Research Article

Wireless Point-Of-Care Health Monitoring Systems for Sleep Disorders

Se Chang Oh1,*, Vijay K. Varadan1,2,3, Robert E. Harbaugh 2,3
1Department of Electrical Engineering, University of Arkansas, Fayetteville, AR 72701, USA
2Department of Engineering Science and Mechanics, Pennsylvania State University, University Park, PA 16802, USA
3Department of Neurosurgery, Pennsylvania State University, Hershey, PA 17033, USA
*Corresponding author:

Se Chang Oh, Department of Electrical Engineering, University of Arkansas, Fayetteville, AR 72701, USA, Email:


Sleep disorder, Biopotentials, Polysomngraphy, Wireless sensor network, ambulatory monitoring, Power consumption

More than 40 million people in the United States have been suffering from chronic sleep disorders. More than 50 % of them remain undiagnosed and untreated, and it results in spending over $22 billion in unnecessary healthcare costs because of undiagnosed and untreated sleep disorders. Therefore, there is a need for introduction of remote patient monitoring/diagnosing systems which provide low cost and reliable monitor/diagnosis without hospitalization. A remote point-of-care system is one of the solutions by enabling to monitor the patient's health condition at local or remote place and providing real-time feedback information from a medical center. Recent progress in mobile platforms has resulted in novel wearable health monitoring systems for sleep disorders. This paper describes characteristics and measurement of the biopotentials (EEG, ECG, EOG, and EMG) for sleep disorder diagnosis and discusses basic platforms in remote health monitoring systems. In this paper, several remote POC systems are introduced to diagnose sleep disorders. One of the remote point-of-care systems used both ZigBee and Wi-Fi to provide flexibility of the number of physiological channels with ZigBee network topology and support network extension from WPAN to WLAN by adopting Wi-Fi. The other point-of-care system removes the environmental restriction in term of wireless coverage by using GSM/WCDMA network. It provides remote real-time monitoring in practical. In addition to introducing system configuration of these systems, the algorithm which reduces power consumption of wireless systems in WPAN by adjusting RF output power based on RSSI.

Rapid economic growth in this decade improves the quality of life, but people have been getting stresses to survive in a competitive society. Most of the people may have experienced sleep disturbance due to stresses. Sleep disorders or sleep deprivation itself may be considered as an insignificant disease. However, cumulative-long term effects of sleep disorders are associated with serious health consequences and they lead to chronic diseases [1]. The healthcare cost increased by 94% from $1, 337 billion in 2000 to $2,594 billion in 2010. They projected the cost will be $4,487 billion in 2020 [2]. This exponential increase in the healthcare costs is an important fact to the states whose budgets constitute an average of 20% from Medicaid [3]. Therefore, the ever increasing costs affect health insurance, mental health problems, and other healthcare problems [4-9]. The potential methods to control the healthcare costs include improved prevention, early detection/identification of the causes of chronic diseases. Improved prevention methods can be introduced by educating people about healthy lifestyle, origin of various chronic diseases and their ill effects. Because preventive effort and improving care for a smaller group of people will be more effective in terms of quality than implementing it on a larger population. Early detection/identification of the causes of chronic diseases leads to a substantial reduction in the cost of treatments. Constant monitoring is one of the significant factors in early detection of chronic diseases, but repetitive visits to the hospital may involve large expenses. As a solution for this, remote Point-of-Care (POC) systems can be introduced. The remote POC systems allow monitoring the physiological signals of the patient and diagnosing the diseases with his/her daily routine instead of visiting a hospital or a physician. Cost of hospitalization and lab tests can be reduced by implementing POC. Identification of causes leading to chronic diseases can also reduce the cost of treatment to a great extent. One of the causes of chronic diseases is sleep disorders. Undiagnosed and untreated sleep disorders can be associated with cardiac problems, hypertension, stroke, and memory loss, to name a few [10, 11].  Various studies have shown the prevalence of the sleep disorders with chronic diseases. The sleep disorders are associated with heart failure by 60% [12], stroke/transient ischemic attack (TIA) by 70% [13, 14], type 2 diabetes by 65 % [15], and hypertension by 37% [16].

2.1 Types of Sleep Disorders

Sleep disorders are related to sleep patterns and characterized by disturbance in the amount, quality or timing of sleep. According to the third edition of the International Classification of Sleep Disorders (ICSD-3), the sleep disorders are classified as 7 categories; Insomnia, Sleep-related breathing disorders, Central disorders of hypersomnolence, Circadian rhythm sleep-wake disorders, Parasomnias, Sleep-related movement disorders, Other sleep disorders [17].

Insomnia - American Academy of Sleep Medicine defines insomnia as "a persistent difficulty with sleep initiation, duration, consolidation, or quality that occurs despite adequate opportunity and circumstances for sleep, and results in some form of daytime impairment". Based on the definition, the criteria for diagnosis of insomnia include problems of initiation or persistence of sleep, adequate opportunity and circumstances for sleep, daytime dysfunction, duration and frequency of insomnia.

Sleep-related breathing disorders - Sleep-related breathing disorders are related to the disturbance of respiration during sleep. The disorders are further categorized as four sections, obstructive sleep apnea (OSA), central sleep apnea (CSA), sleep-related hypoventilation disorders, and sleep-related hypoxemia disorder. Obstructive sleep apnea is caused by the block of the upper airway. Central sleep apnea is related to neurological function. The abnormalities of neurological function do not control contraction and relaxation of muscles for respiration. Sleep-related hypoventilation disorders are caused by elevation of arterial carbon dioxide partial pressure. Obesity hypoventilation syndrome is a distinct hypoventilation diagnosis. Sleep-related hypoxemia is caused by the reduction of oxyhaemoglobin saturation.

Central disorders of hypersomnolence - The characterization of these disorders is daytime sleepiness which is not caused by other disorders. One of the causes is that central nervous system does not control the sleep and wake events properly. ICSD-3 defines excessive sleepiness as "daily episodes of an irrepressible need to sleep or daytime lapses into sleep". Central disorders of hypersomnolence are further classified to 7 sections, Narcolepsy type 1, Narcolepsy type 2, Idiopathic hypersomnia, Kleine-Levin syndrome, Hypersomnia due to medical disorder, Hypersomnia due to medication or substance, Hypersomnia associated with a psychiatric disorder, and Insufficient sleep syndrome.

Circadian rhythm sleep-wake disorders - Circadian rhythms are regulated genetically by circadian regulatory system, but changes in the endogenous circadian system or through environmental and behavioral manipulations can impair the regulation of circadian rhythms. The diagnosis of these disorders uses actigraphy and biomarkers. In addition, morningness-eveningness questionnaires can be used to identify chronotype. The disorders are further classified to delayed sleep-wake phase disorder, advanced sleep wake phase disorder, irregular sleep-wake rhythm disorder, non-24-h sleep-wake rhythm disorder, shift work disorder, jet lag disorder, and circadian sleep-wake disorder not otherwise specified (NOS).

Parasomnias - Parasomnias are another group of sleep disorders which are related to the undesirable physical phenomena that occur at sleep onset, during sleep (NREM, REM, or transitions between sleep stages), or during arousal from sleep. The concept of parasomnias is the results of dissociation of sleep stages. The categories of parasomnias are NREM-related parasomnias, REM-related parasomnias, and other parasomnias. REM-related parasomnias require video polysomnography (PSG) for diagnostic criteria among the types of parasomnias.

