System to stimulate phrenic nerve to treat sleep apnea

Abstract

A medical device for the treatment of Obstructive Sleep Apnea (OSA), the medical device including: a set of electrical stimulation electrodes; a stimulation circuitry configured to generate excitation waveforms; a set of sensors; a power source; and a controller running an adaptive feedback algorithm to monitor and modulate the physiology of the upper airway muscles and the diaphragm by modifying excitation waveforms generated by the stimulation circuity.

Description

SYSTEM TO STIMULATE PHRENIC NERVE TO TREAT SLEEP APNEACROSS-REFERENCE TO RELATED APPLICATION

[0001] This application claims priority to U.S. Provisional Application No. 63 / 600,474, filed November 17, 2023, the entire contents of each being hereby incorporated by reference. This application also incorporates the entirety of WO 2024 / 108078TECHNICAL FIELD

[0002] The techniques herein relate to implantable devices to stimulate phrenic nerves to treat sleep disordered breathing — such as airway collapse in patients with Obstructive Sleep Apnea (OSA).INTRODUCTION

[0003] Healthy sleep is an important part of our lives. It improves physical and mental health. Sleep happens in stages, including REM sleep and non-REM sleep. When humans sleep, their body has a chance to rest and restore energy. A good night’s sleep can help us cope with stress, solve problems, or recover from illness. Not getting enough sleep can lead to many health concerns, affecting how we think and feel.

[0004] Sleep disordered breathing is a common sleep disorder where patients have repetitive episodes of either cessation of breathing (apneas) or periods of reduced flow (hypopneas) during sleep. For patients with sleep disordered breathing, sleep is interrupted with 10 or more second periods without proper airflow, occurring hundreds of times during a typical night’s sleep. Apneas generally originate as either obstructive, central, or some combination of the two etiologies.

[0005] Obstructive Sleep Apnea (OSA) is a form of sleep disordered breathing characterized by periodic interruptions of lung ventilation that disrupt sleep due to a momentary collapse and obstruction of the pharyngeal airway. Obstruction of the pharyngeal airway can be attributed to decreased upper airway muscle tone, relaxing muscles that support the soft tissues in the throat, such as the tongue and / or soft palate. Relaxation of these muscles results in the narrowing of the pharyngeal airway, causing airflow obstruction thus limiting airflow and leading to a decrease in oxygen saturation. Central Sleep Apnea (CSA) is a less common sleep disorder that ischaracterized by apneas due to a lack of signals from the respiratory center. With CSA, thoracic neural receptors fail to send a signal to the respiratory center to initiate inspiration. As a result, airflow ceases due to no respiratory muscle activity. Mixed Sleep Apnea is a combination of OSA and CSA where there is both decreased respiratory drive and decreased upper airway muscle tone. The inventors believe that OSA may be caused by an inadequate reflex mechanism in response to an obstructed airway.

[0006] In healthy individuals, upper airway stability during sleep is ensured by coordinated and synchronized central control of the respiratory system, specifically the airway muscles, that are comprised of about twenty airway dilator and constrictor muscles. The central nervous system (CNS) pattern generator, also referred to as the respiratory control system, in the medulla of the brain receives inputs from physiologic sensors (also called receptors) via various afferent sensory nerve fibers and controls airway muscles via efferent motor fibers. These physiologic sensors provide physiologic feedback used by the medulla to trigger a reflex from the effectors in a closed loop reflex arrangement. These reflexes are known as “autonomic” since they do not depend on consciousness. In some cases, the reflexes become insufficient for optimal health during sleep.

[0007] Respiration during sleep is governed mainly by three systems the Central Neural Controller, the respiratory system itself, and the cardiovascular system. The brainstem, cortex, limbic system, and hypothalamus primarily contribute to respiratory effort from the central neural controller. The central pattern generator from the brainstem controls the periodic nature of inspiration and expiration. Three main groups of neurons located in the pons and medulla aid in generating rhythmic breathing: the medullary respiratory center, apneustic center, and pneumotaxic center. The medulla respiratory center is comprised of different groups of cells that are responsible for the basic rhythm of ventilation, including the generation of respiratory rhythm, inspiration, and expiration. These cells generate repetitive bursts of action potentials without afferent stimuli to send nervous impulses to the diaphragm and other inspiratory muscles. The rhythmic pattern of inspiration begins with an initialization of severalseconds where no activity occurs. Action potentials then occur to create a crescendo for a period of several seconds, causing inspiratory muscle activity to become stronger. The inspiratory action potentials then cease, and inspiratory muscle tone falls to its preinspiratory level. The apneustic center creates impulses that have an excitatory effect on the inspiratory area of the medulla, prolonging the action potentials, causing abnormal breathing. The pneumotaxic center regulates inspiration volume and respiration rate by inhibiting inspiration. However, a normal breathing pattern can exist without the pneumotaxic center, leading scientists to believe that this center’s role is for “fine-tuning” the respiratory rhythm.

[0008] The cortex may override the function of the brainstem in certain situations, such as hyperventilation or voluntary hypoventilation. Other parts of the brain, such as the limbic system and hypothalamic, can alter rhythmic breathing as well, due to different emotional states.

[0009] Sensory inputs to the respiratory center include signals from chemoreceptors and many distributed mechanoreceptors. Central chemoreceptors are involved with the minute-by-minute control of ventilation and react to the amount of CO2 dissolved in the blood (Pco2), but not the amount of oxygen (P02). In additional, central chemoreceptors respond to changes in hydrogen ion (H+) concentrations, where an increase in H+concentration stimulates ventilation, and a decrease inhibits ventilation. Peripheral chemoreceptors are also a type of chemoreceptor that aid in ventilation and are important for maintaining homeostasis during hypoxemia. Peripheral chemoreceptors respond to a decrease in arterial P02, an increase in Pco2, a change in H+, and a change in arterial pH.

[0010] Afferent receptors in the tracheobronchial tree and lungs detect alterations in airway pressure, temperature, air flow, and lung stretch which may be indicators of a collapsed airway. The afferent receptors provide feedback signals to the spinal cord or CNS which may respond to the feedback signals by triggering reflex responses that stimulate the upper airway muscles, which can then mitigate airway obstruction.

[0011] Some of the afferent receptors that aid in ventilation and respiration are mechanoreceptors. Lung receptors are one type of afferent receptor that provide inputscarried via the vagus nerve to the CNS to influence ventilation. Pulmonary stretch receptors are a type of lung receptor located in the smooth muscle of the airway walls that respond to changes in lung inflation. That is, these stretch receptors contribute to switching off inspiration and initiate exhalation based on how inflated the lungs are. Feedback from these stretch receptors inhibit further inspiratory muscle activity as the lungs inflate, and an initiation of inspiratory activity results in a deflation of the lungs.

[0012] Other types of lung receptors include irritant receptors, J-receptors, and bronchial C fibers. Irritant receptors are stimulated by inhaled noxious stimulus such as cigarette smoke or inhaled dust. These receptors are more rapidly adapting than stretch receptors and result in tachypnea. J-receptors also cause tachypnea, dyspnea, and apnea, as a result of events such as pulmonary edema, pulmonary emboli, pneumonia, etc. Bronchial C fibers are supplied by bronchial circulation rather than pulmonary circulation and respond to chemicals injected into bronchial circulation. Stimulation of the bronchial C fibers also results in tachypnea.

[0013] Additional receptors also can impact ventilation and respiration. These include receptors in the nose, nasopharynx, larynx, trachea. These receptors respond to, for example, mechanical and chemical stimulation — e.g., irritants. Joint and muscle receptors impact ventilation by sending signals during exercise. Receptors in the intercostal muscles and diaphragm contain muscle spindles that sense elongation of the muscle. These receptors adjust the output of respiratory muscles if the degree of muscular work has not been met or has been exceeded, helping to control the strength and degree of contraction. When unusually large respiratory efforts are required to move the lung and chest wall, dyspnea occurs due to the discrepancy between the output from the CNS controller and the amount of stretch sensed by these receptors. Arterial baroreceptors can cause reflex hypoventilation or apnea through stimulation of the aortic and carotid sinus baroreceptors. Accordingly, many afferent nerves can induce changes in ventilation.

[0014] In patients with CSA, control loops relying on neurochemical chemoreceptor sensory inputs become deranged and may be hyperactive. In patients with OSA, neuromuscular control loops relying on mechanoreceptors may become insufficientlyactive to maintain airway patency. These control loops may malfunction from decreased activity, increased activity, or delayed activity of one or more portions of the loops. For example, in chronic OSA patients, afferent receptors may gradually desensitize. The patient’s brain may fail to adjust to the gradual development of airflow obstruction. Airway blockage may occur because the brain is not receiving adequate signals from the afferent receptors indicating danger of airway blockage thus leading to an OSA event. Under these circumstances, airway neuromuscular activity no longer compensates for obstructions in the airway occurring during sleep.

[0015] The main purpose of chemoreceptors and mechanoreceptors sending impulses to the brainstem are to provide input into physiological respiratory control response housed in the brainstem, which coordinate appropriate respiration. This coordination is done between dilatory muscles in the upper airway, inspiratory muscles, and expiratory muscles. While many muscle groups contribute to inspiration, the diaphragm is the primary inspiratory muscle. The diaphragm is innervated by the phrenic nerves, which send signals to the diaphragm to contract, and forces the abdominal contents downward and forward, increasing the vertical dimension of the chest cavity. When the abdominal contents are forced downwards the chest wall expands, intrathoracic pressure is lowered, and airflows into the lungs along the pressure gradient. Expiration is more passive, with the lungs and chest wall returning to their equilibrium positions after inspiration due to the elastic stretch of the thoracic cavity.

[0016] The airway muscles that keep the upper airway open are accessory muscles of respiration that maintain pharyngeal patency during tidal inspiration. Basal tone in these muscles generally declines at sleep onset thus making the airway prone to collapse and obstruct airflow during sleep. Prior investigators suggested that patients with obstructive sleep apnea rely heavily on the aforementioned reflexes to maintain upper airway patency during wakefulness, and that the loss or decrease of reflex activation of the airway muscles during inspiration leads to increases in airway collapsibility during sleep.

[0017] Researchers have been searching for many years for the mechanism / control that provides an airway reflex response to keep the airway open during wakefulness.In other words, it is well understood that humans do not typically suffer from airway obstruction during wakefulness — even those with severe obstructive sleep apnea. Accordingly, the brain and / or other neurological systems of the human body provide control to automatically keep the airway patent during wakefulness. However, this same control is not present during sleep (at least not at the same level) — and the lack of this control can cause cases of sleep disordered breathing. It will accordingly be appreciated that techniques that can, for example, supplement the lack of control provided by the body during sleep are sought after to provide for the treatment of sleep disordered breathing.SUMMARY

[0018] Example techniques herein include methods and systems for treatment of sleep disordered breathing. In some example embodiments, the techniques include restoration of airway stability in the context of treating OSA via the stimulation of peripheral nerves involved in respiration and utilizing existing physiologic autonomic control reflex loops. For example, by selectively triggering or otherwise augmenting physiologic and autonomic control reflex loop(s), the described techniques augment and / or restore natural control of the airway stability and / or treat OSA by maintaining airway patency during sleep In certain examples, such techniques may be in contrast to direct muscle stimulation via efferent nerve fibers to open portions of the airway.

[0019] In certain example embodiments, the techniques discussed herein allow for, in the case of CSA, providing respiratory drive via the phrenic nerve to the diaphragm.

[0020] Exploratory research on human subjects was conducted where unilateral phrenic nerve stimulation was applied to sleeping patients. Initially this research was focused on evaluation of increased lung volume and caudal “stretch” of the trachea, with the goal of increased airway patency via this caudal traction What was instead determined was that bursts of neuro stimulation just before and / or during the initial inspiratory phase by a patient was adequate to open the airway and prevent occurrence of obstructive sleep apnea. It was observed that this neuro stimulation response was incredibly fast — measuring less than about 50ms. The earliness, speed, and magnitude of upper airway response were unprecedented for the lungvolume / caudal traction mechanism — so much so that it led to evaluation of physiologic pathways that allow for this stimulation to cause the reflex in the upper airway.

[0021] The body often has redundant mechanisms, especially around such critical functions such as respiration. For example, the body has both diaphragmatic inhalation muscles and the intercostal muscles of the rib cage. Both are capable, in varying amounts, to ventilate the lungs.

