IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS 1

Transcription

1 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS 1 A Sub-10 na DC-Balanced Adaptive Stimulator IC With Multi-Modal Sensor for Compact Electro-Acupuncture Stimulation Kiseok Song, Student Member, IEEE, Hyungwoo Lee, Student Member, IEEE, Sunjoo Hong, Student Member, IEEE, Hyunwoo Cho, Student Member, IEEE, Unsoo Ha, Student Member, IEEE, and Hoi-Jun Yoo, Fellow, IEEE Abstract A compact electro-acupuncture (EA) system is proposed for a multi-modal feedback EA treatment. It is composed of a needle, a compact EA patch, and an interconnecting conductive thread. The 3 cm diameter compact EA patch is implemented with an adaptive stimulator IC and a small coin battery on the planarfashionable circuit board (P-FCB) technology. The adaptive stimulator IC can form a closed current loop for even a single needle, and measure the electromyography (EMG) and the skin temperature to analyze the stimulation status as well as supply programmable stimulation current (40 A 1 ma) with 5 different modes. The large time constant (LTC) sample and hold (S/H) current matching technique achieves the high-precision charge balancing ( 10 na) for the patient s safety. The measured data can be wirelessly transmitted to the external EA analyzer through the body channel communication (BCC) transceiver for the low power consumption. The external EA analyzer can show the patient s status, such as the muscle fatigue and the change of the skin temperature. Based on these analyses, the practitioner can adaptively change the stimulation parameters for the optimal treatment value. A 12.5 mm m RF CMOS stimulator chip consumes 6.8 mw at 1.2 V supporting 32 different current levels. The proposed compact EA system is fully implemented and tested on the human body. Index Terms Adaptive stimulator, biofeedback, charge balancing, current matching, electro-acupuncture, multi-modal sensor, sample and hold. I. INTRODUCTION ACUPUNCTURE has been practiced in the oriental countries for thousands of years. Recently, its effectiveness has been recognized in the western world and now it is accepted as a beneficial form of the complementary medicine [1], [2]. As a combination of the acupuncture and electrical current stimulation, the electro-acupuncture (EA) has been widely used for its effectiveness in pain relief in 1970s [3] and later for the treatment of various diseases [4] such as depression, addiction, and gastrointestinal disorders, and even non-medical applications including obesity treatment [5]. To realize the EA stimulation, the wired EA system uses a pair of thin needles Manuscript received May 27, 2012; revised August 25, 2012 and November 18, 2012; accepted November 21, This paper was recommended by Associate Editor A. Burdett. The authors are with Korea Advanced Institute of Science and Technology (KAIST), Guseong-dong, Yuseong-gu, Daejeon , Korea ( sks8795@eeinfo.kaist.ac.kr; kiseok.song87@gmail.com). Color versions of one or more of the figures in this paper are available online at Digital Object Identifier /TBCAS which are connected to the external EA controller with thick and heavy wires for stimulating the human body. Unfortunately, the previous wired EA system suffers from the following two fatal problems: 1) Inconvenience: The thin needles may be easily bent or even pulled out by the patient s slight motion during the EA treatment. Consequently, the patient s mobility is severely limited, and it increases the stress or even brings dangerous side effects to the patient. Furthermore, the practitioner should continuously pay careful attention and check the wire connectivity for the stable EA treatment. As a result, if there are many needles, it is difficult to continuously supply power to all needles. Recently, the wirelessly-powered EA system [6] was proposed to remove the cumbersome wire connections for the convenient treatment. However, its wirelessly harvesting power level is only 8, which is far too low to be applied in the various EA applications [2] [5], [23] [25]. 2) Danger: Most of EA systems use the biphasic stimulation to reduce the tissue damage, electrolysis, and electrolytic degradation [7]. However, the high-precision balancing of biphasic current pulse was difficult to achieve because the required offset, 10 na [8], is only order of of the stimulation current level, 1 ma. Moreover, the practitioner can monitor the patient only visually with the unaided eye while the patient is taking the EA treatment so that they do not have any scientific feedback parameters to monitor the patient s status. As a result, they have no choice but to empirically and subjectively determine the stimulation parameters, such as the stimulation time, amplitude, frequency, and pulse width. In this paper, we present an adaptive stimulator IC for a compact EA system