A Wireless Implantable System for Facilitating Gastrointestinal Motility

micromachines

Article A Wireless Implantable System for Facilitating Gastrointestinal Motility

Po-Min Wang 1, Genia Dubrovsky 2, James C.Y. Dunn 1,2,3,4, Yi-Kai Lo 5 and Wentai Liu 1,*

1 Department of Bioengineering, University of California, Los Angeles, Los Angeles, CA 90095, USA 2 Division of Pediatric Surgery, Department of Surgery, David Geﬀen School of Medicine, University of California, Los Angeles, Los Angeles, CA 90095, USA 3 Division of Pediatric Surgery, Department of Surgery, Stanford University School of Medicine, Stanford, CA 94305, USA 4 Department of Bioengineering, Stanford University, Stanford, CA 94305, USA 5 Niche Biomedical Inc., Los Angeles, CA 90095, USA * Correspondence: [email protected];

Received: 17 July 2019; Accepted: 7 August 2019; Published: 9 August 2019

Abstract: Gastrointestinal (GI) electrical stimulation has been shown in several studies to be a potential treatment option for GI motility disorders. Despite the promising preliminary research progress, however, its clinical applicability and usability are still unknown and limited due to the lack of a miniaturized versatile implantable stimulator supporting the investigation of effective stimulation patterns for facilitating GI dysmotility. In this paper, we present a wireless implantable GI modulation system to fill this technology gap. The system consists of a wireless extraluminal gastrointestinal modulation device (EGMD) performing GI electrical stimulation, and a rendezvous device (RD) and a custom-made graphical user interface (GUI) outside the body to wirelessly power and configure the EGMD to provide the desired stimuli for modulating GI smooth muscle activities. The system prototype was validated in bench-top and in vivo tests. The GI modulation system demonstrated its potential for facilitating intestinal transit in the preliminary in vivo chronic study using porcine models.

Keywords: gastrointestinal stimulation; implant; motility; neuromodulation; bioelectronics medicine; electroceuticals

1. Introduction The estimated annual healthcare expenditures for gastrointestinal (GI) motility disorders is USD 29 billion, which imposes a significant burden on the U.S. healthcare system [1]. GI motility disorder is defined as the abnormal movement behavior of the GI tract that affects the function of mixing and propelling food. This can happen to any segment of the GI tract, including the esophagus (e.g., dysphagia, achalasia), stomach (e.g., gastroparesis, gastroesophageal reflux disease), and intestines (e.g., diarrhea, constipation). The pathophysiological mechanism of GI motility disorders is not completely understood due to the complex, cooperative mechanisms between the smooth muscle cells (SMC), interstitial cells of Cajal (ICC), enteric nervous system, central nervous system, and hormones [2,3]. This gap of knowledge impedes the development of efficacious therapies for improving GI motility. Although a number of pharmaceutical drugs have been developed, most of them do not completely alleviate GI dysmotility [4]. Moreover, the use of drugs faces the challenge of target specificity, as it is difficult to precisely control the dose of pharmacological agents. A potential treatment alternative for GI motility disorders is GI electrical stimulation. The fundamental principle behind this therapy is to electrically modulate myoelectric activity that controls smooth muscle contraction/relaxation, i.e., slow waves and spikes [5]. A slow wave is

Micromachines 2019, 10, 525; doi:10.3390/mi10080525 www.mdpi.com/journal/micromachines Micromachines 2019, 10, 525 2 of 11 a rhythmical electrical event originating from ICC, while a spike is the action potential that stems from the inflow of calcium ions to the SMC. ICC and SMC are both regulated by the enteric nervous system in the GI tract [6,7]. It is thus feasible to directly induce smooth muscle contraction through modulating SMC or indirectly via activating the ICC/myenteric network using GI electrical stimulation. Several studies have demonstrated the therapeutic potential of GI electrical stimulation in GI dysmotility, including gastric electrical stimulation for pharmaceutically intractable gastroparesis [8,9], intestinal electrical stimulation for accelerating intestinal transit [10,11], and colonic electrical stimulation for constipation [12,13]. The potential mechanisms are either the stimulation of the SMC [8,9] or the activation of the cholinergic or nitrergic pathways in the enteric nervous system [10–13]. Despite promising research results, the clinical applicability and usability of GI electrical stimulation for GI motility disorders is still limited due to the lack of appropriate implantable stimulation devices. In light of the fact that existing commercial implantable pulse generators (e.g., deep brain stimulator, spinal cord implant, vagus nerve stimulator [14–16]) have limited parameter programmability, the implantable GI stimulator should support a wide range of stimulation parameters (i.e., pulse width, current intensity, frequency), especially long-pulse stimuli to activate smooth muscle when necessary. Interestingly, even on the same GI segment, diverse stimulation parameters have been reported for improving GI dysmotility, e.g., colonic transit could be accelerated in rats by stimulation protocols of 4 ms, 10 mA, 40 Hz [17], and 0.3 ms, 5 mA, 10 Hz [18]. It is thus essential to develop a device capable of versatile stimulation parameters in order to customize an optimal stimulation strategy for each individual. This above feature is also critical, as the investigation of an optimal stimulation protocol for effectively activating GI contraction and the sophisticated working mechanism of the GI tract is still ongoing. Another desired feature for GI implants is the miniaturized form factor. This would be clinically beneficial, as surgical invasiveness can be minimized during implantation. Existing commercial GI implants are bulky and only support stimuli with a very limited range of parameters, e.g., pulse width shorter than 1 ms [19,20]. The short pulse width limits their usage in treating nausea and vomiting [19] or increasing the pressure of lower esophageal sphincter for relieving gastroesophageal reflux disease [20]. Both commercial devices might not be applicable to directly altering GI motility as most protocols for inducing GI contraction require a stimulation pulse greater than 1 ms [8–13]. In order to fill the technology gap, a miniaturized implantable GI stimulator supporting a wide range of stimulation parameters is needed. We developed a wireless GI modulation system that fulfills the aforementioned needs (Figure1). The system consists of a wireless extraluminal gastrointestinal modulation device (EGMD) interfacing with the gut, a custom-made graphical user interface (GUI) on a tablet for configuring the EGMD, and a rendezvous device (RD) serving as a relay between the EGMD and the GUI. The core of the EGMD is a system on a chip (SoC) capable of generating versatile stimulation patterns through GUI control to modulate GI motility and wireless power and data transfer via inductive links. A heterogeneous packaging approach integrates the SoC and printed circuit board (PCB) with a polyimide-based electrode array, which achieves device miniaturization. The EGMD is encapsulated by biocompatible epoxy and implanted in the abdominal wall through midline abdominal incision. The RD carried by the subject receives user-defined commands from the GUI via WiFi link and delivers power and command to configure the EGMD through inductive links. Design consideration and the implementation of the EGMD were presented, and preliminary in vivo acute tests were performed [21]. Building upon our prior work, the EGMD was advanced to a fully implantable GI modulation system; a preliminary in vivo chronic study was conducted using porcine models to investigate its potential for accelerating intestinal transit. The remainder of this paper is organized as follows. Section2 presents the microelectronic design of the wireless implantable GI modulation system and the heterogeneous packaging technique that achieves implant miniaturization. Section3 presents experiment results of the GI modulation system in bench-top and in vivo chronic tests. The conclusion is drawn in Section4. Micromachines 2019, 10, 525 3 of 11 MicromachinesMicromachines 2019 2019, 10, x10 , x 3 of3 12of 12