[0022] A reflex response of the upper airway is the negative pressure reflex. Negative pressures in the airway result in a reflex response resulting in increased EMG activity in muscles — such as the genioglossus and other airway muscles. During wakefulness, upper airway muscle EMG signals have been shown to be in phase with respiration. This indicates either brain or local reflex control of these muscles to ensure the airway is patent.

[0023] Native pathway that is believed to be responsible for this behavior. Pressuresensitive nerve endings in the pharyngeal airway send afferent signals via the superior laryngeal branch of vagus, i.e. Cranial Nerve X, to Nucleus Tractus Solitarius (NTS) 88. Resulting efferent signals originating from Nucleus Anbiguous (NA) travel via the Cranial Nerve XII to the muscles of the upper airway, such as the Genioglossus Muscle, the open up the upper airway. The afferent nerve fibers of the phrenic nerve are believed to be directly causing the above-described reflex in the upper airway. The phrenic nerve has afferent fibers. See Nair J, Streeter KA, Turner SMF, et al. Anatomy and physiology of phrenic afferent neurons. J Neurophysiol. 2017; 118(6):2975-2990. doi: 10.1152 / jn.00484.2017. These fibers include several inputs for respiratory control.

[0024] The phrenic nerve fibers originate in the C3, C4, and C5 vertebrae, which is also the (general) area that key upper airway nerves originate. Phrenic nerve stimulation has shown to increase activity of C1-C2 interneurons (Razook JC, Chandler MJ, Foreman RD. Phrenic afferent input excites C1-C2 spinal neurons in rats. Pain. 1995;63(1):117-125. doi: 10.1016 / 0304-3959(95)00026-0) possibly causing increased firing of nerves originating from C1 and C2, such as the ansa cervicalis, which provides efferent innervation to much of the upper airway musculature and is specificallyinvolved in depressing the hyoid bone, such as in swallowing (Waxenbaum JA, Reddy V, Bordoni B. Anatomy, Head and Neck: Cervical Nerves. In: StatPearls. Treasure Island (FL): StatPearls Publishing; 2023 Jan-. Available from: https: / / www.ncbi.nlm.nih.gov / books / NBK538136 / ).

[0025] The upper airway is innervated by several nerves that play roles in controlling various functions such as breathing, swallowing, and vocalization. Some of the key upper airway nerves include:

[0026] 1 . Vagus Nerve (Cranial Nerve X): The vagus nerve is the most important nerve for controlling the upper airway. It is a cranial nerve that provides parasympathetic innervation to the majority of the upper respiratory tract, including the larynx and pharynx. The vagus nerve is responsible for controlling some of the muscles involved in swallowing and vocalization, and it also regulates the constriction of airway smooth muscles, secretions, and the gag reflex.

[0027] 2. Glossopharyngeal Nerve (Cranial Nerve IX): The glossopharyngeal nerve is another cranial nerve that provides sensory and motor innervation to the pharynx and the back of the tongue. It plays a role in the reflexes associated with the gag reflex and the swallowing process.

[0028] 3. Hypoglossal Nerve (Cranial Nerve XII): The hypoglossal nerve primarily controls the muscles of the tongue. While it is not directly involved in upper airway control, it plays an important role in articulation and tongue movement during speech and swallowing.

[0029] 4. Trigeminal Nerve (Cranial Nerve V): The trigeminal nerve provides sensory innervation to the face and the upper airway, including the nasal passages, oral cavity, and part of the pharynx. It is responsible for sensations such as touch, temperature, and pain in these regions.

[0030] 5. Accessory Nerve (Cranial Nerve XI): The accessory nerve controls the muscles of the neck and shoulders. While it does not directly innervate the upper airway, it plays an indirect role in maintaining the stability of the neck and head during breathing and other upper airway functions.

[0031] Additionally, the phrenic nerve is innervated by the Dorsal Respiratory Group (DRG), which also innervates the glossopharyngeal nerve and vagus nerve. Phrenic afferents have been found to project into the reticular formation, where tonic drive for upper airway muscles such as the hypoglossal originate, and the rostral ventral respiratory group (rVRG) which also drives airway dilator functions.

[0032] Neurological reflexes are automatic, involuntary responses that occur in the body in response to specific stimuli. These reflexes help protect the body from harm and maintain basic physiological functions. The following are examples of neurological reflexes:

[0033] 1 . Stretch Reflex: This is also known as the myotatic reflex. When a muscle is stretched, sensory receptors in the muscle (muscle spindles) detect the change in length and send signals to the spinal cord, which then causes the muscle to contract in response. This reflex helps maintain muscle tone and posture.

[0034] 2. Patellar Reflex (Knee-Jerk Reflex): A classic example of a stretch reflex. When a doctor taps the patellar tendon just below the kneecap with a reflex hammer, it causes a quick contraction of the quadriceps muscle, which extends the leg.

[0035] 3. Withdrawal Reflex: When you touch something painful or hot, a withdrawal reflex causes your body to quickly pull away from the source of the pain. This reflex is mediated by interneurons in the spinal cord.

[0036] 4. Gag Reflex: When something touches the back of the throat or the soft palate, the body responds with a gag reflex, causing a protective reflexive contraction of the muscles involved in swallowing.

[0037] 5. Pupillary Light Reflex: When light is shined into one eye, the pupil constricts to reduce the amount of light entering the eye. This reflex helps protect the retina from excessive light exposure.

[0038] 6. Corneal Reflex: When an object touches the cornea of the eye, it triggers a reflexive blink to protect the eye from potential damage.

[0039] 7. Cough Reflex: When foreign particles or irritants enter the airways, a cough reflex is triggered to clear the airways and protect the respiratory system.

[0040] 8. Sneeze Reflex: Similar to the cough reflex, the sneeze reflex helps remove irritants from the nasal passages and respiratory system by forcefully expelling air.

[0041] 9. Babinski Reflex: This reflex is seen in infants and involves the fanning and extension of the toes when the sole of the foot is stroked. In adults, the normal response is toe flexion rather than extension.

[0042] 10. Blink Reflex: When something suddenly approaches the eyes, a blink reflex causes you to close your eyelids rapidly to protect your eyes from potential harm.

[0043] 11 . Swallowing Reflex: This reflex is essential for the process of swallowing, which involves the coordinated activity of muscles in the throat and esophagus to move food and liquids from the mouth to the stomach.

[0044] These are just a few examples of the many neurological reflexes in the body. Reflexes are crucial for our survival and well-being, as they allow us to react quickly to potential dangers and maintain basic bodily functions.

[0045] Reflexes can involve both local neurological signals and the brain, depending on the type of reflex and its complexity.

[0046] 1 . Local Reflexes: Many reflexes occur at the level of the spinal cord and are mediated by local neural circuits, without significant involvement from the brain. These reflexes are often known as “spinal reflexes.” For example, the knee-jerk reflex (patellar reflex) involves a sensory signal from the knee's stretch receptors that is transmitted to the spinal cord, where a motor signal is immediately sent back to contract the quadriceps muscle. The brain's involvement in these types of reflexes is minimal, if at all.

[0047] 2. Brain-Influenced Reflexes: Some reflexes, like the pupillary light reflex and the corneal reflex, involve signals that travel to the brain for handling before generating a response. In the case of the pupillary light reflex, the signal from the retina travels to the brain's visual centers (specifically the pretectal nucleus) before generating a response to constrict the pupil. Similarly, the corneal reflex involves input from the cornea to the brainstem, resulting in a blink response.

[0048] 3. Higher-Level Reflexes: More complex reflexes, such as those involving swallowing, coughing, or sneezing, are influenced by both local reflex arcs at the spinal or brainstem level and higher-level control from the brain. These reflexes often require integration of sensory information and coordination with other body systems. The brain plays a significant role in these more intricate reflexes.

[0049] In summary, the involvement of the brain in a reflex varies depending on the specific reflex and its complexity. Some reflexes are purely local, with little to no brain involvement, while others involve the brain's participation for handling sensory information and generating an appropriate response.

[0050] Accordingly, in certain example embodiments, a technique (which may be embodied in a system that includes a neurostimulation device) is provided for activating the direct afferent fibers of the phrenic nerve in a patient. This stimulation triggers a reflex response of the upper airway muscles, which keeps the upper airway in a mechanical state more similar to that of wakefulness while concurrently / simultaneously allowing the patient to sleep.

[0051] NPR is characterized by a robust and rapid (within 30-50 milliseconds) activation of pharyngeal dilator muscles when a rapid increase in negative pressure occurs via inspiration of ambient air through the nose or mouth. Such activation is presumably a protective reflex that allows the pharynx to resist closure during a potentially collapsing perturbation under conditions of increased ventilatory drive while sniffing, exercising, gasping for air, or challenges of anatomy such as excessive body weight.

[0052] In certain example embodiments, stimulation energy is applied and may include a stimulation burst initiated while the airway is closed. In some examples, a substantial proportion, e.g., greater than 50%, 75% or 85%, of the stimulation bursts may be simulation bursts initiated when the airway is closed. The stimulation burst(s) may be first applied at a first energy level sufficient to generate action potentials in the phrenic nerve and later at second energy level sufficient to evoke reflex opening of the collapsed airway by activation of upper airway muscles via potentiation of a mechanoreflex. The mechanoreflex may be a negative pressure reflex (efferent phrenicdriving a mechanical response leading to afferent negative pressure stimulus and subsequent efferent upper airway muscle response) and / or direct afferent stimulation resulting in an efferent upper airway muscle response. These two different responses may be targeted via different waveforms such as a lower frequency, higher pulse-width efferent waveform or a higher frequency, lower pulse-width afferent waveform.

[0053] The invention may be embodied as a medical device for the treatment of Obstructive Sleep Apnea (OSA), the medical device comprising: stimulation circuitry configured to generate stimulation waveforms, wherein the stimulation waveforms are configured to deliver stimulation energy; at least one electrode configured to apply the stimulation waveforms and the stimulation energy to a phrenic nerve of a patient while the patient is asleep; at least one sensor configured to sense at least one physical characteristic of the patient; and a hardware controller configured to: receive sensor data from the at least one sensor, wherein the sensor data includes information representing the at least one physical characteristic of the patient monitoring, using the sensor data, physiology of at least one respiratory muscle of the patient while asleep, generating sleep data including information determined from the monitoring of the physiology; adaptively modifying the stimulation waveform using the sleep data to generate a modified stimulation waveform to be provided to the simulation circuity to be used as the stimulation waveform.

[0054] The hardware controller may be integral with the stimulation circuitry. The electrical stimulation electrode may be a tri-polar electrode with at least three contact electrodes.

[0055] The stimulation circuitry may be configured to generate a unidirectional excitation by holding two of the contact electrodes of a non-traveling direction of the tri- polar electrode at a negative potential while delivering a positive pulse to another of the contact electrodes of the tri-polar electrode in the traveling direction.

[0056] The stimulation circuitry may be configured to generate a bidirectional excitation by holding a middle contact of the tri-polar electrode at a negative potential while delivering a positive pulse to a plurality of the at least three contact electrodes on either side of the middle contact.

[0057] The set of electrical stimulation electrodes may include at least one noncontacting electrode.

[0058] The stimulation circuitry may be configured to deliver biphasic stimulation.

[0059] The hardware controller may be configured to control the stimulation circuitry to selectively apply different electrical potentials to different contacts of the set of electrical stimulation electrodes to alter morphology of resulting electrical fields used for stimulation of target tissue.

[0060] The hardware controller may be further configured to determine a capture threshold; optionally wherein the hardware controller is configured to execute an algorithm for the detection of the capture threshold comprising: monitoring the human patient, optionally with actigraphy and / or accelerometry, to confirm that patient is supine and resting, and increasing stimulation energy in graded steps, optionally using steps of 0.1 to 0.25 mA between 0.5 and 2.5 mA, to detect the fists distinct rhythmic twitches of the diaphragm muscle.

[0061] The stimulation waveforms may be modified to have amplitudes based on the capture threshold, wherein the amplitudes are above or below the capture threshold. Stimulation waveforms may be modified by modifying at least one parameter for the stimulation waveforms based on the capture threshold, wherein the at least one parameter includes amplitude, pulse width, frequency, and / or duration of the stimulation waveforms, wherein a value of the at least one parameter is above or below the capture threshold.

