Yi-Kai Lo

A Fully-Integrated High-Compliance Voltage SoC for Epi-Retinal and Neural Prostheses

This paper presents a fully functionally integrated 1024-channel mixed-mode and mixed-voltage system-on-a-chip (SoC) for epi-retinal and neural prostheses. Taking an AC input, an integrated power telemetry circuits is capable of generating multiple DC voltages with a voltage conversion efficiency of 83% at a load of 100 mW without external diodes or separate power integrated circuits, reducing the form factor of the prosthetic device. A wireless DPSK receiver with a novel noise reduction scheme supports a data rate of 2 Mb/s at a bit-error-rate of 2 × 10-7. The 1024-channel stimulator array meets an output compliance voltage of ±10 V and provides flexible stimulation waveforms. Through chip-clustering, the stimulator array can be further expanded to 4096 channels. This SoC is designed and fabricated in TSMC 0.18 μm high-voltage 32 V CMOS process and occupies a chip area of 5.7 mm × 6.6 mm. Using this SoC, a retinal implant bench-top test system is set up with real-time visual verification. In-vitro experiment conducted in artificial vitreous humor is designed and set-up to investigate stimulation waveforms for better visual resolution. In our in-vivo experiment, a hind-limb paralyzed rat with spinal cord transection and implanted chronic epidural electrodes has been shown to regain stepping and standing abilities using stimulus provided by the SoC.

A 37.6mm 2 1024-channel high-compliance-voltage SoC for epiretinal prostheses

Retinal implants elicit light perception for people blinded by photoreceptor loss. Commercialized 60-channel retinal prostheses allow patients to perform simple tasks, but several hundreds to a thousand electrodes are required for face recognition/reading [1,2], posing great challenges for the design of next-generation retinal stimulators. Aside from the higher power/data demand, the electrode impedance is also increased. Placing 1000 epiretinal electrodes in the 5mm-diameter macula region reduces the electrode size to less than 0.01mm2, leading to a 30kΩ electrode-tissue impedance [2]. To elicit light perception of various brightness levels, the stimulators for epiretinal prostheses require an output compliance voltage of ±10V [3], thus requiring area-consuming high-voltage (HV) transistors. The stimulator in [4] achieves 1600 channels, but it is designed for subretinal rather than epiretinal prostheses. It has ±2V compliance and needs a separate chip for power telemetry. For epiretinal prostheses, an HV-compliant 1024-channel stimulator in [3] is estimated to occupy 64mm2 and requires off-chip diodes for power rectification. For a space-restricted retinal implant, a small-sized fully integrated SoC with minimal number of off-chip components is preferred.

Precision control of pulse widths for charge balancing in functional electrical stimulation

Maintaining a balance charge is one of the key factors to ensure a safe neural stimulation employing biphasic current stimulus. However, a zero net charge is difficult to achieve due to the fabrication variation of stimulation drivers. Moreover, even with a perfectly matched cathodic and anodic current stimulus, a non-zero residual charge is still built up due to the inter-pulse delay adopted in the stimulation pattern. During this period, the charge injected by the first stimulus pulse is leaked to the surrounding tissue and thus, the remaining charge cannot be completely removed by the following compensating current stimulus even with matched intensity. In this paper, we present a charge-balancing scheme by precisely controlling the pulse width of current stimulus. Charge balance is achieved by using the residual voltage as a feedback signal to control the pulse width of next stimulus. Simulated in Matlab/Simulink, the proposed scheme is shown to mitigate up to 15% intensity mismatch of biphasic current pulses with 1ms inter-pulse delay. For a stimulus with constantly varying intensity, such as retinal and cochlear implants, the proposed scheme is also capable of maintaining a balance charge for a safe neural stimulation.