Sleep-related movement disorders - Sleep-related movement disorders are related to the simple, often stereotyped movement occurring during sleep. Periodic limb movements (PLMs) can occur during sleep or wakefulness and diagnosed as sleep-related movement disorders although PLMs are not accompanied by any complaint or sleep disturbance. Sleep-related movement disorders are categorized as restless leg syndrome, periodic limb movement disorder, sleep-related leg cramps, sleep-related bruxism, sleep-related rhythmic movement disorder, benign sleep myoclonus of infancy, propriospinal myoclonus at sleep onset, sleep-related movement disorder due to a medical disorder, sleep-related movement disorder due to a medication or substance, and sleep-related movement disorder, unspecified.

Other sleep disorders - Sleep disorders which cannot be classified according to the International Classification of Sleep Disorders are defined as other sleep disorders.

2.2 Measurement Systems for Diagnosis of Sleep Disorders

Accurate diagnosis of sleep disorders is imperative to the treatment plan and early recovery. There are four types of sleep study devices according to the Center for Medicare & Medicaid Services (CMS) and the American Academy of Sleep Medicine (AASM). Table 1 shows the required physiological signals for each device type. Type I test is a gold standard method for diagnosing sleep disorders and it measures almost all kinds of physiological signals with around 20 sensors and is performed by a sleep technologist in a sleep lab. For this reason, type I is referred to as a full sleep study and attended study. An In-lab test is when a patient will sleep at the sleep laboratory and a sleep technologist monitors vitals and observes through a video camera in a control room. However, it is expensive, inconvenient, time consuming and labor intensive. Home sleep test (HST) is an unattended sleep study that is performed at home without the oversight of a sleep technologist. It captures only what is necessary for the diagnosis.



Mandatory Channels

Type I

in-lab test,

attended studies with full sleep staging

 Type II + Chin/Limb EMG

Additional channels for CPAP/BiPAp levels, CO2, pH, pressure, etc

Type II

home sleep test

unattended studies, minimum of 7 channels


Type III

home sleep test

unattended studies, minimum of 4 channels

respiratory movement/airflow

ECG/heart rate, oxygen saturation

Type IV

home sleep test

unattended studies, minimum of 3 channels

oxygen saturation, airflow

Table 1. Definitions of the types of sleep studies devices

In 1937, Loomis, Harvey and Hobart firstly described 5 different sleep stages by classifying NREM sleep. Their notes were modified by Dement and Kleitman in 1957. They identified REM sleep with rapid eye movements, and low amplitude and fast EEG waveforms. They also found cyclical sleep patterns between NREM and REM sleep during sleep. They announced the revised sleep stages including REM sleep. In 1968, Allan Rechtschaffen and Anthony Kales completed the criteria for scoring sleep stages and published manual by summarizing the rules and the manual has been used for scoring sleep stages ever since.

3.1 Electrodes Placement for Scoring Sleep Stages for R&K Sleep Scoring

EEG, EOG, and EMG are usually recorded to score the sleep stages. EEG patterns show different regional information according to the placement of electrodes over the skull. However, regional information is not critical in scoring the sleep stages if the key features, alpha, sleep spindles, K complexes, vertex sharp waves and delta waves, are adequately registered. It is necessary to specify the electrode placement to record the optimal EEG patterns to identify the sleep stages. The C3 or C4 placement is preferred because the sleep spindles, K complexes and vertex sharp waves are clearly monitored in this placement. As the placement of a reference electrode, ear lobe or mastoid placement is preferred to maximize inter-electrode distance to monitor the maximum or nearly maximum amplitude of high voltage slow waves. In addition to the C3 or C4 position, the O1 or O2 placement, the portion of the occipital lobe, is used for alpha waves activities to classify the accurate evaluation of sleep onset. EOG patterns for stage REM and wakefulness are measured with two channels. One electrode is placed a few centimeters above and slightly lateral to the outer canthus of one eye and the other electrode is placed a few centimeters below and slightly lateral to the outer canthus of the other eye. The reference electrode is placed on either the ear lobe or mastoid. These arrangements produce in-phase deflection in two EOG channels because the eye movements are binocularly synchronous during the sleep stages. The feature of in-phase deflection helps to discriminate artifacts which have out-of-phase deflection. The electrode placement for EMG is preferred on or beneath the chin to classify stage REM during sleep. Figure 1 shows the arrangement of the electrodes for the sleep stages scoring recommended by R&K manual.

Clyto Access
Figure 1. Placement of electrodes for EEG, EOG, and EMG recording

3.2 Scoring Criteria for Sleep Stages

Sleep stages are mainly classified as wakefulness, non-REM sleep, and REM sleep. Non-REM sleep is further into divided into four stages; stage 1, stage 2, stage 3 and stage 4. Table 2 summarized the sleep-wake scoring criteria according to R&K sleep scoring manual [18, 19].

Clyto Access
Table 2. Summary of Sleep-wake scoring criteria

Stage W (Wakefulness): Stage W means the waking state. The EEG in this stage shows alpha waves and/or low voltage (10-30 uV), mixed frequency EEG. Alpha waves occupy more than 50% of the epoch. In addition, this stage may show a relatively high tonic EMG.

Stage 1: Stage 1 is characterized by low voltage, mixed frequency EEG with a distinction in 2-7 Hz frequency range. The amplitude in 2-7 Hz is about 50-75 uV, and it can reach to 200 uV. Stage 1 does not show K complexes and sleep spindles. Alpha waves in the mixed frequency EEG occupy less than 50% of the epoch.

Stage 2: The main feature of Stage 2 is K complexes and/or sleep spindles. K complexes show a negative high peak, followed by a slower positive wave. Sleep spindles have 12-14 Hz frequency band and are composed of at least six consecutive waves or train duration longer than 0.5 s. The duration of K complexes and sleep spindles should be at least 0.5 sec and the time between K complexes and sleep spindles is less than 3 min to be scored as stage 2. This stage also shows relatively low voltage, mixed frequency EEG on background and waves with amplitude above 75 uV occupy less than 20 % of the epoch.

Stage 3: Stage 3 is characterized by the brain waves with 2 Hz or slower. The waves with amplitude above 75 uV occupy at least 20 % but not more than 50 % of the epoch in stage 3. K complexes and sleep spindles may or may not be present in stage 3.

Stage 4: Stage 4 shows the same waves as stage 3, with 2 Hz or slower waveforms. The classifying criteria between stage 3 and stage 4 is how much the waves with 2 Hz or slower with amplitude above 75 uV occupy over the epoch. The waves appear more than 50 % of the epoch.

Stage REM: Stage REM shows the similar wave patterns of stage 1, low voltage and mixed frequency EEG. The feature of stage REM is that sawtooth wave pattern is frequently present and vertex sharp waves are not distinctive. In addition, stage REM shows the episodic rapid eye movements in EOG.

Remote patient monitoring is a technology where the patient can be effectively monitored outside a hospital. The testing is performed at the patient's own level of comport and at no particular places such as a hospitals or a clinic. It is one of the recent advances in the telemedicine and a prospective cost reducing method in the field of healthcare. This method involves sending the patient’s medical data to a physician from the patients home/work place. Remote patient monitoring can be of much use to the people suffering from sleep disorder problems. The patients need not to be present in the hospital or the sleep lab, when they are required to be tested for a complete sleep cycle. Instead the patient can sleep in his/her home normally, where the recorded data are sent to the physician for diagnosis.