[0062] The adaptive feedback algorithm may adapt to the physiology by applying a search algorithm for stimulation domain parameters.

[0063] The stimulation domain parameters may include one or more of amplitude, frequency, pulse-width, stimulation duration, or timing.

[0064] Feedback for the adaptive feedback algorithm may be one or more of transthoracic impedance, chest acceleration, laryngeal acceleration, electromyogram of upper airway muscles, electromyogram of scalene muscles, audio signals from breathing, or intramuscular pressure.

[0065] The hardware controller may be further configured to, by using the adaptive feedback algorithm, control therapy applied to a patient by reducing towards at least one of: 1) apnea hypopnea index, 2) a number of arousals, 3) a sum of the apnea hypopnea index and the number of arousals, or 4) a difference between physiological parameters measured during an awake state and during a sleep state of the patient.

[0066] The hardware controller may be further configured to, by using the adaptive feedback algorithm, control therapy applied to a patient by maximizing towards at least one of: 1) airflow, 2) upper airway muscle tone, 3) tidal volume, or 4) minute volume.

[0067] The adaptive feedback algorithm may include: determining a respiratory period of the patient; determining a stimulation period based on, but less than, the respiratory period; and controlling a timing of application of the stimulation waveforms to the electrodes based on the determined stimulation period.

[0068] The hardware controller may be further configured to: based on one or more signals from the sensor that indicate detection of obstruction of an airway of the patient, cause, via the stimulation energy to be delivered to the phrenic nerve of the patient.

[0069] The adaptive feedback algorithm may include: determining a respiratory rate of the patient, and controlling the application of the stimulation waveform to the electrode based on the determined respiratory rate, or controlling the application of the stimulation waveform to achieve entrainment at a rate based on the determined respiratory rate.

[0070] The hardware controller may be further configured to cause the stimulation waveform to be delivered at a time that is predicted from a respiratory waveform.

[0071] The hardware controller may be configured to control the stimulation circuitry to generate excitation waveforms having a pulse train frequency in a range of 20 to 50 Hz, optionally 30 Hz; to generate excitation waveforms having a pulse width in a range of 30 to 215 ps, optionally 150 ps, and / or to generate excitation waveforms having an amplitude in a range of 0.4 to 5.0 mA, optionally 1.0 to 4.0 mA.

[0072] The hardware controller may be configured to control the stimulation circuitry to generate excitation waveforms having stimulation duty cycle in a range of 30-70%, optionally 40-50%.

[0073] The invention may be embodied as method to treat Obstructive Sleep Apnea (OSA) including: monitoring a sleeping human patient to detect at least one physical characteristic of the human patient; generating a stimulation signal based on the at least one physical characteristic, and applying the stimulation signal to a nerve in the patient to activate an afferent sensory nerve fiber in the nerve which activation triggers a mechanoreflex in at least one upper airway muscle of the patient, wherein the trigging of the mechanoreflex opens or holds open an airway passage in the human patient.

[0074] The method may further monitor a response of the sleeping patient to the application of the stimulation signal; based on the monitoring, determining whether the application of the stimulation signal triggers the mechanoreflex to open or hold open the airway passage; and applying an adaptive feedback algorithm to modify the simulation signal in response to a determination that the stimulation signal does trigger the mechanoreflex.

[0075] The method may detect a capture threshold by: monitoring the human patient, optionally with actigraphy and / or accelerometry, to confirm that patient is supine and resting, and increasing stimulation energy in graded steps, optionally using steps of 0.1 to 0.25 mA between 0.5 and 2.5 mA, to detect the fists distinct rhythmic twitches of the diaphragm muscle.

[0076] In the method of claim, adaptive feedback algorithm may be configured to monitor and modulate a physiology of the at least one upper airway muscle and / or a diaphragm of the human patient by modifying excitation waveforms of the stimulation signal, or to adapt to the physiology by applying a search algorithm for stimulation domain parameters, optionally wherein the stimulation domain parameters include one or more of amplitude, frequency, pulse-width, stimulation duration, or timing.

[0077] In the method, the feedback for the adaptive feedback algorithm may be one or more of transthoracic impedance, chest acceleration, laryngeal acceleration,electromyogram of upper airway muscles, electromyogram of scalene muscles, audio signals from breathing, or intramuscular pressure.

[0078] In the method the adaptive feedback algorithm may be applied to reduce an apnea hypopnea index of the human patient; reduce a number of arousals of the human patient; reduce a sum of the apnea hypopnea index and the number of arousals, and / or reduce a difference between at least one physiological parameter measured during an awake state and during a sleep state of the patient.

[0079] In the method, the adaptive feedback algorithm may be applied to increase at least one of: 1 ) airflow, 2) upper airway muscle tone, 3) tidal volume, or 4) minute volume.

[0080] In the method, the adaptive feedback algorithm may include: determining a respiratory period of the patient; and determining a stimulation period based on, but less than, the respiratory period.

[0081] In the method, the adaptive feedback algorithm may include controlling a timing of the stimulation signal based on a determined stimulation period.

[0082] In method, the monitoring includes receiving one or more signals from one or more sensors monitoring the patient, wherein the signals indicate detection of obstruction of an airway of the patient, such as indicating transthoracic impedance.

[0083] In the method, the adaptive feedback algorithm may include: determining a respiratory rate of the patient, and controlling the application of the stimulation signal based on the determined respiratory rate.

[0084] In the method, the adaptive feedback algorithm may include controlling the application of the stimulation signal to achieve entrainment at a rate based on the determined respiratory rate.

[0085] In the method, the stimulation signal may include a pulse train frequency in range of 20 to 50 Hz, optionally 30 Hz; stimulation pulses each having a pulse width in a range of 30 to 215 ps, optionally 150 ps; stimulation waveforms having an amplitude in a range of 0.4 to 5.0 mA, optionally 1 .0 to 4.0 mA, and / or stimulation waveforms having stimulation duty cycle in a range of 30-70%, optionally 40-50%.SUMMARY OF FIGURES

[0086] FIG. 1 A illustrates an embodiment of the overall system to stimulate the phrenic nerve, which includes an algorithm, stimulator, and sensory feedback.

[0087] FIG. 1 B illustrates an example system that is used to provide stimulation to a patient according to certain example embodiments.

[0088] FIGS. 2A-2D illustrate different cuff electrodes for bidirectional and unidirectional stimulation of the target nerve.

[0089] FIG. 2A illustrates a unipolar cuff electrode for bidirectional stimulation.

[0090] FIG. 2B shows a multi-contact cuff electrode that is used for bipolar stimulation, and the biphasic stimulation waveforms, resulting in bi-directional excitation of the target nerve.

[0091] FIG. 20 shows the unidirectional nerve stimulation when a multi-contact cuff electrode is used as a nerve blocker.

[0092] FIG. 2D shows unidirectional and bidirectional stimulation phased together on a single cuff, wherein a tripolar electrode configuration is activated in two different phases to produce either high frequency unidirectional stimulation or bidirectional stimulation with high frequency tonic stimulation in one direction and an amplitude- modulated signal in the reverse direction to provide periodic stimulation.

[0093] FIG. 3A illustrates example monophasic pulse trains for stimulation of the phrenic nerve.

[0094] FIG. 3B illustrates example biphasic pulse trains for stimulation of the phrenic nerve.

[0095] Figure 4A shows a diagram of the cross-section of a nerve dissected out of embedded fascia, showing a roughly elliptical shape with additional insulating fascia to both sides, and a roughly homogeneous distribution of fibers within each fascicle.

[0096] Figure 4B shows an implementation of a tripolar cuff electrode for afferent nerve fiber stimulation.

[0097] FIG. 5 shows the resulting electric potential and the depolarization and hyperpolarization of the nerve when subjected to a configuration shown in Figure 2C.

[0098] FIG. 6A illustrates a non-contacting dipole electrode for phrenic nerve stimulation.

[0099] FIG. 6B illustrates a planar electrode array with field steering for phrenic nerve stimulation.

[0100] FIG. 60 illustrates the field steering with bipolar electrodes to capture nerves depending on their orientation with respect to the electrode positions.

[0101] FIG. 7 illustrates an injectable wireless electrode for phrenic nerve stimulation which is powered by radio frequency signal applied externally.

[0102] FIG. 8A illustrates an embodiment that utilizes a sensing electrode and a stimulation electrode.

[0103] FIG. 8B illustrates an embodiment for a leadless electrocardiogram measurement setup.

[0104] FIG. 8C illustrates an embodiment of measuring changes in respiration via impedance plethysmography and techniques for obtaining electrical impedance using existing leads.

[0105] FIGS. 9A-9C illustrate the correlation between the excitation current, the resulting voltage, and respiration waveform when measured with electrical impedancebased sensors, where FIG. 9A shows the excitation current, FIG. 9B shows the voltage waveform, and FIG. 9C shows the respiration waveform. Stimulation for impedance measurement is applied in a unipolar fashion between the electrode and the IPG can.

[0106] FIG. 10 shows the block diagram of an example stimulator.

[0107] FIG. 11A provides the flow chart of a process running on the therapy system.

[0108] FIG. 11 B illustrates a configuration where the therapeutic system is producing a uni-directional stimulation to excite the afferent nerves going to the nucleus tractus solitarius while sensors are providing the feedback.

[0109] FIGS. 12A and 12B show the flow chart of a therapeutic algorithm and the resulting timing diagram for an algorithm in the basic category.

[0110] FIG. 13A illustrates an example implementation in which an algorithm uses treatment parameters that are kept the same while feedback is used for timing.

[0111] FIG. 14A shows the algorithm for the delivery of treatment in advance of the inhalation phase.

[0112] FIG. 14B illustrates resulting timing diagram when the algorithm described in FIG.44A is used.

[0113] FIG. 15A illustrates the configuration of the therapy device where the same nerve is used as the stimulation target and source of sensing.

[0114] FIG. 15B is a flowchart for an example process to estimate the respiratory period of a process.

[0115] FIG. 16 illustrates an embodiment of an algorithm where the therapy is withheld periodically to determine if the patient is in need of continued treatment.

[0116] FIG. 17 illustrates an embodiment of an algorithm to detect and time nerve stimulation to not occur during expiration.

[0117] FIG. 18A illustrates an example implementation in which an algorithm uses treatment parameters that are dynamically optimized using feedback.

[0118] FIG. 18B shows the flowchart of an example intermediate algorithm that carries out a 1 D opti ization of a parameter of the sti ulation system.

[0119] FIG. 180 shows the process that is followed by the algorithm shown FIG.18B in a graphical format.

[0120] FIG. 19A illustrates an example implementation of a 2D optimization algorithm to be used in a multi-parameter optimization of the treatment.

[0121] FIG. 19B illustrates an example optimization of two treatment parameters when subjected to the optimization algorithm in FIG. 19A.

[0122] FIG. 20 shows the output of a faster algorithm for the simultaneous optimization of two stimulation parameters.

[0123] FIG. 21 illustrates an example implementation in which an algorithm uses treatment parameters that are dynamically optimized using an adaptive algorithm and a model.

[0124] FIG. 22 shows data from a clinical study where the respiratory waveforms as well as endoscopy images from the same patient are shown for an intrinsic as well as a stimulated breath.

[0125] FIG. 23A shows data from a clinical study where the respiratory waveforms from the same patient are shown for intrinsic as well as a stimulated breaths.

[0126] FIG. 23B shows data from a clinical study where the respiratory waveforms from the same patient are shown for intrinsic as well as a stimulated breaths.

[0127] FIG. 23C shows data from a clinical study where the respiratory waveforms from the same patient are shown for intrinsic as well as a stimulated breaths.

[0128] FIG. 24 shows exemplary respiratory waveforms for an intrinsic and a stimulated breath.DETAILED DESCRIPTION

[0129] Sections are used in this Detailed Description solely in order to orient the reader as to the general subject matter of each section; as will be seen below, the description of many features spans multiple sections, and headings should not be read as affecting the meaning of the description included in any section.Example Treatment System

[0130] Figures 1A and 1 B show an example system 100. In certain example embodiments, a system is provided that includes one or more electrical stimulation electrodes configured to stimulate an efferent and / or afferent target involved either directly in respiration or innervating a respiratory reflex, one or more sensors configured to obtain data regarding real-time or historical respiratory mechanics, and a controller that is configured to receive the data from the sensors and control the electrical stimulation electrodes based on that data. Each of these, and other associated elements, are discussed in greater detail below.