to solve the abovementioned problems at once. It can form a closed current loop for even a single needle with surface electrodes under an EA patch in Fig. 1. A thin conductive thread is used for the convenient and safe connection between the needle and the EA patch. The proposed stimulator IC provides 5 different programmable stimulation modes and contains a multi-modal sensor circuit to measure the electromyography (EMG) and the temperature from the skin around the needle. The measured data is wirelessly transmitted to the external EA analyzer realized in the PC through the body channel communication (BCC) [9]. As a result, the practitioner can check the patient s status and can adaptively control the stimulation parameters to reflect the patient s status for the /$ IEEE

2 2 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS Fig. 1. Proposed compact electro-acupuncture (EA) system. Fig. 3. Two stimulation types. Fig. 2. Compact EA patch for multi-modal feedback EA treatment. high remedial effect. To guarantee the high-precision current matching ( 10 na offset current), a large time constant (LTC) sample and hold (S/H) circuit is proposed to keep the mismatch current steady for 10 seconds. The rest of this paper is organized as follows. In Section II, the compact EA patch for either the EA stimulation or multimodal sensing will be presented. The adaptive stimulator IC will be shown in Section III. Section IV presents the multi -modal feedback EA treatment between the EA patch and the external EA analyzer. The implementation and system summary results are given in Section V. Finally, the conclusions will be made in Section VI. II. COMPACT ELECTRO-ACUPUNCTURE PATCH The compact EA patch is implemented on the multi-layered fabric patch using the planar-fashionable circuit board (P-FCB) technology [10], [11] as shown in Fig. 2. The adaptive stimulator IC in [12] is integrated on the compact EA patch so that it can monitor the patient s status and adaptively control the stimulation parameters during the EA treatment. The compact EA patch consists of 3 layers: 1) electrode layer: reconfigurable surface electrodes of which diameter is 3 cm, 2) power layer: power plane to supply reliable power to the patch, and 3) circuit layer: P-FCB on which the adaptive stimulator IC is directly wire-bonded and a coin battery is attached. The small coin battery (, ) provides an independent power (1.2 V, 6.2 mah) to the stimulator IC so that it can remove the external wire connections for high usability. The needle electrode is connected to the stimulator IC through the short and light conductive thread and the conductive velcro on the patch for the user s convenience. Firstly, the practitioner attaches the compact EA patch on the skin. Then he/she pricks the skin with an acupuncture needle near the patch and connects the acupuncture needle tip to the patch with a thin conductive thread. Then the practitioner programs the stimulation types (uni- or bi-acupoint) and parameters (mode, amplitude, frequency, and pulse width), and starts up the EA stimulation through the BCC using the external EA analyzer. The proposed compact EA patch provides 2 different stimulation types to form a closed current loop: 1) uni-acupoint (UA) stimulation and 2) bi-acupoint (BA) stimulation as shown in Fig. 3. The UA stimulation type uses one needle electrode and surface electrodes as an anode and cathode, respectively, in the EA stimulation so that it can reduce the number of the required needles for the stimulation. It stimulates the human tissue around the single acupoint [13]. The BA stimulation type uses two needle electrodes and stimulates the human tissue between two acupoints. The load impedances of UA type and BA type are expressed in (1) and (2), respectively. (1) (2) As shown in Fig. 4(a), in the stimulation frequency band (1 Hz 1 khz), the impedance of the dry electrode with smooth surface is significantly larger than the needle electrode impedance so that the maximum stimulation current level is severely limited under only 25 in the fixed stimulation voltage (3.3 V). Consequently, the surface electrode impedance should be reduced to guarantee the required stimulation current level in the UA stimulation type. Fig. 4(a) shows the impedance reduction of the proposed wet electrode with rough surface. In the stimulation frequency band (1 Hz 1 khz), the surface electrode impedance is reduced by 45%, 70%, and 87% using rough surface, conductive gel, and larger size (3 cm diameter), respectively. As a result, the impedance of the proposed 3 cm diameter wet electrode with rough surface have impedance value of less than 2.7 by more than 94% reduction compared to the 1 cm diameter dry electrode with smooth surface. From the magnified photos of the proposed electrode shown in Fig. 4(b), the proposed surface electrode has a surface roughness.

