Wireless device with energy management for mlosed-loop deep brain stimulation (CLDBS)

Citation: Matheus Nordi, T.; Augusto Ginja, G.; Gounella, R.; Talanoni Fonoff, E.; Colombari, E.; Moreira, M.M.A.; Afonso, J.A.; Monteiro, V.; Afonso, J.L.; Carmo, J.P. Wireless Device with Energy Management for Closed-Loop Deep Brain Stimulation (CLDBS). Electronics 2023,12, 3082. https://doi.org/10.3390/ electronics12143082 Academic Editors: Luigi Scarcello, Carlo Mastroianni and Teemu Leppänen Received: 8 June 2023 Revised: 8 July 2023 Accepted: 11 July 2023 Published: 14 July 2023 Copyright: © 2023 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 (https:// creativecommons.org/licenses/by/ 4.0/). electronics Article Wireless Device with Energy Management for Closed-Loop Deep Brain Stimulation (CLDBS) Tiago Matheus Nordi 1, Gabriel Augusto Ginja 1, Rodrigo Gounella 1, Erich Talanoni Fonoff 2, Eduardo Colombari 3, Melkzedekue M. Alcântara Moreira 4, Jose A. Afonso 5,6 , Vitor Monteiro 7, Joao L. Afonso 7,* and João Paulo Carmo 1 1Group of Metamaterials Microwaves and Optics (GMeta), Department of Electrical Engineering (SEL), University of São Paulo (USP), Avenida Trabalhador São-Carlense, Nr. 400, Parque Industrial Arnold Schimidt, São Carlos CEP 13566-590, SP, Brazil; [email protected] (T.M.N.); [email protected] (G.A.G.); [email protected] (R.G.); [email protected] (J.P.C.) 2Department of Neurology, Faculty of Medicine, Avenida Dr. Arnaldo, Nr. 455, Cerqueira César, São Paulo CEP 01246-903, SP, Brazil; [email protected] 3Department of Physiology and Pathology, Faculty of Odontology, São Paulo State University (UNESP), Rua Humaitá, Nr. 1680, Araraquara CEP 14801-385, SP, Brazil; [email protected] 4Department of Mechancial Engineering (SEM), University of São Paulo (USP), Avenida Trabalhador São-Carlense, Nr. 400, Parque Industrial Arnold Schimidt, São Carlos CEP 13566-590, SP, Brazil; [email protected] 5CMEMS-UMinho, University of Minho, 4800-058 Guimarães, Portugal; [email protected] 6LABBELS—Associate Laboratory, 4800-058 Guimarães, Portugal 7ALGORITMI Research Centre/LASI, University of Minho, 4800-058 Guimarães, Portugal; [email protected] *Correspondence: [email protected] Abstract: Deep brain stimulation (DBS) is an effective and safe medical treatment that improves the lives of patients with a wide range of neurological and psychiatric diseases, and has been consolidated as a first-line tool in the last two decades. Closed-loop deep brain stimulation (CLDBS) pushes this tool further by automatically adjusting the stimulation parameters to the brain response in real time. The main contribution of this paper is a low-size/power-controlled, compact and complete CLDBS system with two simultaneous acquisition channels, two simultaneous neurostimulation channels and wireless communication. Each channel has a low-noise amplifier (LNA) buffer in differential configuration to eliminate the DC signal component of the input. Energy management is efficiently done by the control and communication unit. The battery supports almost 9 h with both the acquisition and stimulation circuits active. If only the stimulation circuit is used as an Open Loop DBS, the battery can hold sufficient voltage for 24 h of operation. The whole system is low-cost and portable and therefore it could be used as a wearable device. Keywords: closed-loop deep brain stimulation (CLDBS); neurostimulation; implantable devices; internet of things (IoT); energy management 1. Introduction Most human motor impairment and dysfunctions originate in the nervous system and the signal transmission between neurons is electrochemical. In this context, electrophysiology studies how these interactions work and an important procedure that is used in this branch of neuroscience is deep brain stimulation (DBS). DBS is a treatment that uses an electrode attached to an implantable pulse generator (IPG) or a neurostimulator that generates electrical impulses inside the brain and, therefore, the central nervous system partially or totally recovers motor functions in patients with neurological diseases such as Parkinson’s and essential tremors [ 1 ]. Figure 1depicts the concept and positioning of a DBS system. Electronics 2023,12, 3082. https://doi.org/10.3390/electronics12143082 https://www.mdpi.com/journal/electronics Electronics 2023,12, 3082 2 of 21 Figure 1. DBS system positioning (adapted from [1]). The application of DBS in motor rehabilitation is an alternative to treatments that use remedy and surgery to remove some of the symptoms of motor impairments. The first surgery procedures used to reduce tremors caused damage to the brain and subsequent procedures involved the use of Levodopa which is a remedy that relieves involuntary movements [ 2 ]. However, Levodopa had collateral effects such as psychosis and hallucinations. As a solution that does not harm the brain and does not create psychological dysfunctions, DBS was successfully used to treat Parkinson’s disease [ 3 , 4 ], essential tremors [ 5 , 6 ], dystonia [ 7 , 8 ] and chronic pain [ 9 , 10 ]. Furthermore, DBS is also used to treat behavioral dysfunctions such as obsessive–compulsive disorder [ 11 ], morbid obesity [ 12 ] and depression [13–15]. The DBS system control of the amplitude and frequency is performed either by a user that manually changes the values or by itself through acquiring biomarkers and adjusting the parameters according to the response of the patient. For example, a biomarker may be an electrical signal generated by cells of a human organism and is an indicator of the response of the human body to electrical stimuli. The first described system is called an open-loop DBS whereas the second one is called a closed DBS. Open-loop DBS has around 75% efficacy in motor treatment; however, the use of open-loop DBS in psychiatric diseases does not return good results, which indicates that a sophisticated closed-loop DBS may be more suitable for these applications [16]. Figure 2illustrates the types of control on DBS systems with a comparison between open-loop DBS and closed-loop DBS [ 16 ]. Figure 2also illustrates the concept behind the connections required for stimulation and for the acquisition of biopotentials. Figure 2a is the simple open-loop DBS with fixed parameters and a continuous pulse sequence and Figure 2b is a similar system with bursts of pulses separated by a constant time interval. In both cases, the parameters can only be modified by a healthcare worker and frequency and amplitude are constant. Figure 2c shows an on/off responsive control, which is a closed-loop DBS where the system responds to the activation of brain cells by changing the pulse width. When the biopotential exceeds a threshold value, the DBS system stops the stimulation, and as the biopotential achieves a value below the threshold, the DBS reactivates the stimulation. Bouthour et al. [ 17 ] propose an on/off responsive DBS system by monitoring the frequency components registered by local field potentials produced in the subthalamic nucleus of the brain. A setpoint value is defined, and, if an amplitude signal of the biomarker is greater than the setpoint, the neurostimulation signal is interrupted. Otherwise, if the biomarker signal is lower than the setpoint, the neurostimulation signal is reactivated. Figure 2d depicts a closed-loop DBS named adaptive control, like in Figure 2c, but with a variation in the pulse width and the amplitude of the stimulation signal. Finally, Figure 2e illustrates the dynamic stimulation signal control, which is the most complete closed-loop DBS system, since it uses multiple channels to generate pulses in addition to the other features of the system of Figure 2d. Figure 2f shows the connections required for stimulation and acquisition. Electronics 2023,12, 3082 3 of 21 Figure 2. Types of control on DBS systems with a comparison between ( a , b ) open-loop DBS and ( c – e ) closed-loop DBS, and ( f ) illustration of the connections required for stimulation and for the acquisition of biopotentials. Adapted from Hoang et al., [16]. The majority of works found in the literature and in the market comprise simple neurostimulators for conventional DBS without acquisition. In fact, only a small portion of works allow DBS with adaptive stimulation. The lack of works regarding closed-loop DBS systems (CLDBS) was one of the main motivations behind the proposed system. The main contribution of this paper is a low-size/power-controlled, compact and complete CLDBS system with two simultaneous acquisition channels, two simultaneous neurostimulation channels and wireless communication. This CLDBS system was designed for its low size and easy positioning and at the same time for energy management targeting high efficiency. The rest of this paper is organized as follows: Section 2presents the design of the proposed system, and Section 3presents the obtained results. Finally, the conclusions are presented in Section 4. 