[0131] Techniques to apply stimulation to targets within the body to treat sleep disordered breathing, including obstructive sleep apnea, are numerous and varied. The techniques discussed herein may be based on any or all of: physiologic inputs, stimulation characteristics, stimulation patterns, target nerves, controller and / orelectrode configurations, different algorithm or process implementations, and other aspects as discussed in connection with the example embodiments herein.

[0132] Stimulation characteristics can include pulse amplitude, width, frequency, and electrode polarity with values chosen to enervate or suppress specific afferent and / or efferent nerve fibers within target nerves such as phrenic, vagus, and / or hypoglossal. Stimulation patterns may be constructed comprising repetitive cycles of pulses with varying amplitude profiles in response to physiologic functions. Any number of physiologic data may be monitored by sensors, such as respiration rate, respiration profile, body position, body movement, heart rate, blood oxygen, blood pressure, snoring, airway flow, ECG, etc., Such of physiologic data can be used in connection with example applied stimulation using an algorithm (e.g., to optimize).

[0133] Stimulation output adjustments in response to physiologic data may be made in many ways. The outputs may be fixed and changed only under the control of a physician based on review of accumulated physiologic. In certain example embodiments, a stimulation controller may make changes in real time based on monitored sensor data, or it may monitor sensors over time and adaptively adjust stimulation outputs. Additional system components may be used to provide further data gathering, processing, and / or analysis capabilities. In certain example embodiments, inputs may be extended to include patient compliance data, well-being surveys, and the like. In certain example embodiments, a knowledge database may be constructed using a cloud server such that predictive analytics may be applied / generated based on learnings from large patient populations containing accumulated patient therapies and outcomes.

[0134] Different treatment algorithms are discussed below in connection with the different example embodiments. The treatment delivered in connection with the example embodiment may be optimized via one or more feedback loops. For example, a pattern of stimulation pulses in repeating cycles may be applied based on feedback from a breathing profile. Adaptive feedback loops may extend sensor monitoring for longer periods of time during a sleep session or between sleep sessions. Data gathered may be transferred wirelessly to a secondary controller for additional and / orenhanced analysis and / or long-term cloud storage for trending (e.g., in connection with a knowledge database). Results of comprehensive data analysis from multiple sources may be used to identify adjustments to stimulation outputs and transferred wirelessly back to an example controller for application in connection with therapy applied to a patient. Accordingly, a combination of feedback loops, incorporating multiple sensors and data sets, may optimize therapy efficacy.

[0135] Next, the respective components of the example system will be described in detail.1 ) Description Of Electrical Stimulation Electrodes:

[0136] Electrical stimulation electrodes can be wired or wireless. Wired electrodes are attached to the stimulator via electrical connections while the wireless electrodes are injected to the vicinity of the target stimulation site.

[0137] Wired stimulation electrodes can be one of two types, namely the contacting electrodes and non-contacting electrodes.

[0138] Contact electrodes are usually in the shape of cuff electrodes surrounding the nerve that is being stimulated. They can be configured to be unipolar or multipolar.

[0139] In certain example embodiments, unipolar cuff electrodes 2401 can be used. Such cuff electrodes, as shown in Figure 2A, are designed to provide stimulation to the target tissue, usually the nerve 2402, by a single contact 2404 where the return electrode 2406 is located at a distance. The return electrode 2406 for an implanted system is usually the case of the stimulator, although it can be a different electrode as well. Unipolar electrodes 2401 may have the advantage of being simpler to design and implant while also being smaller in size. However, they can suffer from the disadvantage of not being specific for stimulation or sensing. In other words, inadvertent stimulation of the tissue between the cuff electrode 2401 and the return electrode 2406 — e.g., the implanted device — is possible. Similarly, it is possible to sense signals that are not generated by the target nerve 2402, but caused by the sources that are in the vicinity of the unipolar cuff electrode 2401 or near the return electrode 2406. Furthermore, any stimulation that is applied to the unipolar cuffelectrode 2401 will generate excitation 2407a and 2407b in the target nerve 2402 which will travel bidirectionally, preventing the ability to select the excitation 2407a and 2407b of the afferent or efferent nerve fibers only.

[0140] Multipolar cuff electrodes 2408, as shown in Figure 2B, are designed to provide stimulation to the target tissue, usually a nerve 2402, by multiple contacts. In the case that the multipolar cuff electrode 2408 contains three electrodes 2409a, 2409b, 2410, stimulation is generated such that the two outer contacts 2409a and 2409b carry an electrical potential that is opposite of the potential of the central contact 2410. For example, when the central contact 2 2406 is held at a positive potential, the outer contact 1 2404a and 32404b are held at a potential that is equal in amplitude, but opposite in sign, i.e. , negative. This type of design concentrates the stimulation to the nerve 2402 that is being targeted and can reduce the possibility of the stimulation of unintended tissues.

[0141] Stimulation that is delivered to the contacts can be monophasic, as illustrated in Figure 3A or biphasic, as illustrated in Figure 3B. Figures 3A and 3B also illustrate the parameters of the stimulation waveforms, which as shown as the period 4002, amplitude 4004, pulse width (PW) 4006 and train duration 4008. Furthermore, the stimulation amplitude may be given as the magnitude of the voltage or the current that is being delivered to the target tissue, nerve in this case. In monophasic stimulation, waveform that is applied to each contact has a single phase, and it is not alternated. In a biphasic stimulation, as shown in Figure 3B, the waveform has two phases, and alternates once for each cycle The use of biphasic stimulation can be more advantageous in certain examples since it reduces the chances of electrode corrosion and the potential damage to target tissue by limiting the free radicals and excess of ions that are being gathered around the contacts.

[0142] The phrenic and vagus nerves are located in the neck. Figure 4A shows the cross-section of a nerve 5102 dissected out of embedded fascia 5104, showing a roughly elliptical shape. Figure 4B illustrates an exemplary design of a tripolar cuff electrode which can be used for the stimulation of the target afferent nerve fibers. Since the efferent fibers are generally easier to recruit than afferent fibers owing to theefferent fibers’ larger diameter and thus higher potential difference across the cell for a given electric field gradient, an efficient stimulation cuff design may be key in connection with the selective targeting of afferent fibers. Additionally, in mixed nerves such as the phrenic nerves, afferent and efferent fibers are evenly distributed. This may make techniques related to current steering not as useful.

[0143] In the example design shown in Figure 4B, the cuff has an elliptical interior 5202 with three electrodes placed in the long flap on the bottom of the cuff 5204. For the cervical phrenic nerve application, the long flap of this cuff may lie against the anterior scalene muscle which the phrenic nerve is immediately superficial to. By placing the electrodes on the bottom 5206 of the cuff, the phrenic nerve will tend to lie directly against the electrodes, reducing capture thresholds and extending battery life of the implanted stimulator. The tripolar design with the center electrode of one polarity and the outer electrodes of the opposite polarity reduces leakage current outside of the cuff. This can advantageously extend battery life in certain instances.

[0144] The ellipsoidal interior fit of the electrode to the nerve, as shown in Figure 4B, reduces the variability in capture threshold as the nerve cannot migrate as far from the electrodes as with a circular interior for a given size of nerve. Additionally, the electrodes on the bottom 5206 of the cuff avoid stimulating against the fascia the nerve was dissected out of, further reducing stimulation thresholds in hope of better capturing afferent fibers with low current and voltage and extending stimulator battery life.

[0145] Multipolar cuff electrodes 2408 can be used for bi-directional or unidirectional stimulation. When bidirectional stimulation of a nerve 2402 is desired, electrical potentials as indicated in Figure 2B are applied to the contacts of the multipolar cuff electrodes 2408, where the inner contact 2 2410 is held at a potential 2412 that is opposite of the potential of the potential 2414a, 2414b of the outer contacts 1 2409a and 32409b. Furthermore, the bidirectional stimulation of the nerve 2402 is generally initiated by the negative phase 2414a, 2414b at the outer contacts 2409a, 2409b and the positive phase 2412 at the inner contact 2410, which is further illustrated in Figure 2B. Resulting action potential 2416a and 2416b would travel in either direction, capturing both the afferent and efferent nerves.

[0146] To generate a unidirectional nerve stimulation 2418, electrical potentials 2419a, 2419b, 2419c as illustrated in Figure 20 are applied to the multipolar cuff electrode 2408, which results in the electrical field that is illustrated in Figure 5. For the generation of unidirectional stimulation, two contacts 2420a and 2420b on the nontraveling direction of a three-contact electrode are kept at a negative potential 2419a, 2419b while delivering a positive pulse 2419c to the contact 2422 on the traveling direction 2418, as illustrated in Figure 2C. This pattern allows the depolarization of the nerve 2402 on the travel direction 2418 while keeping the segments of the nerve 2402 on the non-travel direction hyperpolarized. Resulting electrical potentials as well as the segments that hyperpolarized and depolarized are shown on Figure 5, where the action potential would travel only in the direction of depolarization. Unidirectional stimulation of the nerve 2402 allows the selective capture of afferent or efferent nerves in a bundle that the cuff electrode 2417 surrounds. Further variations of the stimulation patterns are illustrated in Figure 2D.

[0147] Non-contacting electrodes 2604 are the ones that are placed in the near vicinity of the target tissue, e.g., the nerve 2602, but they do not get in direct contact with the target. They can be linear, as shown in Figure 6A, or planar 2702 as shown in Figure 6B. It will be appreciated that the two contact embodiment shown in Figure 6A is provided by way of example, and that other example embodiments may include one, two, three or more contacts. Embodiments beyond three contacts are functionally similar and functionally not different from one, two, or three contact embodiments when arranged linearly.

[0148] Advantages of non-contacting electrodes is the ease of placing them during surgery. They allow the electrodes 2604 to be brought to the close vicinity of the target tissue, without being so close that there will be a chance for damaging the target tissue. In the case of the target tissue being a nerve 2602, as the electrodes 2604 do not surround the nerve 2602, long term damage to the nerve 2602, such as nerve pinching or demyelination of the nerve, can be avoided.

[0149] In certain example embodiments, another advantage of non-contacting electrodes is their ability of steering of the electrical field 2606, 2704 that is generatedfor the stimulation of the target tissues, as illustrated in Figures 6A and Figure 6B. By selectively applying different electrical potentials to different contacts 2608, 2610, 2706, 2708 in the electrode 2604, 2702, one can alter the morphology of the resulting electrical fields 2606, 2704 used for the stimulation of the target tissues, such as the nerves 2602, 2708 or the muscles. This brings two clinical advantages: First, the field steering mitigates the issue of electrode migration. If the electrode migrates following the implant, the device can be reprogrammed, either automatically or manually, to generate a different stimulation pattern to restore the ability to capture the target tissue. The second advantage is the field steering the ability to capture nerves according to their relative orientation with respect to the resulting electrical field. For example, nerves that are located parallel to the stimulation electrode can be captured by adjusting the gradient (e.g., volts per meter) of the resulting electrical field. On the other hand, nerves 2708 that are at an oblique angle to the electrode 2706, 2708 can be captured by adjusting the magnitude of the electrical stimulation, as illustrated in Figure 6C.

[0150] Above discussion so far has described the wired electrodes which are connected to the stimulator by electrical wires, which are constructed using metals or conductive polymers. But, alternatively, or additionally, wireless electrodes can be used in connection with certain example embodiments.

[0151] Wireless electrodes, which are sometimes referred to as injectable electrodes or injected electrodes 2802, are those that do not have a wired connection. Instead, they are self-contained and generate stimulation autonomously or under the control of another device 2804, as shown in Figure 7. Injectable electrodes 2802 have at least two electrical contacts and operate in 2806, 2808 bipolar configuration. The stimulation they generate could be mono-phasic or biphasic, and could be fixed or programmable. Although some contain batteries, most of them have a radio frequency (RF) receiver coil 2810 within them and use externally transmitted RF to power themselves and generate the stimulation waveform. The exemplary implementation illustrated in Figure 7 shows a wireless connection to the injectable electrode 2802 is achieved by the use of an external transmitter coil 2812 worn by the patient 2814. Anexternal device 2804 that powers and governs the overall operation of the system is also worn by the patient 281 . Due to their small size and ease of implantation, Injectable electrodes 2804 are preferred for cases where the surgical access to the target organ is challenging.

