Implanted Sensors in Neuromodulation via Electrical Stimulation

Introduction

Neuromodulation systems rely on external control systems to adjust parameters, which control their output. Observation of the patient’s response to stimulation, or the patient’s own perception of stimulation, is used to manually adjust stimulation levels or other parameters to optimize the benefit from the device. Sensor technology can provide data that speed up the optimization process and, in some instances, eliminate the need for user input altogether by providing real-time data for continuous optimization of devices that operate in a closed loop of continuous sensing and adjusting. Sensor data can also be used to detect the onset of symptoms and provide prophylactic, demand-based stimulation. The idea of continuous monitoring and adjusting of neuromodulation in response to the underlying neurophysiological state is a powerful way to improve these devices and enhance performance. However, despite its obvious advantages, closed-loop monitoring is rare in neuromodulation and treatment regimens are informed only by the diagnostic and test information, which is gathered during medical treatment. This is an indication of the level of technical difficulty needed to implement sensors in active devices. In this chapter, we will focus on the requirements for sensors and closed-loop monitoring in neuromodulation via electrical stimulation and review the advances, which have thus far been made in the field.

Electrical neuromodulation of the nervous system with implantable devices historically began with spinal cord stimulation (SCS) for the treatment of chronic, intractable neuropathic pain. The technique for SCS involves the implantation of a pulse generator connected to one or more multipolar leads that is/are placed in the epidural space over the dorsal columns (DCs) of the spinal cord. Activation of the electrodes has the aim of recruiting nerve fibers to produce action potentials (APs) that then propagate orthodromically and antidromically along the DCs SCS has been applied to a number of painful conditions that have neural targets in in both the central (CNS) and peripheral nervous systems (PNS). Deep brain stimulation (DBS) uses similar technology to stimulate deep regions of the brain, most commonly targeting the basal ganglia in Parkinson disease. In this application, electrical activity can downregulate some of the activity in neural pathways controlling motor functions and therefore relieve tremors and rigidity.

The relationship between stimulation parameters (e.g., pulse width, pulse frequency, amplitude, and the spatial relationship between multiple active electrodes) and the neural recruitment that results is, except in a few circumstances, nearly impossible to predict. Neural recruitment is influenced by a large number of variables, and its effectiveness at relieving symptoms is likewise dependent on many factors. Fig. 32.1 catalogues some of these factors.

The integration of direct sensing into ES neuromodulation systems is the obvious solution to address the therapy-limiting effects of these factors. In essence, the integration of direct sensing into ES neuromodulation systems can reduce some of the ambiguity of the intervention. To give an example, it is known that neurons that are in their refractory period after an AP cannot respond to a therapeutic pulse from a neuromodulation system. On-board sensing could be used to adjust the timing of stimulation in order to better recruit neurons near their resting potential. To give another example, it is known that the gap between SCS stimulating electrodes and the DCs frequently widens and narrows due to changes in body posture; The narrowing of the gap between the DCs and the epidurally placed electrodes is perceived by the patient as uncomfortable changes in therapy. Integrated sensors in SCS systems could be used to automatically maintain the ideal relative stimulation amplitude.

In vivo neural systems, particularly in pathologic conditions that require treatment, are so complex that “blind” prediction of neural recruitment to any given stimulation parameter is practically impossible. Hence, sensors that measure the neural activity, the therapeutic consequences of the activity, or the physical changes which affect recruitment are an essential component for effective neuromodulation. Direct objective measurements are required to optimize the therapy for a population and for individual patients.

Sensing and Using Sensed Data

Three main categories of sensors are applicable for neuromodulation systems: physical, chemical, and electrical. Physical sensors determine a property of the environment of the sensor; accelerometers that detect movement and position are a prime example. Other physical sensors include temperature and pressure measurement. Chemical sensors measure the concentration of constituents in its surrounding (typically, the extracellular matrix). Chemical sensors may be highly specific for a single compound or may have a broad spectrum to many compounds. Electrical sensors detect an electrical property such as voltage gradient or impedance.

Currently, integrated sensors in neuromodulation systems are used for a variety of applications. For example, short-term sensing can assess the performance of a therapy and optimize neuromodulation stimulation parameters. It can also be used over long periods of time in either a continuous or an intermittent manner. Alternately, the sensors may be built into the neuromodulation system and be used for real-time control of the device. Table 32.1 lists a variety of sensors used in neuromodulation therapies, and some prominent examples are described next.

Physical sensors : The advent of microelectromechanical systems has allowed the incorporation of suitably miniaturized accelerometers into SCS systems ( ). These accelerometers detected changes in position (e.g., from standing to reclining) and automatically adjusted the amplitude of stimulation in order to maintain comfortable levels of stimulation. With position-adaptive stimulation, the number of manual amplitude changes that patients had to make each day was reduced from 31 to 18 per day. Thus, although promising, this technology has not reached the goal of truly automatic stimulation maintenance.

Chemical sensors : The most advanced chemical sensor application in neuromodulation has been pioneered by . In an attempt to look beyond the electrophysiology, they characterize the neurochemical effects of the stimulation therapy. A wireless instantaneous neurotransmitter concentration sensor has been developed that allows the measurement of both glutamate and dopamine levels with fast-scan cyclic voltammetry in real time during DBS in anesthetized rats and in awake monkeys. Because the in vivo use of chemical sensors is highly technically challenging, further hurdles regarding materials used and long-term stability must be crossed before routine incorporation into implanted neuromodulation systems is possible.

Electrical sensors : Because electroencephalographic (EEG) recordings from implanted electrodes can be used to detect the onset of epileptic seizures ( ), responsive neuromodulation for the treatment of epilepsy has been developed. The RNS system developed by Neuropace uses the neurologic markers for the onset of seizure to trigger stimulation at the epileptic focus to abort or lessen the impact of seizures. With the RNS technology, a 53% reduction in seizures has been reported ( ). This constitutes the first application of the use of sensors to detect a change in state and trigger a response.

In addition to implanted sensors outlined, implanted devices may interact with external sensors. For example, accelerometers and wireless technologies are ubiquitous due to the widespread acceptance of smartphones; their use for measuring the extent of motor disorders at first appears to be relatively simple. Smartphone accelerometer-based measurements and support vector machine analysis have indeed been used to determine the state of DBS stimulation for essential tremor ( ). Determining the difference in magnitude between intentional and nonintentional movement can, however, present significant challenges. Specialized technology is available in the form of wrist-mounted devices ( ) and a combination of sensor bands and Android tablets ( ). The burgeoning interplay of implanted devices, externally worn sensors, and smartphones posits a new horizon for large scale data analytics that will revolutionize health care. The resultant data sets will form the foundation on which new therapies and improvement to existing therapies will be based.

Incorporation of sensors into implanted ES systems has been limited to date by technical barriers; sensors must be sufficiently reliable and have low power demands. Electrical sensors may be the most attractive option for robust sensing in implanted neuromodulation. That is because the already existing electronic subsystems within the implant can be leveraged for this role. After all, some electrical sensing—impedance or resistivity between electrode contacts—is already routinely measured to provide objective measurement of system and lead integrity.