Figure 1. Illustration of wireless gastrointestinal (GI) modulation system interfacing with gut. FigureFigure 1. Illustration 1. Illustration of wireless of wireless gastrointestinal gastrointestinal (GI) modulationmodulation system system interfacing interfacing with with gut. gut. 2.2. Wireless Wireless Implantable Implantable GI GI Modulation Modulation System 2. Wireless Implantable GI Modulation System 2.1.2.1. Microelectronic Microelectronic Design Design 2.1. Microelectronic Design TheThe custom-designed custom-designed SoC in the wireless EGMDEGMD adaptedadapted from from our our prior prior work work supports supports wireless wireless powerpowerThe custom-designed and and data data transfer, transfer, SoC and in highlythe wireless programmableprogrammable EGMD ad current-modecurrent-modeapted from our stimulationstimulation prior work for for supports activating activating wireless GI GI contraction (Figure2)[ 22]. The quad-voltage power converter consisting of rectifiers and voltage powercontraction and data (Figure transfer, 2) [22]. and The highly quad-voltage programmable power convertercurrent-mode consisting stimulation of rectifiers for andactivating voltage GI regulators receives inductive power signals from the RD and provides regulated +/–1.8 V to supply contractionregulators (Figure receives 2) inductive [22]. The power quad-voltage signals from power the RDconverter and provides consisting regulated of rectifiers +/–1.8 V andto supply voltage the data receiver and the control logic, and +/–12 V to the current driver in the stimulator. The high regulatorsthe data receives receiver inductiveand the control power logic, signals and from+/–12 theV to RD the and current provides driver regulated in the stimulator. +/–1.8 V The to supplyhigh compliancecompliance voltage voltage forfor thethe stimulator stimulator is is critical critical as itas allows it allows a wide a wide range range of parameters of parameters for current-mode for current- the data receiver and the control logic, and +/–12 V to the current driver in the stimulator. The high modestimulation. stimulation. The di Thefferential differential phase shift phase keying shift (DPSK) keying receiver (DPSK) that receiver is immune that tois theimmune interference to the compliancein the amplitude voltage for domain the stimulator demodulates is critical the incoming as it allows DPSK a signalwide range back toof theparameters digital command for current- interference in the amplitude domain demodulates the incoming DPSK signal back to the digital modeissued stimulation. by the user. The The differential command isphase then decoded shift ke byying the digital(DPSK) controller, receiver and that the stimulatoris immune is setto the command issued by the user. The command is then decoded by the digital controller, and the interferenceto accordingly in the generateamplitude the desireddomain electrical demodulates stimuli. the The incoming highly programmable DPSK signal digital back to controller the digital stimulator is set to accordingly generate the desired electrical stimuli. The highly programmable enables the stimulator to support intensities from 0.3 µA to 20 mA, with a pulse width from 10 µs to commanddigital controllerissued by enables the user. the stimulatorThe command to support is then intensities decoded from by 0.3the μ Adigital to 20 controller,mA, with a andpulse the a user-defined length, and a frequency from 0.001 to 200 Hz. stimulatorwidth from is set 10 toμs accordinglyto a user-defined generate length, the and de asired frequency electrical from 0.001stimuli. to 200The Hz. highly programmable digital controller enables the stimulator to support intensities from 0.3 μA to 20 mA, with a pulse width from 10 μs to a user-defined length, and a frequency from 0.001 to 200 Hz.

Figure 2. Functional block diagram of wireless GI modulation system. Figure 2. Functional block diagram of wireless GI modulation system. The RD made by off-the-shelf components receives the command from the GUI on the tablet via WiFiThe and, RD accordingly, made by off-the-shelf powers and components configures the receives EGMD the through command dual-band from the inductive GUI on telemetries.the tablet via WiFiCarrierFigure and, 2. frequencies Functionalaccordingly, ofblock powers the diagram power and and ofconfigures wireless data links GI the modulation are EGMD 2 and through 22 system. MHz, dual-band respectively. inductive This frequencytelemetries. Carrierseparation, frequencies one decade of apart,the power allows and the data use of links signal are filtering 2 and in 22 the MHz, DPSK respectively. receiver to relieve This powerfrequency to separation,dataThe RD interference. made one decadeby In off-the-shelf the apart, power allows transmitter, components the use a highlyof receivessignal effi filteringcient the class-Ecommand in the power DPSK from amplifier receiver the GUI was to onrelieve designed the tabletpower to via WiFitodrive and, data theaccordingly,interference. primary-side In powers the power power and coil transmitter, configures [23]. The power a the highly EGMD amplifier efficient through is class-E supplied dual-band power by an amplifier adjustable inductive was regulator telemetries.designed Carriertowith drive frequencies voltage the primary-side ranging of the from powerpower 2 to coil 10and V. [23]. data This The adjustabilitylinks powe arer amplifier 2 enablesand 22 is thesuppliedMHz, power respectively. by transmitter an adjustable This to tune regulator frequency the separation,withoutput-power voltage one decaderanging level apart, based from onallows2 to the 10 variation theV. Thisuse ofadjust of signal couplingability filtering coeenablesfficients in thethe between DPSKpower receivertransmitter coil pairs to in relieveto order tune power tothe to dataoutput-poweroptimize interference. power-transfer level In thebased power effi onciency. the transmitter, variation On the other ofa hi hand,couplighly inngefficient the coefficients data class-E transmitter, between power the amplifiercoil same pairs architecture wasin order designed of to to driveoptimizeclass-E the powerprimary-side power-transfer amplifier power isefficiency. used, coil and [23]. On the the The amplifier other powe hand isr amplifier driven, in the by data ais noncoherent supplied transmitter, by DPSK anthe adjustable same modulator architecture regulator [24]. withof voltage class-E powerranging amplifier from 2 isto used, 10 V. and This the adjust amplifierability is drivenenables by the a noncoherentpower transmitter DPSK modulator to tune the output-power[24]. This modulation level based scheme on the encodes variation the of command coupling bit coefficients in the relative between phase coil shift pairs between in order two to optimizeconsecutive power-transfer symbols rather efficiency. than theOn phasethe other shift hand to a, referencein the data clock, transmitter, which avoids the same a complicated architecture of class-E power amplifier is used, and the amplifier is driven by a noncoherent DPSK modulator [24]. This modulation scheme encodes the command bit in the relative phase shift between two consecutive symbols rather than the phase shift to a reference clock, which avoids a complicated

