The Evolving Landscape of Sleep Disorder Diagnosis and Treatment: Integrating AI and Personalized Medicine

Abstract

Sleep disorders represent a significant global health concern, impacting not only individual well-being but also societal productivity and healthcare costs. Characterized by disturbances in sleep patterns, duration, and quality, these disorders encompass a broad spectrum of conditions, ranging from insomnia and obstructive sleep apnea (OSA) to more complex parasomnias and circadian rhythm disorders. Traditional diagnostic approaches, often relying on polysomnography (PSG) and subjective patient reports, present limitations in accessibility, cost-effectiveness, and ecological validity. This research report delves into the current state of sleep disorder diagnosis and treatment, highlighting the challenges and opportunities associated with current methodologies. It explores the emerging role of artificial intelligence (AI) and machine learning (ML) in revolutionizing sleep medicine, focusing on the potential of AI-driven tools like PFTSleep (a hypothetical AI sleep solution) to enhance diagnostic accuracy, personalize treatment strategies, and improve patient outcomes. Furthermore, this report addresses the ethical considerations and future research directions necessary to ensure the responsible and effective integration of AI into the clinical management of sleep disorders, alongside the growing potential of personalized medicine based on genetic and phenotypic data.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

1. Introduction

Sleep is a fundamental biological need, essential for physical and cognitive restoration, immune function, and overall health. Chronic sleep deprivation and untreated sleep disorders are associated with a myriad of adverse health outcomes, including cardiovascular disease, metabolic dysfunction, cognitive impairment, mood disorders, and increased mortality risk [1, 2]. The prevalence of sleep disorders is substantial, with estimates suggesting that a significant proportion of the adult population experiences chronic insomnia symptoms [3] and a substantial number suffer from OSA [4]. These disorders impose a considerable economic burden on healthcare systems, primarily due to increased healthcare utilization, reduced productivity, and accident risk [5].

Traditional approaches to sleep disorder diagnosis and treatment face several challenges. Polysomnography (PSG), the gold standard for sleep assessment, is resource-intensive, time-consuming, and often conducted in specialized sleep laboratories, limiting accessibility for many patients [6]. Moreover, PSG provides a snapshot of sleep patterns on a single night, which may not accurately reflect an individual’s typical sleep behavior. Subjective assessments, such as sleep diaries and questionnaires, are prone to recall bias and may not capture the full complexity of sleep disturbances. Current treatment options, including cognitive behavioral therapy for insomnia (CBT-I), continuous positive airway pressure (CPAP) therapy for OSA, and pharmacological interventions, are not universally effective and may be associated with side effects or adherence issues [7].

This report aims to provide a comprehensive overview of the current state of sleep disorder diagnosis and treatment, highlighting the limitations of existing methods and exploring the potential of AI and personalized medicine to address these challenges. We will discuss the different types of sleep disorders, their prevalence, current diagnostic methods, and the potential impact of AI-driven tools like PFTSleep on improving diagnosis, treatment options, and patient outcomes. Furthermore, we will delve into the ethical considerations and future research directions necessary to ensure the responsible and effective integration of AI and precision medicine into the clinical management of sleep disorders.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

2. Overview of Sleep Disorders

Sleep disorders are classified into several categories, each with distinct etiologies, clinical presentations, and diagnostic criteria. The International Classification of Sleep Disorders (ICSD) is the most widely used diagnostic manual, providing standardized definitions and criteria for over 80 different sleep disorders [8]. Key categories of sleep disorders include:

• Insomnia Disorders: Characterized by difficulty initiating or maintaining sleep, or experiencing non-restorative sleep, despite adequate opportunity for sleep [9]. Insomnia can be acute or chronic and is often associated with daytime fatigue, impaired cognitive function, and mood disturbances.
• Sleep-Related Breathing Disorders: Primarily represented by Obstructive Sleep Apnea (OSA), characterized by repetitive episodes of upper airway collapse during sleep, leading to intermittent hypoxia and sleep fragmentation [10]. Central Sleep Apnea (CSA) is another breathing disorder involving a lack of respiratory effort during sleep.
• Central Disorders of Hypersomnolence: Including narcolepsy, idiopathic hypersomnia, and Kleine-Levin syndrome, characterized by excessive daytime sleepiness and difficulty staying awake [11]. Narcolepsy is often associated with cataplexy, a sudden loss of muscle tone triggered by strong emotions.
• Circadian Rhythm Sleep-Wake Disorders: Resulting from misalignment between the individual’s internal biological clock and the external environment [12]. Examples include shift work disorder, jet lag, and delayed sleep phase disorder.
• Parasomnias: Characterized by abnormal behaviors or physiological events that occur during sleep [13]. Examples include sleepwalking, sleep terrors, REM sleep behavior disorder (RBD), and sleep paralysis.
• Sleep-Related Movement Disorders: Characterized by repetitive movements that occur during sleep or wakefulness. Restless legs syndrome (RLS) and periodic limb movement disorder (PLMD) are the most common examples [14].