4.1 Configuration of the Remote Patient Monitoring System

In general, the remote point-of-care system includes a device for testing associated with the patient, a central server, and a monitoring device for a physician. Figure 2 shows the data flow of the remote point-of-care system. The device for testing the patient health condition consists of sensors, amplifiers/filters, a processing unit, graphic user interface, data storage and a communication module. The physiological signals are detected by the sensors. The signals are small in magnitude and contain undesired signals such as noises. The amplifier/filter unit modifies the raw signals to the appropriate level and quality of signals for data processing in the processor unit. The analog signals are converted to digital signals with an analog-to-digital converter. The sampling rate and resolution are decided according to the characteristic of the physiological signals. The processing unit manages the data flow in the peripheral units, a display, a communication, and a data storage unit. The display unit provides a graphic user interface through which the patient can manage the device and displays the sensed physiological signals. The data storage unit records the signals. The communication unit, either wired or wireless communication, sends the signals to the central sever which is in a service provider site or a medical center. The physician can monitor the signals in real time or diagnose disease with stored data by connecting to the central server.

Clyto Access
Figure 2. Data flow of the remote point-of-care system

4.2 Fundamental Platforms in Wireless Health Monitoring System

Research Institute - This section focuses on the fundamental platforms in the physiological signal acquisition system architectures developed by various research institutions around the world. Apart from the number of input channels and the type of physiological signals acquired, the general trends in wireless communication and electronics have evolved. The current trends include: cloud infrastructure, mobile applications, Bluetooth Low Energy, Server based analytics, Software as a Service (SaaS), Software as Medical Device (SaMD) and so on. However, the acquisition methods and physiological parameters serve as a fundamental platform upon which these feature additions and improvements become a viable value addition. Table 3 shows a comparison of different systems having a fundamental platform developed by research institutions. The systems have been developed with the concept of low power consumption and communication distance. The systems measure the vital signals, ECG, SpO2, temperature and so on. UCLA developed a system with 8 analog channels and ZigBee as the communication module [20]. It adopted on board signal processing, thereby reducing the power consumption comparatively. NASA developed a 10-channel, Bluetooth enabled device [21]. Kansas State University also developed a Bluetooth enabled system with 5 analog channels [22]. Harvard University developed a 2-channel device which used ZigBee for communication [23]. Duke University developed a 12-channel device which used 802.11b as its communication method, but had higher power consumption because of its high performance control unit and the Wi-Fi [24]. ZigBee and Bluetooth are commonly used, but Duke used Wi-Fi. While the systems using ZigBee and Bluetooth consume about 100 mW and 350 mW respectively, the system using Wi-Fi consumes more power, 4 W. Even though the power consumption shown in the Table 3 reflects the whole system, most of the power is consumed by RF communication. ZigBee consumes less power and Wi-Fi consumes more power among ZigBee, Bluetooth and Wi-Fi. They built wireless personnel area network (WPAN) or wireless body area network (WBAN) and it has limited coverage area as the transmission and reception is only possible between short distances. Therefore, effective remote patient monitoring is not feasible with these types of communication methodologies.

Institution UCLA NASA Kansas State Harvard Duke TU-Berlin (Germany)
# of input Channels 8 10 5 2 12 -
Physiological signals ECG
Blood Pressure
ECG Neural signals ECG
Wireless ZigBee BT BT ZigBee Wi-Fi ZigBee
Transmission range 9 m 10 m - 9 m 30 m -
Power supply 3V
(2x AA)
(2x AA)
(4x AA)
(2x AA)
3.7V (1200mAh)
Ploy lithium-ion
3.7V (850mAh)
Ploy lithium-ion
Power consumption 97 mW 360 mW 344 mW 100 mW 4W -
Table 3. Comparison of the wireless health monitoring systems from research institutes

Commercially available products for diagnosis of sleep disorders - Table 4 shows the features of the commercially available HST devices for sleep disorders. Most systems use external or internal memory device to record sleep data. In this case, real time monitoring is not feasible.

Patients should upload the recorded sleep data to the hospital server or deliver the memory card storing the data to their sleep lab the next day. Some devices use wireless communication to alleviate the inconveniences from wired connections. However, wireless communication is used for saving sleep data to a local storage devices instead of sending sleep data to the remote server or the central server because these devices build wireless personal area network or wireless body area network. The wireless HST devices use Bluetooth or their own wireless protocol in wireless personal area network. In addition, these devices have the fixed number of channels so that they cannot increase or decrease the number of physiological channels flexibly. Consequently, the devices cannot extend the number of channels to measure additional physiological signals. For instance, the type III device for diagnosis sleep apnea cannot be upgraded to the type II device for diagnosing sleep apnea and evaluating sleep stages by simply adding the physiological channels of EEG, EOG and EMG to the type III device. Patients should additionally purchase the type II device to diagnose both sleep apnea and sleep stages even though they have the type III device. As a result, having a fixed number of channels per type is not a cost-effective solution. In the following section, two home sleep test systems are introduced. Implementations of home sleep test systems are described in terms of wide are coverage and the cost effectiveness. First, the system using a combination of two communication methods, ZigBee and Wi-Fi, provides remote patient monitoring in real time and flexibility of the number of physiological channels. Second, the system using GSM/WCDMA communication provides seamless remote health monitoring due to wide area, continuous and transmission of data over long distances of the GSM/WCDMA.

Manufactures Product Name Total number of channels Storage Wireless
Philips /Respironics [25] Alice PDx Up to 21channel with optional ECG and ExG Yokes Channel 1GB SD card No
Embla [26] Embletta X100 12-channel with X100 proxy 128MB internal memory No
Compumedics [27] Somte PSG 25 available channels 2GB Compact Flash Bluetooth
Compumedics [28] Siesta 802 32-amplified channel Compact Flash Siesta's Ethernet radio link
Cleveland Medical [29] SleepScout 9-channel SD card 2.4-2.484 GHz
ResMed [30] ApneaLink Plus 4-channel 15MB internal memory No
BD Medical Technology [31] CareFusion Nox-T3 14-channel 1GB SD card No
Table 4. Features of the commercially available home sleep test devices

The first introduced system is designed with ZigBee star network and combination of two wireless communication standards, ZigBee and Wi-Fi. The system mainly consists of two wireless HST devices, a receiver and a monitoring unit. One of the HST devices has 5 input channels as a type III HST device which measures ECG, nasal airflow, chest/abdomen effort, body movement and oxygen saturation. The other HST device, as a supplementary device, has 5 channels - 2x EOG, 1x EMG and 2x EEG. The receiver unit adopts two wireless communication standards, Wi-Fi and ZigBee. The combination of two wireless standards in the receiver unit satisfies both, low power consumption and high data transfer rate of the overall system. Each HST device sends its measured signals to the receiver unit with a ZigBee star network. Each device is regarded as a standalone type III and supplementary HST device, but on the receiver unit side, the whole system is extended to a type II device by collecting and combining the signals from the type III device and the supplementary device. The sleep data in the receiver unit are retransmitted to the monitoring unit or the remote server through the local Wi-Fi network. Figure 3 shows the data flow of the system [44, 45].