3 SONG et al.: A SUB-10 na DC-BALANCED ADAPTIVE STIMULATOR IC 3 Fig. 4. (a) Impedance reduction of wet electrode with rough surface. (b) Magnified photo of rough surface. Fig. 6. Overall block diagram of adaptive stimulator IC architecture. Fig. 5. Reconfigurable surface electrodes for multi-modal feedback EA treatment. During the EA stimulation, the compact EA patch measures the EMG and the skin temperature signal from the patient and transmits the measured data to the external EA analyzer with the help of the reconfigurable surface electrodes as shown in Fig. 5. The compact EA patch is operated in 3 phases. In the EA stimulation phase (Phase 1), the surface electrodes are used as the cathode of the closed current loop for the EA stimulation, and also as the communication electrodes to receive the stimulation parameters through the BCC [9]. In every minute, the multi-modal sensor in the adaptive stimulator IC measures the EMG and the skin temperature for 5 seconds in the multi-modal sensing phase (Phase 2). In the Phase 2, Inner circular electrode plays the role of a reference electrode. The upper 2 electrodes and the lower 2 electrodes act as differential sensing electrodes. After that, within 0.01 seconds (Phase 3), the body channel transceiver [9] transmits the measured data to the external EA analyzer with low power consumption (1.6 mw) and high data rate (1.25 Mbps) using all surface electrodes as data communication electrodes. The external EA analyzer extracts the patient s status, such as the muscle fatigue and the change of the skin temperature. Then, the stimulation parameters can be properly updated in the two different methods: 1) in the automatic mode, the stimulation parameters can be properly adjusted by the update policy which was previously programmed by the practitioner. 2) In the manual mode, the stimulation parameters should be manually adjusted after the practitioner check the patient s status. The updated control signals of the stimulation parameters are uploaded to the compact EA patch through the body channel. III. HIGH-PRECISION DC-BALANCED ADAPTIVE STIMULATOR IC Fig. 6 shows the overall block diagram of the proposed highprecision DC-balanced adaptive stimulator IC architecture. It

4 4 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS Fig. 7. (a) Programmable electro-acupuncture stimulator front-end (EASFE). (b) Load impedance modeling results. (c) Stimulation current range. consists of five key building blocks for the multi-modal feedback EA treatment: 1) power management unit (PMU) to generate the high stimulation voltage of 3.3 V from a 1.2 V coin battery, 2) high-precision DC-balanced EA stimulator front-end (EASFE) for the programmable EA stimulation, 3) multi-modal sensor front-end for the EMG and the skin temperature acquisition and digitization, and 4) BCC transceiver for the external communication, and 5) system controller with 10 kb on-chip SRAM and system clock generator. Fig. 7(a) shows the programmable EASFE circuit. The low frequency (16 khz) stimulation clock is generated by an integrator-based oscillator with capacitor bank for the calibration. The frequency and pulse width of the stimulation pulse can be digitally programmed in the range of Hz and s, respectively. The reference stimulation current of 40 can be also calibrated by a R-2R ladder. The amplitude of the stimulation current is controlled in the range of 40 1 ma by a pull-up and pull-down current source with digital controller. From the measurement results, the load impedance of UA type and BA type can be modeled as shown in Fig. 7(b). The proposed EASFE can provide the output current up to 0.32 ma and 1 ma with the UA type load and BA type load, respectively, as shown in Fig. 7(c). As shown in Fig. 8, the proposed EASFE provides 5 different stimulation modes which are used in the commercial EA system; 1) continuous mode with constant amplitude, frequency, and pulse width, 2) burst mode with periodic concentration and rest, 3) fast-slow mode with frequency alternation, 4) surge mode with amplitude sweeping, and 5) sweep mode with frequency sweeping. In each stimulation mode, the stimulation parameters can be programmed in the range as shown in the Table I. Fig Different simulation modes. To avoid the charge accumulation in the human body, the biphasic stimulation is required. In the ideal case, biphasic stimulation pulses do not have net DC charge transfer. However, it suffers from the net DC charge transfer due to the mismatch current between the pull-up current and pull-down current in the practical implementation. Consequently, two charge balancing types were proposed in order to generate matched biphasic stimulation pulses; passive

