Pitch jump probability measures for the analysis of snoring sounds in apnea

Obstructive sleep apnea (OSA) is a highly prevalent disease in which upper airways are collapsed during sleep, leading to serious consequences. The gold standard of diagnosis, called polysomnography (PSG), requires a full-night hospital stay connected to over ten channels of measurements requiring physical contact with sensors. PSG is inconvenient, expensive and unsuited for community screening. Snoring is the earliest symptom of OSA, but its potential in clinical diagnosis is not fully recognized yet. Diagnostic systems intent on using snore-related sounds (SRS) face the tough problem of how to define a snore. In this paper, we present a working definition of a snore, and propose algorithms to segment SRS into classes of pure breathing, silence and voiced/unvoiced snores. We propose a novel feature termed the 'intra-snore-pitch-jump' (ISPJ) to diagnose OSA. Working on clinical data, we show that ISPJ delivers OSA detection sensitivities of 86-100% while holding specificity at 50-80%. These numbers indicate that snore sounds and the ISPJ have the potential to be good candidates for a take-home device for OSA screening. Snore sounds have the significant advantage in that they can be conveniently acquired with low-cost non-contact equipment. The segmentation results presented in this paper have been derived using data from eight patients as the training set and another eight patients as the testing set. ISPJ-based OSA detection results have been derived using training data from 16 subjects and testing data from 29 subjects.     

Signal quality indices and data fusion for determining clinical acceptability of electrocardiograms

A completely automated algorithm to detect poor-quality electrocardiograms (ECGs) is described. The algorithm is based on both novel and previously published signal quality metrics, originally designed for intensive care monitoring. The algorithms have been adapted for use on short (5-10?s) single- and multi-lead ECGs. The metrics quantify spectral energy distribution, higher order moments and inter-channel and inter-algorithm agreement. Seven metrics were calculated for each channel (84?features in all) and presented to either a multi-layer perceptron artificial neural network or a support vector machine (SVM) for training on a multiple-annotator labelled and adjudicated training dataset. A single-lead version of the algorithm was also developed in a similar manner. Data were drawn from the PhysioNet Challenge 2011 dataset where binary labels were available, on 1500 12-lead ECGs indicating whether the entire recording was acceptable or unacceptable for clinical interpretation. We re-annotated all the leads in both the training set (1000 labelled ECGs) and test dataset (500 12-lead ECGs where labels were not publicly available) using two independent annotators, and a third for adjudication of differences. We found that low-quality data accounted for only 16% of the ECG leads. To balance the classes (between high and low quality), we created extra noisy data samples by adding noise from PhysioNet's noise stress test database to some of the clean 12-lead ECGs. No data were shared between training and test sets. A classification accuracy of 98% on the training data and 97% on the test data were achieved. Upon inspection, incorrectly classified data were found to be borderline cases which could be classified either way. If these cases were more consistently labelled, we expect our approach to achieve an accuracy closer to 100%.     

A method to screen obstructive sleep apnea using multi-variable non-intrusive measurements

Obstructive sleep apnea (OSA) is a serious sleep disorder. The current standard OSA diagnosis method is polysomnography (PSG) testing. PSG requires an overnight hospital stay while physically connected to 10-15 channels of measurement. PSG is expensive, inconvenient and requires the extensive involvement of a sleep technologist. As such, it is not suitable for community screening. OSA is a widespread disease and more than 80% of sufferers remain undiagnosed. Simplified, unattended and cheap OSA screening methods are urgently needed. Snoring is commonly associated with OSA but is not fully utilized in clinical diagnosis. Snoring contains pseudo-periodic packets of energy that produce characteristic vibrating sounds familiar to humans. In this paper, we propose a multi-feature vector that represents pitch information, formant information, a measure of periodic structure existence in snore episodes and the neck circumference of the subject to characterize OSA condition. Snore features were estimated from snore signals recorded in a sleep laboratory. The multi-feature vector was applied to a neural network for OSA/non-OSA classification and K-fold cross-validated using a random sub-sampling technique. We also propose a simple method to remove a specific class of background interference. Our method resulted in a sensitivity of 91 ± 6% and a specificity of 89 ± 5% for test data for AHI(THRESHOLD) = 15 for a database consisting of 51 subjects. This method has the potential as a non-intrusive, unattended technique to screen OSA using snore sound as the primary signal.     