Clyto Access
Figure 3. Data flow of the system to diagnose sleep disorders with ZigBee and Wi-Fi

The second introduced HST system is designed with a skull cap installing electronics/wires and GSM/WCDMA network to reduce a patient's discomfort and monitor physiological signals at a remote place. A skull cap with equipped electronics on the top senses biopotential signals and transmits them to a remote server through the mobile networks. The device consists of an amplifier module, a power management circuit, a processor module and a GSM/WCDMA module. 5 channels of biopotential signals, 2-channel of EOG to detect right and left eyes movement, 1-channel EMG from the chin and 2-channel EEG on C3 and O2 position, are monitored to evaluate sleep stages, same as the function of the supplementary device of the first system. After transmission, based on TCP/IP (Transmission Control Protocol/Internet Protocol), the signals are processed and saved in the server. The physician connects to the server and monitors signals in real time or analyzes/diagnoses with the stored data. Figure 4 shows the data flow of the second system, the wireless telemedicine system with GSM/WCDMA network [34]. The deployment of the all-in-one sleep monitoring system provides the comfort and cost reduction by removing long wire connection and installation/maintenance of dedicated sleep lab. In addition, the system based on mobile network enables the monitoring of signals in real time at any place.

Clyto Access
Figure 4. Data flow of the system to evaluate sleep stages with GSM/WCDMA

5.1 System Design and Implementation

The first introduced system with ZigBee and Wi-Fi network consists of four components; sensors, two HST devices, a receiver, and monitoring unit. Sensors detect the physiological signals. The HST devices which consist of an amplifier module, a microprocessor and a ZigBee module sense the physiological signals and send them to the receiver through ZigBee network. The receiver which consists of two RF modules, ZigBee and Wi-Fi, a microprocessor and memories re-transmits the signals to the monitoring unit through Wi-Fi network. The monitoring unit displays real time sleep data and records sleep data to the local or the remote server.

The system with GSM/WCDMA network is composed of sensors and a data acquisition/transmission unit which is placed on top of the skull cap. The sensed 5-channel signals are sent to the remote server which has the static IP address through the mobile network. The signals are processed and monitored in the remote server.

5.1.1 Sensors

Sensors are selected and characterized to measure different types of physiological signals. Biopotential signals; EEG, EOG, EMG and ECG, can be sensed with conventional wet Ag/AgCl electrodes which use the conductive gels to reduce the skin impedance or make conductive path through hairs. However, these gels can cause skin irritation. Moreover, when the gels dry out, it changes the impedance and introduces erroneous signals. As a remedy, vertically aligned gold nanowire electrodes [35] are used. As dry electrodes, the gold nanowire electrodes obviate the use of conductive gel, and minimize the impedance changes and skin irritation due to dry-out of the conductive gel. Body position is used to analyze the pattern of body movement during sleep. The gyroscope is used for body position, which is mounted on the chest belt. The amount and the direction of motions can be derived with the sensitivity of the gyroscope over yaw, pitch and roll axis. The chest respiration effort is measured by detecting the up/down movements of chest during breathing. The changes of waveforms indicate the number of movements from inspiration and expiration. The movement of chest is measured with the piezoelectric sensor which is embedded in the chest belt. Nasal airflow is to measure the rate of respiration and identity interruptions in breathing. Nasal airflow is measured by detecting pressure changes from respiration volume through the nose with a pressure sensor which is connected to the nose prongs. By measuring chest respiration effort and nasal airflow, the case of sleep apnea that has chest respiration effort without airflow can be detected. Oxygen saturation is measured and derived with a pulse oximeter. It is a relative measurement of amount of hemoglobin and the combination of hemoglobin and oxygen. The transmittance mode of the pulse oximeter is used, in which the light emitted by LEDs is detected by a photodiode which is placed opposite to the LEDs. Arterial pulsatile waveforms are detected with the RED/IR REDs and the photodiode, and the SpO2 is derived with the absorption ratio of RED and IR LEDs. Figure 5 shows the sensors which are used to diagnose sleep disorders with the systems.

Clyto Access
Figure 5. Sensors to diagnose sleep disorders [36]

5.1.2 Amplifier Module

The raw physiological signals from the sensors are very weak, ranging from micro-volts to milli-volts, and contaminated by noises, especially 50/60 Hz power line interference. Therefore, the signals need to be amplified and filtered to improve the signal to noise ratio. The biopotential signals for sleep studies are differential signals. The body impedance and contact impedance between skin and the electrodes might vary under different skin conditions leading to impedance mismatches. Therefore, amplifier should have common mode rejection ratio (CMRR) of at least 60 dB and input impedance of at least 10 MW to maintain signal integrity. The 5-channel amplifier modules for the supplementary HST device and the device with GSM/WCDMA consist of 3 stages and each amplification channel has the same configuration. Figure 6 shows the schematic diagram of the amplification circuit for the biopotential signals. The gain and bandwidth of each channel were tuned according to the AASM manual [37]. The instrumentation amplifier is adopted as the first stage to guarantee the high input impedance and high CMRR. The second stage, includes a RC high pass filter (HPF) designed with non-inverting operational amplifier to remove DC offset from the accumulated electric charge on the surface of the electrodes. The final stage is an active filter and a HPF.  The second order Sallen-Key low pass filter is implemented. The other 5-channel amplifier for the type III HST device, for measuring different types of physiological signals, was designed according to the type of sensor used. The channels for the body position and SpO2 have different configurations as compared to the other channels for ECG, nasal airflow and chest effort. The gyroscope for measuring body position processes data internally. Thus, the amplifier channel is not required for the gyroscope. The output of the gyroscope is directly connected to the input of the analog-to-digital converter (ADC) in the microprocessor. The first stage of the amplifier for the SpO2 was implemented with a trans-impedance amplifier to convert the photocurrent from the oximeter sensor to voltage signals which was then amplified with a non-inverting amplifier. The amplifier channels for nasal airflow and chest effort have the same configuration as those used for the biopotential signals. The amplifier module for the device with GSM/WCDMA has the same configuration as that of the supplementary HST device because they measure the same biopotential signals, EOG, EMG and EEG. Table 5 summarizes the specification of each amplifier channel.

Clyto Access
Figure 6. Schematic diagram of the amplification circuit for biopotential signals



Device type

Type III HST device

Supplement HST device / device with GSM/WCDMA

Gain &












54 dB

0.3 ~ 71 Hz



76 dB

0.3 ~ 35 Hz


Body Position





76 dB

0.3 ~ 35 Hz


Nasal Airflow

59 dB

0.14 ~ 18 Hz



76 dB

 0.3 ~ 100 Hz


Chest effort

34 dB

0.13 ~ 16.3 Hz



80 dB

0.3 ~ 35 Hz



6 dB

 0.15 ~ 30 Hz



80 dB

0.3 ~ 35 Hz

Virtual ground


Table 5. Specification of the amplifiers for physiological signals

5.1.3 Data Acquisition and Wireless Units

Devices with ZigBee and Wi-Fi

The data acquisition/wireless unit with ZigBee and Wi-Fi consists of two HST transmitters and a receiver. The HST transmitters and the receiver build the ZigBee star network in WPAN. In addition, the network is extended to WLAN with Wi-Fi module in the receiver and the network is further extended to WAN with an access point which is connected to the Ethernet network. Therefore, the real time monitoring of the physiological signals at remote places can be realized by extending WPAN with ZigBee to WAN through WLAN with Ethernet network and Wi-Fi network respectively.

