Abstract
Measuring the quality of life of an animal in a production system is difficult, time-consuming, and expensive. We tested the ability of both NIR and MIR spectroscopy, each combined with machine learning, to predict the prior exposure of pigs to long- and short-term life challenges when they are raised in an intensive system. Samples were obtained post-mortem from two locations in the prefrontal cortex. The analysis showed a clear separation between the gray and white matter from the prefrontal cortex with MIR spectroscopy. Exposure to long-term challenge was poorly predicted by the MIR or NIR spectra (< 45% correct classifications). By contrast, the correct classification of samples according to the exposure to a short-tern challenge before death was higher than 65%. These rates of classification, considering the complexity of the stimulus and the small sample size, support that vibrational spectroscopy could be used to assess the exposure to challenging events of an animal on post-mortem brain tissue.