Toward a soft computing-based correlation between oxygen toxicity seizures and hyperoxic hyperpnea

Exposure to high levels of hyperbaric oxygen (HBO 2) can induce central nervous system oxygen toxicity in humans and animals, a phenomenon characterized by repeated tonic–clonic seizures. The risk of developing this type of convulsions represents the limiting factor in using HBO 2 for a number of clinical and diving applications. Previously, using radio-telemetry in awake rats, a significant increase in the mean value of the ventilatory responses in rats to HBO 2 approximately 5–8 min before onset of seizures has been observed. This response has been termed “hyperoxic hyperpnea,” and it has been hypothesized it may serve as a predictor of an impending seizure while breathing HBO 2. The purpose of the present study was to apply soft computing methods to determine whether there is a direct correlation between the onset of hyperoxic hyperpnea (i.e., the ventilatory response as defined by tidal volume and respiratory frequency) and the onset of seizure. In our experiments, we used Multilayer Perceptron, Naive Bayes, J48, and U-BRAIN aimed at evidencing the correlation between the respiratory feature vectors and the onset of an impending seizure in unanesthetized, freely behaving rats breathing 4, 5 or 6 atmospheres absolute (ATA) of oxygen. This strategy was also aimed to finding a set of rules relating physiological parameters and convulsive phenomena through inductive inference.