2. Proposed Design 2.1. Low-Noise Amplifier (LNA) The DBS system presented in this paper is a closed-loop system with both acquisition and stimulation circuits. Figure 3a shows the block diagram of the proposed DBS system. To provide a constant electrical current and to control the waveform parameters, a Howland current pump (HCP) was used alongside a microcontroller. The HCP receives a control signal from a digital-to-analog converter (DAC) [ 18 , 19 ]. The DAC that generates the waveform is internal to the ESP32 microcontroller, which will receive the information from the signals acquired through an analog-to-digital converter (ADC), which is also internal to the ESP32. The choice of the ESP32 development platform is also justified based on its wireless communication interfaces, such as Wi-Fi and Bluetooth. Moreover, the ESP32 has two cores, which allows the DBS system to control the stimulation and acquisition tasks individually. An internal program receives the input signal data and controls the values of the stimulation waveform parameters. Figure 3b shows the block diagram of the stimulation system, composed of the wireless module (ESP32 board), the DAC converters, the buffers, the constant current HCP circuits and the stimulation electrodes. Electronics 2023,12, 3082 4 of 21 Figure 3. ( a ) Block Diagram for the stimulator circuit of the DBS system. ( b ) Neurostimulator block diagram. (c) Acquisition system block diagram. The acquisition circuit receives the signals delivered by electrodes inserted in brain regions of interest monitoring during stimulation of the brain, sites that can provide biomarker signals for a better analysis of the performance of the DBS and be the feedback loop in a closed-loop stimulation system. The first stage of the circuit has two channels with low-noise amplifiers (LNAs) in the differential configuration and is insensitive to DC levels present in the input signal. Both LNAs have the function of improving the signal from the electrodes and blocking the DC levels that are often present on the acquired signal. The second stage has two acquisition channels that conduct the output signals to a multiplex (MUX) and a buffer to promote a good impedance match between the stages. The next stage is an analog notch filter (ANF) to remove the noise component from the supply. The subsequent stage of the circuit consists of an active low-pass filter (ALPF) that will promote the attenuation of high-frequency noise and provide more gain to the signal. Finally, the signal is conducted to an ADC from the microcontroller ESP32. Figure 3c shows the diagram of blocks of the biopotential acquisition circuit. The operational amplifier chosen for the design of the signal conditioning is the AD8609 manufactured by Analog Devices, a low-power precision amplifier with low-noise rail-to-rail input technology, a typical common-mode rejection ratio (CMRR) of 100 dB at the output and noise density of 22 nVHz-1/2. The gain of the first stage was designed to be Electronics 2023,12, 3082 5 of 21 around 33 dB and the gain of the ALPF at approximately 21 dB, so the total loop gain is 54 dB. The frequency of ANF was designed for 60 Hz, while the cutoff frequency of the ALPF was designed to be at approximately 10.6 kHz. 2.2. Neurostimulator The control unit generates the stimulation signal by an analog output created by a digital-to-analog converter (DAC) of the control unit, which is the input of each channel of the neurostimulator. Even though the pulse amplitude is defined by the control unit the amplitude value could be affected by the impedance of the brain tissue which could vary according to the region where the electrode is implanted. To bypass this problem, the stimulation circuit generates an electrical current with a value proportional to the voltage input and constant regarding the brain impedance. This system is the HCP represented in Figure 4. The quantity I out is the output current and Z L is the impedance of the brain tissue represented as the load impedance. Figure 4. Howland current pump (HCP) circuit diagram. The quantity V in is the input voltage given by the control unit and R 1 ,R 2 ,R 3 and R 4 are fixed