more detailed understanding of the exact location and extent of changes occurring during recovery. We propose to augment the analysis of segmental motor recovery with a qualitative descriptor of local motor score patterns, which is defined for all upper and lower extremity myotomes below the neurological level of injury (NLI) and based on a categorization of the changes along the rostrocaudal motor score sequence. Our hypothesis is that recovery of segmental motor scores depends on the residual function as described by the newly proposed descriptors of local motor score patterns. Using data of 1385 patients from the European Multicenter Study about SCI, we analyze differences in recovery at approximately 6 months after injury between local motor score patterns and find an increased probability of full motor recovery for myotomes associated with increased motor scores in the caudal direction. We further use an aggregated descriptor of motor score patterns focusing on increases of motor scores in the caudal direction as an alternative or complementary feature to the AIS grade in prediction models for segmental motor scores at recovery. We observe equivalent predictive performance as measured by the root mean square error between actual and predicted motor scores below the NLI for models using the same set of features and additionally either the AIS grade (median = 0.79) or local pattern (median = 0.80). This is noteworthy as the definition of local motor score patterns requires only the examination of the 10 key muscles of the International Standards for Neurological Classification of SCI on each side of the body, while the AIS grade can only be reliably determined if extensive sensory testing is performed in addition. These results indicate the potential benefits of considering information inherent to the rostrocaudal sequence of motor scores for a better understanding of motor recovery. Furthermore, it supports the development and use of abbreviated sensorimotor examinations specifically in the early phase after SCI.