Micromachines 2019, 10, x 4 of 12 synchronization circuit in the implant side, e.g., phase-locked loop. The scheme is realized by a XOR gate and a D flip-flop, such that the D flip-flop toggles when the input bit is logic one and remains unchanged when the input bit is logic zero (Figure 2). The clocks and data command required by the power and dataMicromachines transmitters2019, 10, 525 are generated by a microcontroller (CC3200, Texas Instruments.,4 of 11 Dallas, TX, USA) that can be wirelessly controlled by the tablet through WiFi. The DC–DC converter and low-dropoutThis regulators modulation take scheme input encodes from the command a 3.7 V bitlithium in the relative ion rechargeable phase shift between battery two consecutive to power circuits in the RD. Thesymbols wireless rather and than battery-power the phase shift to features a reference allow clock, whichthe RD avoids to be a complicated carried by synchronization freely moving subjects circuit in the implant side, e.g., phase-locked loop. The scheme is realized by a XOR gate and a D receiving theflip-flop, implant. such that the D flip-flop toggles when the input bit is logic one and remains unchanged when the input bit is logic zero (Figure2). The clocks and data command required by the power and data 2.2. Heterogeneoustransmitters Packaging are generated and bySystem a microcontroller Integration (CC3200, Texas Instruments., Dallas, TX, USA) that can be wirelessly controlled by the tablet through WiFi. The DC–DC converter and low-dropout regulators The heterogeneous-packagingtake input from a 3.7 V lithium iontechnique rechargeable utilizes battery an to power8 μm circuits thick inpolyimide the RD. The wirelesssubstrate and to integrate the SoC, PCB,battery-power off-chip featurescomponents, allow the and RD to electrode be carried by array freely movinginto a subjectsminiaturized receiving theEGMD implant. (Figure 3). The electrodes, 2.2.bond Heterogeneous pads, and Packaging the interconnections and System Integration on the substrate are made of platinum/titanium (200/10 nm thickness)The heterogeneous-packaging fabricated by the technique e-beam-e utilizesvaporated an 8 µm thick deposition polyimide technique substrate to integrate [25]. All required electrical connectionsthe SoC, PCB, for off -chipenabling components, system and func electrodetions array (e.g., into the a miniaturizedconnection EGMD between (Figure electrodes3). and outputs of stimulationThe electrodes, drivers) bond pads, are and achieved the interconnections by the patterned on the substrate metal are madetrace ofon platinum the polyimide/titanium substrate, (200/10 nm thickness) fabricated by the e-beam-evaporated deposition technique [25]. All required as well as theelectrical gold connections bumps and for enabling ball-bonded system functions gold wires. (e.g., the The connection gold bumps between electrodesconnect andthe polyimide substrate withoutputs the ofSoC, stimulation while drivers)the connection are achieved among by the patterned the substrate, metal trace SoC, on the and polyimide the PCB substrate, are made by the ball-bondedas gold well aswires. the gold The bumps use of and a small-footprin ball-bonded gold wires.t PCB Thesignificantly gold bumps facilitated connect the polyimidethe development of the EGMD substrateprototype. with theThis SoC, configuration while the connection enables among de thesign substrate, debugging SoC, and theand PCB the are monitoring made by the of critical ball-bonded gold wires. The use of a small-footprint PCB significantly facilitated the development signals in theof theSoC EGMD by prototype.directly probing This configuration PCB pads enables du designring debuggingthe system and bench the monitoring top test. of critical Nevertheless, it should be notedsignals that in the the SoC PCB by directly is not probing a necessary PCB pads pa duringrt in the the system final bench EGMD top test.design. Nevertheless, Indeed, it all off-chip componentsshould can be be notedconnected that the to PCB pads is not on a necessary the polyim part inide the substrate final EGMD via design. the Indeed,silver allconductive off-chip paste in components can be connected to pads on the polyimide substrate via the silver conductive paste in the the absenceabsence of a PCB. of a After PCB. After the theassembly assembly process, process, the the EGMD EGMD is encapsulated is encapsulated by the epoxy by the (3M epoxy DP100 (3M DP100 Plus Clear, Plus3M Clear,Corp., 3M Maplewood, Corp., Maplewood, MN, MN, USA). USA). The The materialsmaterials interfacing interfacing with the with biological the biological tissue, tissue, including theincluding polyimide, the polyimide, platinum, platinum, and and epoxy, epoxy, are are all all biocompatible. biocompatible.

Figure 3. Illustration of heterogeneous packaging for wireless extraluminal gastrointestinal modulation Figure 3. deviceIllustration (EGMD). of heterogeneous packaging for wireless extraluminal gastrointestinal modulation device (EGMD). Figure4 shows the miniaturized EGMD prototype. The coaxial arrangement of power and data coils reduces EGMD size. The separation between the coils and the PCB were suggested by the surgeon Figure in4 favorshows of EGMDthe miniaturized implantation during EGMD the surgicalprototyp procedure.e. The Suturingcoaxial holes arrangement were designed of onpower the and data coils reducestip ofEGMD the electrode size. arrayThe for separation anchoring the between array on the the gut coils by suture. and The the miniaturized PCB were EGMD suggested has by the surgeon in favora weight of ofEGMD ~4.2 g and implantation a volume of ~4.9 during cm3, which the surgical is at least procedure. an order lighter Suturing and 7.7 times holes smaller were designed than existing commercialized GI implantable stimulators [19,20]. on the tip of the electrode array for anchoring the array on the gut by suture. The miniaturized EGMD has a weight of ~4.2 g and a volume of ~4.9 cm3, which is at least an order lighter and 7.7 times smaller than existing commercialized GI implantable stimulators [19,20].

Micromachines 2019, 10, x 4 of 12

synchronization circuit in the implant side, e.g., phase-locked loop. The scheme is realized by a XOR gate and a D flip-flop, such that the D flip-flop toggles when the input bit is logic one and remains unchanged when the input bit is logic zero (Figure 2). The clocks and data command required by the power and data transmitters are generated by a microcontroller (CC3200, Texas Instruments., Dallas, TX, USA) that can be wirelessly controlled by the tablet through WiFi. The DC–DC converter and low-dropout regulators take input from a 3.7 V lithium ion rechargeable battery to power circuits in the RD. The wireless and battery-power features allow the RD to be carried by freely moving subjects receiving the implant.