An efficient method for snore/nonsnore classification of sleep sounds

A new method to detect snoring episodes in sleep sound recordings is proposed. Sleep sound segments (i.e., 'sound episodes' or simply 'episodes') are classified as snores and nonsnores according to their subband energy distributions. The similarity of inter- and intra-individual spectral energy distributions motivated the representation of the feature vectors in a lower dimensional space. Episodes have been efficiently represented in two dimensions using principal component analysis, and classified as snores or nonsnores. The sound recordings were obtained from individuals who are suspected of OSAS pathology while they were connected to the polysomnography in Gülhane Military Medical Academy Sleep Studies Laboratory (GMMA-SSL), Ankara, Turkey. The data from 30 subjects (18 simple snorers and 12 OSA patients) with different apnoea/hypopnea indices were classified using the proposed algorithm. The system was tested by using the manual annotations of an ENT specialist as a reference. The accuracy for simple snorers was found to be 97.3% when the system was trained using only simple snorers' data. It drops to 90.2% when the training data contain both simple snorers' and OSA patients' data. (Both of these results were obtained by using training and testing sets of different individuals.) In the case of snore episode detection with OSA patients the accuracy is 86.8%. All these results can be considered as acceptable values to use the system for clinical purposes including the diagnosis and treatment of OSAS. The method proposed here has been used to develop a tool for the ENT clinic of GMMA-SSL that provides information for objective evaluation of sleep sounds.     

Impact of gender on snore-based obstructive sleep apnea screening

Obstructive sleep apnea syndrome (OSA) is a serious widespread disease in which upper airways (UA) are collapsed during sleep. OSA has marked male predominance in prevalence. Although women are less vulnerable to OSA, under-diagnosed OSA in women may associate with serious consequences. Snoring is commonly associated with OSA and one of the earliest symptoms. Snore sounds (SS) are generated due to vibration of the collapsing soft tissues of the UA. Structural and functional properties of the UA are gender dependent. SS capture these time varying gender attributed UA properties and those could be embedded in the acoustic properties of SS. In this paper, we investigate the gender-specific acoustic property differences of SS and try to exploit these differences to enhance the snore-based OSA detection performance. We developed a snore-based multi-feature vector for OSA screening and one time-measured neck circumference was augmented. Snore features were estimated from SS recorded in a sleep laboratory from 35 females and 51 males and multi-layer neural network-based pattern recognition algorithms were used for OSA/non-OSA classification. The results were K-fold cross-validated. Gender-dependent modeling resulted in an increase of around 7% in sensitivity and 6% in specificity at the decision threshold AHI = 15 against a gender-neutral model. These results established the importance of adopting gender-specific models for the snore-based OSA screening technique.     

Discriminative spatial-spectral manifold embedding for hyperspectral image classification

In hyperspectral images (HSI) classification, it is important to combine multiple features of a certain pixel in both spatial and spectral domains to improve the classification accuracy. To achieve this goal, this article proposes a novel spatial-spectral feature dimensionality reduction algorithm based on manifold learning. For each feature, a graph Laplacian matrix is constructed based on discriminative information from training samples, and then the graph Laplacian matrices of the various features are linearly combined using a set of empirically defined weights. Finally, the feature mapping is obtained by an eigen-decomposition problem. Based on the classification results of the public Indiana Airborne Visible Infrared Imaging Spectrometer dataset and Texas Hyperspectral Digital Imagery Collection Experiment data set, the technical accuracies show that our method achieves superior performance compared to some representative HSI feature extraction and dimensionality reduction algorithms.     

Frequency modulation stimulated Raman scattering scheme for real-time background correction with a single light source

A frequency modulation (FM) scheme for stimulated Raman scattering (SRS) is presented with a single fiber-based light source. Pulse-to-pulse wavelength-switching allows real-time subtraction of parasitic signals leaving only the resonant SRS signal with a noise reduction of up to 30 % compared to digital subtraction schemes, leading effectively to a contrast improvement by a factor of up to 8.3. The wide tuning range of the light source from 1500 cm⁻¹ to 3000 cm⁻¹ and the possibility to separately adjust the resonant and the nonresonant wavenumber for every specimen allow to investigate a variety of samples with high contrast and high signal-to-noise ratio, e. g., for medical diagnostics.     