The prevalence of specific sleep disorders varies depending on the population studied, diagnostic criteria used, and data collection methods employed. However, studies consistently demonstrate that insomnia and OSA are highly prevalent, affecting a substantial proportion of the adult population [3, 4]. The prevalence of other sleep disorders, such as narcolepsy and parasomnias, is generally lower but still clinically significant.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

3. Current Diagnostic Methods for Sleep Disorders

The diagnosis of sleep disorders typically involves a combination of clinical history, physical examination, subjective assessments, and objective sleep studies. Key diagnostic methods include:

• Clinical History and Physical Examination: A thorough clinical history should include detailed information about the patient’s sleep patterns, daytime symptoms, medical and psychiatric history, medication use, and lifestyle factors. A physical examination may reveal signs suggestive of specific sleep disorders, such as obesity, enlarged tonsils, or craniofacial abnormalities in patients with OSA.
• Subjective Assessments: Sleep diaries and questionnaires are commonly used to assess sleep patterns, sleep quality, and daytime symptoms. The Pittsburgh Sleep Quality Index (PSQI) [15], Epworth Sleepiness Scale (ESS) [16], and Insomnia Severity Index (ISI) [17] are examples of standardized questionnaires used in clinical practice and research.
• Polysomnography (PSG): PSG is the gold standard for objective sleep assessment. It involves the continuous monitoring of multiple physiological parameters during sleep, including brain activity (electroencephalography, EEG), eye movements (electrooculography, EOG), muscle activity (electromyography, EMG), heart rate (electrocardiography, ECG), respiratory effort, and blood oxygen saturation. PSG allows for the identification of sleep stages, sleep architecture, and the detection of sleep-related events, such as apneas, hypopneas, and leg movements.
• Home Sleep Apnea Testing (HSAT): HSAT is a simplified version of PSG that can be performed in the patient’s home. HSAT typically involves monitoring respiratory effort, airflow, and blood oxygen saturation. HSAT is primarily used for the diagnosis of OSA in patients with a high pre-test probability of the disorder [18].
• Actigraphy: Actigraphy involves the use of a wrist-worn device that measures activity levels. Actigraphy can be used to assess sleep patterns, circadian rhythms, and sleep duration over extended periods. It is particularly useful for evaluating patients with circadian rhythm disorders and for monitoring treatment response.
• Multiple Sleep Latency Test (MSLT): The MSLT is a daytime nap study used to assess daytime sleepiness. It involves a series of scheduled naps, typically spaced two hours apart, and measures the time it takes for the patient to fall asleep (sleep latency) and the occurrence of REM sleep during the naps. The MSLT is used to diagnose narcolepsy and idiopathic hypersomnia [19].
• Maintenance of Wakefulness Test (MWT): The MWT is a daytime test used to assess the ability to stay awake in a quiet, stimulus-free environment. It is used to evaluate the effectiveness of treatment for sleep disorders and to assess fitness for duty in occupations that require sustained wakefulness.