2.2. Heterogeneous Packaging and System Integration The heterogeneous-packaging technique utilizes an 8 μm thick polyimide substrate to integrate the SoC, PCB, off-chip components, and electrode array into a miniaturized EGMD (Figure 3). The electrodes, bond pads, and the interconnections on the substrate are made of platinum/titanium (200/10 nm thickness) fabricated by the e-beam-evaporated deposition technique [25]. All required electrical connections for enabling system functions (e.g., the connection between electrodes and outputs of stimulation drivers) are achieved by the patterned metal trace on the polyimide substrate, as well as the gold bumps and ball-bonded gold wires. The gold bumps connect the polyimide substrate with the SoC, while the connection among the substrate, SoC, and the PCB are made by the ball-bonded gold wires. The use of a small-footprint PCB significantly facilitated the development of the EGMD prototype. This configuration enables design debugging and the monitoring of critical signals in the SoC by directly probing PCB pads during the system bench top test. Nevertheless, it should be noted that the PCB is not a necessary part in the final EGMD design. Indeed, all off-chip components can be connected to pads on the polyimide substrate via the silver conductive paste in the absence of a PCB. After the assembly process, the EGMD is encapsulated by the epoxy (3M DP100 Plus Clear, 3M Corp., Maplewood, MN, USA). The materials interfacing with the biological tissue, including the polyimide, platinum, and epoxy, are all biocompatible.

Figure 3. Illustration of heterogeneous packaging for wireless extraluminal gastrointestinal modulation device (EGMD).

Figure 4 shows the miniaturized EGMD prototype. The coaxial arrangement of power and data coils reduces EGMD size. The separation between the coils and the PCB were suggested by the surgeon in favor of EGMD implantation during the surgical procedure. Suturing holes were designed on the tip of the electrode array for anchoring the array on the gut by suture. The miniaturized EGMD Micromachines 2019, 10, 525 5 of 11 has a weight of ~4.2 g and a volume of ~4.9 cm3, which is at least an order lighter and 7.7 times smaller than existing commercialized GI implantable stimulators [19,20].

Figure 4. (a) Photo of EGMD prototype for chronic study. Zoom-in view of (b) system on a chip (SoC) part under microscope and (c) electrode array. 3. Experiment Results and Discussion The proposed wireless implantable GI modulation system was used to study its potential to improve intestinal motility in porcine models. In order to ensure effective and safe intestinal stimulation, the experiment protocol was designed as follows: 1) The charge injection limit and impedance of the electrode–tissue interface were first characterized using cyclic voltammetry (CV) and two different impedance-measurement methods. This provides a guideline for a safe stimulation protocol that avoids any electrochemical tissue-damaging reaction in the electrode–tissue interface [26]. 2) An acute in vivo test using the electrode array was then performed to investigate the stimulation protocol that is also electrochemically safe for facilitating intestinal motility. 3) The wireless GI modulation system was tested on the bench top to ensure its capability of generating the desired stimulation pattern identified in the acute test. 4) A preliminary in vivo chronic test was finally conducted to demonstrate the wireless GI modulation system as an effective tool for accelerating intestinal transit in porcine models. The use of all animals was approved by the UCLA Animal Research Committee (Institutional Review Board no. 2014-142-03E).

3.1. Electrode Characterization The electrode–saline interface and electrode–intestinal fluid interface were both characterized for stimulation-safety evaluation, as both extraluminal and intraluminal intestinal stimulation have been reported as potential treatments for intestinal dysmotility [10,27]. A 0.1 M NaCl solution was used to mimic the extraluminal environment, while intraluminal intestinal fluid was extracted from a juvenile mini Yucatan pig. A customized electrode array that was fabricated by the same process and had the same electrode dimension with the electrode array on the EGMD was soaked separately in both saline and intestinal fluid. The charge injection limit, bioimpedance, and Randles cell model of both electrode–electrolyte interfaces were characterized using CV, electrochemical impedance spectroscopy (EIS), and the time-domain Randles cell model characterization method [28], respectively. The measured cyclic voltammogram is shown in Figure5, and all measured electrochemical properties are summarized in Table1. The electrochemical window of the electrode was [–0.9, 1 V] in both saline and intestinal fluid environments. Charge storage capacity (CSC) of the electrode, which defines the charge injection limit per stimulation phase, of the intestinal fluid group was 1.16 times higher than the saline group (10.68 vs. 9.19 µC). The EIS-measured impedance of the electrode at 1 kHz of the intestinal fluid group was 2.45 times higher than the saline (6.39 vs. 2.61 kΩ), which implies that the maximum allowable stimulation current without exceeding the compliance voltage of the stimulator was ~2.45 times smaller in intestinal fluid than in saline. The measured Randles cell model of the electrode also provided the same insight, as RS was 2.26 times larger in the intestinal-fluid group (3.69 vs. 1.63 kΩ). The product of RCT and Cdl, i.e., RC time constant, was comparable between the two groups (29.70 vs. 29.11 µs), suggesting that the required time to discharge residual electrical charge at the stimulation electrode was similar. According to this experiment result, extraluminal stimulation is more beneficial compared to intraluminal, as it allows a 2.45-fold higher stimulation current, with an acceptable penalty of 1.16-fold decrease in charge injection limit. Micromachines 2019, 10, 525 6 of 11 Micromachines 2019, 10, x 6 of 12