Spectral–spatial classification of hyperspectral images using deep convolutional neural networks

In this letter, a novel deep learning framework for hyperspectral image classification using both spectral and spatial features is presented. The framework is a hybrid of principal component analysis, deep convolutional neural networks (DCNNs) and logistic regression (LR). The DCNNs for hierarchically extract deep features is introduced into hyperspectral image classification for the first time. The proposed technique consists of two steps. First, feature map generation algorithm is presented to generate the spectral and spatial feature maps. Second, the DCNNs-LR classifier is trained to get useful high-level features and to fine-tune the whole model. Comparative experiments conducted over widely used hyperspectral data indicate that DCNNs-LR classifier built in this proposed deep learning framework provides better classification accuracy than previous hyperspectral classification methods.     

Resolution of dyshidrotic dermatitis of the hand after treatment with continuous positive airway pressure for obstructive sleep apnea

Eczema of the palm and obstructive sleep apnea (OSA) are common disorders. Proinflammatory cytokines and cell adhesion molecules are elevated in both of these disorders. We describe an unusual patient with OSA who had recurrent free remission of dermatitis after treatment with continuous positive airway pressure. We speculate that the resolution of the patient's skin condition may reflect the effects of increased tissue oxygenation during sleep, reduced sleep fragmentation, and/or a reduction in sympathetic tone associated with successful sleep apnea treatment.     

Investigation of sequential properties of snoring episodes for obstructive sleep apnoea identification

In this paper, 'snore regularity' is studied in terms of the variations of snoring sound episode durations, separations and average powers in simple snorers and in obstructive sleep apnoea (OSA) patients. The goal was to explore the possibility of distinguishing among simple snorers and OSA patients using only sleep sound recordings of individuals and to ultimately eliminate the need for spending a whole night in the clinic for polysomnographic recording. Sequences that contain snoring episode durations (SED), snoring episode separations (SES) and average snoring episode powers (SEP) were constructed from snoring sound recordings of 30 individuals (18 simple snorers and 12 OSA patients) who were also under polysomnographic recording in Gülhane Military Medical Academy Sleep Studies Laboratory (GMMA-SSL), Ankara, Turkey. Snore regularity is quantified in terms of mean, standard deviation and coefficient of variation values for the SED, SES and SEP sequences. In all three of these sequences, OSA patients' data displayed a higher variation than those of simple snorers. To exclude the effects of slow variations in the base-line of these sequences, new sequences that contain the coefficient of variation of the sample values in a 'short' signal frame, i.e., short time coefficient of variation (STCV) sequences, were defined. The mean, the standard deviation and the coefficient of variation values calculated from the STCV sequences displayed a stronger potential to distinguish among simple snorers and OSA patients than those obtained from the SED, SES and SEP sequences themselves. Spider charts were used to jointly visualize the three parameters, i.e., the mean, the standard deviation and the coefficient of variation values of the SED, SES and SEP sequences, and the corresponding STCV sequences as two-dimensional plots. Our observations showed that the statistical parameters obtained from the SED and SES sequences, and the corresponding STCV sequences, possessed a strong potential to distinguish among simple snorers and OSA patients, both marginally, i.e., when the parameters are examined individually, and jointly. The parameters obtained from the SEP sequences and the corresponding STCV sequences, on the other hand, did not have a strong discrimination capability. However, the joint behaviour of these parameters showed some potential to distinguish among simple snorers and OSA patients.     

Abdominal fat and sleep apnea: the chicken or the egg?

Obstructive sleep apnea (OSA) syndrome is a disorder characterized by repetitive episodes of upper airway obstruction that occur during sleep. Associated features include loud snoring, fragmented sleep, repetitive hypoxemia/hypercapnia, daytime sleepiness, and cardiovascular complications. The prevalence of OSA is 2-3% and 4-5% in middle-aged women and men, respectively. The prevalence of OSA among obese patients exceeds 30%, reaching as high as 50-98% in the morbidly obese population. Obesity is probably the most important risk factor for the development of OSA. Some 60-90% of adults with OSA are overweight, and the relative risk of OSA in obesity (BMI >29 kg/m(2)) is >or=10. Numerous studies have shown the development or worsening of OSA with increasing weight, as opposed to substantial improvement with weight reduction. There are several mechanisms responsible for the increased risk of OSA with obesity. These include reduced pharyngeal lumen size due to fatty tissue within the airway or in its lateral walls, decreased upper airway muscle protective force due to fatty deposits in the muscle, and reduced upper airway size secondary to mass effect of the large abdomen on the chest wall and tracheal traction. These mechanisms emphasize the great importance of fat accumulated in the abdomen and neck regions compared with the peripheral one. It is the abdomen much more than the thighs that affect the upper airway size and function. Hence, obesity is associated with increased upper airway collapsibility (even in nonapneic subjects), with dramatic improvement after weight reduction. Conversely, OSA may itself predispose individuals to worsening obesity because of sleep deprivation, daytime somnolence, and disrupted metabolism. OSA is associated with increased sympathetic activation, sleep fragmentation, ineffective sleep, and insulin resistance, potentially leading to diabetes and aggravation of obesity. Furthermore, OSA may be associated with changes in leptin, ghrelin, and orexin levels; increased appetite and caloric intake; and again exacerbating obesity. Thus, it appears that obesity and OSA form a vicious cycle where each results in worsening of the other.     