The main functions of the wireless HST transmitters are data processing, data flow management and data transmission with the ZigBee module. The analog signals from the amplifier output are digitized using the AD converter onboard the microprocessor and timer registers were used to fix the sampling rate, 100 Hz, for channels. The register value is initially set and interrupt signals are generated whenever the value crosses zero based on the frequency of count. The interrupt signal is generated at a fixed interval and is used to trigger the AD conversion. The output of the ADC is stored in the buffer every one second and transferred to the ZigBee module through Universal Asynchronous Receive/ Transmit (UART) interface. The data from the microprocessor are transmitted to the wireless receiver unit through ZigBee. Programs were developed to drive the ZigBee module and transmit the sleep data to the wireless receiver unit. The ZigBee module is controlled using the integrated command set. There are two different kinds of devices in a ZigBee network: A fully functional device (FFD) and reduced functional device (RFD). FFD can be a coordinator, a router or an end device. FFD can communicate with other FFD or RFD. On the other hand, RFD can act only as an end device and can communicate only with FFD in the network. ZigBee network allows only one coordinator and a number of routers and end devices. The function of an end device to two HST devices and the function of a coordinator to the wireless receiver unit were assigned while building the ZigBee star network. Upon device power up, the boot loader is loaded, following which the initialization functions are executed for the AD converter and the ZigBee module. The routine for ZigBee initialization sets up the connection between the ZigBee module and the microprocessor and configures the PAN ID and self ID of the ZigBee module. After this step, AD conversion is performed and the data are reconstituted as packets with sequential hex code including node number, sampling rate, the number of channel and the data according to the defined communication protocol. These packets are saved in the data buffer temporarily and transmitted when receiving the data request. Finally, ZigBee interface routinely checks the wireless connection between the HST transmitters and the receiver unit, and then transmits the sleep data to the receiver unit. Figure 7 shows the data flow in the microcontroller of the HST transmitters.

Clyto Access
Figure 7. Data flow in the microcontroller of the HST transmitters

Both the wireless HST devices use a 3.7 V poly lithium battery as a power supply. 3.7 V is boosted up to 5 V by DC/DC converter to provide power to the microprocessor and the amplifier. 5 V is regulated to 3.3 V for the ZigBee module by a low-dropout regulator (LDO). Figure 8 shows the block diagram and image of the wireless HST device. While the function of the wireless transmitter is to build the wireless personnel area network with ZigBee, the role of the wireless receiver is to extend the ZigBee network to the wireless local area network with Wi-Fi. The sleep data need to be monitored at remote places. Therefore, the network should be extended to use Internet. Although there are two ways that ZigBee has advantages on power consumption and WPAN/WBAN construction, the communication distance of ZigBee is at most a few hundred meters and ZigBee is not as prevalent as Wi-Fi. It will be expensive to build a local area network using ZigBee alone. Additionally, the amount of sleep data in the receiver from two transmitters is twice the amount of the sleep data in each transmitter, 8 kbps. The receiver should handle 10 channels of data while each transmitter processes 5 channels of data. Therefore, Wi-Fi facilitates extension of the number of channels by coordinating the data received from the transmitters and additional transmitters can simply be added on.

Clyto Access
Figure 8. Wireless HST device: (a) block diagram and (b) image

The processor board in the wireless HST receiver manages the ZigBee and Wi-Fi module and consists of a 16 MB SDRAM, a 64 MB NAND Flash memory, a 32-bit microcontroller and an Ethernet controller. The memories are needed to process and store data. SDRAM is used as the buffer for the sensor data, and the flash memory is used for the storage of programs. The Wi-Fi module and the Ethernet controller enable the receiver to connect to the Internet through wireless or wired communication. The Wi-Fi module, adopting IEEE 802.11 b/g, is connected to the microcontroller through USB 2.0 interface. USB 2.0 interface whose data transfer rate is 480 Mbps assures the full speed of the Wi-Fi module, 54 Mbps. The supplied 5 V is regulated to 3.3V for the ZigBee module, the external memory and I/O ports of the microcontroller. The other regulated voltages are 2.7V and 1.2V, is for the internal memory and operation of the microcontroller, respectively. The receiver includes the ZigBee and Wi-Fi module management and the server management based on TCP/IP. After performing devices initialization, ZigBee module management task is performed. The functions of this task are setting up the ZigBee configuration and monitoring the ZigBee connection status. After confirming the ZigBee connection between the receiver and the transmitters, the sleep data from the transmitter are saved in a ring buffer. The Wi-Fi module management task configures the Wi-Fi module to connect to the access point. After connecting to the access point, an IP address is acquired from the access point. Finally, the server management task is executed after finishing the Wi-Fi module management task. The server management task enables the receiver to access the monitoring unit based on the TCP/IP and the receiver can start sending sleep data in the ring buffer to the monitoring unit or the server every 0.5 second. Figure 9 shows the data flow in the wireless receiver and Figure 10 shows the block diagram and image of the wireless receiver.

Clyto Access
Figure 9. Data flow in the wireless receiver

Clyto Access
Figure 10. Wireless HST receiver; (a) block diagram and (b) image

Device with GSM/WCDMA

The data acquisition/wireless device with GSM/WCDMA includes a microcontroller, a power management circuit and a GSM/WCDMA module. Figure 11 shows the block diagram of the system.

Microprocessor - The main functions of a microcontroller in the system are data acquisition and management of data flow to the GSM/WCDMA module. ATmega328P from Atmel is used as a microprocessor, which has 8-bit reduced instruction set computer (RISC) architecture, 32-kB flash memory, 2-kB SRAM, 8-channel 10-bit A/D converter and other peripherals. The operating voltage ranges from 1.8 to 5.5V and the operating voltage of the processor is set at 5V with a 16 MHz external resonator. The purpose of setting the operating voltage at 5V is to match the voltage level of the amplifier module. The signals from the amplifier module are sampled and quantized by the A/D converter of the microprocessor. The data are saved to a buffer and then transferred to the GSM/WCDMA module through universal asynchronous receiver transmitter (UART) interface.

Power Management Circuit - A poly lithium battery with a current capacity of 1800 mAh and output of 3.7V is used as a power source. Because each block of the system operates at different voltage level, two DC/DC converters are implemented to step up to 3.8V and 5V. 3.8V is for the GSM/WCDMA module and 5V for the microprocessor and amplifier module. The DC/DC converter for the GSM/WCDMA module should guarantee that the input voltage of the module never drops below 3.3V even during transmission when the current consumption rises up to 2A, to prevent the module from being shut down. LTC3113 from Linear Technology is used as the DC/DC converter for the GSM/WCDMA module, which can deliver up to 3A of continuous output current and operate from input voltage above, below or equal to the output voltage. LTC1751-5 from Linear Technology is used to provide 5V to the amplifier module and the microprocessor. The DC/DC converter can produce up to 100mA at fixed 5V of output voltage. The DC/DC converter can operate only one flying capacitor and two bypass capacitors at input and output pins without an external inductor.