[0152] Other types of electrodes and stimulation devices, such as magnetic stimulation devices may be used in connection with certain example embodiments.2) Description Of Sensors:

[0153] The example system includes sensors for the monitoring of the physiological signals from the patient. These sensors can include, but are not limited to, electrocardiogram (ECG) sensors, electromyogram (EMG) sensors, nerve monitors, inertial sensors such as accelerometers, auscultatory sensors and microphones, electrical impedance sensors, ultrasonic sensors, temperature sensors, pressure sensors, microwave sensors, and the like.

[0154] ECG sensors are used for the detection of the cardiac rhythms and eventual extraction of information, such as the heart rate. ECG signals can be obtained from the leads 2202a, 2202b of the stimulator 2204, as shown in Figure 8C, or from the sense electrodes 2002a, 2002b, 2002c placed on the stimulator 2004, as shown in Figure 8A, which is referred to as leadless ECG. In all cases, the ECG signal is used to assess the cardiovascular component of the respiratory effect, as the beta sympathetic and para sympathetic efferent pathways from the central nervous system (CNS) directly affect the heart rate. ECG signals are usually in the frequency range of 0.05 Hz to 100 Hz, and can be detected using analog or digital circuits. ECG signals can also be used to determine heart rate that can indicate periodic breathing, sleep, and rest state.

[0155] Electromyogram (EMG) sensors detect the electrical signals resulting from the muscle activity and muscle contractions. This information is used for the detection of the contraction and relaxation of the diaphragm, rib muscles and muscles in the neck (axillary breathing muscles) and muscles of the upper airway. Accessory muscle use, defined as inspiratory contraction of the sternocleidomastoid and scalene muscles, is associated with severe obstructive disease as well as hyperpnea andexcessive effort associated with airway occlusion. The EMG frequency ranges vary from 0.01 Hz to 10 kHz, but the most useful and important frequency ranges are within the range from 50 to 150 Hz.

[0156] Nerve monitors that are discussed herein are for the detection of nerve activity in the nerves of interest, including, but not limited to, the vagus nerve, phrenic nerve, and the hypoglossal nerve. Amplitude of the signals measured are in the microVolt range and correspond to the action potentials. Since the nerve signals — e.g., the evoked compound action potential (ECAP) generated by the motor neurons — are easier to detect, certain example embodiments use them more commonly. For example, nerve signals may be used in connection with the detection of the activity of the phrenic nerve for the contraction of the diaphragm and subsequent application of the stimulation to excite the Nucleus Tractus Solitarius (NTS). This, in turn activates Nucleus Anbiguous (NA) and increases the tone in Genioglossus Muscle via the activation of Cranial Nerve XII, also known as the Hypoglossal Nerve.

[0157] In certain example embodiments, accelerometers may be used. Accelerometers are used for the detection of onset of sleep, sleep position, body motion during sleep and respiratory activity as well as respiratory effort. Patient activity and position during sleep is different during sleep state versus awake state. During sleep, torso of the patient is positioned horizontally, which is detected by a 3D accelerometer monitoring Earth’s gravity. This information is used to detect the sleep state.

[0158] Some patients suffer from a condition known as the positional sleep apnea, meaning that their apneic events occur more frequently during certain body positions during sleep. Accelerometers can sense the position of sleep, such as on side, supine or prone, and allow the therapy to be turned on or turned off depending on patient need. Ability to turn off the therapy when not needed not only increases patient comfort and the tolerability of the therapy, but also prolongs the battery life of the implanted device.

[0159] In certain example embodiments, accelerometers are used to gather information about the respiratory activity or respiratory effort of the patients. Anaccelerometer that is within the implanted device would pick up the chest motion resulting from the respiratory activity. It would also allow the ability to distinguish between respiratory effort resulting in inhalation versus breathless activity, which would be an apnea. Accelerometers that are incorporated into the stimulation electrode and placed in the neck region would gather the activity of upper airway muscles, which is main target according to certain example embodiments, and provide feedback for the therapy.

[0160] In certain example embodiments, microphones are used to allow the detection of the sounds relating to the sleep. For example, if the patient is snoring, the implanted device interprets this as signs of successful inhalation and exhalation.Microphones can be placed on the chest or the neck region, and can be part of the implanted system or external. Analysis of the snoring sounds provides additional information that can be used by the treatment system. The frequency range of simple waveform snoring typically starts at 180 Hz and peaks at 300 Hz The frequency range of complex waveform snoring typically begins at 60 to 130 Hz, with internal oscillations ranging up to 1 KHz. The higher the frequency, the greater the obstruction of the upper airways. Hence, in certain example embodiments, the stimulator interprets the higher frequency content of the snoring sounds as it adjusts the stimulation parameters to reduce upper airway obstructions.

[0161] Figure 8B shows the example implementation of an electrical impedance sensor 2102 for the detection of transthoracic impedance 2104 of the patient.Electrical impedance 2104 is measured between the sense electrode 2106b and the case of the device 2102. As the patient 2108 inhales, the lungs fill with air, which is not conductive, and the transthoracic impedance 2104 increases. Upon exhalation, the transthoracic impedance 2104 decreases, as the amount of air decreases.

[0162] Figures 9A, 9B, and 9C show data output graphs from an exemplary implementation of an electrical impedance sensing system with current, voltage, or impedance graphed (as all are related via Ohm’s Law). Either voltage or current will be set by the pulse generator, while the other one will be measured, and the impedance is able to be derived with no additional information needed. Figure 9A shows theelectrical current waveform that is used as excitation. This excitation is below the stimulation threshold of the muscles and the nerves in the chest — and accordingly does not cause any contractions. It can be applied at a low frequency, such as about 16 Hz, and its amplitude is kept low, such as around 50 micro-Am peres, with a short duration, such 0.5 milli-seconds or less.

[0163] Figure 9B shows the resulting voltage waveform that can be observed, caused by the respiration signal illustrated in Figure 9C. The amplitude of the detected voltage shown in Figure 9B would be positively correlated to the respiratory waveform 2302. The nominal value of the chest impedance is about 500 Ohms and the respiration causes a change in the impedance in the range of 0.5 Ohms to 3 Ohms, or more, depending on the electrode configuration.

[0164] Pressure sensors can be used for the monitoring of the pressures in the chest or in the neck region, and can be used to determine if and when a contraction and relaxation begins Pressure sensors placed in the neck region provide additional information regarding the muscle tone. Furthermore, the high frequency content of the pressure signals can be used as sound signals, to detect breathing or snoring sounds, as it was described above for the case of the microphones.

[0165] In certain example embodiments, ultrasonic sensors may be used for the detection of distance between the sensor using the time-of-flight technique, where the sensors measure the time for a short burst produced by one sensor to arrive at the other sensor. This signal conveys two forms of information, namely the distance between the sensors and the media between the sensors. Distance between the sensors is proportional to the time that the transmitted ultrasound wave takes to travel from one transducer to the other. Attenuation that the received signal experiences is related to the nature of the tissue in between the transducers. For example, inhalation would expand the chest and increase distance between the transducers, which in turn would increase the time for the ultrasound pulse to travel from one transducer to the other. Transducers placed in the neck region would detect the muscle tone as changes in the signal attenuation, as the ultrasound signal would travel with less attenuation through muscles that are in contraction.

[0166] Microwave signals that are generated by external devices can be used for the imaging of the tissues, and the results can be conveyed to the stimulator, which in turn uses the information to further optimize an example treatment algorithm.

[0167] Additional sensors that are not listed above can be used in certain example embodiments for the detection of the tissue oxygen saturation, respiratory activity, nerve activity as well as muscle activity.3) Description Of Stimulation Circuitry:

[0168] Figure 10 is a block diagram of an example of stimulation circuitry 4902 that may be used in connection with the example treatment system.

[0169] Pattern generator 4904 produces the output waveform that is specified by the controller. A sample output waveform is shown in Figure 2B. Pattern consists of a train of pulses where the amplitude, pulse width, pulse frequency and number of pulses are specified by the controller. Upon triggering by the controller, the pattern generator 2904 produces the pattern and delivers it to the output circuitry 4906.

[0170] Charge pump 4908 generates the voltage that is necessary for the stimulation. Although the battery of the implant may be set at 2.8 Volts or 3.7 Volts, pattern generator 4904 may need a different stimulation voltage. Hence the charge pump 4908 acts as a DC-to-DC power converter to increase or decrease the voltage and to produce the necessary power to the output circuitry 4906. Charge pumps 4908 are capacitive in nature, but the inductive type Buck convertors may also be used according to certain example embodiments. Utilization of the charge pump 4908 to generate the exact voltage needed by the output circuitry reduces the power waste and increases the battery life of the implant.

[0171] Safety limiter 4910 is used to assure that the output stimulation that is delivered to the patient adheres to the requirements. For example, if a charge balance waveform is desired, then the safety limiter 4910 circuit contains one or more capacitors to make sure that no DC current can reach to the patient. Similarly, the safety limiter may have Zener diodes to prevent voltage overload, or timer circuits, such as monostable vibrators, to prevent stimulation at too high frequencies or for extended periods, due to runaway firmware in the controller.

[0172] Other features of the stimulation circuitry 4902, such as defibrillation protection, electrostatic discharge (ESD) protection, electromagnetic immunity (EMI) and compatibility with magnetic resonance imaging (MRI) can be used in connection with certain example embedments.

[0173] Stimulation can be applied during any or all of the following instances: 1 ) Continuously; 2) Continuously except during exhalation; 3) Only during the inhalation time (Positive Delay); 4) Only during the pre-inhalation time (Approach or Negative Delay).

[0174] Stimulation can be applied to any of the following: Phrenic nerve; 2) Vagus nerve; and / or 3) Hypoglossal nerve.

[0175] Stimulation can be applied in any or all of the following formats: 1 ) Unipolar or Bipolar; 2) Uniphasic or BiPhasic; 3) Subthreshold or Over Threshold.

[0176] Stimulation can be applied in the following applications: 1 ) Bidirectionally to capture both the afferent and efferent fibers; and / or 2) Unidirectionally to capture only the afferent or only the efferent fibers.4) Description Of Controller:

[0177] In certain example embodiments, the controller governs the operation of the implant. In some examples the controller may include or be coupled to the stimulation circuitry that is discussed herein. The controller monitors the signals coming from the sensors, interprets and / or otherwise processes such signals, and then instructs the stimulation circuitry to generate the desired stimulation waveforms based on an algorithm (e.g., process) that is being perform ed / executed by the controller.Examples of different algorithms that may be implemented in conjunction with an example controller include those described in connection with figures discussed herein.

[0178] The controller may include a hardware processor, memory, and program instructions stored therein. The program instructions may include instructions for causing the processor to perform operations in connection with any or all of the steps associated with, as noted above, any of the algorithms discussed herein.

[0179] Figure 1 B is a block diagram of the system 100. The system 100 can include sensors 1604, the implantable pulse generator (IPG) 1606, electrodes 1608,patient programmer 1610, controller 1612, signal processor 1614, and / or stimulator 1616. The system 1610 operates to receive signals in connection with the patient 1618 and applies therapy to the patient 1618. Illustrative examples of sensors 1604 and electrodes 1608 of the system 100 are described above, and the following sections will describe example algorithms that may be programmed, run, or implemented on the controller 1612.

[0180] It will be appreciated that the example controller 1612 shown in Figure 1 B is provided as part of an IPG 1606 by way of example. In other examples, the controller and processing performed thereby may span multiple hardware processors that operate to carry out one or more steps of any or all of the algorithmic processes described herein. As an illustrative example, an example IPG may include a processor that performs some processing and that processor may be in communication (e.g., wireless communication) with another processor that is outside the IPG. In certain example embodiment, the controller may include multiple separate hardware processors that are distributed from each other. For example, a first processor may be included along with an IPG, another processor may be provided via a bedside computing device, another process in a mobile device (e.g., a mobile phone), and another processor may be part of a cloud computing platform. Collectively these multiple processors may form a controller according to certain example embodiments — e.g., a distributed controller. Any or all of the processing that is performed by a controller as discussed herein may be performed by such a controller that is distributed between two or more different processors.