Despite their utility, current diagnostic methods have limitations. PSG is expensive, time-consuming, and requires specialized expertise. HSAT is less comprehensive than PSG and may not be suitable for all patients. Subjective assessments are prone to recall bias and may not accurately reflect objective sleep measures. These limitations highlight the need for more accessible, cost-effective, and accurate diagnostic tools for sleep disorders.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

4. The Potential of AI in Sleep Disorder Diagnosis and Treatment

The increasing availability of large datasets, coupled with advancements in AI and machine learning (ML) algorithms, has created new opportunities for revolutionizing sleep disorder diagnosis and treatment. AI-driven tools have the potential to:

• Enhance Diagnostic Accuracy: AI algorithms can analyze complex sleep data, such as EEG signals, actigraphy data, and patient-reported symptoms, to identify patterns and features that are indicative of specific sleep disorders. ML models can be trained to differentiate between different sleep disorders with high accuracy, potentially improving diagnostic precision and reducing the need for manual interpretation of sleep studies [20]. For example, deep learning models have shown promise in automatically scoring sleep stages from EEG data [21], and in identifying apnea events from respiratory signals [22].
• Improve Accessibility and Cost-Effectiveness: AI-powered diagnostic tools can be deployed on mobile devices and wearable sensors, allowing for remote sleep monitoring and assessment. This can improve accessibility to sleep medicine services, particularly in underserved areas, and reduce the cost of traditional sleep studies. AI can also automate certain aspects of the diagnostic process, such as sleep stage scoring and apnea detection, reducing the workload on sleep technologists and clinicians.
• Personalize Treatment Strategies: AI algorithms can be used to predict treatment response based on individual patient characteristics, such as age, gender, disease severity, and genetic factors. This can enable clinicians to tailor treatment strategies to the specific needs of each patient, improving treatment efficacy and reducing the risk of adverse effects. For example, ML models can be trained to predict CPAP adherence in patients with OSA [23], and to identify patients who are more likely to benefit from CBT-I [24].
• Facilitate Early Detection and Prevention: AI can be used to identify individuals at high risk of developing sleep disorders based on their medical history, lifestyle factors, and genetic predisposition. This can enable early intervention and preventive strategies, such as lifestyle modifications and behavioral interventions, to reduce the incidence of sleep disorders. For example, AI can be used to predict the risk of developing OSA based on demographic and anthropometric data [25].

PFTSleep: A Hypothetical AI-Driven Sleep Solution

To illustrate the potential of AI in sleep medicine, consider PFTSleep, a hypothetical AI-driven sleep solution. PFTSleep is a comprehensive platform that integrates multiple AI-powered tools to enhance sleep disorder diagnosis and treatment. Key features of PFTSleep include:

• AI-Powered Sleep Analysis: PFTSleep utilizes advanced machine learning algorithms to analyze sleep data from various sources, including wearable sensors, smartphone apps, and PSG recordings. The platform can automatically score sleep stages, detect sleep-related events, and generate comprehensive sleep reports.
• Personalized Treatment Recommendations: Based on the patient’s sleep data, medical history, and genetic information, PFTSleep provides personalized treatment recommendations, including CBT-I, CPAP therapy, and medication management. The platform also offers virtual coaching and support to help patients adhere to their treatment plans.
• Remote Sleep Monitoring: PFTSleep allows for remote sleep monitoring using wearable sensors and smartphone apps. This enables clinicians to track patient progress, identify potential problems, and adjust treatment plans as needed. The platform also provides patients with personalized feedback and insights into their sleep patterns.
• Predictive Analytics: PFTSleep utilizes predictive analytics to identify individuals at high risk of developing sleep disorders and to predict treatment response. This enables early intervention and preventive strategies to improve patient outcomes.

While PFTSleep is a hypothetical example, it illustrates the potential of AI to transform sleep medicine. By integrating multiple AI-powered tools into a comprehensive platform, clinicians can improve diagnostic accuracy, personalize treatment strategies, and improve patient outcomes.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

5. Ethical Considerations and Challenges of AI in Sleep Medicine

While AI holds tremendous promise for improving sleep disorder diagnosis and treatment, it is essential to address the ethical considerations and challenges associated with its implementation. Key ethical considerations include:

• Data Privacy and Security: AI algorithms rely on large datasets of patient information. It is crucial to ensure that patient data is protected from unauthorized access and misuse. Data privacy and security protocols must be implemented to comply with relevant regulations, such as HIPAA and GDPR [26].
• Algorithmic Bias: AI algorithms can perpetuate and amplify existing biases in the data they are trained on. It is important to ensure that AI algorithms are trained on diverse datasets that accurately represent the population they will be used on. Algorithmic bias can lead to inaccurate diagnoses and unfair treatment recommendations for certain patient groups [27].
• Transparency and Explainability: AI algorithms can be complex and opaque, making it difficult to understand how they arrive at their decisions. It is important to develop AI algorithms that are transparent and explainable, so that clinicians and patients can understand the rationale behind the AI’s recommendations. This is particularly important in high-stakes situations, such as treatment decisions [28].
• Clinical Validation and Regulation: AI-driven diagnostic and treatment tools must be rigorously validated in clinical trials before they are deployed in clinical practice. Regulatory agencies, such as the FDA, must establish clear guidelines for the development and approval of AI-based medical devices [29].
• Human Oversight: AI should be used to augment, not replace, human clinicians. Clinicians should retain ultimate responsibility for patient care and should use their clinical judgment to interpret AI’s recommendations. AI should be viewed as a tool to assist clinicians in making better decisions, not as a replacement for clinical expertise [30].

In addition to ethical considerations, there are also technical challenges associated with the implementation of AI in sleep medicine. These challenges include:

• Data Quality and Standardization: AI algorithms require high-quality, standardized data to function effectively. Sleep data can be noisy and variable, making it difficult to train accurate AI models. Efforts are needed to improve data quality and standardization across different sleep laboratories and data sources.
• Integration with Existing Systems: AI-driven tools must be seamlessly integrated with existing electronic health record (EHR) systems and clinical workflows. This requires interoperability standards and collaboration between AI developers and healthcare providers.
• User Acceptance: Clinicians and patients must be comfortable using AI-driven tools in clinical practice. This requires training and education to ensure that users understand how the tools work and how to interpret their results. Addressing clinician and patient concerns about AI is essential for successful adoption.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

6. The Future of Sleep Medicine: Personalized Medicine and Beyond

The future of sleep medicine is likely to be shaped by the integration of AI with personalized medicine approaches. Personalized medicine aims to tailor treatment strategies to the individual patient based on their unique genetic, phenotypic, and environmental characteristics [31]. Key areas of personalized medicine in sleep medicine include:

• Genetic Predisposition: Identifying genetic variants that are associated with an increased risk of developing sleep disorders. This can enable early detection and preventive strategies for individuals at high risk. For example, studies have identified genetic variants associated with restless legs syndrome [32] and narcolepsy [33]. Further research is needed to identify genetic factors that contribute to other sleep disorders, such as insomnia and OSA.
• Biomarkers: Identifying biomarkers that can be used to predict treatment response. This can enable clinicians to select the most effective treatment for each patient. For example, studies have identified biomarkers that predict CPAP adherence in patients with OSA [34]. Further research is needed to identify biomarkers that predict treatment response for other sleep disorders, such as insomnia and circadian rhythm disorders.
• Phenotyping: Characterizing sleep disorders based on their underlying physiological mechanisms. This can enable the development of targeted therapies that address the specific underlying causes of the disorder. For example, studies have identified different phenotypes of insomnia based on EEG activity and cognitive function [35]. Further research is needed to identify phenotypes of other sleep disorders, such as OSA and narcolepsy.
• Chronotherapy: Tailoring treatment strategies to the individual’s circadian rhythm. This can improve treatment efficacy and reduce side effects. For example, studies have shown that the timing of light exposure can affect sleep quality and circadian rhythms [36]. Further research is needed to develop chronotherapeutic interventions for other sleep disorders, such as shift work disorder and delayed sleep phase disorder.

The integration of AI with personalized medicine approaches has the potential to revolutionize sleep disorder diagnosis and treatment. AI can be used to analyze large datasets of genetic, phenotypic, and environmental data to identify patterns and features that are indicative of specific sleep disorders and to predict treatment response. This can enable clinicians to tailor treatment strategies to the specific needs of each patient, improving treatment efficacy and reducing the risk of adverse effects.

Furthermore, the future may see the development of closed-loop systems that dynamically adjust treatment parameters based on real-time feedback from wearable sensors and AI algorithms. Imagine a CPAP machine that automatically adjusts pressure settings based on the patient’s respiratory patterns and sleep stage, or a sleep tracking app that delivers personalized CBT-I interventions based on the individual’s sleep diary data and AI-driven analysis of their sleep patterns.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

7. Conclusion

Sleep disorders are a significant global health concern, impacting individual well-being, societal productivity, and healthcare costs. Current diagnostic and treatment methods have limitations in accessibility, cost-effectiveness, and personalization. AI and personalized medicine offer tremendous potential to address these challenges and revolutionize sleep medicine.