5 SONG et al.: A SUB-10 na DC-BALANCED ADAPTIVE STIMULATOR IC 5 TABLE I STIMULATION PARAMETER RANGES Fig. 9. (a) Large time constant (LTC) S/H current matching circuit, and (b) its measurement results. charge balancing, and active charge balancing techniques. Passive charge balancing technique inserts DC blocking capacitor in series with the stimulation electrode [14] or uses electrode discharge circuit [15]. However, it requires much large DC blocking capacitor than electrode capacitance or long discharge time ( 3 ms) by the RC time constant of the electrode impedance. To solve these problems, active charge balancing technique is adopted to insert short current pulses [16] or to regulate offset current [17] after each stimulation pulse. Unfortunately, it suffers from the unwanted stimulation effect due to the inserted current spike or long offset nulling time. The LTC S/H current matching technique is proposed in the EASFE to achieve the instantaneous charge balancing without any large passive element as shown in Fig. 9. The leakage compensated LTC S/H circuit is proposed to hold the steady mismatch current without degradation during the negative phase for high-precision balancing. The stimulates the human body during the positive phase. The control code of the pull-up current source is always higher than the control code of the pull-down current source by one as shown in Fig. 7. As a result, the amplitude of is always higher than by. In such circumstance, the mismatch current between the and is sampled by the gate voltage of through the LTC S/H circuit during the interphase period. After the sampling, is held during the negative phase. Consequently, the stimulation current in the negative phase is a sum of and, which is equal to in the positive phase. In such circumstance, the negative phase can be extended up to 0.5 seconds when the stimulation frequency is 1 Hz. Consequently, the S/H circuit should hold the Fig. 10. High-precision charge balancing results using large time constant (LTC) sample and hold (S/H) current matching circuit. during the negative phase without variation for the high-precision current matching. The proposed LTC S/H circuit has a dummy cell,, of 1/5 storage capacitance with the same CMOS switch as the main cell. The minimum CMOS switches (, ) are used to minimize the effect of charge injection. The and are 1 pf and 5 pf, respectively, so that the parasitic capacitance is negligible compared to and. Since the leakage current can be detected by the voltage difference between the main cell and dummy cell, the proposed S/H circuit can compensate for the by the compensation current of the negative feedback loops. The measurement results of the S/H circuit are shown in Fig. 9(b). Moreover, the effects of the charge injection are same in the main cell and dummy cell so that the charge injections are also compensated by the negative feedback loops. It can maintain the sampled voltage during 0.8 seconds. Compared to the S/H circuit without dummy cell, the proposed LTC S/H circuit achieves only 1% of the voltage variation. The offset current regulation [17] is also used for the patient s safety as shown in Fig. 9(a). Following the current matched biphasic stimulation, the residual electrode potential is measured and compared to the reference potential.if the voltage difference exceeds a safe threshold ( 100 mv), the offset current is added for steering the electrode potential to the balanced condition. Moreover, it can detect the failure of the stimulator output stage and prohibit the EA stimulation for the patient s safety. Fig. 10 shows the output current waveforms which are high-precision charge-balanced by the proposed LTC S/H current matching circuit. From the measurement results shown in Fig. 7(b), the BA type load (series RC with and ) is used for the DC offset current measurement. The level of DC flowing through the load was obtained from

6 6 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS Fig. 12. (a) Multi-modal feedback EA signal flow. (b) PC GUI. Fig. 11. DC offset current (I ) measurement results with (a) different stimulation modes and (b) stimulation frequency. Where, level of DC offset current; generated offset voltage due to the DC charge transfer; series capacitance of load impedance ; EA stimulation time. Without the proposed current matching technique, there is an of 480 na which generates of 240 mv during of 5 seconds. However, the proposed LTC S/H current matching technique successfully reduces the to less than 10 na resulting in of only 5 mv during of 5 seconds. Consequently, the proposed LTC S/H current matching circuit achieves 48-times improvement. If the stimulation persists, the increased