[0181] A flow chart of an example algorithm is shown in Figure 11 A. When the algorithm begins 3201 , the device briefly alternates between two states, namely SLEEP 3202 and NO SLEEP 3204. Program starts in the NO SLEEP state 3204, and remains there until it is in the SLEEP state 3202. Switching from the NO SLEEP to SLEEP state 3206 can occur under different conditions. In some example embodiments, the patient can indicate that he or she is entering a SLEEP state or exiting it. This may be accomplished by using, for example, the patient programmer. In certain example embodiments, alternatively, or additionally, sensors can be used forthe detection of the SLEEP state. For example, when the patient remains at a horizontal position with minimal activity for a given period of time, e.g., 10 minutes, SLEEP algorithm 3206 enters the SLEEP 3202 state. Upon entering into the SLEEP 3202 state, device produces therapy 3208 and continues to deliver it until the patient exits the SLEEP state 3210. Exiting the SLEEP state 3210 can be indicated by the patient using the patient programmer or detected via sensors.

[0182] An exemplary therapy configuration is shown in Figure 11 B. Signals obtained from one or more sensors 3212 (examples of which are discussed herein) are used as the input and / or feedback for the stimulator 3214.

[0183] As an illustrative example, in the case of using an electrical impedance sensor, the onset of inhalation 2304 can be detected as illustrated in Figure 9B (e.g., as the inhalation manifests itself as an increase in the transthoracic impedance). This data may be communicated to the signal processor 3216, which may be an amplitude detector in the case of electrical impedance sensor, to produce the inhalation indicator. This indication is then communicated to the controller 3218 which may then instruct the stimulator 3214 to generate the desired stimulation 3220 to the nerve 3222 by way of an electrode 3224.

[0184] Figures 12A and 12B further illustrate the details of operation according to certain example embodiments. In Figure 12A, the onset of inhalation 3202 is detected, and a burst stimulation 3904 is delivered following a delay period of T1 3906. In some examples, the processing shown in Figure 12A allows for the controller to be kept in a lower power (or off) state until inhalation is detected. Once the stimulation is delivered and exhalation 3908 of the patient is detected, then the cycle repeats. Figure 12B illustrates the resulting stimulation pattern 3912 and its relationship (e.g., time dependance) to the respiration waveform 3910 over time.

[0185] Different modes of operation for the controller can be used depending on which algorithm is to be used. Algorithms described herein include those that have been labeled, for ease of reference, as basic, intermediate, and advanced.

[0186] Figure 1A illustrates the operation of the system 100 using a basic (e.g., first) version of the algorithm, as illustrated in Figure 13A. In this mode of operation,the system parameters, such as the delay and amplitude for the stimulation are preprogrammed, and the system operates in a monotonic fashion, as illustrated in Figure 1A. Although the stimulation parameters of the system remain unaltered, the system still uses the feedback 1604 from the sensors sent to the controller 1605, as shown in Figure 13A to adjust the timing of the stimulation 1606 as well as the entry and exit from the SLEEP state, as described above and illustrated in Figure 12A, while monitoring various parameters, such as AH1 1610

[0187] In certain example embodiments, stimulation may be delivered in advance of the inhalation 4402a, 402b by the patient. In other words, the delay 4404 (e.g., shown in Fig. 12A) may be a negative number. This approach is illustrated in Figure 14A and Figure 14B. In this implementation, the algorithm continuously monitors the respiration cycle, and delivers the therapy 4406 in advance of the next inhalation period 4402b, as shown as the time interval APPROACH 4408 in Figure 14B.

[0188] As shown in Figure 14A, the process starts 4410 with setting / initializing a Period parameter 4412 to, as an illustrative example, 5 seconds the onset, although any other value in the range of 1 to 20 seconds could be used. Afterwards, the algorithm waits until detecting an inhalation 4402a, and records the time as T1 4414.

[0189] When an elapsed time reaches to T1 + Period - APPROACH 4404, the algorithm delivers 4406 the therapy. Note that the trigger point in time in relation to the breathing cycle of the patient will usually (e g., always) be in advance of the patient’s next inhalation cycle.

[0190] Next, the algorithm waits to detect the onset of next inhalation cycle 4402b. Based on detection of the next inhalation cycle 4402b, the estimate of the Period is corrected / updated as Current time - T1 4416 before entering the next cycle. It will be appreciated that this type of implementation may be advantageous in certain example embodiments as it can proactively enhance the muscles of the upper airway in advance of the inhalation.

[0191] Figure 15 shows the implementation of the system using the signals from nerve monitors to generate the feedback. Although the illustration shows the nervethat is monitored is also the nerve that is being stimulated, other implementations are also possible in which, for example, a first nerve is stimulated while monitoring the action potentials or nerve activity in another nerve. Advantage of this embodiment is the reduced number of sensors and direct monitoring of the nervous activity, which is the target of the therapy.

[0192] Figure 16 illustrates an embodiment where the therapy is halted periodically to determine if the patient is continuing to have apneas. Although algorithm specifies that the therapy is halted every 20 minutes for monitoring the native activity of the patient for 20 seconds 4602, use of other durations is also possible.Algorithm allows the setting of a timer 4604, T1 , to deliver therapy for 4 minutes in this example. Therapy is turned off 4606 if patient activity is detected 4608 or the timer T 1 runs out 2609. Once the therapy is turned off 4606, patient is monitored for the next 20 seconds using the timer T2 4602 to see if he or she is able to breathe on his or her own 4610, or starting to have apneas. If successful respiration without apneas is observed, timer T2 is reset 4612 for additional monitoring without intervention. If the timer T2 runs out, then the situation is interpreted as an apneic event, and the therapy is restarted for another four minutes 4604. This embodiment can reduce the amount of therapy delivered, which in turn enhances battery longevity while reducing patient discomfort.

[0193] Yet another embodiment of the algorithm is shown in Figure 17. In this configuration, therapy is delivered at all times of the respiratory cycle except when it coincides with the period of exhalation by tracking the onset of exhalation 4702 and completion of exhalation 4704. This mode of operation maximizes the amount of therapy 4706 delivered to a patient to maintain the patency of the upper airways while avoiding the stimulation of the phrenic nerve and the resulting contraction of the diaphragm during the exhalation period. Such an implementation is suitable for subjects who need almost continual treatment due to their disease state.

[0194] Algorithms mentioned up to this point have described the delivery of stimulation with fixed values of parameters used, such as the delay time, approach time or the stimulation amplitude, even though there were methods to activate and de-activate the therapy, either by patient control, or by automatic means, such as sleep detection or apnea detection. While this type of implementation using fixed parameter values is categorized as the basic category —other implementations are also possible.

[0195] When the parameters of the stimulation, such as the stimulation amplitude, stimulation duration, onset delay or the approach time are governed by the algorithm itself, they are grouped into the category of intermediate and / or advanced algorithms.

[0196] For implementation in the intermediate category, the parameters of the stimulation waveform are modified by the algorithm running on the controller 4802, as shown in Figure 18A. This is done by implementing an optimization algorithm on the controller 4802. For optimization, one can choose one of many objective functions to optimize over a set of variables. Some of the objective functions that can be used for optimization are listed below:Objective function 1 = Apnea Hypopnea Index (AHI);Objective function 2 = Apnea Hypopnea Index (AHI) - Muscle tone;Objective function 3 = Number of arousals in a given period; and Objective function 4 = Patient reported discomfort.

[0197] List shown above is for illustrative purposes, and other possible objective functions can be used.

[0198] Figure 18B provides the flow chart for an example implementation of an optimization algorithm with optimization of a single parameter. Although the example is given for the optimization of the delay time before the stimulation applied following the onset of the inhalation. Similar optimizations may be performed for stimulation amplitude, stimulation duration, approach time, and the like.

[0199] Algorithm shown in Figure 18B starts with the definition of a range for the variable to optimize (the delay in this example), which is indicated as D_A and D_B 4102, representing the minimum and maximum values for the parameter. Therapy 4103a is applied at both values for the subject parameter, and the resulting outcome — e.g., the values of the objective function — are stored in variables labelled as F_A 4104a and F_B 4104b respectively. At this point, the algorithm calculates a new valuefor the delay parameter as the mid value between D_A and D_B, which is labelled as D_C 4106, and the corresponding outcome value of F_C 4108 is determined by applying the therapy 4103c at the delay setting of D_C. Now the algorithm decides if the delay value of D_A or D_B should be eliminated. If resulting outcome at D_A, which is F_A, is larger than the resulting outcome at D_B, then the F_A is larger 4110, hence the delay value of D_A should be eliminated by replacing it with D_C 4112. In the alternate situation, D_B is replaced with D_C 4114, since B is the largest 4116. In either case, the width of the search interval is reduced to half of what it was before, and the steps are repeated until there is no further improvement. Figure 18C shows the operation of the 1 -dimensional optimization in a graphical format 4102. Although the above algorithm described the operation in the form of a bisection method — e.g. , D_C 4106 is set to the half way point between D_A and D_B, other ratios can be chosen according to certain example embodiments.

[0200] Figure 19A shows an algorithm where multiple parameters of the stimulation system are optimized. In this case of 2D optimization, one of the 1 D optimization techniques, such as the bisection search as illustrated in Figures 18B and 18C will be used iteratively, as illustrated in Figures 19A and 19B. Figure 19A shows the flow chart of a 2D linear optimization algorithm where the two parameters of the stimulation system, namely the delay and the amplitude are optimized successively. The system may keep the delay constant, and optimize the amplitude 4202, by bringing the operating point from one that is labelled as 1 4204a in Figure 42B to the one that is labelled as 2 4204b. Then, the amplitude is kept at the optimal value while optimizing the delay 4206, which brings the operating point from that is labelled as 2 4204b to the one that is labelled as 3 4204c in Figure 42B. Algorithm repeats the steps until the improvement in the objective function falls below a preset value, shown as MaxLimit 4208 in the flow chart that is represented in Figure 19A.

[0201] Algorithm that is illustrated in Figure 19B tends to find the optimal operating point by optimizing the values of the stimulation parameters, but may be considered to be slow. Operation of an alternative algorithm is shown in Figure 20 where operating point 4302 is moved in both directions simultaneously. In this case,the direction of change, or the downward slope is estimated around the operating 4302 point by slightly changing the operating parameters. Then, a search is carried out along the direction of the downward slope 4304. Steps for such an algorithm are shown below in the form of pseudo code:00202] Above examples were given for implementation of algorithms with 1 D anc 2D optimization techniques as illustrated in Figures 18A, 18B, 19A, 19B, and 20. Other 1 D optimization techniques, such exhaustive search and quadratic approximation search, as well as other 2D optimization techniques, such as Ameoba and random search may be implemented according to certain example embodiments. Algorithms in the intermediate category differ from the ones in the basic category for their ability to optimize the parameters of the stimulation based on the patient outcomes which were listed as objective functions above.

[0203] Algorithms in the advanced category may include additional tasks related to optimization of the stimulation parameters step further by adding a model of the overall system and tracking it as it evolves. Such algorithms can be adaptive in nature and may be implemented in different ways. Overall structure of the system is shown in Figure 21.

[0204] Algorithms running in the advanced category utilize a model of the system, which uses various methods to model the system 4801 as closely as possible. As shown in Figure 21 , output of the stimulator 4802 is applied to the patient 4804 and at the same time, to the model system 4806. If the model 4806 were perfectly emulating the patient 4804 response, then the difference between the outcomes, shown as ACTUAL AHI 4808 and PREDICTED AHI 4810 would have been zero, which is shown as the ERROR signal 4812. However, since the model 4806 would not be perfect, there will be some residual error, which is minimized by the optimizer to modulating the model coefficients.

[0205] Optim izer for algorithms in the advanced category carries out two functions, optimization of the model and the optimization of the stimulation waveform. Once the model is optimized — e.g., the ERROR signal is minimized — then the model is believed to be representing the physiological system closely, and can be used to predict the outcomes to be obtained based on the stimulation to be applied.

[0206] Optimization of the model parameters can be as follows: Optimizer monitors and records the stimulation parameters and the patient outcomes. Once there is a sufficient data set in hand, for example 20 sets of different stimulation parameters and the resulting outcomes, the model parameters are determined. This optimization can be done by using artificial intelligence (Al) tools or statical tools. To determine model parameters using statistical tools, one can minimize the mean squared error between the actual outcome (ACTUAL AHI) and model outcome (PREDICTED AHI). This can be done using following pseudo code:

[0207] The above example has used the patient model as a simple quadratic equation model 4806 as shown in Figure 21 . It is possible to use an empirical model as a polynomial equation, or a more sophisticated based on the understanding of the physiology. Other adaptive algorithms, such as Kalman Filtering could also be used.