22.2 A 176-channel 0.5cm3 0.7g wireless implant for motor function recovery after spinal cord injury

Epidural spinal stimulation has shown effectiveness in recovering the motor function of spinal cord transected rats by modulating neural networks in lumbosacral spinal segments [1, 2]. The state-of-the-art neuromodulation implant [3] reports a 4-channel stimulator with wireless data and power links for small animal experiments, yet weighs 6g and has a volume of 3cm3. It is preferable that the implant package has a comparable size to its bioelectronics and a high-density stimulator to support stimulation with high spatial resolution. Furthermore, the epidural electrode should be soft and flexible because a mechanical mismatch exists at the tissue-electrode interface [1]. Unlike other implant/SoCs that stimulate with pre-loaded patterns [4-5], the implant for motor function recovery should be capable of adaptively adjusting its stimulation patterns at run time in response to the subjects varying physiological states [2]. Measuring the electrode-tissue impedance is also critical to ensure safe stimulation. Deriving the equivalent circuit model of the electrode-tissue interface determines the safe stimulation boundary (i.e. pulse width and intensity) to ensure the electrode overpotential is within the water window [6]. However, an SoC implementation of this function has not been reported.

An On-Chip Multi-Voltage Power Converter With Leakage Current Prevention Using 0.18

In this paper, we present an on-chip multi-voltage power converter incorporating of a quad-voltage timing-control rectifier and regulators to produce ±12 V and ±1.8 V simultaneously through inductive powering. The power converter achieves a PCE of 77.3% with the delivery of more than 100 mW to the implant. The proposed rectifier adopts a two-phase start-up scheme and mixed-voltage gate controller to avoid substrate leakage current. This current cannot be prevented by the conventional dynamic substrate biasing technique when using the high-voltage CMOS process with transistor threshold voltage higher than the turn-on voltage of parasitic diodes. High power conversion efficiency is achieved by 1) substrate leakage current prevention, 2) operating all rectifying transistors as switches with boosted gate control voltages, and 3) compensating the delayed turn-on and preventing reverse leakage current of rectifying switches with the proposed look-ahead comparator. This chip occupies an area of 970 μm × 4500 μm in a 0.18 μm 32 V HV CMOS process. The quad-voltage timing-control rectifier alone is able to output a high DC voltage at the range of [2.5 V, 25 V]. With this power converter, both bench-top experiment and in-vivo power link test using a rat model were validated.

\mu

Knowledge of the bio-impedance and its equivalent circuit model at the electrode-electrolyte/tissue interface is important in the application of functional electrical stimulation. Impedance can be used as a merit to evaluate the proximity between electrodes and targeted tissues. Understanding the equivalent circuit parameters of the electrode can further be leveraged to set a safe boundary for stimulus parameters in order not to exceed the water window of electrodes. In this paper, we present an impedance characterization technique and implement a proof-of-concept system using an implantable neural stimulator and an off-the-shelf microcontroller. The proposed technique yields the parameters of the equivalent circuit of an electrode through large signal analysis by injecting a single low-intensity biphasic current stimulus with deliberately inserted inter-pulse delay and by acquiring the transient electrode voltage at three well-specified timings. Using low-intensity stimulus allows the derivation of electrode double layer capacitance since capacitive charge-injection dominates when electrode overpotential is small. Insertion of the inter-pulse delay creates a controlled discharge time to estimate the Faradic resistance. The proposed method has been validated by measuring the impedance of a) an emulated Randles cells made of discrete circuit components and b) a custom-made platinum electrode array in-vitro, and comparing estimated parameters with the results derived from an impedance analyzer. The proposed technique can be integrated into implantable or commercial neural stimulator system at low extra power consumption, low extra-hardware cost, and light computation.

m High-Voltage CMOS Process

A detailed design, fabrication, characterization and test of a flexible multi-site platinum/polyimide based electrode array for electrical epidural stimulation in spinal cord prosthesis is described in this paper. Carefully designed 8.4 μm-thick structure fabrication flow achieves an electrode surface modification with 3.8 times enhanced effective surface area without extra process needed. Measured impedance and phase of two type of electrodes are 2.35±0.21 KΩ and 2.10±0.11 KΩ, -34.25±8.07° and -27.71±8.27° at 1K Hz, respectively. The fabricated arrays were then in-vitro tested by a multichannel neural stimulation system in physiological saline to validate the capability for electrical stimulation. The measured channel isolation on adjacent electrode is about -34dB. Randles cell model was used to investigate the charging waveforms, the model parameters were then extracted by various methods. The measured charge transfer resistance, double layer capacitance, and solution resistance are 1.9 KΩ, 220 nF and 15 KΩ, respectively. The results show that the fabricated array is applicable for electrical stimulation with well characterized parameters. Combined with a multichannel stimulator, this system provides a full solution for versatile neural stimulation applications.