reduces mismatch current between the pull-up and pull-down current source. As a result, the is saturated after the tens of seconds due to the zero. Fig. 11(a) shows the measured DC offset current,, in 5 different stimulation modes. To measure the maximum case of, the maximum pulse width is used for the maximum amount of stimulation charge in each stimulation mode. As shown in Fig. 11(a), increases with increasing stimulation current and the worst is measured in the continuous mode. However, is never higher than 10 na. The stimulation frequency dependency of is shown in Fig. 11(b). In the (3) higher stimulation frequency, the stimulation voltage swing decreases because the load impedance decreases. As a result, the unwanted effect of the finite output impedance of the pull-up and pull-down current source is reduced. Consequently, decreases with increasing stimulation frequency as shown in Fig. 11(b). IV. MULTI-MODAL FEEDBACK EA TREATMENT During the EA stimulation, the compact EA patch measures the EMG and skin temperature signal from the human body and updates the stimulation parameters to reflect the patient s status with the external EA analyzer. In the proposed compact EA system, the external EA analyzer is implemented on the PC with the 1:N BCC interface [9] for the multi-channel EA treatment. The ground of each stimulation channel is separated by its independent power source so that there is no crosstalk between the channels. In the proposed EA system, the communication interface is implemented using TI MSP-EXP430F5438 (MSP430 F5438 Experimental Board) [18]. Fig. 12(a) shows the signal flow in the proposed compact EA system. The multi-modal signals are recorded during the EA stimulation in the compact EA patch. And then, they are transmitted using a BCC interface to the MSP-EXP430F5438 which converts the data format to the USB data. After the feature extraction processing, the PC GUI shows the patient s status, such as the muscle fatigue and the change of the skin temperature as shown in Fig. 12(b). In the proposed EA system, the median frequency (MF) method is used for the muscle fatigue extraction. The MF is defined as the frequency that divides the power spectrum into two of equal area. The low frequency shift of the MF can be interpreted as a sign of the local muscle fatigue [19]. After the practitioner checks the patient s status, the stimulation parameters are properly updated. And then, the updated stimulation parameters are transmitted to the compact EA patch.

7 SONG et al.: A SUB-10 na DC-BALANCED ADAPTIVE STIMULATOR IC 7 Fig. 14. Chip micrograph and its performance summary. Fig. 13. Example of multi-modal feedback EA treatment. Fig. 13 shows the example of the multi-modal feedback EA treatment. The stimulation parameters are adaptively controlled by the previously programmed policy. The muscle fatigue is accumulated by the repeated EA stimulation. As shown in Fig. 13, the MF of the 3rd EMG is shifted to the low frequency (43 Hz) from the MF (83 Hz) of the 1st EMG due to the repeated EA stimulation. To avoid the excessively accumulated muscular fatigue, the stimulation current amplitude and the pulse width are reduced from 200 to 120 and 9.5 ms to 5.6 ms, respectively. V. IMPLEMENTATION RESULTS AND SYSTEM SUMMARY Fig. 14 shows the adaptive stimulator chip micrograph. The proposed stimulator IC is fabricated in P8M RF CMOS process, and it occupies 2.5 mm 5.0 mm chip area including pads. The performance of the adaptive stimulator chip is summarized in the Table II. It dissipates 6.8 mw peak power when the EASFE operates when the stimulation current is maximized to 1 ma, and the BCC receiver is activated at 1 Mbps with sensitivity of 60 dbm. The battery life of the proposed EA system is longer than 1 hour which guarantees the nominal EA treatment time ( 30 minutes). The coin battery is replaceable and rechargeable so that it can be reused for other EA treatment after recharging. The EASFE provides 5 different stimulation modes which are used in the commercial EA system. For the high-precision charge balancing, the LTC S/H current matching technique achieves less than 10 na DC offset current. The multi-modal sensor front-end measures the EMG and the skin temperature with only 40 power consumption. Fig. 15. Performance comparison results in (a) charge balancing techniques and (b) EA applications. The performance comparison results of the proposed charge balancing techniques are shown in Fig. 15(a). Compared to the previous charge balancing techniques [15], [20] [22], the proposed LTC S/H current matching technique removes 1) large off-chip blocking capacitor, 2) long discharge time, 3) additional current spike, and 4) long offset nulling time. The system summary results compared to various EA applications are shown in Fig. 15(b). The proposed EA system provides wide stimulation parameter ranges to cover the both medical and non-medical applications, such as nerve stimulation [23], pain relief [24], and obesity treatment [25]. With the small coin battery (6.2 mah),