Figure 5. Measured cyclic voltammogram of electrode array in saline and intestinal fluid. Figure 5. Measured cyclic voltammogram of electrode array in saline and intestinal fluid. Table 1. Summary of measured electrochemical properties of electrode–intestinal fluid and Tableelectrode–saline 1. Summary of interface. measured electrochemical properties of electrode–intestinal fluid and electrode– saline interface. Electrochemical Property Saline Intestinal Fluid ElectrochemicalCyclic voltammogram Property Saline Intestinal Fluid Electrochemical window (V) [ 0.9, 1] [ 0.9, 1] − − Cyclic voltammogramCharge storage capacity (µC) 9.19 10.68 ElectrochemicalElectrochemical window (V) Impedance Spectroscopy [–0.9, (EIS) 1] [–0.9, 1] Charge storageMagnitude@ capacity 1 ( kHzμC) (k Ω) 2.619.19 6.39 10.68 ElectrochemicalPhase@ Impedance 1 kHz (◦) Spectroscopy (EIS) –52.1 –53.9 Magnitude@ Randles1 kHz (k cellΩ model) characterization 2.61 6.39 R (kΩ) 1.63 3.69 Phase@ 1 kHzS (°) –52.1 –53.9 RCT (kΩ) 2.12 4.7 Randles cell Cmodeldl (nF) characterization 13.73 6.32 RS (kΩ) 1.63 3.69 3.2. Stimulation-ParameterRCT (kΩ) Identification 2.12 4.7 InC vivodl (nF)acute tests using juvenile mini Yucatan pigs were 13.73 conducted to investigate 6.32 the effective and safe intestinal-stimulation protocol for improving intestinal dysmotility. The animal preparation 3.2. Stimulation-Paraand surgical proceduresmeter Identification were described in our work [11]. During the experiment, a midline laparotomy was performed, and a short segment of jejunum was identified and externalized. The jejunum was In vivo acute tests using juvenile mini Yucatan pigs were conducted to investigate the effective transected, and 5 mL of ultrasound gel was placed inside (Figure6). The segment of jejunum was then and safe intestinal-stimulation protocol for improving intestinal dysmotility. The animal preparation monitored for 20 minutes, first under no stimulation, and then under electrical stimulation generated and bysurgical our SoC procedures via the customized were electrodedescribed array in laidour on work top of [11]. the exposed During segment. the experiment, The gel forced a backmidline laparotomyout of the was intestine performed, via peristalsis and a wasshort collected segment and of weighed jejunum for was each identified 20-minute timeand interval.externalized. The jejunum Ewasffective transected, stimulation and parameters 5 mL of ultrasound for increasing gel the was rate placed of peristalsis inside were(Figure investigated 6). The segment in our of jejunumprevious was work then [ 11monitored]. One set offor biphasic 20 minutes, current stimulifirst under (100 Hz,no 2stimulation, mA, and 2 ms) and was then obtained under and electrical was stimulationable to induce generated both localby our contraction SoC via and the peristalsis customiz (Figureed electrode6). The result array showed laid on that top in aof 20-minute the exposed segment.time period,The gel an forced average back of 0.51 out grams of the of intestine gel were via expelled peristalsis without was intestinal collected stimulation, and weighed while 1.67for each 20-minutegrams oftime gel interval. were expelled with stimulation (n = 6, p < 0.05). The potential mechanism is the activation ofEffective ICC or thestimulation local enteric parameters nervous systemfor increasing by stimulation. the rate The of peristalsis identified stimulationwere investigated protocol in is our previouselectrochemically work [11]. One safe inset terms of biphasic of charge current injection stimuli limit. (100 The stimulusHz, 2 mA, with and 2 mA 2 ms) intensity was obtained and 2 ms and pulse width lead to a 4 µC charge per stimulation phase, which was smaller than the charge-storage was able to induce both local contraction and peristalsis (Figure 6). The result showed that in a 20- capacity of the electrode (9.19 µC) characterized in the CV test. In addition, based on the Randles minutecell time model period, of the electrode–tissue an average of interface0.51 grams characterized of gel were in the expelled impedance without measurement, intestinal it couldstimulation, be whileestimated 1.67 grams that theof gel peak-to-peak were expelled voltage with induced stimulation by this current (n = 6, stimulation p < 0.05). patternThe potential was around mechanism+/–7.5 is the activationV, which was of withinICC or the the compliance local enteric voltage nervous of the system stimulator. by stimulation. The identified stimulation protocol is electrochemically safe in terms of charge injection limit. The stimulus with 2 mA intensity and 2 ms pulse width lead to a 4 μC charge per stimulation phase, which was smaller than the charge- storage capacity of the electrode (9.19 μC) characterized in the CV test. In addition, based on the Randles cell model of the electrode–tissue interface characterized in the impedance measurement, it could be estimated that the peak-to-peak voltage induced by this current stimulation pattern was around +/–7.5 V, which was within the compliance voltage of the stimulator.

Micromachines 2019, 10, 525 7 of 11 Micromachines 2019, 10, x 7 of 12

Micromachines 2019, 10, x 7 of 12

Figure 6. In vivo acute test showing that intestinal local contraction and peristalsis were initiated Figure 6. InFigure vivo 6.acute In vivo test acute showing test showing that that intestinal intestinal lolocalcal contraction contraction and peristalsis and peristalsis were initiated were initiated through the identiﬁed stimulation protocol. through thethrough identified the identified stimulation stimulation protocol. protocol. 3.3. System Bench-Top Test 3.3. System Bench-Top Test 3.3. System Bench-TopThe wireless Test GI modulation device was tested on the bench top to verify its capability of generating the desiredThe wireless stimulation GI modulation protocol. The device EGMD was was tested powered on andthe conﬁguredbench top by to the verify RD throughits capability inductive of The wirelessgeneratinglinks with GI athe separation modulation desired ofstimulation 2 cm device between protocol. thewas primary-and Thtestede EGMD on secondary-side was the powered bench coils.and top configured A GUIto verify on anby Android theits RD capability of generating thethroughtablet desired wirelessly inductive stimulation controlled links with the a RDseparationprotocol. to generate of Th2 the cme power betweenEGMD and the was data primary-and signalspowered required secondary-side and by theconfigured EGMD coils. [29 A]. by the RD GUI on an Android tablet wirelessly controlled the RD to generate the power and data signals through inductiveThe output links of the with stimulator a separation was connected of 2 to cm a 1 kbetweenΩ resistor throughthe primary-and PCB pads for currentsecondary-side intensity coils. A requiredcharacterization. by the EGMD An exemplary [29]. The waveform output of to the demonstrate stimulator the was operation connected of theto a wireless 1 kΩ resistor implantable through GI GUI on an PCBmodulationAndroid pads for systemtablet current is wirelessly shown intensity in Figure characterization controlled7. .the An RDexemplary to generate waveform the to demonstratepower and the data signals required byoperation the EGMD of the [29].wireless The implantable output GIof mothedulation stimulator system iswas shown connected in Figure 7. to a 1 kΩ resistor through PCB pads for current intensity characterization. An exemplary waveform to demonstrate the operation of the wireless implantable GI modulation system is shown in Figure 7.

FigureFigure 7. Stimuli 7. Stimuli of 100 ofHz, 100 2 mA, Hz, 2and mA, 2 andms were 2 ms generated were generated by the bywireless the wireless GI modulation GI modulation system systemon the benchon top. the Cyan, bench pink, top. Cyan,and blue pink, tr andaces blueare command traces are commandsent from sentmicrocontroller from microcontroller unit (MCU) unit in (MCU) the RD, in commandthe RD, demodulated command demodulated by differential by phase diﬀerential shift phasekeying shift (DPSK) keying receiver (DPSK) in receiver EGMD, in and EGMD, the induced and the outputinduced voltage output waveform, voltage respectively. waveform, Tr respectively.aces presented Traces in presentedtime period in of time (a) period40 ms, of(b ()a 10) 40 ms, ms, and (b) ( 10c) 200 ms, μs. and (c) 200 µs.