GSM/WCDMA Module - Mobile communication network is used instead of building local area networks to send physiological signals to the remote server. To connect the system to the mobile network, the GSM/WCDMA module, SIM5320 from SIM Com, is used. SIM5320 supports a quad-band GSM/GPRS/EDGE and dual-band UMTS/HSDPA [38]. The operating voltage of the module is set to 3.8V. The power from LTC3113 is supplied to the baseband and the RF power amplifier of the module. The ripple from the emission burst of GSM/GPRS, 2A of current consumption, may cause voltage drop, which may lead to the shutdown of the module. Because LTC3113 delivers current up to 3A, the power supply of both the baseband and the RF amplifier are tied up to the output of LTC3113. A large capacitor (100 mF) and by-pass capacitor is placed close to the input voltage pins of the module for the system stability and EMC. The turning on/off of the module is controlled by the power on/off logic in the module. The logic is pulled up internally and the logic operates at active low. The logic is pulled down to ground by an external resistor in circuit directly to turn on the module automatically when the main switch is on.

Clyto Access
Figure 11. Block diagram of the data acquisition/wireless device with GSM/WCDMA

The module communicates with the microprocessor over UART port. It consists of a flexible 7-wire serial interface. The module is referred as the data communication equipment (DCE) and the microprocessor as the data terminal equipment (DTE). The module supports both full modem and null modem topologies, and the null modem is used in the system. The baud rate of UART interface is set to 115,200 bit per second. The universal subscriber identity module (USIM) is implemented on the board for authentication and association with the mobile network. A transient-voltage-suppression (TVS) diode array is inserted between USIM and the GSM/WCDMA module for ESD protection. SIM5320 provides several general purposes input/output (GPIO) and the GPIO status is read or written with AT commands. One of GPIOs was used to indicate the network status with a light emitting diode (LED). The turn on/off intervals of LED represents the status of network; power on/searching network, data transmit, network registration and power off. The module supports a quad-band of GSM 850 MHz, EGSM 900 MHz, DCS 1800 MHz and PCS 1900 MHz and dual-band of WCDMA 850 MHz and 1900 MHz. A penta-band SMD antenna, Reflexus from Antenova, is used to cover all frequency bands. The average gain is -1.3 dBi for the 824-960 MHz and -1.5 dBi for 1710-2170 MHz. A microstrip line which has impedance of 50 Ohm is designed as a transmission line between the module and the antenna. A matching circuit is placed on the transmission line and a separate DC blocking capacitor is added between the module and the antenna matching circuit. The antenna is placed on the corner of the board to achieve maximum antenna performance. A ceramic passive chip antenna, W3011A form Pulse Electronics, is used to receive the position information at GPS L1 frequency (1.575 GHz).

The amplifier module, the processor board, the DC/DC circuit (LTC1751-5) and the battery are stacked on the back side of the GSM/WCDMA board. Figure 12 shows the images of top and bottom side of the system. The connection between electrodes and the amplifier module is made by snap button type of adaptors.

Clyto Access
Figure 12. Images of the wireless HST system with GSM/WCDMA; (a) top and (b) bottom side of the system

5.2 Test Results

Regarding to the system with ZigBee and Wi-Fi, in particular, the tests were performed to examine whether each transmitter can send the sleep data to the receiver without loss through the ZigBee star network, thereby verifying the feasibility of channel extension through wireless. The five different types of sensors were placed on the body for the wireless type III HST device. Two gold nanowire electrodes were attached to the left and right upper chest according to the Manson-Likar modification. A chest belt equipped with a piezoelectric sensor and a gyroscope were worn to measure the chest respiration effort and the body position. The nasal prongs including the pressure sensor were placed in the nostrils to monitor the nasal airflow. Finally, the oximeter sensor was clipped on to the fingertip to measure SpO2. For the wireless supplementary HST device, the gold nanowire electrodes were used to detect the three different types of the biopotential signals, EOG, EMG and EEG. The reference electrode to measure the differential biopotential signals was placed on the left ear lobe. Two electrodes were placed on the left upper and right lower side of the eyes to measure 2-channel EOG. With these positions of the electrodes, movements of the left and right eye ball produce the opposite polarities of left and right EOG. One of electrodes was placed on the chin to measure EMG and to detect grinding of teeth or mouth movement. The last two electrodes were placed on C3 and O2 position on the scalp according to the international 10-20 system. Figure 13 shows an image of a subject under sleep study with the system.

Clyto Access
Figure 13. Experimental test set-up for the system with ZigBee and Wi-Fi [33]

The 5-channel signals from the type III device and other 5-channel signals from the supplementary device were recorded and monitored in real time with the utility program at the monitoring unit/ the server under different test conditions. The test condition was established to classify the changes of physiological signals. P, QRS complex and T waves should be identified in ECG. Body position was measured as the subject moved. The chest effort and nasal airflow were monitored under no breathing, to emulate sleep apnea, fast breathing and normal breathing. In case of EOG, five different types of eye motions, blinking, left, right, up and down movement of eyeballs, were tested to emulate REM and wakeful sleep stage. EMG was produced by the movement of the chin. The EEG acquisition performance was verified by discerning between beta waves and alpha waves. The beta waves are associated with normal waking consciousness and the alpha waves with wakeful relaxation with eyes closed. By opening and closing eyes alternatively, the changes of brain waves were examined. Regarding to the system with GSM/WCDMA, the electrodes were placed on the same positions as the supplementary HST device to monitor EOG, EEG and EMG, and test method is also same as the method involved in the supplementary HST device because the function of the system is same as that of the supplementary device with ZigBee and Wi-Fi. Figure 14 shows experimental test set up to evaluate the system with GSM/WCDMA.

Clyto Access
Figure 14. Experimental test set up for the system with GSM/CDMA [34]

Figure 15 shows the waveforms of each physiological signal which were measured with the type III HST device and recorded to the server. Figure 15 (a) shows the measured lead I ECG. The ECG waves show the P wave, QRS complex and T waves clearly. Generally, the components of ECG, P wave, QRS complex, T wave, PR interval and QT interval are used to diagnose heart conditions. Furthermore, heart rate variability (HRV) is derived with the components of ECG and HRV might be used to diagnose sleep apnea [39]. Figure 15 (b) shows the waveforms from the gyroscope according to the three levels of body motion: no motion, slow motion and fast motion. Corresponding to the body movements, the output voltage of the gyroscope changed. The positive or negative slope of the waveforms represents the direction of rotation of the subject and the output voltage of the gyroscope reveals the velocity of the body movement. Based on the sensitivity of the sensor and the polarity of the slope, the velocity of the body movement is derived. By analyzing the waveform from the gyroscope, it can be diagnosed that how often and how much the subject moves during sleep. Chest respiration effort and nasal airflow were also measured under three different types of breathing. Positive and negative slope of the waveforms means inhalation and exhalation, respectively. Respiration cycles can be discerned by tracking the positive and negative slopes. In addition, the frequency of the waveforms is related to the speed of breathing. Oxygen saturation is derived from the output of the oximeter sensor, amount of transmission of the IR and RED LEDs through fingertip. The monitoring utility program displays the derived SpO2 instead of displaying absorption coefficient of the IR and RED LEDs.