8 8 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS programmable stimulation as well as safe and convenient EA treatment. Fig. 16. Stimulation current verification with ES-160 [26]. it also guarantees the operating time more than 1 hour which is enough EA treatment time for various EA applications. Besides, it can remove the cumbersome wire connections and reduce the number of required needles per stimulation channel for user s convenience. The optimal treatment effect also can be obtained by the multi-modal feedback EA treatment. The electrical characteristics of the proposed EA system are compared with the commercial EA system (ES-160) [26] to verify the clinical effectiveness of the proposed EA system. The clinical effectiveness of the commercial EA system has been successfully verified in the previous researches [3], [4], [23] [25]. As shown in Fig. 16, if the stimulation parameters of the proposed EA system and the ES-160 are equal, the output current waveforms are also identical to each other. As a result, it is observed that the proposed EA system can achieve the same stimulation effect as the commercial EA system. VI. CONCLUSION A very convenient and effective electro-acupuncture (EA) system is developed using the compact EA patch. For the user s convenience, we remove the external wire connections and reduce the number of the required needles for the stimulation with the proposed EA patch. The adaptive stimulator IC is implemented in P8M RF CMOS process, and it is integrated on the compact EA patch with the help of planar-fashionable circuit board (P-FCB) technology. It provides 5 different stimulation current modes, and wide stimulation parameter ranges to cover various EA applications. To guarantee the high-precision charge balancing, the offset current is reduced to 10 na by the large time constant (LTC) ( 10 seconds) sample and hold (S/H) current matching technique. The proposed stimulator IC contains the multi-modal sensor so that it can measure the electromyography (EMG) and the skin temperature during the EA stimulation. The body channel communication (BCC) is also used for the wireless communication between the compact EA patch and the external EA analyzer with low power consumption. The external EA analyzer shows the patient s status, such as the muscle fatigue and the change of the skin temperature, for the multi-modal feedback. Consequently, the practitioner checks the patient s status during the EA stimulation, and aptly updates the stimulation parameters for the safe and efficient EA treatment. Compared to the commercial EA system, the proposed EA system injects the same stimulation current with the same temporal parameters and achieves the same stimulation effects in the various EA applications. As a result, this system is effective for the optimal and interactive acupuncture treatment with the real-time multi-modal feedback signals and REFERENCES [1] G. Stux, B. Berman, and B. Pomeranz, Basics of Acupuncture, 5th ed. Berlin, Germany: Springer, [2] A. Sarkova and M. Sarek, EAV and gemmotherapy Medicine for the next millennium? (Technique as a means to link eastern and western medicine), in Proc. 27th IEEE Engineering in Medicine and Biology Society Conf., Sep. 2005, pp [3] R. S. S. Cheng and B. H. Pomeranz, Electroacupuncture analgesia is mediated by stereo specific opiate receptors and is reversed by antagonists of type I receptors, Life Sci., vol. 26, no. 8, pp , Feb [4] G. A. Ulett, S. Han, and J.-S. Han, Electroacupuncture: Mechanisms and clinical application, Biol. Psych., vol. 44, no. 2, pp , Jul [5] F. Wang, D.-R. Tian, and J.-S. Han, Electroacupuncture in the treatment of obesity, Neurochem. Res., vol. 33, no. 10, pp , Aug [6] K. Song, L. Yan, S. Lee, J. Yoo, and H.-J. Yoo, A wirelessly powered electro-acupuncture based on adaptive pulsewidth monophase stimulation, IEEE Trans. Biomed. Circuits Syst., vol. 5, no. 2, pp , Apr [7] A. Scheiner, J. T. Mortimer, and U. Roessmann, Imbalance biphasic electrical stimulation: Muscle tissue damage, Ann. Biomed. Eng., vol. 18, no. 4, pp , [8] D. F. Mayer, Electroacupuncture: A Practical Manual and Resource, 1st ed. New York: Elsevier Health Science, 2007, vol. 1. [9] S.-J. Song et al., A 0.9 V 2.6 mw Body-coupled scalable PHY transceiver for body sensor applications, in Proc. IEEE ISSCC Dig. Tech. Papers, Feb. 2007, pp [10] H. Kim, Y. Kim, Y.