Computer Methods and Programs in Biomedicine 140 (2017) 265–274 Contents lists available at ScienceDirect Computer Methods and Programs in Biomedicine journal homepage: www.elsevier.com/locate/cmpb Classification techniques on computerized systems to predict and/or to detect Apnea: A systematic review Nuno Pombo a,∗ , Nuno Garcia a , Kouamana Bousson b a Research Units: Instituto de Telecomunicações and ALLab Assisted Living Computing and Telecommunications Laboratory, Department of Informatics, Universidade da Beira Interior, Covilhã, Portugal and Universidade Lusófona de Humanidades e Tecnologias, Lisbon, Portugal b Research Unit: LAETA/UBI-AEROG, Department of Aerospace Sciences, Universidade da Beira Interior, Covilhã, Portugal a r t i c l e i n f o Article history: Received 13 June 2016 Revised 28 December 2016 Accepted 3 January 2017 Keywords: Sleep apnea Machine learning Classification Threshold-based classification Systematic review a b s t r a c t Background and objective: Sleep apnea syndrome (SAS), which can significantly decrease the quality of life is associated with a major risk factor of health implications such as increased cardiovascular dis- ease, sudden death, depression, irritability, hypertension, and learning difficulties. Thus, it is relevant and timely to present a systematic review describing significant applications in the framework of computa- tional intelligence-based SAS, including its performance, beneficial and challenging effects, and modeling for the decision-making on multiple scenarios. Methods: This study aims to systematically review the literature on systems for the detection and/or prediction of apnea events using a classification model. Results: Forty-five included studies revealed a combination of classification techniques for the diagno- sis of apnea, such as threshold-based (14.75%) and machine learning (ML) models (85.25%). In addition, the ML models, were clustered in a mind map, include neural networks (44.26%), regression (4.91%), instance-based (11.47%), Bayesian algorithms (1.63%), reinforcement learning (4.91%), dimensionality re- duction (8.19%), ensemble learning (6.55%), and decision trees (3.27%). Conclusions: A classification model should provide an auto-adaptive and no external-human action de- pendency. In addition, the accuracy of the classification models is related with the effective features se- lection. New high-quality studies based on randomized controlled trials and validation of models using a large and multiple sample of data are recommended. © 2017 Published by Elsevier Ireland Ltd. Abbreviations: AI, Apnea Index; AHI, Apnea and Hypopnea Index; AIRS, Arti- ficial Immune Recognition System; ANN, Artificial Neural Network; ANFIS, Adap- tive Neuro-Fuzzy Inference System; AUC, Area Under receiver operating characteris- tic Curve; BHC, Binary Hierarchical Classification; BNN, Bayesian Neural Network; CAS, Central Sleep Apnea; ECOC, Error Correcting Output Code; ECG, Electrocar- diogram; EEG, Electroencephalogram; EMG, Electromyography; EOG, electrooculog- raphy; FP, False Positive; FN, False Negative; HI, Hypopnea Index; HMM, Hidden Markov Model; KNN, K-Nearest Neighbor; LDA, Linear Discriminant Analysis; LS- SVM, Least Squares Support Vector Machine; LR, Logistic Regression; LVQ, Learning Vector Quantization; ML, Machine Learning; MLR, Multi-Linear Regression; MSA, Mixed Sleep Apnea; NARX, Nonlinear AutoRegressive network with eXogenous; OSA, Obstructive Sleep Apnea; PNN, Probabilistic Neural Network; NPV, Negative Predictive Value; PPG, Photoplethysmogram; PPV, Positive Predictive Value; PSG, Polysomnogram; RBFNN, Radial Basis Function Neural Network; RCT, Randomized Controlled Trial; RDI, Respiratory Disturbance Index; ROC, receiver operating char- acteristic; SAS, Sleep Apnea Syndrome; Sp02, Oxygen Saturation; SRN, Simple Re- current Network; SVM, Support Vector Machine; TP, True Positive; TN, True Nega- tive; VDA, Voice Activity Detection. ∗ Corresponding author. E-mail addresses: ngpombo@ubi.pt (N. Pombo), ngarcia@di.ubi.pt (N. Garcia), bousson@ubi.pt (K. Bousson). 1. Introduction Sleep apnea syndrome (SAS) is defined as a temporary closure of the upper airway during sleep when air is prevented from enter- ing lungs which may results on the complete cessation of breath- ing for more than 10 s in adults [1]. This is typically, accompanied by a reduction in blood oxygen saturation and leads to arousal from sleep in order to breathe. In addition, repetitive obstructive events during sleep are hypothesized to cause intermittent hy- poxia, resulting in activation of oxygen free radicals and an oxida- tive stress response. As SAS events are classified according to whether the patient exhibits respiratory effort, then the presence of abdominal and thoracic effort for continuing breathing while air flow completely stops, is called Obstructive Sleep Apnea (OSA) representing the most common pattern of SAS. On the contrary, when a complete cessation of both respiratory movements and airflow during at least 10 s, is considered as Central Sleep Apnea (CSA). Finally, the combination of these two symptoms, defined by a central respira- tory pause followed, in a relatively short interval of time, by an http://dx.doi.org/10.1016/j.cmpb.2017.01.001 0169-2607/© 2017 Published by Elsevier Ireland Ltd.