A few words on sensor sensitivity...
Talking with technicians and sensor sales-people, you often hear claims about one sensor being more sensitive than another. The truth is that with today's high performance polysomnographic amplifiers, any sensor can produce a readable signal. But which sensor has a better signal to noise ratio, or "SNR"?
A properly designed sensor aims at achieving the best possible SNR by carefully designing out as many noise sources as possible, while maintaining good response to as many components of the desired signal as possible. Of course, the differences between a properly designed and built sensor and one of lower quality will not be so evident when comparing them on the bench, where interference and artifacts are low to non-existent. The better sensors will excel when it matters most—when strapped to an agitated, sweaty, 400 pound patient who's twisting and rocking in bed. In these cases, movement artifacts, noise from static charges generated by friction against bed linen, cable movement noise, attenuation of the desired signal due to sweat and many more factors can make the difference between a workable recording, and a tired technician who needs to come in and readjust the sensor every hour.Sensors cable's selection:
All sensors are connected with a cable to the system. The cable poses conflicting design requirements: It must be very strong, so as not to break, even if repeatedly pulled by patients, but it must be very thin, light and flexible to allow for patient comfort and handling.
All SleepSense sensors include the best version of Tinsel cables, employing Kevlar® fiber core (the same material used for bullet proof vests). This type of design is the most suitable and stable in order to maintain conductivity.
Effort: Piezo Crystal Technology
Piezo crystals are silicone elements, generating an electrical signal when force is applied to them. They are therefore very effective in making respiration effort sensors, where respiration movements of the chest or abdomen stress the crystal to generate a small voltage proportional to the movement or tension in the elastic belt around the body. Although any Piezo sensor will produce an acceptable signal in bench tests, designing a truly effective device is a significant technical challenge.
The first and most important issue is to protect the Piezo active element from any interference that might degrade the signal and produce artifacts. Piezo crystals are extremely sensitive, so a high degree of insulation from the environment is needed, since even small temperature changes, acoustic signals or moisture can generate significant artifacts.
The best way to protect the crystal is to enclose it in a rigid, sealed box slightly larger than the crystal itself, and suspending the crystal inside so that it is “floating” within the box without touching any of the walls. All of the SleepSense Piezo Crystal Effort Sensors are designed in this exact way to ensure that sleep technicians are getting the most accurate signal out of their Piezo Crystal. In fact, the way that SleepSense has isolated the Piezo Crystal element in the small black box is so durable, that a person can literally walk on the sensor, re-install it, and it will work with no problems at all.
Respiration movements are fed to the crystal through a special “belt” that has high stiffness in the axial direction but which is relatively flexible for bending. This maximizes transmission of respiration movements, while damping other, unrelated movements.
It is also important that the sensor will produce the same signal polarity when the belt is pulled, as well as when the box is compressed. This is needed to avoid a known artifact in Piezo sensors where, if the belt is not tight around the body or the sensor is caught between the body and mattress. A properly designed Piezo Effort Sensor, like the SleepSense Effort Sensor, will insure the signal polarity is not reversed and paradoxical breathing is not erroneously presented to the system.
Effort: Piezo Film
Plastic films with Piezo-electric propreties were first produced over 15 years ago, and are in general used in many types of sensors. The same guidelines that are applicable to Piezo Crystal Effort Sensors also apply to Piezo Film.
It is important that the film is isolated from the environment as much as possible. The demand for totally flexible sensors dictates that the film can not be enclosed in a rigid box (which would be the best option). Thick elastic should therefore be used both under and over the Piezo Film element, and the elastic should be sealed as much as possible to prevent any exposure of the film.
Piezo-films are significantly less sensitive than crystals, so the “collection gap” should be made as long as possible to get the most robust and reliable signal. “Collection gap” means the length of the non-elastic part of the sensor, across which the tension is measured. The longer this gap, the stronger and more reliable the signal, since there is a lower loss of useable force across elastic parts of the sensor. Larger “collection gaps” require the use of a larger, and more expensive film, so other manufacturers tend to disregard this requirement. The same applies to the width of the “collection gap”, the wider the encore points the stronger the signal, and of course wider elastic belts are always better, because they transmit the respiration forces more reliably, and they maintain a better position on the body.
It is extremely important that the Piezo Film is anchored in such a way to produce a linear signal starting at a very weak tension all the way to very tight pulls. The SleepSense Piezo Film Sensor is totally flat under low tension, which results in a more accurate output, even when the belt is not tight around the body.