Robust electrocardiogram (ECG) beat classification using discrete wavelet transform

This paper presents a robust technique for the classification of six types of heartbeats through an electrocardiogram (ECG). Features extracted from the QRS complex of the ECG using a wavelet transform along with the instantaneous RR-interval are used for beat classification. The wavelet transform utilized for feature extraction in this paper can also be employed for QRS delineation, leading to reduction in overall system complexity as no separate feature extraction stage would be required in the practical implementation of the system. Only 11 features are used for beat classification with the classification accuracy of approximately 99.5% through a KNN classifier. Another main advantage of this method is its robustness to noise, which is illustrated in this paper through experimental results. Furthermore, principal component analysis (PCA) has been used for feature reduction, which reduces the number of features from 11 to 6 while retaining the high beat classification accuracy. Due to reduction in computational complexity (using six features, the time required is approximately 4 ms per beat), a simple classifier and noise robustness (at 10 dB signal-to-noise ratio, accuracy is 95%), this method offers substantial advantages over previous techniques for implementation in a practical ECG analyzer.     

Implementation of a Wearable Ultrasound Device for the Overnight Monitoring of Tongue Base Deformation during Obstructive Sleep Apnea Events

Obstructive sleep apnea (OSA) is a breathing disorder characterized by the repeated collapse of the pharyngeal airway during sleep. Previous studies have reported that tongue base deformation may be a major contributing factor. However, overnight monitoring of tongue motion in patients with OSA has previously been impracticable. We developed a wearable ultrasound device for prolonged recording during natural sleep of the changes in tongue base thickness (TBT) in patients with OSA. The maximum TBT was fed into a polysomnography system so that physiologic signals and TBT data were simultaneously monitored. Subject trials revealed that TBT increased significantly during snoring, hypopnea and apnea events during natural sleep in patients with OSA. Moreover, the data revealed that the location of the maximum TBT during normal breathing was significantly different compared with the location during obstructive respiratory events, which implies a posterior or inferior displacement of the tongue base during sleep apnea.     

Spectral-spatial classification of hyperspectral imagery using a dual-channel convolutional neural network

In this article, a novel dual-channel convolutional neural network (DC-CNN) framework is proposed for accurate spectral-spatial classification of hyperspectral image (HSI). In this framework, one-dimensional CNN is utilized to automatically extract the hierarchical spectral features and two-dimensional CNN is applied to extract the hierarchical space-related features, and then a softmax regression classifier is used to combine the spectral and spatial features together and predict classification results eventually. To overcome the problem of the limited available training samples in HSIs, we propose a simple data augmentation method which is efficient and effective for improving HSI classification accuracy. For comparison and validation, we test the proposed method along with three other deep-learning-based HSI classification methods on two real-world HSI data sets. Experimental results demonstrate that our DC-CNN-based method outperforms the state-of-the-art methods by a considerable margin.     

Intensive care unit depth of sleep: proof of concept of a simple electroencephalography index in the non-sedated

Introduction

Intensive care unit (ICU) patients are known to experience severely disturbed sleep, with possible detrimental effects on short- and long- term outcomes. Investigation into the exact causes and effects of disturbed sleep has been hampered by cumbersome and time consuming methods of measuring and staging sleep. We introduce a novel method for ICU depth of sleep analysis, the ICU depth of sleep index (IDOS index), using single channel electroencephalography (EEG) and apply it to outpatient recordings. A proof of concept is shown in non-sedated ICU patients.