AI-driven tools can enhance diagnostic accuracy, improve accessibility and cost-effectiveness, personalize treatment strategies, and facilitate early detection and prevention of sleep disorders. However, it is essential to address the ethical considerations and challenges associated with the implementation of AI in sleep medicine, including data privacy and security, algorithmic bias, transparency and explainability, clinical validation and regulation, and human oversight.

The future of sleep medicine is likely to be shaped by the integration of AI with personalized medicine approaches. By tailoring treatment strategies to the individual patient based on their unique genetic, phenotypic, and environmental characteristics, we can improve treatment efficacy and reduce the risk of adverse effects. Further research is needed to identify genetic variants, biomarkers, and phenotypes associated with sleep disorders and to develop targeted therapies that address the specific underlying causes of the disorder. The convergence of AI and personalized medicine promises to usher in a new era of precision sleep medicine, ultimately leading to improved sleep health and overall well-being for individuals worldwide.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

References

[1] Grandner MA, Hale L, Moore M, Patel NP. Mortality associated with short sleep duration: the evidence, the possible mechanisms, and the future. Sleep Med Rev. 2010;14(3):191-203.
[2] Irwin MR. Why sleep is important for health: a psychoneuroimmunology perspective. Annu Rev Psychol. 2015;66:143-172.
[3] Ohayon MM. Epidemiology of insomnia: what do we know? Sleep Med Rev. 2002;6(2):97-111.
[4] Peppard PE, Young T, Barnet JH, Palta M, Hagen EW, Hla KM. Increased prevalence of sleep-disordered breathing in adults. Am J Epidemiol. 2013;177(9):1006-1014.
[5] Hafner M, Stepanek M, Taylor J, Troxel WM, van Stolk C. The economic burden of insomnia: direct and indirect costs for individuals with insomnia, comorbid diseases, and sleep-related accidents. J Clin Psychiatry. 2017;78(9):1268-1276.
[6] Iber C, Ancoli-Israel S, Chesson A, Quan SF. The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications. Westchester, IL: American Academy of Sleep Medicine; 2007.
[7] Qaseem A, Kansagara D, Forciea MA, et al. Management of Chronic Insomnia Disorder in Adults: A Clinical Practice Guideline From the American College of Physicians. Ann Intern Med. 2016;165(2):125-133.
[8] American Academy of Sleep Medicine. International Classification of Sleep Disorders. 3rd ed. Darien, IL: American Academy of Sleep Medicine; 2014.
[9] Morin CM, Belleville G, Bélanger L, Ivers H. The Insomnia Severity Index: psychometric indicators to detect insomnia cases and evaluate treatment response. Sleep. 2011;34(5):601-608.
[10] Kapur VK, Auckley D, Chowdhuri S, et al. Clinical practice guideline for diagnostic testing for adult obstructive sleep apnea: an American Academy of Sleep Medicine clinical practice guideline. J Clin Sleep Med. 2017;13(3):479-504.
[11] Bassetti CLA, Adamantidis A, Burdakov D, et al. Narcolepsy—clinical spectrum, aetiopathology, diagnosis and management. Nat Rev Neurol. 2019;15(9):519-539.
[12] Auger RR, Burgess HJ, Emens JS, Deriy LV, Thomas SM, Zee PC. Clinical Practice Guideline for the Treatment of Intrinsic Circadian Rhythm Sleep-Wake Disorders: An American Academy of Sleep Medicine Clinical Practice Guideline. J Clin Sleep Med. 2015;11(10):1199-1236.
[13] Howell MJ. Parasomnias: an updated review. Neurotherapeutics. 2015;12(4):796-813.
[14] Allen RP, Picchietti DL, Garcia-Borreguero D, et al. Restless legs syndrome/Willis-Ekbom disease diagnostic criteria: updated International Restless Legs Syndrome Study Group (IRLSSG) consensus criteria–history, rationale, description, and significance. Sleep Med. 2014;15(6):644-654.
[15] Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28(2):193-213.