Inductive Respiration Effort Sensors
Inductive bands depend on electrical “inductance”, a property of any electrical conductive wire, by which it limits the flow of high frequency current as a result of a change of magnetic flux. Maxwell equations show that the inductance of a closed loop of conductive wire is linearly proportional to the area enclosed in the loop. The sensors are simply closed loops of conductive wire, placed around the chest and abdomen. The wires are usually woven in an elastic band in a zig-zag pattern. The allows the band to stretch and conract. A specially designed electronic circuit measures the impedance of the loops using a high frequency voltage source, and generates an analogue signal proportional to the area enclosed in the loop.
The bands can be wiped with a mild detergent or washed in a gentle machine cycle. To sterilize the bands, use autoclave sterilization or standard gas or solution sterilization procedures. Wipe the cable with a non-corrosive (to plastic) cleanser to clean before use. Make sure the bands and cable are thoroughly dry before reusing them.
Inductive bands can be more accurate than Piezo bands because inductive bands produce an output that is proportional only to changes in the area enclosed by the wire loop. Piezo sensors, on the other hand, produce an output proportional to the tension in an elastic band strapped around the body. This tension depends on many factors, such as the placement of the band, initial tension, slipping of the band, locking of parts of the elastic band under the weight of the patient, and many more—all of which produce errors and artifacts. In the inductive band, any part of the band contributes to the output, whereas with the Piezo bands there is one sensor which is the sole sensitive area, making the elastic band itself passive.
SleepSense respiration effort sensors provide a qualitative measure of respiration movements for recording onto a data acquisition system. They will provide clear, strong signals.
Snoring sensor design is one of the most challenging of all devices. The sensor must be attached on the skin, exposed to sweat, pressure, accidental high pressure, and other interference. At the same time it must pick up even minute levels of acoustic energy, and transmit them reliably to the system. Low sensitivity to external noises and good compatibility with both patient and technician are additional requirements.
Many users refer to the snoring sensor’s output as decibels, or dB. Noise in dB can only be measured by specially calibrated equipment at a fixed distance from the noise source—something impossible to achieve in a sleep lab setting. It is therefore important to remember that all indications by any patient attached sensors are relative to that particular patient at that particular night, and can not be compared across patients or nights. Sound intensity decreases as the square of the distance between noise source and pickup increases, so even very small changes in the location of the sensor can change the output dramatically.
SleepSense offers three types of snoring sensors to match the specific requirements of the different recording systems on the market. All of the SleepSense snore sensors are small, but not too small, and have high sensitivity to snoring and reduced sensitivity to artifacts and external noises. All of the SleepSense sensors have unique design features that make them superior to other, seemingly similar sensors on the market.
All of SleepSense's contact snoring sensors have a specially designed “bubble” on the skin side. This design feature, found only on SleepSense sensors, insures that even if the sensor’s attachment on the skin loosens-up during the night, and much of the sensor’s body doesn’t touch the skin, the bubble will continue to make contact with the skin and send snoring vibrations and sounds straight to the central of the crystal disk, which is the most sensitive part of the crystal.
Pressure flow sensing is quickly becoming the leading pickup modality for respiration flow monitoring. When designing a good pressure-flow sensor, aspects of aerodynamics and fluid-mechanics come into play, which, if ignored, can produce severe artifacts or measurement errors.
Just as with thermal flow sensors, it must be remembered, that the signal produced indicates pressure, not flow. A good design can make these two physical parameters related, correlating as much as possible, but they are not identical. Many factors exist that can affect this relationship which are not under the designer or the user’s control. For example, the cannula used to monitor nasal pressure during respiration can generate either a positive or negative pressure, or any value in between, depending on the angle between the tube axis and the airstreams, thanks to the Bernoulli Effect found in moving fluids.
One of the important design aspects is the dynamics of the air column locked in the tube leading from the tip of the cannula to the sensor. Since the tube is very narrow, only minute quantities of air can be moved through it in a short time. In order for the sensor, located at the distal end of the tube to “feel” the incrase pressure, you need to push air into it. The less air you have to push, the faster and more accurate the reading will be. It is therefore very important to keep the volume of the sensor and its related tubing to a minimum. Using the smallest sensor, and avoiding the use of filters which have a very large volume, will increase overall accuracy. Since the tube is totally blocked on the sensor side, there is no flow of air in either direction, only changes in compression.
Just as with any sensor, insulating the sensitive element from any interference that is not related to the measured signal is important. SleepSense AC Pressure Sensor achieves this goal by suspending the Piezo Crystal on flexible mounts inside a rigid, protective box. It is also important to minimize signal filtering inside the sensor, allowing as much flexibility in selecting post filtering of the raw signal, and not eliminating parts of the wave which may become important in analysis. Any extra filtering or switches in the signal path only lead to additional electrical noise.