parameters, which means that the microcontroller can directly alter the current even if the impedance of the body tissue changes during the application of the DBS. The output electrical current I out produced by the HCP circuit is given by the following equation: Iout =vin R2+R3 ×1+2R3 R4(1) The values used for the built circuit are R 1 =R 2 = 12.1 k Ω ,R 3 = 10 k Ω and R 4 = 2.2 k Ω , and the output current Iout is 1.4 mA for the maximum output voltage of the DAC. Electronics 2023,12, 3082 6 of 21 2.3. Acquisition Circuit The first stage has an LNA with an elevated gain to amplify low-amplitude signals from biomarkers and improve those signals for the following stages. The LNA also has a high entrance impedance. The LNA circuit has a differential configuration to remove DC components from the input signal. Figure 5shows the circuit diagram of the LNA with C 1 and C 2 capacitances to remove the DC component alongside two resistances R 1 to polarize the circuit. The electrical supply of the LNA is provided by a single V DD , while the common-mode voltage V cm is equal to half of the voltage supply, i.e., Vcm =VDD/2 = (3.3 V)/2 = 1.65 V. Figure 5. LNA circuit diagram. The schematic circuit of Figure 6has both the inverter and non-inverter configurations that generate V 1 and V 2 , respectively. Both the input and feedback impedances Z 1 and Z 2 are given by: Z1=1 sC1(2) and Z2=R2//1 sC2=R2×1 sC2 R2+1 sC2=R2 sR2C2+1(3) The ratio of the impedances Z1and Z2with (Z1+ Z2) is given by: Z1 Z1+Z2=sR2C2+1 sR2(C1+C2) + 1(4) and Z2 Z1+Z2=sR2C1 sR2(C1+C2) + 1(5) The complete LNA forms a difference amplifier; thus, the voltage V out at the output of LNA is given by: Electronics 2023,12, 3082 7 of 21 Vout =−(β−1)×A(s) 1−βA(s)×(VIM)−A(s) 1−βA(s)×(V+) =−(β−1)×A(s) 1−βA(s)×(VIM)−A(s) 1−βA(s)×Z2 Z1+Z2×(VIP) =−(β−1)×A(s) 1−βA(s)×−Vin 2−A(s) 1−βA(s)×Z2 Z1+Z2×Vin 2 (6) Figure 6. Closed − loop gains for a few combinations of C 2 = {10, 100} nF, and R 1 = {1, 10, 100} M Ω , and their comparison with the open-loop gain A(f) of a generic OpAmp with only one pole. The feedback factor βand its inverse are defined, respectively, as: β=Z1 Z1+Z2(7) and 1 β=Z1+Z2 Z1(8) Therefore, the gain of the LNA (i.e., the feedback gain) Af(s) is given by: Af(s) = Vout Vin =A(s) βA(s)−1×Z2 Z1+Z2 =−A(s) βA(s)−1×sR2C1 sR2(C1+C2)+1 (9) At the medium frequencies, C 1 +C 2≈ C 1 , and [sR 2 (C 1 +C 2 ) + 1] ≈ sR 2 C 1 , making the midband voltage gain equal to: Af(s) = Vout Vin =A(s) βA(s)−1(10) Normally, in the majority of commercial operational amplifiers, A(s) can be expressed with a single pole, or at least with a dominant pole next to the origin. However, the additional poles must be taken into account if they exist. This particular case is a good and general example to deduce the feedback gain and to understand the respective frequency behavior. In the particular case of only one pole: A(s) = A0 s1 2πfp+1,A0ˇ1 (11) Electronics 2023,12, 3082 8 of 21 At the medium frequencies, A(s) » 1, and the midband voltage gain results on: Af(s)≈1 β=Z1+Z2 Z2=sR2(C1+C2) + 1 sR2C2+11(12) At the medium frequencies, it is already known that [sR 2 (C 1 +C 2 ) + 1] ≈ sR 2 C 1 , which is combined with [sR 2 C 2 + 1] ≈ sR 2 C 2 at these frequencies, so the midband voltage gain is then equal to: Af(s) = C1 C2(13) Figure 6illustrates plots of generic transfer functions of the closed-loop gains, where it is possible to observe the existence of one zero and two poles, as well as their relative positions. These positions are similar to the zeros and poles of the LNA of this paper, because Equation (11), i.e., the full equation of the gain of the LNA for all frequencies, also contains a zero f z = 0 Hz in the origin. This zero is canceled by the first pole f L = 1/(2 π R 2 C 2 ), flattening the feedback gain. The feedback gain starts to decrease from the midband gain of A1[dB] at the frequency of: f1=GBW ×10−A1 20 = (fp×10A0 20 )×10−A2 20 =fp×10A0−A1 20 .(14) Figure 7shows the active notch filter (ANF) used to remove the 60 Hz frequency from the energy supply. The Rvalue used was 10 MΩand the Cvalue was 270 pF. Figure 7. Active notch filter (ANF) circuit diagram. A low-pass filter with gain (ALPF) was built as shown in Figure 8with R 1 = 100 k Ω , R2=1MΩ,R3= 15 kΩand C= 1 nF. The gain and cutoff frequency are then given by: GALPF =1+R2 R1(15) and fH=1 2πR3C3(16) This means that the cutoff frequency was settled to 10.6 kHz and the gain settled to 11×or ≈20.8 dB. Electronics 2023,12, 3082 9 of 21 Figure 8. Active low-pass filter circuit diagram. 