Figure 7. Stimuli3.4. Preliminary of 100 Hz, In 2Vivo mA, Chronic and 2Test ms were generated by the wireless GI modulation system on the bench top. Cyan,A pink, preliminary and blue in vivo traces chronic are studycommand was performed sent from to microcontrollervalidate the wireless unit implantable (MCU) in GI the RD, command demodulatedmodulation system by differential and the identified phase stimulation shift keying protocol (DPSK) for accelerating receiver inintestinal EGMD, motility. and the Four induced output voltagejuvenile waveform, mini Yucatan respectively. pigs underwent Traces presented a laparoto myin time to receive period EGMDs of (a) 40that ms, were (b )sterilized 10 ms, and by (c) 200 ethylene oxide. Two of the four pigs were the control group that received the EGMDs without turning μs. on the stimulation, while the other two received active EGMDs delivering the stimulation. The EGMD was placed inside a pouch created in the abdominal wall, and the electrode array was sutured 3.4. Preliminaryon top In of Vivo the intestine Chronic that Test was the proximal end of the jejunum identified by the ligament of Treitz (Figure 8a) [11]. Fifteen metal beads were placed inside the intestine, about 5 cm above the segment A preliminary in vivo chronic study was performed to validate the wireless implantable GI modulation system and the identified stimulation protocol for accelerating intestinal motility. Four juvenile mini Yucatan pigs underwent a laparotomy to receive EGMDs that were sterilized by ethylene oxide. Two of the four pigs were the control group that received the EGMDs without turning on the stimulation, while the other two received active EGMDs delivering the stimulation. The EGMD was placed inside a pouch created in the abdominal wall, and the electrode array was sutured on top of the intestine that was the proximal end of the jejunum identified by the ligament of Treitz (Figure 8a) [11]. Fifteen metal beads were placed inside the intestine, about 5 cm above the segment

Micromachines 2019, 10, 525 8 of 11

3.4. Preliminary In Vivo Chronic Test A preliminary in vivo chronic study was performed to validate the wireless implantable GI modulation system and the identified stimulation protocol for accelerating intestinal motility. Four juvenile mini Yucatan pigs underwent a laparotomy to receive EGMDs that were sterilized by ethylene oxide. Two of the four pigs were the control group that received the EGMDs without turning on the stimulation, while the other two received active EGMDs delivering the stimulation. The EGMD Micromachines 2019, 10, x 8 of 12 was placed inside a pouch created in the abdominal wall, and the electrode array was sutured on top of the intestine that was the proximal end of the jejunum identified by the ligament of Treitz where the electrode(Figure8a) array [ 11]. Fifteen was metalsutured. beads wereRight placed after inside surgery, the intestine, an X-ray about 5 contrast cm above the was segment injected into the pig’s stomachwhere through the electrode a gastric array tube was sutured.and X-ray Right images after surgery, were an acquired. X-ray contrast The was EGMD injected was into thethen wirelessly pig’s stomach through a gastric tube and X-ray images were acquired. The EGMD was then wirelessly powered andpowered configured and configured through through the RD the RD to todeliver deliver thethe stimulation stimulation to the electrodeto the electrode that was in thethat was in the middle columnmiddle and column closest and closestto the to tip the tipof ofthe the array. array. Each Ea pigch waspig trained was trained to wear a jacketto wear that carried a jacket the that carried the RD beforeRD the before surgery the surgery (Figure (Figure 88b).b). The The pigs pigs were examinedwere examined three times using three X-ray: times right using after surgery, X-ray: right after one day postoperative, and two days to monitor the transit of the X-ray contrast. surgery, one day postoperative, and two days to monitor the transit of the X-ray contrast.

Figure 8. ExperimentFigure 8. Experiment setup of setup chronic of chronic test. test. (a ()a )EGMD EGMD was was placed placed inside inside a pouch a created pouch in the created in the abdominal wall with its electrode sutured on top of the intestine. Note that this deployment of abdominal walleach with component its electrode of the EGMD sutured helps mitigate on top signal of the loss intestine. of power and Note data that signals this due deployment to tissue of each component ofabsorption, the EGMD as separation helps mitigate between external signal and loss internal of power coils is merely and abdominal-walldata signals thickness,due to tissue while absorption, as separationthe between electrode array external can be extended and internal to access thecoil intestine.s is merely Same deployment abdominal-wall method is also thickness, applicable while the to future human studies to mitigate the issue of signal loss due to tissue absorption. (b) RD was carried electrode arrayby thecan pig. be extended to access the intestine. Same deployment method is also applicable to future human studies to mitigate the issue of signal loss due to tissue absorption. (b) RD was carried X-ray results showed that electrical stimulation likely facilitated intestinal transit as the contrast by the pig.appeared in the rectum in Pig 4 on Day 1, earlier than in the other control pigs (Figure9). Defecation of feces mixed with the X-ray contrast was also observed in Pig 4 on Day 1, but not clearly in other X-ray resultspigs. However, showed the that movement electrical of metal stimulation beads did not likely show coherentfacilitated results intestinal in X-ray photography. transit as the contrast The potential root cause is the small size and slippery surface of the metal beads that make the appeared in thepropelling rectum of the in beads Pig 4 inside on Day the GI 1, tract earlier challenging. than Althoughin the other our preliminary control datapigs using (Figure EGMD 9). Defecation of feces mixedshow with encouraging the X-ray results, contrast further studieswas also involving obse morerved subjects in Pig should 4 on be conducted.Day 1, but not clearly in other pigs. However, theIntestinal-tissue movement histology of metal was also beads performed did to not examine show for anycoherent tissue damage results due toin stimulation. X-ray photography. Three sections of jejunum were compared: normal jejunum distal to the surgical site (Section1), jejunum The potentialthat root was attachedcause tois thethe planar small electrode size arrayand but slippery did not receive surface stimulation of the (Section metal2), and beads jejunum that make the propelling ofthat the was beads attached inside to the the planar GI tract electrode challenging. array and received Although stimulation our preliminary (Section3) (Figure data 10). using EGMD show encouragingSections results,2 and3 showed further some studies mild inflammation involving and more hemorrhage subjects in the should serosa and be longitudinalconducted. muscle layer, likely related to the surgical manipulation of these intestinal sections instead of the stimulation. These changes did not appear to be clinically significant. The three different sections appeared otherwise comparable. There was no difference in the villi or crypts seen on these sections, and the submucosa and muscle layers also appeared similar.

Micromachines 2019, 10, 525 9 of 11 Micromachines 2019, 10, x 9 of 12

FigureFigure 9. Intestinal 9. Intestinal transit transit examined examined throughthrough X-ray X-ray photography. photography. Pigs 1Pigs and 21 areand control 2 are group, control while group, while3 and3 and 4 are 4 are the the experimental experimental group. group. (a) Fifteen (a) Fifteen metal beads metal were beads inside were the inside intestine, the and intestine, X-ray contrast and X-ray was in the pigs’ stomach right after the surgery. (b) On postoperative Day 1, the contrast appeared contrast was in the pigs’ stomach right after the surgery. (b) On postoperative Day 1, the contrast in the rectum in Pig 4 with electrical stimulation, while the contrast had only moved to the intestines appeared in the rectum in Pig 4 with electrical stimulation, while the contrast had only moved to the in other pigs, suggesting that electrical stimulation likely facilitates intestinal transit. Defecation of intestinesfeces was in alsoother observable pigs, suggesting in Pig 4 but that not in electrical other pigs. stimulation (c) Contrast movedlikely tofacilitates the rectum intestinal in all pigs transit. on Defecationpostoperative of feces Day was 2. also Note observable that each image in Pig is 4 combined but not in with other photos pigs. taken (c) Contrast by the camera moved as to the the X-ray rectum Micromachinesin allmachine pigs 2019 on, does10 postoperative, x not support Day an image-capture 2. Note that function,each image and is its combined ﬁeld of view with is not photos broad taken enough by to the show camera10 of 12 as thethe X-ray whole machine GI tract ondoes the not screen. support an image-capture function, and its field of view is not broad enough to show the whole GI tract on the screen.