[0208] In some embodiments, the optimization is done via one or more cloud computing platforms such as AWS, Azure, etc. The optimization uses data from the patient or data from a group of patients. In other embodiments, the device monitors the patient physiology during the awake state to determine the model and delivers stimulation during the sleep state to replicate the awake physiology.

[0209] Algorithms as described above can be used in isolation or in combination. For example, in stimulation algorithm termed as Entrainment, the stimulation 4502 is delivered in advance of the respiration while monitoring the patient 4504 response to stimulation 4502, which is combination of algorithms shown in Figure 15A and 15B. Entrainment algorithm in the beginning determines the respiratory period by monitoring few intrinsic breaths. Afterwards, it delivers the stimulation 4502 in a fashion as illustrated in Figure 14B, where the stimulation 4502 starts a time interval labelled as APPROACH 4506 in advance of the onset of inhalation 4508. However, this stimulation is applied at a rate that is 1-5 breaths per minute less than the intrinsic respiratory rate. As long as the patient’s breathing remains as the stimulation rate and the upper airway remains open, algorithm concludes that the entrainment between the algorithm and the patient 4504 is maintained. Algorithm monitors the airway opening using the sensors 4510, such as the pressure sensor, and if it detects that the entrainment is lost, resets itself. Pseudo-code shown below further illustrates the implementation:

[0210] As appreciated in the ENT field, sleep disordered breathing (SDB) encompasses a wide spectrum of sleep-related conditions including increased resistance to airflow through the upper airway, heavy snoring, marked reduction in airflow (hypopnea), and complete cessation of breathing (apnea). These episodes of SDB can be either obstructive, central, or some combination thereof (mixed) in origin. Obstructive sleep apnea occurs due to dysfunction of the upper airway musculature wherein the airway loses patency leading to airway collapse.

[0211] Normal human physiology has autonomic mechanisms in place to mitigate against airway collapse. These mechanisms are in the form of reflexes within the body that detect conditions such as negative pressure or flow limitations and through activation of various upper airway muscles maintain patency and allow air flow. Additionally, the coordination of the UA musculature throughout the breath is also managed by these reflexes. Two such mechanisms are highlighted in the analysis below. The inventors hypothesize that airway patency and breath coordination is due to either one or a combination of these two mechanisms. First is the negative pressure reflex (NPR) which occurs during inspiration where a negative pressure is generated by air flowing from the nose through the airway and into the lungs. Second, the diaphragm is the primary driver for inspiratory activity and is innervated by the phrenicnerve. Afferent signals from the phrenic nerve likely coordinate inspiration activity with the upper airway to ensure airway patency is maintained.

[0212] During a clinical study, anesthetized subjects were instrumented with flow sensors to measure air flow and pressure sensors as well as belts around their thorax and abdomen. They were also subjected to transcutaneous electrical stimulation and observed using an endoscope, as it is done during a procedure known as Drug Induces Sleep Endoscopy (DISE). Figure 16 shows recorded waveforms and DISE images of two consecutive breaths, first one with stimulation of the phrenic nerve and subsequent one without stimulation.

[0213] Examining the waveforms shown in Figure 22, in the first breath with stimulation, it can be seen that stimulation 1002 was delivered in the early inspiratory phase. One can immediately notice that the flow waveform 1004 is not flow-limited and has much larger inspiration and expiration indicating the airway was more open and a larger volume of air was moved. Additionally, the abdominal 1006 and thoracic belts 1008 are out of phase. From the endoscope images during DISE, it can be seen that the airway while minimal during stimulation (1) 1010 is flow limited however as the inspiration cycle progresses the airway opens (2) 1012. In contrast for the nonstimulated breath, it can be seen that the breath is severely flow limited throughout the inspiratory cycle (3,4) 1014, 1016. It may be that both the NPR and coordination of the UA musculature via afferent signals from the phrenic nerve play a role in maintaining airway stability and coordination.

[0214] Figures 17A-17C show additional examples of stimulated and unstimulated breaths. In general, the opening of the airway 1102, 1202 occurred between 0.3 second and 1 .5 second following the onset of stimulation 1104, 1204. Furthermore, Figure 17C shows the effect of the stimulation on snoring, as the stimulated breath has less snoring 1302, and a non-stimulated breath has more snoring 1304, which provides additional evidence that the stimulation caused a better opening of the upper airway.

[0215] Figure 18 illustrates the air flow 1402 as well as the air pressure 1404 on the proximal side of the obstruction in a graphical format. The first breath 1406 is anintrinsic one without stimulation, and the air flow 1408 is li ited due to remaining restriction of the upper airway. Electrical stimulation 1410 resulted in the excitation of the afferent nerve pathways which in turn enhanced the air flow 1412 during the second breath by reducing the constriction of the upper airways.5) Description Of Additional System Elements:

[0216] Different types of power sources 1620 may be used in connection with the system 100 and the components thereof. Any or all of the components of the treatment system may be powered using a power source 1620 that is different or the same. In other words, each component may have its owner power source or all components (or some subset thereof) may share a power source. The power source can be a primary battery, a secondary battery, such as a rechargeable battery, a wireless charging system, a kinetic-based charging system, and / or a combination of any of these.

[0217] Primary batteries usually provide 2.8 Volts while the rechargeable batteries provide 3.7 Volts. An example wireless charging system could be powered by an RF connection operating in the frequency range of 10 KHz to 64 MHz. It may be designed to use the principles of resonant energy transfer and be used by the patient to periodically, or continuously, supply the electrical stimulation electrodes with electrical power.

[0218] The following is an illustrative table that shows how the different elements discussed herein may be combined in connection with various example embodiments.Additional Embodiments

[0219] The following are additional embodiments that may be used in conjunction with the example system 100 discussed above and its associated components.

[0220] The artificial stimulation of phrenic nerves described here may target the afferent pathways that relay signals from peripheral nerves to the central nervous system. There are several ways to express the selective stimulation of an afferent limb of a mechanoreflex including:

[0221] A) Stimulation of proprioceptive fibers since proprioception is only sensory input from joints and muscles to tell us about our movements and body position, and

[0222] B) Stimulation of large sensory fibers that conduct proprioception and other non-painful sensations (such as warmth and haptic sensations).

[0223] Stimulation of alpha fibers: intrafusal muscle fibers have high conduction velocity of 80-120 m / s. Type la is large fibers innervating intrafusal fibers. They are related to muscle spindle primary endings and determine the velocity of muscle stretch. Type lb fibers are related to golgi tendon organs. Their cell bodies are located in the dorsal root ganglia, adjacent to the spinal cord.Energy Application

[0224] In certain example embodiments, the inventors selected applications of 80-150 Hz, preferred about 100Hz pulse trains, to stimulate the nerve below the mechanical contraction level. This stimulation is believed to activate nerves from proprioceptive sensors to the spine and brain where they have inhibitory affect or reflex activating affect (such as well-known H-reflex). A mixed nerve trunk stimulation such as a mixed sensory-motor nerve has never been used this way to treat airway patency.

[0225] Waveform of a series of biphasic pulses at 100-150 Hz, 50-300 ps pulse width. Stimulation periods (pulse trains) at a frequency approximately equal to the breathing frequency (6-20 / min) to evoke reflex tightening of airway muscles.Phase locking with Breathing

[0226] The phase-locked synchronization means that sub motor stimulation of sensory nerves can entrain respiratory centers without forcing contraction of respiratory muscles.

[0227] The pattern of stimulation is based on known properties of nerves conducting afferent and efferent signals in a mixed nerve such as a median nerve. For example, the well-known gate control theory of pain first proposed in 1965, widely used to suppress pain, states that selective stimulation of large “fast conducting” afferent A- beta fibers can interfere with and suppress signals of pain conducted by larger, slower A-delta fibers when they converge at the spinal neurons and further reduce the pain conducted to the brain by projections of these neurons.

[0228] Selective stimulation of fast A-alpha and A-beta nerve fibers is possible because these neurons have lower excitation thresholds than larger A-delta fibers and C-fibers that conduct pain. Stimulation of A-alpha and A-beta fibers may create a relatively pleasant buzzing sensation in the patient’s brain.

[0229] Specifically, in the targeted muscle anatomy, the fastest, most excitable A-alpha fibers innervate primary muscle spindles, (conducting muscle rate of length change) and Golgi tendon organs (muscle tension). The slightly less excitable A-beta fibers innervate secondary muscle spindles (static length of muscle and skinmechanoreceptors). Together these groups of sensory nerve fibers are called “proprioceptive fibers”.

[0230] The much slower A-delta fibers conduct signals of sharp pain and other unpleasant sensations. Those are called “nociception” fibers. The efferent nerve fibers that are present in the same nerve bundle innervate skeletal muscles and conduct excitatory signals that command contraction. These fibers are called motor neurons.

[0231] High Frequency TENS devices use the same “gate control” concept to control pain by stimulating skin at 50-250Hz. Muscle motor fibers are excited at a much lower frequency of 8-15 Hz. Thus, it is possible to excite afferent sensory fibers in the main trunks of mixed nerves without causing motion or pain.

[0232] An alternative way to stimulate proprioception fibers without affecting nociception fibers or motor neurons is by stimulating the entire nerve bundle with a high frequency alternating current. For example, electric current frequency higher than, for example, 75-80 Hz and lower than 150-200 Hz may be suitable for such selective stimulation. Electrodes can be implanted near the nerve thus achieving selectivity and low energy consumption. The targeted nerve bundle can be a:

[0233] A) phrenic nerve,

[0234] B) pharyngeal nerves, such as likely the superior laryngeal nerve

[0235] glossopharyngeal nerve, and

[0236] C) hypoglossal nerve, vagus nerve, glossopharyngeal nerve, or trigeminal nerve (e.g., at a mandibular bra

[0237] Brain Plasticity

[0238] General idea behind motor rehabilitation induced plasticity is that muscle contraction (voluntary or electrically induced) generates reflected afferent proprioceptive inputs to the motor centers of the brain that stimulate development of voluntary or autonomic motor function.

[0239] Afferent information travels to the CNS via mixed motor nerves from biological sensors. Muscle spindles signal information about the length and velocity of a muscle, Golgi tendon organs signal information about the load or force applied to amuscle. Stress and strain receptors are involved in the control of breathing and located in airways, chest wall, diaphragm and pleura.

[0240] Nerve fibers that carry this information are the largest, fastest fibers in the nerve bundle, such as type la / lb (Aa) afferents with conduction velocities greater than 70 m / sec (72-120 m / s). In contrast muscle fiber conduction velocities are 3-5 m / s.Skin pain and tactile receptors A5 and C fibers have conduction velocity of 0.5 to 30 m / s.

[0241] Stimulation of a mixed nerve bundle at 100 Hz activates indiscriminately la and lb afferents from the muscle belly stretch receptors and tendon Golgi organs. In the past, experiments conducted with thin wire electrodes activated only stretch receptors. For example, when Gogli tendon organs were activated vibration or tapping of tendons were used. Input from Gogli organs plays crucial role in maintaining pastural stability. In summary, sensors and proprioceptors fire at a low rate in the normal range of chest motion during sleep but can be artificially stimulated at a higher, saturated rate if other fibers in the nerve bundle are not engaged. This may lead to rehabilitation of CNS components involved in sleep by induced plasticity and restoration of specific motion functionality in patients when aging or disease affects specific control of airway motor function.

[0242] Certain example embodiments may be embodied to target the afferent pathways that relay signals from peripheral nerves to the central nervous system. Such targeting may not include targeting (at least directly), efferent pathways. There are several ways to express the selective stimulation of an afferent limb of a mechanoreflex: stimulation of afferent fibers also called sensory fibers (but these also may include pain fibers); stimulation of proprioceptive fibers but only provide sensory input from joints and muscles to tell us about our movements and body position; stimulation of large, myelinated fibers, and stimulation of large sensory fibers that conduct proprioception and other non-painful sensations (such as warmth and haptic sensations) are considered the best for stimulation for OSA treatment.