Figure 16 shows the recorded biopotential signals from the supplement HST device. The EOGs from left and right eyes are classified with five different movements of eyeballs. The left and right EOGs show opposite polarity of slope due to the placement of the electrodes over the reference electrode. The blinks are represented by the sharp and high peak amplitude due to the nature of the quick eye motion while left, right, up and down movement of eyeballs are represented by wider slower waveforms. The waveforms, representing EMG, show low amplitude and flat baseline when the chin is at rest, but the waveforms are changed to high amplitude and high frequency ones during movement of chin. EEGs from C3 and O2 positions show mixed brain waves of beta and alpha waves when eyes are opened, but alpha waves are dominant when eyes are closed. EEG from position O2 shows more strong alpha waves than from position C3 even though both the channels of the amplifier are set to the same gain. It verifies that the alpha waves predominantly originate from the occipital lobe during wakeful relaxation with closed eyes. Each HST device is operated with the receiver simultaneously by building the ZigBee star network. There was no data loss in the monitoring unit/ server due to interference between the two HST devices. Therefore, the wireless type III HST device could be upgraded to type II HST device by adding 5 channels of biopotential signals which are measured with the supplement HST device.

Clyto Access
Clyto Access
Figure 15. Monitored physiological signals from the wireless type III HST device; (a) ECG, (b) Body position, (c) Nasal airflow, (d) Chest respiration effort, and (e) Oxygen saturation

Clyto Access
Clyto Access
(c) EEGs from C3 and O2
Figure 16. Monitored physiological signals from the wireless supplement HST device

Figure 17 shows the monitored biopotential signals from the system with GSM/WCDMA. They show the similar waveforms of those of the supplement device with ZigBee/Wi-Fi under the same test conditions for each physiological signal, EOG, EMG and EEG.

Clyto Access
Clyto Access
(c) EEGs from C3 and O2
Figure 17. Monitored physiological signals from the system with GSM/WCDMA

In a case study on a human subject from 10 pm to 4 am, 5-channel signals were monitored and recorded in real-time with the HST device. The EEG signals are plotted in Figures 30-32. Based on the criteria specified in Table 2, a transition between sleep stages can be seen in the EEG data from C3 and O2.

Clyto Access
Figure 18. The record is clearly Stage W in that more than half the epoch is occupied by alpha activities. In the middle of the epoch, there is a high and wide range of signal which comes from eye movement
Clyto Access
Figure 19. This epoch illustrates a transition between Stage 2 and Stage REM. In the beginning of the epoch, there is sleep spindle followed by a K complex. Following the K complex are saw-tooth waves which herald the appearance of REM. The last half of epoch shows relatively low voltage, mixed frequency.

Clyto Access
Figure 20. This epoch illustrates the Stage 2.In the beginning of the epoch, slow waves are shown. There is a burst of activity in the 12-14 cps range (Sleep Spindle). High amplitude waves are also shown. Its occupation is less than 20 %.

Normally, in the present communication systems for WPAN and WLAN, the output power is not varied according to the transmission distance. This requires a constant and steady power supply. Adaptive RF output power control algorithm implements a concept of controlling the RF output power with respect to the transmission distance involved in the process of communication. The power consumption when the distance between the transmitter and receiver is reduced. It is based on the received signal strength indicator (RSSI). RSSI is the measurement of power present in the signal received in the receiver.

The transmitter and receiver are initialized when the system is turned on. The analog data are converted by the processing unit in the transmitter and the data are buffered. The transmitter starts transmission every one second when the buffer is full. After the transmission process is initiated and the data are transmitted to the receiver, the receiver goes to command mode and acquires the RSSI value. After getting the RSSI value, the receiver goes back to transparent mode, enabling two-way communication between the transmitter and the receiver. The RSSI value is fed back to the transmitter and receiver starts performing its usual function of receiving, displaying and data saving. Once the transmitter receives the RSSI value from the receiver, it goes to the command mode and changes the output power according to the received RSSI value. As soon as the output power is adopted according to the RSSI value, the transmitter once again gets back to the transparent mode to enable normal transmission of data. The regular data transmission takes place again till the distance between the transmitter and receiver changes. When there is a change in the distance between the communicating units, the same process takes place again to determine the required output power. Figure 21 shows the data flow between the transmitter and the receiver.

Clyto Access
Figure 21. Data flow between the transmitter and the receiver with adaptive RF output power control

An experiment was carried on to verify the adaptive power algorithm and to measure RSSI and data loss at different output power level and at a fixed distance between the transmitter and receiver with ZigBee. The experiment was performed with four different RF power output levels (10dBm, 12dBm, 14dBm and 18dBm). As shown in the Figure 22(a), increasing the distance with constant output power reduces the value of RSSI. Results also show that, a RF output power of 18 dBm can communicate up to 350 meters and a RF output power of 10 dBm can transmit up to 150 meters. This indicates that RF output power and the communicating distance is proportional. In addition, test results also show that the power consumption is inversely proportional to the RF output power as shown in the Figure 22(b). The transmitter can transmit up to 150 m at 10 dBm rather than using 18 dBm of output power. Thereby, the power consumption can be reduced up to 21 %. Hence, based on the experimental results, it is evident that controlling the output power based on the transmission distance can considerably reduce power consumption.

Wireless health monitoring systems are a new class of unobtrusive continuous health monitoring with significant benefits for the patients suffering from chronic diseases. This paper introduces various design considerations to improve the systems in terms of wireless communication and the number of physiological channels by comparing the basic platforms in health monitoring systems. A relationship between the combination of two wireless communication systems and the remote health monitoring system is established and the algorithm, an adaptive RF output power control, to reduce the power consumption is discussed. Two wireless health monitoring systems for diagnosis of sleep disorder are described. The HST system with ZigBee and Wi-Fi is introduced to improve the present monitoring systems which have the fixed number of physiological channels. The system allows adding supplementary devices and extending WPAN to WLAN/WMAN. By the implementation of star network topology in a ZigBee network, type III HST device is upgraded to a type II HST device by adding the supplementary device. The combination of two wireless communication standards provides flexibility in data transmission and distance, as WPAN of the ZigBee network is extended to the WLAN with Wi-Fi. These incorporations provided successful results in remote health monitoring. The other HST system with GSM/WCDMA enables the acquired physiological signals to be transmitted over long distances and to be monitored from anywhere. In addition, to reduce the power consumption of the system, the RF output power control algorithm controls the transmission power of the transmitter according to the distance between the transmitter and the receiver based on the RSSI. It shows a significant reduction up to 21 % in the power consumption.

