Bayesian Model Averaging (BMA) methodology is applied for the assessment of newborn brain maturity from sleep EEG. In theory, this methodology provides the most accurate assessments of uncertainty in decisions. However, existing BMA techniques have been shown to provide biased assessments in the absence of some prior information allowing the model parameter space to be explored in detail within a reasonable time frame. Lack of detail leads to disproportionate sampling of the posterior distribution. In the case of EEG assessment of brain maturity, BMA results may be biased due to the absence of information on the importance of EEG features. In this paper, we explore how posterior information about EEG features can be used to reduce the negative impact of disproportionate sampling on BMA performance. We use EEG data recorded from sleeping newborns to test the effectiveness of the proposed BMA technique. The assessment of brain maturity can be obtained by estimating the age of the newborn from the sleep EEG [1] - [3]. This approach is based on clinical evidence that the post-conception and EEG-estimated ages of healthy newborns generally correspond to each other, and the brain maturity of the newborn is most likely abnormal if the ages do not match. not [2], [4]. Thus, the mismatch alerts to abnormal brain development. Established assessment methodologies are based on learning models from EEGs recorded in sleeping newborns whose brain maturity has already been assessed by clinicians. Regression models are capable of mapping brain maturity into an EEG-based index [5]. Classification models are made capable of distinguishing maturity levels: at least one with normal brain maturity and another with abnormal brain maturity [4], [6]. The established method...... middle of document ...... their impact on the result is negligible. On the contrary, when the number of weak attributes is large, the disproportion of the models becomes significant. Therefore, we could improve the BMA results by reducing disproportionate sampling. In this research, we aim to determine whether abandoning models using low EEG attributes will reduce bias in the assessment of brain maturity. A trivial strategy to use posterior information for feature selection in BMA is to use this information to learn a new set of a dataset in which weak attributes have been removed. This strategy reduces the model parameter space and therefore allows this space to be explored in more detail. The other possible strategy consists of refining the whole by discarding models that use weak attributes. We hope that such refinement can improve the performance of BMA..