Stimulator consisted of a Micro controller (MSP430F2013, TI), a stimulating waveform generator, and a data receiver (IR LED). The stimulating waveform was generated by the following process and Fig. 3A shows block diagram for biphasic pulse stimulation. First, the pulse-width modulation (PWM) waveform was generated by the Micro controller. Then, a primary discharging capacitor is charged by an op-amp buffer, and simultaneously, a secondary discharging capacitor is also charged by a voltage divider. The charge of the capacitors is converted into an asymmetric biphasic pulse by analog switches. Since an unbalanced stimulus pulse results in a net ion flow, tissue damage, and corrosion of anode electrode surface, a charge-balanced waveform is important. As shown in Fig. 3B, one channel of the circuit generated a square pulse (Vcathode) and another channel generated a delayed square pulse (Vanode : amplitude of 1/50, duration of 50 pulses). This resulted in the generation of a charge balanced biphasic stimulus pulse (Vstim) from Vcathode and Vanode. In the cathode pulse, stimulus current depolarized nearby axons and initiated the action potential, and the anodic pulse cancelled the charge accumulated by the cathode pulse. By controlling the PWM signal, the amplitude of the waveform can be controlled. The external IR remote enabled voltage control with a 0.1 V increase. We configured a basic setting of waveform as 60 µs, 130 Hz, and 1.0 V.

The wireless power transmission system consisted of a full-bridge rectifier and a secondary coil, which could receive the power based on the principle of inductive coupling¹³⁾. Fig. 4 illustrates the receiver, consisting of L2 (7 mH), C2 (4.021 nF), and the full-bridge rectifier. To maximize the transmission efficiency, the resonance frequency was chosen as 30 kHz.

Wireless power transmission system was designed to have a Class-D power amplifier, which enables the primary coil to generate magnetic field and the implanted secondary coil to generate current. PWM signal was generated by the Micro controller (MSP430F2013, TI) and a Class-D switching amplifier chip (TC4420, TI)^5,13). As shown in Fig. 4, PWM frequency was 30 kHz, which was chosen as the resonant frequency of the L1 (621 µH) and C1 (45.322 nF).

We induced PD by injecting 6-hydroxydopamine (6-OHDA) (12 µg/3 µL) into the right medial forebrain bundle in the rat brain using a stereotaxic frame. 6-OHDA is known to show selective neurotoxicity to catecholaminergic neurons, and it is reported that 6-OHDA also induces PD-like behavioral similarity in primates when injected into the substantia nigra. The PD-rat model was prepared by the following procedure : 1) Rats (250-300 g) were placed on a stereotaxic frame plate. A hemostatic machine was used to maintain the body temperature throughout the procedure. 2) The horizontal alignment was checked holding the ear bar of the frame in the rat's ears. 3) A skin incision was made and the membrane on the skull was stripped away by using a swab. 4) Indicate the injection site (AP : -4.4 mm, ML : -2.1 mm, DV : -7.8 mm) and drill a hole. 5) A Hamilton syringe filled with the required amount of 6-OHDA (3 µL) was delivered to the target depth (-7.8 mm) and the drug was infused slowly (0.25 µL/min). 6) Apomorphine-induced rotation tests, which checks for loss of dopaminergic neuron, were performed from the first week on alternate weeks. Uniaxial rotational movement was observed after hypodermic injection to the back of the neck in PD-animals.

At the completion of the last behavior experiments, each rat was deeply anesthetized and perfused transcardially with about 50 mL of 0.9% NaCl (with heparine), followed by about 50 mL of ice-cold 4% paraformaldehyde. After perfusion, the rat brains were removed and placed, for one day, in a 50 mL conical tube containing ice-cold 4% paraformaldehyde. The rat brains were then transferred sequentially at one-day intervals into 10%, 20%, and 30% sucrose until they sank to the bottom of their container. The rat brains were then entirely cut into 30 µm-thick coronal sections on a freezing cryostat (Leica CM 3000, Leica, Solms, Germany). Every fifth section was collected for staining with cresyl violet to confirm the electrode trajectory at the nuclei [subthalamic nucleus (STN)].

The performance tests for the circuits were conducted twice. We could get the same test results of the output waveform before and after encapsulation. We observed the output waveform when only resister components existed by connecting a 1 kΩ resister and measured the power that the circuit consumed. Initial setting of asymmetric biphasic pulse was generated for 130 Hz of frequency, 60 µs pulse width, and 1 V amplitude.

To expect the in vivo output waveform, we observed the output waveform after connecting the experimental electrodes and dipping them in physiological saline. In the physiological saline test, generated waveform distortion did not occur, which was similar to the result obtained with the 1 kΩ resister test. Moreover, there were no side effects, such as air bubbles on the electrode surface, over a one-hour test period (Fig. 6). The output increased linearly with a 0.1 V increment, according to the external remote control.

As mentioned before, the wireless power transmission system consists of transmission unit and receiving unit. The inductor (L) and capacitor (C) were selected to set the resonance frequency as 30 kHz (Table 1). Neodymium magnet was used to solve misalignment of coils by fixing them⁶⁾. In addition, we tested the uptime of the system based on 500 mAh lithium-polymer battery. The test was conducted while the distance between the transmission unit and the receiving unit was fixed as 2 mm. The results of the test showed that the system was operable for 24 h based on a 500 mAh lithium-polymer battery.

We tested the behavior and implanted device status one week after implanting the device in three PD models wearing the jacket with the external device and battery. The PD models were moving freely as illustrated in Fig. 7A and B. Through the red LED light of the implanted device, the working status of the device could be confirmed (Fig. 7C, D).

The effectiveness of the device was verified through the rotation test, as shown in Table 2 and the possibility of long-term experiment was confirmed through the four-week experiment. We implanted the device in two (PD_#2 and PD_#3) of the three PD models, but gave stimulus to only one (PD_#3) of the two models (1.0 V, 130 Hz, 60 µs). After four weeks, the rotation tests were conducted on each model. As illustrated in Fig. 8, PD_#1, which did not get the device, had the greatest number of turns. We also confirmed that PD_#3, under stimulation for four weeks, had the lowest number of turns. In addition, based on the number of turns of PD_#2, it was clear that the symptom could be relieved just by implanting the device.

To test the safety of the implantable device, we checked if brain region stimulated by the implanted electrodes was damaged four weeks after the implantation. From the analyses of the brain sections from the test animals, corresponding to the site of the electrodes, there was no damaged tissue near the region of implantation, as shown in Fig. 9.

The stability of the device was verified four weeks after the implantation by checking the status of the device. Fig. 10 shows that the capsule did not have any problem after long-term implantation. In addition, we confirmed the effectiveness of the output performance by comparing the output waveform before and after the implantation.