2.4. Control and Communication Unit The resolution of the ADC of the ESP32 microcontroller is defined as the minimum voltage that the ESP32 could read. The signal that the ADC receives is amplified by the LNA and the ANF, which improves the resolution of the ADC, as the real value of the biological signal is lower than the read value. The resolution and the total gain of the circuit Gtotal are given by: ∆Vstep =VADC (2Nbits −1)×Gtotal (17) and Gtotal =GLNA ×GALPF (18) The LNA gain G LNA is 45 v/v, the G ALPF is 11 v/vand N bits is 12 for the ESP32. The maximum voltage input V ADC that the ESP32 can read is 3.3 V. Substituting these values in Equations (15) and (16) the resolution of the DBS acquisition is 1.628 µ V/bit. The resolution is in the same order as the amplitude of biopotentials. 2.5. Energy Control The power supply of the system consists of a 3.7 V lithium polymer battery that supplies +5 V step-up from an MT3608 integrated circuit that generates 5 V. A battery with a capacity of 1800 mAh was used; however, other batteries with smaller capacity could be used at the cost of a small useful life. The resulting voltage supplies both a +3.3 V step-down and a +12 V step-up. The +12 V output supplies the circuit and an inverter from an ICL7660A integrated circuit, which results in a symmetrical supply of +12 V and − 12 V. The power conversion efficiency for a maximum supply current of 20 mA provided by the ICL7660A is higher than 95%. It is recommended to not exceed this value with the penalty to decrease the efficiency. The block diagram in Figure 9a illustrates the battery and circuits used to supply the DBS system. Figure 9b illustrates the individual voltages used to supply the neurostimulator, the control and communication unit and the acquisition circuit. Figure 9shows the configuration used to supply the system. The 3.3 V output is used in the ESP32 microcontroller and the symmetrical output is used to polarize the amplifiers. To optimize the use of the battery, and, therefore, to increase the lifetime of the system, the ESP32 can activate the stimulation and the acquisition portion separately. 2.6. Prototype For the design and building of the printed circuit board, Altium software (Student License) was used, due to its versatility in designing flexible printed circuit boards (PCBs). The PCB has the acquisition and stimulation, power management and wireless communication circuits. To miniaturize the system, the design configuration was flex–rigid, which has flexible parts and reduces the area of the PCB. Electronics 2023,12, 3082 16 of 21 Figure 18. Pulses generated with several shapes by the program and collected on the oscilloscope to illustrate the ability of the proposed stimulator. These signals are currents with ( a ) sinusoidal shape with amplitudes of ± 322 µ A with interpulse interval of 0 s, positive and negative pulse widths of 10 ms (50 Hz), and frequency of 10 Hz; ( b ) sinusoidal shape with amplitudes of ± 322 µ A with interpulse interval of 10 ms, positive and negative pulse widths of 10 ms (50 Hz), and frequency of 10 Hz, ( c ) sinusoidal shape with amplitudes of ± 322 µ A with interpulse interval of 0 s, positive and negative pulse widths of 10 ms (50 Hz), and frequency of 50 Hz; and ( d ) pulsed shape with positive amplitude of − 322 µ A, with negative amplitude of − 32 µ A, positive pulse width of 10 ms, negative pulse width of 20 ms, and frequency of 10 Hz. Figure 19. Cont. Electronics 2023,12, 3082 17 of 21 Figure 19. Shapes of the current signals that were injected in the saline solution during the loop-back tests. These current signals present positive and negative amplitudes of 8 µ A and frequencies of (a) 1 Hz, (b) 0.625 Hz, (c) 0.375 Hz and (d) 27 Hz. Figure 20. Signals acquired during the loop-back tests in response to the current signals that were injected into the saline solution. The acquired signals also present frequencies of ( a ) 1 Hz, ( b ) 0.625 Hz, ( c ) 0.375 Hz and ( d ) 27 Hz. These acquired signals correspond to the injected signals illustrated in Figure 19a–d, respectively. The current signals that were injected in the saline solution present positive and negative amplitudes of 8 µ A and frequencies of (a) 1 Hz, (b) 0.625 Hz, (c) 0.375 Hz and (d) 27 Hz, respectively. The first three signals have rectangular shapes, while the fourth is a sinusoid. It is important to not forget that the acquisition channels limit the bandwidth of the acquired and amplified signal to 7.5 kHz. This effect is observable when the signals with square shapes are injected into the saline solution rather than the sinusoidal shape. 