It is customary in sleep studies to use thermal sensors to monitor oral and nasal flow,. Exhaled air is at or near core body temperature, while inhaled air is at room temperature. Any component sensitive to temperature will produce a readable signal if placed in the respiration air stream emanating from the nose and mouth. Such elements in common use are thermistors (resistors that change their resistance when temperature changes), Thermocouples (elements which produce a small voltage in response to temperature changes) or Piezo Film elements (producing a large output due to change in temperature). The signal produced must be reliable in that it replicates the respiration flow as accurately as possible throughout the night.
When using thermal flow sensors, the user must always remember that the signal is temperature, not flow, and refrain from taking quantitative measurements or referring to the amplitude changes as directly representing flow value changes.
Most companies producing thermal flow sensors strive to produce the “strongest” signal, as if their sensors are “more sensitive”. This requires the use of very small and light thermal sensors which will heat up quickly to body temperature on exhale, and quickly cool to room temperature on inhale, thus achieving the biggest temperature swing, or output signal. Although signal output is indeed “strong”, this approach minimizes the relationship between actual flow value and signal output, making the signal far less sensitive to the actual parameter that the user needs to see: flow changes!
SleepSense thermal flow sensors (both thermistors and thermocouples), on the other hand, employ a unique design that allow them to better track FLOW changes, rather than just temperature. The design is based on the use of high-thermal-mass elements to sense the temperature. Such heavy metal elements take a lot of hot air to heat, and a lot of cold air to cool. If flow is reduced, the rate of heat energy entering or leaving the sensor is also reduced, resulting a lower output “swing” or peak to peak value. SleepSense flow sensors use heavy and thick wires to make them sturdy and insure that once bent into shape they will stay in place for the rest of the night.
Body Position Sensors
Monitoring body position is actually monitoring the sensor in relation to the direction of gravity. This is a simple physical entity and poses none of the difficulties associated with physiological signals. There is, however, a major issue with body position monitoring and that is the fact that the sensor reports the position of the sensor—not of the body. It is therefore extremely important to make sure that both are oriented in space in the same direction. When talking of body position, it usually refers to the orientation of the chest vs. the bed, in 4 categories—supine, prone, left and right. Looking at the sleeping person via CCTV can make it difficult to see the sleeping position, and is certainly very cumbersome for the technician to have to note the position constantly.
The only way to make the sensor align with the body is to make it as flat and wide as possible. A flat sensor is less affected by lateral forces which may create a rotation because the rotational movement is smaller. A wide sensor will better resist these forces and maintain orientation relative to the sternum because the holding forces can generate a larger angular movement. It is important that the sensor will be secured on both sides with high strength straps for the same reason.
The human body is not round, but oval. When thinking about what would be considered Supine and Prone, vs. Left of Right sleeping positions it becomes clear that the 360 rotational angle should not be divided equally between the four positions. The prone and supine positions should cover a larger angle, while left and right should be reported by the sensor only if the patient is indeed balanced on that side within the limits of a narrower angle.
SleepSense position sensors were designed to do just that. The Prone and Supine positions cover about 110 degrees each, while the left and right cover about 70 degrees each. This way, the sensor’s output more accurately tracks a visual report by the technician. SleepSense Body Position Sensors employ separate gravity sensing elements (3 sensors altogether) to detect sitting up, thereby providing a more accurate and reliable indication than other sensors.
Limb Movement Sensors
Although the “gold standard” for monitoring leg jerks is the monitoring of leg EMG, many systems employ movement sensors for the same purpose. Sensors are easier to apply and give a very clear signal that is easier to score, but for reasons unknown EMG still dominates.
Most leg movement sensors on the market are acceleration sensors, which sense the actual movement of the leg. Because legs are large and heavy, and the twitches are many times small, it is possible that many events will be missed by such a sensor. This is especially true if the sensor is not positioned on the leg in its preferred, more sensitive orientation. SleepSense Leg Movement Sensors are flex sensors which are strategically strapped to the front of the foot. At that location, all the tendons that operate the toes run just under the skin, and any movement, even the smallest one, causes them to tighten and push against the skin.
SleepSense sensors can therefore produce output even when there are minute movements of a single toe and, of course, any stronger movement. This very wide range of movement amplitude measured introduce a new problem: how can you show both large movements without causing saturation and overshoot, and movements that are 100 times smaller, on the same sensitivity setting? SleepSense sensors employ a built-in non-linear network that attenuates strong signals much more than weak signals, resulting in a “compressed” view of the signal. A low-pass filter turned to leg movements spectrum in the sub-Hz range insures a clean signal.