Methods

Polysomnographic (PSG) recordings of five ICU patients and 15 healthy outpatients were analyzed using the IDOS index, based on the ratio between gamma and delta band power. Manual selection of thresholds was used to classify data as either wake, sleep or slow wave sleep (SWS). This classification was compared to visual sleep scoring by Rechtschaffen & Kales criteria in normal outpatient recordings and ICU recordings to illustrate face validity of the IDOS index.

Results

When reduced to two or three classes, the scoring of sleep by IDOS index and manual scoring show high agreement for normal sleep recordings. The obtained overall agreements, as quantified by the kappa coefficient, were 0.84 for sleep/wake classification and 0.82 for classification into three classes (wake, non-SWS and SWS). Sensitivity and specificity were highest for the wake state (93% and 93%, respectively) and lowest for SWS (82% and 76%, respectively). For ICU recordings, agreement was similar to agreement between visual scorers previously reported in literature.

Conclusions

Besides the most satisfying visual resemblance with manually scored normal PSG recordings, the established face-validity of the IDOS index as an estimator of depth of sleep was excellent. This technique enables real-time, automated, single channel visualization of depth of sleep, facilitating the monitoring of sleep in the ICU.     

Obstructive sleep apnea and insight into mechanisms of sympathetic overactivity

Nearly two decades ago, we evaluated ten patients with obstructive sleep apnea (OSA). We determined that alarming nocturnal oscillations in arterial pressure and sympathetic nerve activity (SNA) were caused by regulatory coupling and neural interactions among SNA, apnea, and ventilation. Patients with OSA exhibited high levels of SNA when awake, during normal ventilation, and during normoxia, which contributed to hypertension and organ damage. Additionally, we achieved a beneficial and potentially lifesaving reduction in SNA through the application of continuous positive airway pressure (CPAP), which remains a primary therapeutic approach for patients with OSA. With these results in hindsight, we herein discuss three concepts with functional and therapeutic relevance to the integrative neurobiology of autonomic cardiovascular control and to the mechanisms involved in excessive sympathoexcitation in OSA.     

睡眠体位对睡眠呼吸暂停低通气综合征患者呼吸功能的影响

背景阻塞性睡眠呼吸暂停(obstructive sleep apnea,OSA)低通气患者的呼吸紊乱严重度与睡眠体位有关,根据呼吸暂停低通气指数(AHI)将OSA患者分为体位型和非体位型两种类型.目的比较体位型与非体位型OSA患者的临床特征,探讨睡眠体位与OSA患者呼吸功能的关系.设计以OSA患者为研究对象的观察对比研究.单位南京军区南京总医院的睡眠呼吸监测中心.对象选择1998-10/2002-05在南京军区南京总医院呼吸内科睡眠呼吸监测室就诊并行整夜多导睡眠仪检查的患者共225例.纳入标准①AHI≥l0次/h者,②平卧位及侧卧位睡眠时间≥30 min者,③年龄≥20岁者.排除标准①年龄＜20岁者,②患重大躯体疾病者.其中63例为体位型,162例为非体位型.方法所有患者均接受整夜多导睡眠仪检查,同时详询患者病史和进行体格检查,应用体积描计仪进行呼吸功能测定.比较体位型与非体位型两组患者的一般临床特征、肺功能参数及多导睡眠图资料.应用多元逐步回归分析探讨决定患者为体位型或非体位型的因子.主要观察指标主要结局两组患者的临床资料、多导睡眠图资料及肺功能比较.次要结局OSA患者的分型与临床资料各项目的相关性.结果225例OSA患者中体位型患者共63例(28%),体位组和非体位组患者体质量指数分别(27.97±3.21),(26.22±2.72)kg/m 2(t=3.977,P＜0.01).两组多导睡眠图资料比较,体质量指数匹配前除侧卧位睡眠时间、平卧位睡眠时间、侧卧位最低血氧饱和度及基础血氧饱和度外,其他项目两组间差异均有显著性意义(P＜0.05或P＜0 01);匹配后除两组夜间肢体运动次数差异无显著意义外,其余结果与原始组比较结果相同.两组肺功能参数间差异无显著性意义(P＞0.05).多元逐步回归分析显示,决定为体位或非体位患者的主要因子为AHI和体质量指数,两因子的预测能力为26.2%;直线相关分析结果表明,在全组及非体位组患者,睡眠呼吸紊乱度(AHI及血氧饱和度)与体质量指数显著相关(P＜0.05或P＜0.01),而在体位组,两者无相关关系(P＞0.05).结论约1/3的OSA患者可归类为体位型,这类患者应对睡眠姿势训练治疗反应良好,可以保留足够的上气道通畅性,减轻睡眠呼吸阻塞程度,改善患者的呼吸功能.     