3.5. Power Consumption Tests The battery was tested by allowing the CLDBS to work until the battery output value reached below 2.3 V and, therefore, the control unit stopped working. Three scenarios Electronics 2023,12, 3082 18 of 21 were experimented with: only the stimulation channel on and acquisition off; both stimulation and acquisition on; and only the acquisition on and stimulation off. The results are presented in Figure 21, showing the behavior of the voltage at the terminals of the lithium polymer battery that supplied the CLDBS for the three power consumption profiles. The nominal voltage and capacity of the battery used in these tests are, respectively, 3.7 V and 1800 mAh. When both circuits were on, the battery could hold up to 10 h with the ESP32 working well. For more than 8 h, the voltage remained above 3.5 V. When only the stimulation was turned on, the battery could hold the voltage over 3.5 V for more than 20 h. Finally, when only the acquisition was on, the battery voltage also decreased, but at a later time, after around 13 h. It must be noted that a second experimental setup based on an Arduino board was used to read the voltage that supplied the ESP32, and, hence, the supply voltage of the proposed CLDBS system. Figure 21. Voltage at the terminals of the lithium polymer battery that supplied the CLDBS during the consumption tests. 3.6. Comparison with the State-of-the-Art Table 1shows the properties of the present work compared with other DBS systems. The main difference between this work and most of the other works is the integration of two acquisition channels that are used to adjust the parameters of the generated wave for stimulation. Another innovation is that the frequency of operation for this work is significantly higher than that of most other applications. The proposed CLDBS system is the one that presents the best current range, i.e., a quasi-symmetrical current range in both directions to maintain electrical safety. The neurostimulator circuit was designed to offer the capability of generating current with a biphasic waveform, which can invert the direction of charge injection in the neuronal tissue. The phenomenon to nullify charge accumulation is called charge balance [ 21 ]. Table 1also contains two DBS systems with active and continuous neurostimulation circuits based on Howland current pumps. However, only the proposed CLDBS system contains two stimulation channels, acquisition and power control. Moreover, the majority of the listed works ensure the inversion of the current direction with H-topology bridges, with the disadvantage of requiring transistors for current inversion, increasing the number of necessary components and the programming complexity. However, the biggest disadvantage of the H-topology is that it requires access to two different contact points on the electrodes, which are normally unipolar. For these reasons, the option of the circuit responsible for injecting the current into the electrodes fell on the Howland current pump (HCP). Electronics 2023,12, 3082 19 of 21 Table 1. Comparison of the present work with other DBS systems. Ref. Number of Channels Acquisition/Stimulation Stimulation Current [µA] Maximum Pulse Frequency [Hz]/ Minimum Pulse Duration [ms] Stimulation Form System Size This work 2/2 −325 to +318 1.5 ×106/25 Active/Continuous (Howland Current-Pump) 58 mm ×37 mm ×29 mm [22] 2 * 20 to 2000 500/10 Active/Switched (H-bridge) 12.5 diameter ×5 mm [23] 2 * −200 to +200 185/90 Active/Switched 24 mm ×16.8 mm [24] 2 * 0 to 200 130/90 Passive/Switched 12.5 diameter [25] 2 * 13 to 1000 500/24 Active/Switched 3 mm ×20 mm ×8 mm [26] 1 * −375 to +250 5000/20 Active/Continuous (Howland Current-Pump) 32.5 mm ×28 mm ×8 mm [27] 1 * 20 to 2000 300/40 Active/Switched (H-bridge) 22.2 mm ×32.8 mm ×23 mm [28] 1 * 10 to 500 200/60 Passive/Switched 21 mm ×11 mm ×7 mm [29] 4/2 30-1500 5×103/0.01 Active/Constant Current-Generator 28 ×17 ×7 mm * The referred article only mentioned stimulation channels. 