Bio-impedance characterization technique with implantable neural stimulator using biphasic current stimulus.

Implantable functional electrical stimulation (IFES) has demonstrated its effectiveness as an alternative treatment option for diseases incurable pharmaceutically (e.g., retinal prosthesis, cochlear implant, spinal cord implant for pain relief). However, the development of IFES for gastrointestinal (GI) tract modulation is still limited due to the poorly understood GI neural network (gut–brain axis) and the fundamental difference among activating/monitoring smooth muscles, skeletal muscles and neurons. This inevitably imposes different design specifications for GI implants. This paper thus addresses the design requirements for an implant to treat GI dysmotility and presents a miniaturized wireless implant capable of modulating and recording GI motility. This implant incorporates a custom-made system-on-a-chip (SoC) and a heterogeneous system-in-a-package (SiP) for device miniaturization and integration. An in vivo experiment using both rodent and porcine models is further conducted to validate the effectiveness of the implant.

Design and fabrication of a multi-electrode array for spinal cord epidural stimulation

Recovering neural responses from electrode recordings is fundamental for understanding the dynamics of neural networks. This effort is often obscured by stimulus artifacts in the recordings, which result from stimuli injected into the electrode-tissue interface. Stimulus artifacts, which can be orders of magnitude larger than the neural responses of interest, can mask short-latency evoked responses. Furthermore, simultaneous neural stimulation and recording on the same electrode generates artifacts with larger amplitudes compared to a separate electrode setup, which inevitably overwhelm the amplifier operation and cause unrecoverable neural signal loss. This paper proposes an end-to-end system combining hardware and software techniques for actively cancelling stimulus artifacts, avoiding amplifier saturation, and recovering neural responses during current-controlled in-vivo neural stimulation and recording. The proposed system is tested in-vitro under various stimulation settings by stimulating and recording on the same electrode with a superimposed pre-recorded neural signal. Experimental results show that neural responses can be recovered with minimal distortion even during stimulus artifacts that are several orders greater in magnitude.

A Wireless Implant for Gastrointestinal Motility Disorders

We report a transparent, bilayer-nanomesh microelectrode array for concurrent electrophysiology recording and two-photon imaging. Transparent microelectrode arrays have emerged as increasingly important tools for neuroscience by allowing simultaneous coupling of big and time-resolved electrophysiology data with optically measured, spatially and type resolved single neuron activity. Scaling down transparent electrodes to the length scale of a single neuron is challenging since conventional transparent conductors are limited by their capacitive electrode/electrolyte interface. In this study, we establish transparent microelectrode arrays with high performance, great biocompatibility, and comprehensive in vivo validations from a recently developed, bilayer-nanomesh material composite, where a metal layer and a low-impedance faradaic interfacial layer are stacked reliably together in a same transparent nanomesh pattern. Specifically, flexible arrays from 32 bilayer-nanomesh microelectrodes demonstrated near-unity yield with high uniformity, excellent biocompatibility, and great compatibility with state-of-the-art wireless recording and real-time artifact rejection system. The electrodes are highly scalable, with 130 kilohms at 1 kHz at 20 μm in diameter, comparable to the performance of microelectrodes in nontransparent Michigan arrays. The highly transparent, bilayer-nanomesh microelectrode arrays allowed in vivo two-photon imaging of single neurons in layer 2/3 of the visual cortex of awake mice, along with high-fidelity, simultaneous electrical recordings of visual-evoked activity, both in the multi-unit activity band and at lower frequencies by measuring the visual-evoked potential in the time domain. Together, these advances reveal the great potential of transparent arrays from bilayer-nanomesh microelectrodes for a broad range of utility in neuroscience and medical practices.