Successful treatment of obstructive sleep apnea with use of nasal continuous positive airway pressure in three patients with mucosal hemangiomas of the oral cavity

Cysts and benign tumors are uncommon causes of obstructive sleep apnea (OSA), and surgical removal is usually favored. In patients in whom an operation poses a high risk, however, nasal continuous positive airway pressure (CPAP) may prove beneficial. We describe three patients with hemangiomas of the oral cavity in whom polysomnography revealed moderate to severe OSA. In all three patients, nasal CPAP effectively decreased sleep-related disordered breathing events and dramatically improved their sleep. To our knowledge, this is the first report of OSA associated with hemangiomas involving the upper airway. Our experience suggests that nasal CPAP therapy is effective and well tolerated in such patients.     

Tracking Dynamic Tongue Motion in Ultrasound Images for Obstructive Sleep Apnea

Obstructive sleep apnea (OSA), a breathing disorder characterized by repetitive collapse of the pharyngeal airway during sleep, can cause intermittent hypoxemia and frequent arousal. The evaluation of dynamic tongue motion not only provides the biomechanics and pathophysiology for OSA diagnosis, but also helps doctors to determine treatment strategies for these patients with OSA. The purpose of this study was to develop and verify a dedicated tracking algorithm, called the modified optical flow (OF)-based method, for monitoring the dynamic motion of the tongue base in ultrasound image sequences derived from controls and patients with OSA. The performance of the proposed method was verified by phantom and synthetic data. A common tracking method, the normalized cross-correlation method, was included for comparison. The efficacy of the algorithms was evaluated by calculating the estimated displacement error. All results indicated that the modified OF-based method exhibited higher accuracy in verification experiments. In the human subject experiment, all participants performed the Müller maneuver (MM) to simulate the contour changes of the tongue base with a negative pharyngeal airway pressure in sleep apnea. Ultrasound image sequences of the tongue were obtained during 10 s of a transition from normal breathing to the MM, and these were measured using the modified OF-based method. The results indicated that the displacement of the tongue base during the MM was larger in the controls than in the patients with OSA (p < 0.05); the calculated areas of the tongue in the controls and patients with OSA were 24.9 ± 3.0 and 27.6 ± 3.3 cm², respectively, during normal breathing (p < 0.05), and 24.7 ± 3.6 and 27.3 ± 3.8 cm², respectively, at the end of the MM. The percentage changes in the tongue area were 2.2% and 1.3% in the controls and patients with OSA, respectively. We found that quantitative assessment of tongue motion by ultrasound imaging is suitable for evaluating pharyngeal airway behavior in OSA patients with minimal invasiveness and easy accessibility.     

Oximetry spectral analysis in the diagnosis of obstructive sleep apnoea.

Using spectral analysis of oximetry data, we prospectively evaluated the validity of this methodology in patients clinically suspected of suffering from obstructive sleep apnoea (OSA). A total of 233 outpatients were studied. Nocturnal oximetry was performed simultaneously with conventional polysomnography for all participants. The power density of oxygen saturation was analysed using Fast-Fourier transformation of the oximetric signal. Nocturnal oximetry test results were considered as abnormal (suspicion of OSA) if a peak in the spectrum between the period boundaries 30 and 70 s was observed. A normal test result was defined as the absence of the 30-70 s peak from the spectrum. Single-blind evaluation was performed by three independent observers, and agreement of two or more of these was considered definitive. The peak amplitude and the ratio of the area enclosed in the 30-70 s peak to the total area of the spectrum (r(S)) were measured. The presence of a peak has a sensitivity of 78%, a specificity of 89%, a positive predictive value of 89% and a negative predictive value of 78%. Apnoea-hypopnoea indexes were correlated significantly with peak amplitude (r=0.74; P<0.001) and with r(S) (r=0.69; P<0.001). For a peak amplitude threshold of 0.7%(2), the sensitivity was 94% and the specificity was 65% for OSA diagnosis. Using a threshold for r(S) of 0.15, the sensitivity was 91% and the specificity was 67%. Thus the spectral analysis of nocturnal oximetry and identification of a peak at 30-70 s could be useful as a diagnostic technique for OSA subjects.