[0243] An aspect of certain examples includes a method of treating obstructive sleep apnea (OSA) in a human patient comprising artificially stimulating at least onephrenic nerve of a patient at sub-motor threshold. In an aspect, the method of treating OSA may operate such that wherein the sub-motor threshold is a simulation which does not directly cause the diaphragm to contract. In an aspect, the method of treating OSA may operate such that wherein the stimulation directly stimulates afferent fibers of the phrenic nerve.

[0244] In an aspect, the method of treating OSA may operate such that wherein the stimulation is a unidirectional electrical signal applied to the phrenic nerve. In an aspect, the method of treating OSA may operate such that wherein the unidirectional electrical signal is configured to target for direct stimulation afferent fibers of the phrenic nerve. In an aspect, the method of treating OSA may operate such that wherein the stimulation is a hyperpolarized electrical stimulation. In an aspect, the method of treating OSA may operate such that wherein the hyperpolarized electrical signal is configured to target for direct stimulation afferent fibers of the phrenic nerve. In an aspect, the method of treating OSA may operate such that wherein the stimulation is synchronized with entrainment of breathing by the patient. In an aspect, the method of treating OSA may operate such that wherein the stimulation is continuous over a period of at least 10 minutes, 30 minutes, 1 hour and / or four hours. In an aspect, the method of treating OSA may operate such that wherein the stimulation is continuous and substantially constant over a period of at least 10 minutes, 30 minutes, 1 hour and / or four hours. In an aspect, the method of treating OSA may operate such that wherein the stimulation is configured to stimulate a reflex to activate airway dilators. In an aspect, the method of treating OSA may operate such that wherein the stimulation is configured to stimulate airway dilators. In an aspect, the method of treating OSA may operate by including determining the sub-motor threshold. In an aspect, the method of treating OSA may operate by including receiving, via at least one sensor, data of at least one physical characteristic of the patient. In an aspect, the method of treating OSA may operate such that wherein the sub-motor threshold is determined based on the data of at least one physical characteristic of the patient. In an aspect, the method of treating OSA may operate such that wherein thestimulation directly stimulates afferent fibers of the phrenic nerve without directly stimulating the efferent fibers that cause contraction of the diaphragm.

[0245] It is appreciated that different strategies identified in this application may apply differently and in different combinations to patients with different disease, underling physiology and anatomy. Some patients may benefit from some lung volume increase, and some may not. Some patients may benefit from phase locking their breathing to a rate higher than their natural respiration at rest.

[0246] While at least one exemplary embodiment of the present invention(s) is disclosed herein, it should be understood that modifications, substitutions and alternatives may be apparent to one of ordinary skill in the art and can be made without departing from the scope of this disclosure. This disclosure is intended to cover any adaptations or variations of the exemplary embodiment(s). In addition, in this disclosure, the terms “comprise” or "comprising" do not exclude other elements or steps, the terms "a" or "one" do not exclude a plural number, and the term “or” means either or both, unless the disclosure states otherwise. Furthermore, characteristics or steps which have been described may also be used in combination with other characteristics or steps and in any order unless the disclosure or context suggests otherwise. This disclosure hereby incorporates by reference the complete disclosure of any patent or application from which it claims benefit or priority.

Claims

CLAIMS:1 . A medical device for the treatment of Obstructive Sleep Apnea (OSA) in a human patient, the medical device comprising: a set of electrical stimulation electrodes; a stimulation circuitry configured to generate excitation waveforms to be applied as electrical stimulation to the stimulation electrodes; a set of sensors; a power source; and a controller running an adaptive feedback algorithm to modulate the physiology of upper airway muscles and diaphragm of the human patient by modifying excitation waveforms generated by the stimulation circuity.

2. The medical device of claim 1 , wherein the set of electrical stimulation electrodes comprises three contact electrodes.

3. The medical device of claim 1 or 2, wherein the set of electrical stimulation electrodes comprises three cuff electrodes with three electrical contacts.

4. The medical device of claim 1 or 2 or 3, wherein the set of electrical stimulation electrodes is included in a tri-polar electrode with at least three contact electrodes.

5. The medical device of any one of claims 2 to 4, wherein the stimulation circuitry is configured to generate a unidirectional excitation by holding two of the contact electrodes of a non-traveling direction at a negative potential while delivering a positive pulse to another of the contact electrodes in the traveling direction,6. The medical device of any one of claims 2 to 5, when combined with claim 3, wherein the stimulation circuitry is configured to generate a unidirectionalexcitation by holding two of the contact electrodes of a non-traveling direction of the tri- polar electrode at a negative potential while delivering a positive pulse to another of the contact electrodes of the tri-polar electrode in the traveling direction,7. The medical device of any one of claims 2 to 6, wherein the stimulation circuitry is configured to generate a bidirectional excitation by holding a middle contact at a negative potential while delivering a positive pulse to a plurality of the at least three contact electrodes on either side of the middle contact.

8. The medical device of any one of claims 2 to 7, when combined with claim 3, wherein the stimulation circuitry is configured to generate a bidirectional excitation by holding a middle contact of the tri-polar electrode at a negative potential while delivering a positive pulse to a plurality of the at least three contact electrodes on either side of the middle contact.

9. The medical device of any one of claims 1 to 8, wherein the set of electrical stimulation electrodes includes at least one non-contacting electrode.

10. The medical device of any one of claims 1 to 9, wherein the stimulation circuitry is further configured to deliver biphasic stimulation.11 . The medical device of any one of claims 1 to 10, wherein the controller is configured to control the stimulation circuitry to selectively apply different electrical potentials to different contacts of the set of electrical stimulation electrodes to alter morphology of resulting electrical fields used for stimulation of target tissue.

12. The medical device of any one of claims 1 to 11 , wherein the controller is further configured to determine a capture threshold.

13. The medical device of claim 12, wherein at least one parameter for the excitation waveforms is modified based on the capture threshold, wherein the at least one parameter that includes amplitude, pulse width, frequency, and / or duration of the excitation waveforms, wherein a value of the at least one parameter is above or below the capture threshold.

14. The medical device of claim 12 or 13, wherein the controller is configured to execute an algorithm for the detection of the capture threshold comprising: monitoring the human patient, optionally with actigraphy and / or accelerometry, to confirm that patient is supine and resting, and increasing stimulation energy in graded steps, optionally using steps of 0.1 to 0.25 mA between 0.5 and 2.5 mA, to detect the fists distinct rhythmic twitches of the diaphragm muscle.

15. The medical device of any one of claims 1 to 14, wherein the adaptive feedback algorithm is configured to monitor and modulate the physiology of the upper airway muscles and the diaphragm of the human patient by modifying excitation waveforms generated by the stimulation circuity.

16. The medical device of any one of claims 1 to 15, wherein the adaptive feedback algorithm is configured to adapt to patient physiology by applying a search algorithm for stimulation domain parameters, optionally wherein the stimulation domain parameters include one or more of amplitude, frequency, pulse-width, stimulation duration, or timing.

17. The medical device of any one of claims 1 to 16, wherein said set of sensor comprises multiple sensors configured to provide sensory feedback to the controller.

18. The medical device according to any one of claims 1 to 17, wherein the controller is configured to receive sensory feedback from said sensors for periods of time during a sleep session or between sleep sessions.

19. The medical device according to any one of claims 1 to 18, wherein the controller is configured to receive sensory feedback from multiple sensors.

20. The medical device of any one of claims 1 to 19, wherein feedback for the adaptive feedback algorithm is one or more of transthoracic impedance, chest acceleration, laryngeal acceleration, electromyogram of upper airway muscles, electromyogram of scalene muscles, audio signals from breathing, or intramuscular pressure.21 . The medical device of any one of claims 1 to 20, wherein the controller is further configured to, by using the adaptive feedback algorithm, control therapy applied to a patient by reducing apnea hypopnea index.

22. The medical device of any one of claims 1 to 21 , wherein the controller is further configured to, by using the adaptive feedback algorithm, control therapy applied to a patient by reducing a number of arousals.

23. The medical device of any one of claims 1 to 22, wherein the controller is further configured to, by using the adaptive feedback algorithm, control therapy applied to a patient by reducing a sum of the apnea hypopnea index and the number of arousals.

24. The medical device of any one of claims 1 to 23, wherein the controller is further configured to, by using the adaptive feedback algorithm, control therapy applied to a patient by reducing a difference between physiological parameters measured during an awake state and during a sleep state of the patient.

25. The medical device of any one of claims 1 to 24, wherein the controller is further configured to, by using the adaptive feedback algorithm, control therapy applied to a patient by maximizing at least one of: 1 ) airflow, 2) upper airway muscle tone, 3) tidal volume, or 4) minute volume .

26. The medical device of any one of claims 1 to 25, wherein the adaptive feedback algorithm includes: determining a respiratory period of the patient; determining a stimulation period based on, but less than, the respiratory period.

27. The medical device of claim 26, wherein the adaptive feedback algorithm includes: controlling a timing of the excitation waveforms based on the determined stimulation period.

28. The medical device of any one of claims 1 to 27, wherein the controller is further configured to: based on one or more signals from the set of sensors that indicate detection of obstruction of an airway of the patient, cause, in particular via the stimulation circuit and set of stimulation electrodes, stimulation energy to be delivered to the phrenic nerve of the patient.

29. The medical device of any one of claims 1 to 28, wherein the controller is further configured to, based on one or more signals from the set of sensors that indicate transthoracic impedance and using the adaptive feedback algorithm, adapting stimulation energy to be delivered to the phrenic nerve of the patient.

30. The medical device of any one of claims 1 to 29, wherein the controller is further configured to, based on one or more signals from the set of sensors that indicatechest acceleration and using the adaptive feedback algorithm, adapting stimulation energy to be delivered to the phrenic nerve of the patient.31 . The medical device of any one of claims 1 to 30, wherein the controller is further configured to, based on one or more signals from the set of sensors that indicate electromyogram of upper airway muscles and using the adaptive feedback algorithm, adapting stimulation energy to be delivered to the phrenic nerve of the patient.

32. The medical device of any one of claims 1 to 31 , wherein the controller is further configured to, based on one or more signals from the set of sensors that indicate laryngeal acceleration and using the adaptive feedback algorithm, adapting stimulation energy to be delivered to the phrenic nerve of the patient.

33. The medical device of any one of claims 1 to 32, wherein the controller is further configured to, based on one or more audio signals from the set of sensors that indicate breathing and using the adaptive feedback algorithm, adapting stimulation energy to be delivered to the phrenic nerve of the patient.

34. The medical device of any one of claims 1 to 33, wherein the controller is further configured to, based on one or more signals from the set of sensors that indicate electromyogram of scalene muscles and using the adaptive feedback algorithm, adapting stimulation energy to be delivered to the phrenic nerve of the patient.

35. The medical device of any one of claims 1 to 34, wherein the controller is further configured to, based on one or more signals from the set of sensors that indicate intramuscular pressure and using the adaptive feedback algorithm, adapting stimulation energy to be delivered to the phrenic nerve of the patient.

36. The medical device of any one of claims 1 to 35, wherein the adaptive feedback algorithm includes: determining a respiratory rate of the patient, and controlling the application of the electrical stimulation to the electrodes based on the determined respiratory rate.

37. The medical device of claim 36, wherein the adaptive feedback algorithm includes: controlling the application of the electrical stimulation to achieve entrainment at a rate based on the determined respiratory rate.

19. The medical device of any one of claims 1 to 18, wherein the controller is further configured to cause the electrical stimulation to be delivered at a time that is predicted from a respiratory waveform.

39. The medical device of any one of claims 1 to 38, wherein the controller is configured to control the stimulation circuitry to generate excitation waveforms having: pulse train frequency in range of 20 to 50 Hz, optionally 30 Hz.

40. The medical device of any one of claims 1 to 39, wherein the controller is configured to control the stimulation circuitry to generate excitation waveforms having: pulse width in a range of 30 to 215 ps, optionally 150 ps.41 . The medical device of any one of claims 1 to 40, wherein the controller is configured to control the stimulation circuitry to generate excitation waveforms having: an amplitude in a range of 0.4 to 5.0 mA, optionally 1 .0 to 4.0 mA.

42. The medical device of any one of claims 1 to 41 , wherein the controller is configured to control the stimulation circuitry to generate excitation waveforms having: stimulation duty cycle in a range of 30-70%, optionally 40-50%.

Images