[16] Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14(6):540-545.
[17] Morin CM. Insomnia: Psychological Assessment and Management. New York: Guilford Press; 1993.
[18] Kuna ST, Gurubhagavatula I, Maislin G, et al. Objective measurement of sleep apnea severity with a portable diagnostic device. Am J Respir Crit Care Med. 2008;177(10):1147-1155.
[19] Littner MR, Kushida CA, Wise M, Davila D, Morgenthaler T, Lee-Chiong T, Hirshkowitz M, Daniel C, American Academy of Sleep Medicine. Practice parameters for clinical use of the multiple sleep latency test and the maintenance of wakefulness test. Sleep. 2005;28(1):113-121.
[20] Radha, M., & Raghavan, B. (2021). Artificial intelligence in sleep medicine: A review. Sleep Medicine Reviews, 57, 101460.
[21] Chambon, S., Galtier, M. N., Arnal, P. J., Senoussi, M., Perucca, L., & Saporta, L. (2018). A deep learning architecture for temporal sleep stage classification using multivariate and multimodal time series. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 26(4), 758-769.
[22] Biswas, D., Zhang, X., & Mannino, D. M. (2019). Automatic detection of sleep apnea events from single-lead ECG using deep learning techniques. Physiological Measurement, 40(3), 035006.
[23] Yaffe, K., Laffont, B., Roche, N., Goupil, F., Tamisier, R., Borel, A. L., … & Pépin, J. L. (2020). Prediction of adherence to continuous positive airway pressure in patients with obstructive sleep apnea. Journal of Clinical Sleep Medicine, 16(8), 1285-1292.
[24] Zachariae, R., Lyby, K., Ritterband, L. M., O’Toole, M. S., & Christensen, M. G. (2016). Predictors of outcome following internet-delivered cognitive behavior therapy for insomnia: A systematic review. Sleep Medicine Reviews, 30, 51-61.
[25] Mesarwi, O. A., Polotsky, V. Y., & Pack, A. I. (2019). Sleep apnea: A common and costly disorder. American Journal of Managed Care, 25(3), S69-S74.
[26] Price, W. N., Cohen, I. G., & Reif, J. H. (2015). Big data, big liabilities. Science, 349(6247), 509-510.
[27] Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019). Dissecting racial bias in an algorithm used to manage the health of populations. Science, 366(6464), 447-453.
[28] London, A. J. (2019). Artificial intelligence and black-box medical decisions: Accuracy versus explainability. Hastings Center Report, 49(1), 15-21.
[29] Benjamens, S., Dhunnoo, P., & Meskó, B. (2020). The state of artificial intelligence-based FDA-approved medical devices and algorithms: an updated review. NPJ digital medicine, 3(1), 1-8.
[30] Wachter, S., Mittelstadt, B., & Russell, C. (2018). Counterfactual explanations without opening the black box: Automated decisions and the GDPR. Harvard Journal of Law & Technology, 31(2), 841-887.
[31] Hamburg, M. A., & Collins, F. S. (2010). The path to personalized medicine. New England Journal of Medicine, 363(4), 301-304.
[32] Winkelmann, J., Lichtner, P., Schormair, B., et al. Genome-wide association study identifies novel restless legs syndrome susceptibility loci on 2p14 and 16q12. 1. PLoS genetics. 2011;7(2):e1001426.
[33] Kornum, B. R., Gammeltoft, S., Knudsen, S., Jennum, P., & Rye, D. (2017). The neurobiology of narcolepsy. Journal of Sleep Research, 26(5), 547-565.
[34] Bakker, J. P., Veenhuis, J. P., van Dijk, J. P., Kerkhof, G. A., de Vries, G. E., & Renken, R. J. (2013). Predicting CPAP adherence in obstructive sleep apnea syndrome: a machine learning approach. Sleep Medicine, 14(11), 1145-1152.
[35] Plante, D. T., Plante, A. S., Demers, M., & Landry, S. A. (2011). EEG spectral and nonlinear dynamics differentiate good sleepers from those with chronic primary insomnia. Journal of Sleep Research, 20(3), 371-382.
[36] Dijk, D. J., & Czeisler, C. A. (1995). Contribution of the circadian pacemaker and the sleep homeostat to sleep propensity, sleep structure, electroencephalographic slow waves, and sleep spindle activity in humans. Journal of Neuroscience, 15(5 Pt 1), 3526-3538.