4. Conclusions This paper presented a portable and low-cost CLDBS system with two channels for stimulation and two channels for acquisition. The control and communication unit can improve the durability of the battery by activating individually the acquisition and stimulation circuit. As seen in Figure 20, the battery voltage only decreases significantly after 8 h when both the acquisition and circuit are active. If only the acquisition circuit is active the system works for around 13 h. This result may lead to the conclusion that the communication of the ESP32 with the computer via Bluetooth is the most energy-consuming attribute of the circuit, even though, in a real application, it would require very few battery exchanges or recharges during the day. Additionally, if only the stimulation channels are used, the lifespan of the battery would be longer than a day. Therefore, the system could operate as a CLDBS for 8 h straight or as an OLDBS for more than a day. In the future, the proposed CLDBS system can be further improved by including safety mechanisms. This goal can be achieved by adding new software features in the data link layer protocol to enhance the confidentiality, integrity, availability and authenticity of the wireless link. Author Contributions: Conceptualization, T.M.N., G.A.G., R.G. and J.P.C.; methodology, T.M.N., G.A.G., R.G., M.M.A.M. and J.P.C.; validation, T.M.N., M.M.A.M. and R.G.; writing—original draft preparation, T.M.N., G.A.G., M.M.A.M., R.G. and J.P.C.; writing—review and editing, J.A.A., V.M. and J.L.A.; supervision, E.T.F. and E.C.; project administration, J.P.C., J.A.A., V.M. and J.L.A.; funding acquisition, J.P.C., J.A.A., V.M. and J.L.A. All authors have read and agreed to the published version of the manuscript. Funding: This work was partially supported by the FAPESP agency (Fundação de Amparo àPesquisa do Estado de São Paulo) through the project with the reference 2019/05248-7. Professor João Paulo Carmo was supported by a PQ scholarship with the reference CNPq 304312/2020-7. Conflicts of Interest: The authors declare no conflict of interest. Electronics 2023,12, 3082 20 of 21 References 1. Mayfield Clinic. Deep Brain Stimulation for Movement Disorders. Mayfield Clinic. 2018. Available online: https://mayfieldclinic. com/pe-dbs.htm (accessed on 10 July 2023). 2. Hickey, P.; Stacy, M. Deep brain stimulation: A paradigm shifting approach to treat Parkinson’s disease. Front. Neurosci. 2016 , 10, 173. [CrossRef] [PubMed] 3. Bittar, R.G.; Burn, S.C.; Bain, P.G.; Owen, S.L.; Joint, C.; Shlugman, D.; Aziz, T.Z. Deep brain stimulation for movement disorders and pain. J. Clin. Neurosci. 2005,12, 457–463. [CrossRef] 4. Cury, R.G.; Galhardoni, R.; Fonoff, E.T.; Lloret, S.P.; Ghilardi, M.G.S.; Barbosa, E.R.; Teixeira, M.J.; de Andrade, D.C. Sensory abnormalities and pain in Parkinson disease and its modulation by treatment of motor symptoms. Eur. J. Pain 2016 ,20, 151–165. [CrossRef] [PubMed] 5. Rehncrona, S.; Johnels, B.; Widner, H.; Törnqvist, A.L.; Hariz, M.; Sydow, O. Long-term efficacy of thalamic deep brain stimulation for tremor: Double blind assessments. Mov. Disord. 2003,18, 163–170. [CrossRef] 6. Ghilardi, M.G.S.; Ibarra, M.; Alho, E.J.L.; Reis, P.R.; Contreras, W.O.L.; Hamani, C.; Fonoff, E.T. Double-target DBS for essential tremor: 8-contact lead for cZI and Vim aligned in the same trajectory. Neurology 2018,90, 476–478. [CrossRef] [PubMed] 7. Fonoff, E.T.; Ghilardi, M.G.S.; Cury, R.G. Neurocirurgia funcional para o Clínico: Estimulação Cerebral Profunda em Doença de Parkinson, Distonia e Outros Distúrbios do movimento. In Capítulo de Livro, Condutas em Neurologia, 11th ed.; Nitrini, R., Ed.; Manole Editora: Sao Paulo, Brazil, 2016; pp. 53–67. (In Portuguese) 8. Vidailhet, M.; Vercueil, L.; Houeto, J.L.; Krystkowiak, P.; Benabid, A.L.; Cornu, P.; Lagrange, C.; Tézenas du Montcel, S.; Dormont, D.; Grand, S.; et al. Bilateral deep-brain stimulation of the globus pallidus in primary generalized dystonia. N. Engl. J. Med. 2005 , 352, 459–467. [CrossRef] [PubMed] 9. Owen, S.L.; Green, A.L.; Stein, J.F.; Aziz, T.Z. Deep brain stimulation for the alleviation of poststroke neuropathic pain. Pain 2006 , 120, 202–206. [CrossRef] [PubMed] 10. Marchand, S.; Kupers, R.C.; Bushnell, M.C.; Duncan, G.H. Analgesic and placebo effects of thalamic stimulation. Pain 2003 ,105, 481–488. [CrossRef] 