Clyto Access

Clyto Access
Figure 22. Experimental results; (a) relation between communication distance and RSSI with different RF output power values and (b) current consumption with different RF output power values

[1] Daniel Foley, Sonia Ancoli-Israel, Patricia Britz and James Walsh, "Sleep disturbances and chronic disease in older adults Results of the 2003 National Sleep Foundation Sleep in America Survey," Journal of Psychosomatic Research, vol. 56, pp. 497-502, 2004.
[2] California HealthCare Foundation, “Snapshot Health Care Costs 101”,, 2012.
[3] The Council of State Governments, "State Official's Guide to Chronic Illness," 2003.
[4] D. Foley, A. Monjan, E. Simonsick, R. Wallance and D. Blazer, "Incidence and remission of insomnia among elderly adults: an epidemiologic study of 6,800 persons over three years," Sleep, vol. 22, pp. 366-372, 1999.
[5] S. Maggi, J. Langlois, N. Minicuci, F. Gringoletto, M. Pavan and D. Foley, "Sleep complaints in community-dwelling older persons: prevalence, associated factors, and reported causes," Journal of the American Geriatrics Society, vol. 46, pp. 161-168, 1998.
[6] M. Vitiello, "Sleep disorders and aging: understanding the causes," J Gerontol A Biol Sci Med Sci, vol. 52, pp. 189-191, 1997.
[7] D. Foley, A. Monjan, S. Brown, E. Simonsick, R. Wallace and D. Blazer, "Sleep complaints among elderly persons: an epidemiologic study of three communities," Sleep, vol. 18, pp. 425-432, 1995.
[8] S. Ancoli-Israel, D. Kripke, M. Klauber, W. Manson, R. Fell and O. Kaplan, "Sleep disordered breathing in community dwelling elderly," Sleep, vol. 14, pp. 486-495, 1991.
[9] S. Ancoli-Israel, D. Kripke, M. Llauber, W. Manson, R. Fell and O. Kaplan, "Periodic limb movements in sleep in community-dwelling elderly," Sleep, vol. 14, pp. 496-500, 1991.
[10] E. Kramarow, H. Lentzner, R. Rooks, J. Weeks and S. Saydah, "Health and aging chartbook," National Center for Health Statistics, pp. 40-41, 1999.
[11] J. Guranlnik, "Assessing the impact of comorbidity in the older population," Ann Epidermiol, vol. 6. pp. 376-80, 1996.
[12] D. Lipkin, "Sleep-disordered breathing in chronic stable heart failure," Lancet, vol. 354, pp. 531-532, 1999.
[13] Claudio Bassetti and Michael Aldrich, "Sleep Apnea in Acute Cerebrovascular Diseases: Final Report on 128 Patients," Sleep, vol. 22, no. 2, pp. 217-223, 1999.
[14] O. Parra, A. Arboix, S. Bechich and L. Garcia-Eroles et al., "Time course of sleep-related breathing disorders in first-ever stroke or transient ischemic attack," American Journal of Respiratory Care Medicine, vol. 161, pp. 375-380, 2000.
[15] H. Resnick, S. Redlin, E. Shahar, A. Gilpin and A. Newman et al., "Sleep Heart Health Study. Diabetes and sleep disturbances: findings from the Sleep Heart Health Study," Diabetes Care, vol. 26, no. 3, pp. 702-709, 2003.
[16] C. Sjöström, E. Lindberg, A. Elmasry, A. Hägg, K. Svärdsudd and C. Janson, "Prevalence of sleep apnea and snoring in hypertensive men: a population based study," Thorax, vol. 57, pp. 602-607, 2002.
[17] Michael J. Sateia, "International Classification of Sleep Disorders-Third Edition: Highlights and Modifications", Chest Journal, vol. 146, no. 5, pp 1387-1394, 2014.
[18] A. Rechtschaffen and A. Kales, "A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects," UCLA, Los Angeles, 1968.
[19] K. Susmakova, "Human Sleep and Sleep EEG," Measurement Science Review, vol. 4, pp. 59-74, 2004.
[20] Shahin Farshchi, Aleksey Pesterve, Paul H. Nuyujukian, Istvan Mody, and Jack W. Judy, “Bi-Fi: An Embedded Sensor/System Architecture for Remote Biological Monitoring,” IEEE Trans. Information Technology in Biomedicine, vol. 11, no. 6, pp. 611-618, 2007.
[21] C. Mundt, K. Montgomery, U. Udoh, V. Barker, G. Thonier, A. Tellier, R. Ricks, R. Darling, Y. Cagle, N. Cabrol et al., “A multiparameter wearable physiologic monitoring system for space and terrestrial applications,” IEEE Transactions Information Technology in Biomedicine, vol. 9, no. 3, pp. 382–391, 2005.
[22] J. Yao, R. Schmitz, and S. Warren, “A wearable point-of-care system for home use that incorporates plug-and-play and wireless standards,” IEEE Trans. Information Technology in Biomedicine, vol. 9, no. 3, pp. 363–371, 2005.
[23] T. Fulford-Jones, G.Wei, and M.Welsh, “A portable, low-power, wireless two-lead EKG system,” in Proc. 26th Annu. Int. Conf. IEEE Eng. Med. Biol. Soc., vol. 1, pp. 2141–2144, 2004.
[24] I. Obeid, M. Nicolelis, and P. Wolf, “A multichannel telemetry system for single unit neural recordings,” J. Neurosci. Methods, vol. 133, pp. 33–38, 2004.
[25] "Phillips - Alice PDx" Available: [Accessed 18 July 2016].
[26] "Natus - Embletta X100" Available: [Accessed 18 July 2016].
[27] "Compumedics - Somte PSG" Available: [Accessed 18 July 2016].
[28] "Compumedics - Siesta 802" Available: [Accessed 18 July 2016]. [29] "Cleveland Medical Devices - SleepScout" Available: [Accessed 18 July 2016].
[30] "ResMed - ApneaLink Plus" Available: rochure/1013404_apnealink-plus_product-brochure_amer_eng.pdf [Accessed 18 July 2016].
[31] "BD Medical Technology - CareFusion Nox-T3" Available: [Accessed 18 July 2016].
[32] Sechang Oh, Hyeokjun Kwon, Robert Harbaugh and Vijay K. Varadan, "Wireless Point-of-Care Diagnosis for Sleep Disorders in WPAN and WLAN," Smart Nanosystems in Engineering and Medicine, vol. 2, pp. 46-54, 2013.
[33] Sechang Oh, Hyeokjun Kwon and Vijay K. Varadan, "Wireless telemedicine systems for diagnosing sleep disorders with Zigbee star network topology," SPIE Proceedings, vol. 8548, DOI: 10.1117/12.946115, 2012.
[34] Sechang Oh, Hyeokjun Kwon, Prashanth S. Kumar and Vijay K. Varadan, "A Wearable Wireless Monitoring System for Sleep Disorders through GSM/WCDMA Networks," Smart Nanosystems in Engineering and Medicine, vol. 1, pp. 123-134, 2012.
[35] Vijay K. Varadan, Sechang Oh, Hyeokjun Kwon and Phillip Hankins, “Wireless Point-of-Care Diagnosis for Sleep Disorder With Dry Nanowire Electrode,” J. Nanotechnol. Eng. Med., vol. 1(031012), pp. 1-11, 2010.
[36] Sechang Oh, Hyeokjun Kwon and Vijay K. Varadan, "Ubiquitous Health Monitoring System for Diagnosis of Sleep Apnea With Zigbee Network and Wireless LAN," Journal of Nanotechnology in Engineering and Medicine, vol. 2, 021008, 2011.
[37] C. Iber, S. Ancoli-Israel, A. Chesson, S. Quan, "The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology, and Technical Specifications," Amer. Acad. Sleep Med, Westchester, IL, 2007.
[38] SIM Com. SIM5320. Available: [Accessed 11 April 2016].
[39] Phyllis K. Stein, Yachuan Pu, "Heart rate variability, sleep and sleep disorders," Sleep Medicine Reviews, vol. 16, Issue 1, pp. 47-66, 2012.

Published: 28 March 2017

Reviewed By : Dr. Piotr Augustyniak,


© 2017 Oh et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.