Nowadays, polysomnography (PSG) is the gold-standard to diagnose obstructive sleep apnoea syndrome (OSAS). However, it is complex, time-consuming and expensive. Nocturnal pulse oximetry, which provides oxygen saturation (SaO 2 ) recordings, allows us to overcome these difficulties and could be an alternative to PSG. In the present study, multilayer perceptron (MLP) neural networks were applied to help in OSAS diagnosis using information from SaO 2 signals. We performed time and spectral analysis of these recordings to extract 14 features related to OSAS. According to the principle used for network optimisation, wWe compared the performance of two different MLP classifiers: maximum likelihood (ML) and Bayesian (BY) MLP networks. A total of 187 subjects suspected of suffering from OSAS took part in the study. Their SaO 2 signals were divided into a training set with 74 recordings and a test set with 113 recordings to develop and validate the classifiers. BY-MLP networks achieved the best performance on the test set with 85.58% accuracy (87.76% sensitivity and 82.39% specificity). These results were substantially better than those provided by ML-MLP networks, which were affected by overfitting and achieved an accuracy of 76.81% (86.42% sensitivity and 62.83% specificity). Our results suggest that the Bayesian framework is preferred to implement our MLP classifiers. The proposed BY-MLP networks could be used for early OSAS detection, contributing to and thus reduce the number of required PSGs.