11. Gadot, R.; Najera, R.; Hiran, S.; Anand, A.; Storch, E.; Goodman, W.K.; Shofty, B.; Sheth, S.A. Efficacy of deep brain stimulation for treatment-resistant obsessive-compulsive disorder: Systematic review and meta-analysis. J. Neurol. Neurosurg. Psychiatry 2022 , 93, 1166–1173. [CrossRef] 12. Franco, R.; Fonoff, E.T.; Alvarenga, P.; Lopes, A.C.; Miguel, E.C.; Teixeira, M.J.; Damiani, D.; Hamani, C. DBS for Obesity. Brain Sci. 2016,6, 21. [CrossRef] 13. Drobisz, D.; Damborská, A. Deep brain stimulation targets for treating depression. Behav. Brain Res. 2019 ,359, 266–273. [CrossRef] [PubMed] 14. Figee, M.; Riva-Posse, P.; Choi, K.S.; Bederson, L.; Mayberg, H.S.; Kopell, B.H. Deep Brain Stimulation for Depression. Neurotherapeutics 2022,19, 1229–1245. [CrossRef] [PubMed] 15. Sheth, S.A.; Bijanki, K.R.; Metzger, B.; Allawala, A.; Pirtle, V.; Adkinson, J.A.; Myers, J.; Mathura, R.K.; Oswalt, D.; Tsolaki, E.; et al. Deep Brain Stimulation for Depression Informed by Intracranial Recordings. Biol. Psychiatry 2022,92, 246–251. [CrossRef] [PubMed] 16. Hoang, K.B.; Cassar, I.R.; Grill, W.M.; Turner, D.A. Biomarkers and stimulation algorithms for adaptive brain stimulation. Front. Neurosci. 2017,11, 564. [CrossRef] [PubMed] 17. Bouthour, W.; Mégevand, P.; Donoghue, J.; Lüscher, C.; Birbaumer, N.; Krack, P. Biomarkers for closed-loop deep brain stimulation in Parkinson disease and beyond. Nat. Rev. Neurol. 2019,15, 343–352. [CrossRef] [PubMed] 18. AN-1515, A Comprehensive Study of the Howland Current Pump, 26 April 2013, Texas Instruments. Available online: https://www.ti.com/lit/an/snoa474a/snoa474a.pdf (accessed on 10 July 2023). 19. Nordi, T.M.; Barbosa, V.M.; Gounella, R.H.; Assan, G.; Luppe, M.; Junior, J.N.S.; Carmo, J.P.; Fonoff, E.T.; Colombari, E. Charge Pump Circuit in 65nm CMOS for Neural Stimulation on Deep Brain Stimulation. In Proceedings of the XXXVI Conference on Design of Circuits and Integrated Circuits (DCIS 2021), Vila do Conde, Portugal, 24–26 November 2021. 20. Štanfel, D.; Kalogjera, L.; Ryazantsev, S.V.; Hlaˇca, K.; Radtsig, E.Y.; Teimuraz, R.; Hrabaˇc, P. The Role of Seawater and Saline Solutions in Treatment of Upper Respiratory Conditions. Mar. Drugs 2022,20, 330. [CrossRef] 21. Kölbl, F.; N’Kaoua, G.; Naudet, F.; Berthier, F.; Faggiani, E.; Renaud, S.; Benazzouz, A.; Lewis, N. An Embedded Deep Brain Stimulator for Biphasic Chronic Experiments in Freely Moving Rodents. IEEE Trans. Biomed. Circuits Syst. 2016 ,10, 72–84. [CrossRef] 22. Pinnell, R.C.; de Vasconcelos, A.P.; Cassel, J.C.; Hofmann, U.G. A miniaturized programmable deep-brain stimulator for group-housing and water maze use. Front. Neurosci. 2018,12, 231. [CrossRef] [PubMed] 23. Ewing, S.G.; Lipski, W.J.; Grace, A.A.; Winter, C. An inexpensive charge-balanced rodent deep brain stimulation device a step-by-step guide to its procurement and construction. J. Neurosci. Methods 2013,219, 324–330. [CrossRef] [PubMed] 24. Kouzani, A.Z.; Abulseoud, O.A.; Tye, S.J.; Hosain, M.D.K.; Berk, M. A low power micro deep brain stimulation device for murine preclinical research. IEEE J. Translacional Eng. Health Med. 2013,1, 1500109. [CrossRef] [PubMed] 25. Ewing, S.G.; Porr, B.; Riddell, J.; Winter, C.; Grace, A.A. SaBer DBS: A fully programmable, rechargeable, bilateral, charge;balanced preclinical microstimulator for long-term neural stimulation. J. Neurosci. Methods 2013,213, 228–235. [CrossRef] [PubMed] 26. Adams, S.D.; Bennet, K.E.; Tye, S.J.; Berk, M.; Kouzani, A.Z. Development of a miniature device for emerging deep brain stimulation paradigms. PLoS ONE 2019,14, e0212554. [CrossRef] [PubMed] Electronics 2023,12, 3082 21 of 21 27. Tibara, H.; Naudeta, F.; Kölblc, F.; Ribota, B.; Faggiania, E.; Kaouac, G.N.; Renaudc, S.; Lewisc, N.; Benazzouza, A. In vivo validation of a new portable stimulator for chronic deep brain stimulation in freely moving rats. J. Neurosci. Methods 2020 , 333, 108577. [CrossRef] [PubMed] 28. Fluri, F.; Mützel, T.; Schuhmann, M.K.; Krsti´c, M.; Endres, H.; Volkmann, J. Development of a head-mounted wireless microstimulator for deep brain stimulation in rats. J. Neurosci. Methods 2017,291, 249–256. [CrossRef] [PubMed] 29. Pinnell, R.C.; Dempster, J.; Pratt, J. Miniature wireless recording and stimulation system for rodent behavioural testing. J. Neural Eng. 2015,12, 066015. [CrossRef] [PubMed] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.