ABSTRACT
BACKGROUND: Spinal cord injury (SCI) is a debilitating event with multiple mechanisms of degeneration leading to life-long paralysis. Biomaterial strategies, including bridges that span the injury and provide a pathway to reconnect severed regions of the spinal cord, can promote partial restoration of motor function following SCI. Axon growth through the bridge is essential to characterizing regeneration, as recovery can occur via other mechanisms such as plasticity. Quantitative analysis of axons by manual counting of histological sections can be slow, which can limit the number of bridge designs evaluated. In this study, we report a semi-automated process to resolve axon numbers in histological sections, which allows for efficient analysis of large data sets. NEW METHOD: Axon numbers were estimated in SCI cross-sections from animals implanted with poly(lactide co-glycolide) (PLG) bridges with multiple channels for guiding axons. Immunofluorescence images of histological sections were filtered using a Hessian-based approach prior to threshold detection to improve the signal-to-noise ratio and filter out background staining associated with PLG polymer. RESULTS: Semi-automated counting successfully recapitulated average axon densities and myelination in a blinded PLG bridge implantation study. COMPARISON WITH EXISTING METHODS: Axon counts obtained with the semi-automated technique correlated well with manual axon counts from blinded independent observers across sections with a wide range of total axons. CONCLUSIONS: This semi-automated detection of Hessian-filtered axons provides an accurate and significantly faster alternative to manual counting of axons for quantitative analysis of regeneration following SCI.
