Apnea Detection Based on Time Delay Neural Network

Abstract

Sleep apnea syndrome (SAS) is a very common sleep disorder disease. Reliable detection of apnea is very crucial for subsequent treatment. In this article, a novel method based on artificial neural network is proposed for such purpose. With its time-invariant property the time delay neural network (TDNN) is adopted in this system to employ the temporal trend of apnea event. As airflow and SaO2 take the most important roles in sleep apnea syndrome diagnosis, features extracted from both of them are simultaneously fed into the neural network. The proposed algorithm was tested with 15 overnight polysomnographic (PSG) records, and with a sensitivity rate of 90.7% and 80.8%, a specificity rate of 86.4% and 81.4% for apnea and hypopnea detection, respectively. Furthermore, the proposed algorithm can accommodate in some manner the airflow sensor failure due to technical errors. But, as the SaO2 changes are commonly delayed by 10 or more seconds compared to the airflow signal, integration of SaO2 make this method only suited for offline detection. In conclusion, systems based on this algorithm can be used as a valuable timesaving adjunct for PSG SAS diagnosis.

EMG classification in obstructive sleep apnea syndrome and periodic limb movement syndrome patients by using wavelet packet transform and extreme learning machine

“…They reported a sensitivity of 99.6% for differentiating between normal and apnea events. Tian et al [24] used the feature extraction of airflow and oxygen saturation for feeding the neural network input. They found a sensitivity rate of 90.7% and 80.8% and a specificity rate of 86.4% and 81.4% for apnea and hypopnea detection, respectively.…”

Intelligent Computing Techniques for the Detection of Sleep Disorders: A Review

“…ANN has a few advantages over KBS in exhibiting a complementary approach to RBR in terms of knowledge illustration which requires a long time to construct such a system from rule based approach that extracts features from the original recordings like EEG, PSG and then make the rules according to human knowledge (Tian, 2005). ANN possesses a very attractive property for automated recognition of sleep EEG patterns which doesn't require any elaborate classification rules or complex domain knowledge (Ventouras, 2005).…”

“…The semi-automatic arousal detection system was implemented by Sorensen using FFNN (Sorensena, 2011) to overcome the limitations of manual system. The abstraction and repetitiveness of the task also direct to inaccuracies and low inter-scorer agreement (Tian, 2005) wherein no knowledge of probability distribution is required. NN is skillful in estimating the posterior probabilities, providing the base for implementing classification rules (Marcos, 2008).…”

“…Canisius used the bio-signals processing algorithms for the detection of sleep disorders breathing from ECG signals with the accuracy of 77% (Canisius, 2008). Much time is required to create a system from RBR which extract features from original information to construct the rules as per human brain while in opposite numerical classification doesn't need any rule base and features are extracted by power spectrum (Tian, 2005). The chromosomes having variable length structure and fitness function are used to find the optimal input features and recognition of network specifications (Kim, 2000).…”

“…The chromosomes having variable length structure and fitness function are used to find the optimal input features and recognition of network specifications (Kim, 2000). In numerical classification method, no such human knowledge and rules are required (Tian, 2005). The advancement of a portable microcontroller based device can be better proved for longterm and home monitoring of snoring.…”