Intestinal-tissue histology was also performed to examine for any tissue damage due to stimulation. Three sections of jejunum were compared: normal jejunum distal to the surgical site (Section 1), jejunum that was attached to the planar electrode array but did not receive stimulation (Section 2), and jejunum that was attached to the planar electrode array and received stimulation (Section 3) (Figure 10). Sections 2 and 3 showed some mild inflammation and hemorrhage in the serosa and longitudinal muscle layer, likely related to the surgical manipulation of these intestinal sections instead of the stimulation. These changes did not appear to be clinically significant. The three Figure 10. Intestinal-tissue histology. (a) Normal jejunum distal to the surgical site (Section1), differentFigure sections 10. Intestinal-tissue appeared otherwise histology. comparable. (a) Normal jejunum There was distal no to difference the surgical in sitethe villi(Section or crypts 1), (b) seen (b) jejunum that was attached to planar electrode array but did not receive stimulation (Section2), on thesejejunum andsections, (thatc) jejunum was and attached thatthe wassubmucosa to attached planar to andelectrode planar muscle electrode arra layersy but array didalso and not receivedappeared receive stimulation stimulation similar. (Section (Section3). 2), and (c) jejunum that was attached to planar electrode array and received stimulation (Section 3).

4. Conclusions Unlike well-established spinal cord implants and their successful commercialization, there are very few developed GI implantable stimulators, or that have received Food and Drug Administration (FDA) clearance. This might be attributed to the inconclusive GI working mechanism under electrical stimulation and the challenge of developing a versatile implant system with high flexibility. To further the advancement of treating GI dysmotility through electrical neuromodulation (i.e., electroceuticals), we developed a wireless implantable GI modulation system. The system has demonstrated its potential for improving GI motility. The SoC in the implantable EGMD supports wireless power and data telemetries, and a highly programmable stimulator that can generate versatile stimuli. The heterogeneous packaging technique achieves a miniaturized EGMD that is at least an order lighter and 7.7 times smaller than existing commercialized GI implantable stimulators. The RD and the GUI outside the body allow users to wirelessly power and control the EGMD to deliver a tailored stimulation protocol for GI modulation. The system was validated on the bench top and also in preliminary in vivo tests using porcine models. A stimulation protocol for accelerating intestinal transit was obtained and validated in the acute test. The wireless system delivering the proposed protocol likely facilitated GI motility in the preliminary chronic study. In addition, the protocol has shorter pulse width and smaller intensity compared with other works for facilitating GI motility [8–10,12,13], which is beneficial for the implantable device in terms of power consumption and stimulation safety. With the versatile stimulation pattern feature, the wireless implantable system not only serves as a research tool for studying an effective stimulation protocol on various GI segments, but also potential personalized medicine that can customize optimal stimulation strategies for each individual. Its versatility and miniaturized size also make it applicable not only to GI electrical stimulation, but also various biomedical applications, including vagus nerve, spinal cord, and nerve regeneration stimulation.

Funding: This research was partially funded by the U.S. National Science Foundation Small Business Innovation Research Award (award number 1647917) and Endowment of California Capital Equity LLC.

Conflicts of Interest: Y-K.L. and W.L. hold shareholder interest in Niche Biomedical.

References

Micromachines 2019, 10, 525 10 of 11

4. Conclusions Unlike well-established spinal cord implants and their successful commercialization, there are very few developed GI implantable stimulators, or that have received Food and Drug Administration (FDA) clearance. This might be attributed to the inconclusive GI working mechanism under electrical stimulation and the challenge of developing a versatile implant system with high flexibility. To further the advancement of treating GI dysmotility through electrical neuromodulation (i.e., electroceuticals), we developed a wireless implantable GI modulation system. The system has demonstrated its potential for improving GI motility. The SoC in the implantable EGMD supports wireless power and data telemetries, and a highly programmable stimulator that can generate versatile stimuli. The heterogeneous packaging technique achieves a miniaturized EGMD that is at least an order lighter and 7.7 times smaller than existing commercialized GI implantable stimulators. The RD and the GUI outside the body allow users to wirelessly power and control the EGMD to deliver a tailored stimulation protocol for GI modulation. The system was validated on the bench top and also in preliminary in vivo tests using porcine models. A stimulation protocol for accelerating intestinal transit was obtained and validated in the acute test. The wireless system delivering the proposed protocol likely facilitated GI motility in the preliminary chronic study. In addition, the protocol has shorter pulse width and smaller intensity compared with other works for facilitating GI motility [8–10,12,13], which is beneficial for the implantable device in terms of power consumption and stimulation safety. With the versatile stimulation pattern feature, the wireless implantable system not only serves as a research tool for studying an effective stimulation protocol on various GI segments, but also potential personalized medicine that can customize optimal stimulation strategies for each individual. Its versatility and miniaturized size also make it applicable not only to GI electrical stimulation, but also various biomedical applications, including vagus nerve, spinal cord, and nerve regeneration stimulation.

Author Contributions: Y-K.L., W.L., and J.D. study conception and design; P-M.W. and Y-K.L., system development; P-M.W., Y-K.L., and G.D., system validation; P-M.W., article drafting; P-M.W., Y-K.L., and W.L., initiating and leading the study, reviewing, and editing of the article. Funding: This research was partially funded by the U.S. National Science Foundation Small Business Innovation Research Award (award number 1647917) and Endowment of California Capital Equity LLC. Conﬂicts of Interest: Y-K.L. and W.L. hold shareholder interest in Niche Biomedical.

References

1. Peery, A.F.; Crockett, S.D.; Murphy, C.C.; Lund, J.L.; Dellon, E.S.; Williams, J.L.; Jensen, E.T.; Shaheen, N.J.; Barritt, A.S.; Lieber, S.R. Burden and cost of gastrointestinal, liver, and pancreatic diseases in the United States: update 2018. Gastroenterology 2018, 156, 254–272. [CrossRef][PubMed] 2. Sanders, K. Regulation of smooth muscle excitation and contraction. Neurogastroenterol. Motil. 2008, 20, 39–53. [CrossRef][PubMed] 3. Huizinga, J.D.; Lammers, W.J. Gut peristalsis is governed by a multitude of cooperating mechanisms. Am. J. Physiol. Gastrointest. Liver Physiol. 2009, 296, G1–G8. [CrossRef][PubMed] 4. Valentin, N.; Acosta, A.; Camilleri, M. Early investigational therapeutics for gastrointestinal motility disorders: from animal studies to Phase II trials. Expert Opin. Investig. Drugs 2015, 24, 769–779. [CrossRef][PubMed] 5. Sanders, K.M.; Koh, S.D.; Ward, S.M. Interstitial cells of Cajal as pacemakers in the gastrointestinal tract. Annu. Rev. Physiol. 2006, 68, 307–343. [CrossRef] 6. Goyal, R.K.; Hirano, I. The enteric nervous system. New Engl. J. Medicine 1996, 334, 1106–1115. [CrossRef] 7. Furness, J.B. The enteric nervous system and neurogastroenterology. Nat. Rev. Gastroenterol. Hepatol. 2012, 9, 286. [CrossRef] 8. McCallum, R.W.; Chen, J.D.Z.; Lin, Z.; Schirmer, B.D.; Williams, R.D.; Ross, R.A. Gastric pacing improves emptying and symptoms in patients with gastroparesis. Gastroenterology 1998, 114, 456–461. [CrossRef] 9. Lin, Z.; Sarosiek, I.; Forster, J.; Ross, R.A.; Chen, J.D.; McCallum, R.W. Two-channel gastric pacing in patients with diabetic gastroparesis. Neurogastroenterol. Motil 2011, 23, 912-e396. [CrossRef] Micromachines 2019, 10, 525 11 of 11

