Symbolic Fusion: A Novel Decision Support Algorithm for Sleep Staging Application

Abstract

With the rapid extension of clinical data and knowledge, decision making becomes a complex task for manual sleep staging. In this process, there is a need for integrating and analyzing information from heterogeneous data sources with high accuracy. This paper proposes a novel decision support algorithm-Symbolic Fusion for sleep staging application. The proposed algorithm provides high accuracy by combining data from heterogeneous sources, like EEG, EOG and EMG. This algorithm is developed for implementation in portable embedded systems for automatic sleep staging at low complexity and cost. The proposed algorithm proved to be an efficient design support method and achieved up to 76% overall agreement rate on our database of 12 patients.

Cross entropy-based automatic thresholds setting-up method for sleep staging system

“…To overcome the limitations in [7], [8], Cross Entropybased Automatic Thresholds Setting-Up (CEATS) is proposed as shown in Figure 2. It is dedicated to (1) provide optimal thresholds for sleep staging system (2) release the burden of manual interpretation and (3) allow individual flexibility.…”

“…In [7], [8], a sleep staging system is proposed using symbolic fusion which consists three levels: data fusion, feature fusion and decision fusion as shown in Figure 1. In data fusion, digital parameters are extracted from PSG recordings to maximize the useful information and to minimize noise and artifacts.…”

“…In decision fusion, inference method is used to fulfill sleep staging on the basis of feature parameters. In [7], [8], thresholds are used for translating nine digital parameters into 24 features parameters as shown in Table I. However, to define these thresholds, manual interpretation was performed.…”

“…To release the burden of physicians, different sleep staging systems have been proposed in the past decades [4], [5], [6], [7], [8]. Among existing sleep staging systems, thresholds are widely applied in transforming digital parameters into linguistic or symbolic features to model inference process under the guidance of medical knowledge [5], [6], [7], [8].…”

“…However, as far as we know, there are no fully satisfying automatic setting-up thresholds methods in existing sleep staging systems. Manually predefined values of thresholds have been widely applied in [5], [6], [7], [8] due to the following reasons: 1) To build a mathematical model or a threshold function to differentiate boundary requires a set of data with sufficient quantity and adequate quality; 2) There is lack of uniformity between subjects and thresholds variability exists [9].…”