10. Yin, J.; Chen, J.D. Excitatory effects of synchronized intestinal electrical stimulation on small intestinal motility in dogs. Am. J. Physiol. Gastrointest. Liver Physiol. 2007, 293, G1190–G1195. [CrossRef] 11. Dubrovsky, G.; Lo, Y.-K.; Wang, P.-M.; Wu, M.-D.; Huynh, N.; Liu, W.; Dunn, J.C. Intestinal electrical stimulation to increase the rate of peristalsis. J. Surgical Res. 2019, 236, 153–158. [CrossRef][PubMed] 12. Chen, S.; Li, Y.; Yao, S.; Zhang, Y.; Liu, L.; Guo, X.; Chen, W.; Chen, Y.; Du, Y. Implantable colonic electrical stimulation improves gastrointestinal transit and defecation in a canine constipation model. Neuromodulation: Technol. Neural Interface 2016, 19, 108–115. [CrossRef][PubMed] 13. Sallam, H.; Chen, J. Colonic electrical stimulation: potential use for treatment of delayed colonic transit. Colorectal Dis. 2013, 15, e244–e249. [CrossRef][PubMed] 14. Volkmann, J.; Herzog, J.; Kopper, F.; Deuschl, G. Introduction to the programming of deep brain stimulators. Mov. Disord.: Off. J. Mov. Disord. Soc. 2002, 17, S181–S187. [CrossRef][PubMed] 15. Hershey, B.; Valencia, C.; Yearwood, T.L. Pulse width programming in spinal cord stimulation: a clinical study. Pain Physician 2010, 13, 321–335. 16. Mertens, A.; Raedt, R.; Gadeyne, S.; Carrette, E.; Boon, P.; Vonck, K. Recent advances in devices for vagus nerve stimulation. Expert Rev. Med. Devices 2018, 15, 527–539. [CrossRef] 17. Liu, S.; Chen, J. Colonic electrical stimulation regulates colonic transit via the nitrergic pathway in rats. Digestive diseases and sciences 2006, 51, 502–505. [CrossRef] 18. Sevcencu, C.; Rijkhoff, N.J.; Sinkjaer, T. Colon emptying induced by sequential electrical stimulation in rats. IEEE Trans. Neural Syst. Rehabil. Eng. 2005, 13, 516–523. [CrossRef] 19. McCallum, R.W.; Snape, W.; Brody, F.; Wo, J.; Parkman, H.P.; Nowak, T. Gastric electrical stimulation with Enterra therapy improves symptoms from diabetic gastroparesis in a prospective study. Clin. Gastroenterol. Hepatol. 2010, 8, 947–954. [CrossRef] 20. Rodríguez, L.; Rodriguez, P.; Gómez, B.; Ayala, J.C.; Oxenberg, D.; Perez-Castilla, A.; Netto, M.G.; Soffer, E.; Boscardin, W.J.; Crowell, M.D. Two-year results of intermittent electrical stimulation of the lower esophageal sphincter treatment of gastroesophageal reflux disease. Surgery 2015, 157, 556–567. [CrossRef] 21. Lo, Y.-K.; Wang, P.-M.; Dubrovsky, G.; Wu, M.-D.; Chan, M.; Dunn, J.C.; Liu, W. A wireless implant for gastrointestinal motility disorders. Micromachines 2018, 9, 17. [CrossRef][PubMed] 22. Lo, Y.-K.; Kuan, Y.-C.; Culaclii, S.; Kim, B.; Wang, P.-M.; Chang, C.-W.; Massachi, J.A.; Zhu, M.; Chen, K.; Gad, P. A fully integrated wireless SoC for motor function recovery after spinal cord injury. IEEE Transactions on Biomedical Circuits and Systems 2017, 11, 497–509. [CrossRef][PubMed] 23. Kendir, G.A.; Liu, W.; Wang, G.; Sivaprakasam, M.; Bashirullah, R.; Humayun, M.S.; Weiland, J.D. An optimal design methodology for inductive power link with class-E amplifier. IEEE Trans. Circuits Syst. I: Regul. Pap. 2005, 52, 857–866. [CrossRef] 24. Wang, G.; Liu, W.; Sivaprakasam, M.; Zhou, M.; Weiland, J.D.; Humayun, M.S. A wireless phase shift keying transmitter with Q-independent phase transition time. In Proceedings of the Engineering in Medicine and Biology Society, Shanghai, China, 17–18 January 2006; pp. 5238–5241. 25. Chang, C.-W.; Lo, Y.-K.; Gad, P.; Edgerton, R.; Liu, W. Design and fabrication of a multi-electrode array for spinal cord epidural stimulation. In Proceedings of the Engineering in Medicine and Biology Society (EMBC), Chicago, IL, USA, 26–30 August 2014; pp. 6834–6837. 26. Liu, W.; Wang, P.-M.; Lo, Y.-K. Towards closed-loop neuromodulation: a wireless miniaturized neural implant SoC. In Proceedings of the SPIE Defense+ Security, Anaheim, CA, USA, 9–13 April 2017; pp. 1019414–1019418. 27. Lin, X.; Hayes, J.; Peters, L.; Chen, J. Entrainment of intestinal slow waves with electrical stimulation using intraluminal electrodes. Ann. Biomed. Eng. 2000, 28, 582–587. [CrossRef][PubMed] 28. Lo, Y.-K.; Chang, C.-W.; Liu, W. Bio-impedance characterization technique with implantable neural stimulator using biphasic current stimulus. In Proceedings of the Engineering in Medicine and Biology Society (EMBC), Chicago, IL, USA, 26–30 August 2014; pp. 474–477. 29. Massachi, J.; Lo, Y.-K.; Wang, P.-M.; Liu, W. A Wireless Platform to Support Pre-Clinical Trial of Neural Implant for Spinal Cord Injury. In Proceedings of the 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Honolulu, HI, USA, 18–21 July 2018; pp. 5487–5490.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).