-S. Kwon, and H.-J. Yoo, A 1.12 mw continuous healthcare monitoring chip integrated on a planar fashionable circuit board, in Proc. IEEE ISSCC Dig. Tech. Papers, Feb. 2008, pp [11] S. Lee, B. Kim, and H.-J. Yoo, Planar fashionable circuit board and its applications, J. Semicond. Technol. Sci., vol. 9, no. 3, pp , Sep [12] K. Song, H. Lee, S. Hong, H. Cho, and H.-J. Yoo, A Sub-10 na DC-balanced adaptive stimulator IC with multimodal sensor for compact electro-acupuncture system, in Proc. IEEE ISSCC Dig. Tech. Papers, Feb. 2012, pp [13] C. Niu et al., A novel uni-acupoint electroacupuncture stimulation method for pain relief, Evidence-Based Complementary and Alternative Medicine (ecam), vol. 2011, 2011, /ecam/nep104, Article ID , 6 pages. [14] G. J. Suaning and N. H. Lovell, CMOS neurostimulation ASIC with 100 channels, scaleable output, and bidirectional radio-frequency telemetry, IEEE Trans. Biomed. Eng., vol. 48, no. 2, pp , Feb [15] M. Sivaprakasan, W. Liu, M. S. Humayun, and J. D. Wieland, A variable range bi-phasic current stimulus driver circuitry for an implantable retinal prosthetic device, IEEE J. Solid-State Circuits, vol. 40, no. 3, pp , Mar [16] M. Ortmanns, A. Rocke, M. Gehrke, and H.-J. Tiedtke, A 232-channel epiretinal stimulator ASIC, IEEE J. Solid-State Circuits, vol. 42, no. 12, pp , Dec [17] K. Sooksood, T. Stieglitz, and M. Ortmanns, An active approach for charge balancing in functional electrical stimulation, IEEE Trans. Biomed. Circuits Syst., vol. 4, no. 3, pp , Jun [18] MSP-EXP430F5438 Experimental Board-User Guide, Texas Instruments, Jan revised Apr [Online]. Available: [19] G. P. Y. Szeto, L. M. Straker, and P. B. O Sullivan, EMG median frequency changes in the neck-shoulder stabilizers of symptomatic office workers when challenged by different physical stressors, J. Electromyogr. Kinesiol., vol. 15, no. 6, pp , Dec [20] X. Liu, A. Demosthenous, and N. Donaldson, An integrated implantable stimulator that is fail-safe without off-chip blocking-capacitor, IEEE Trans. Biomed. Circuits Syst., vol. 2, no. 3, pp , Sep [21] J.-J. Sit and R. Sarpeshkar, A low-power blocking-capacitor-free charge-balanced electrode-stimulator chip with less than 6 na DC error for 1-mA full-scale stimulation, IEEE Trans. Biomed. Circuits Syst., vol. 1, no. 3, pp , Sep

9 SONG et al.: A SUB-10 na DC-BALANCED ADAPTIVE STIMULATOR IC 9 [22] E. Noorsal et al., A neural stimulator frontend with high-voltage compliance and programmable pulse shape for epiretinal implants, IEEE J. Solid-State Circuits, vol. 47, no. 1, pp , Jan [23] P. Li, K. F. Pitsillides, S. V. Rendig, H.-L. Pan, and J. C. Longhurst, Reversal of relax-induced myocardial ischemia by median nerve stimulation: A feline model of electroacupuncture, Circ.-J. Amer. Heart Assoc., pp , [24] R. S. S. Cheng and B. Pomeranz, Monoaminergic mechanism of electroacupuncture analgesia, Brain Res., vol. 215, no. 1 2, pp , Jun [25] M. T. Cabioglu, N. Ergene, and U. Tan, Electroacupuncture treatment of obesity with psychological symptoms, Int. J. Neurosci., vol. 117, no. 5, pp , May [26] General Catalog for Electrotherapy, Japan Ito Company, Ltd., 2007 [Online]. Available: Kiseok Song (S 09) received the B.S. and M.S. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2009 and 2011, respectively. He is currently working toward the Ph.D. degree in electrical engineering at KAIST. His current research interests include biomedical SoC design focusing on the electrical stimulator. He is also interested in body channel analysis for electrical field-coupled communication. As a Chief Researcher at the Semiconductor System Laboratory at KAIST, he developed a wirelessly-powered electro-acupuncture stimulator and multi-modal feedback electro-acupuncture stimulator for convenient and effective electro-acupuncture treatment. Hyungwoo Lee (S 10) received the B.S. and M.S. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2010 and 2012, respectively. He is currently working toward the Ph.D. degree in electrical engineering at KAIST. His current research interests include developing an electro-acupuncture stimulator system and a body coupled electrical field communication. Sunjoo Hong (S 10) received the B.S. and M.S. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2010 and 2012, respectively. She is currently working toward the Ph.D. degree at KAIST. Her current research interests include lowpower biomedical sensor design for wearable healthcare system and printed circuit board techniques on the fabric. Hyunwoo Cho (S 10) received the B.S. degree in physics and the M.S. degree in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2010, and 2012, respectively. He is currently working toward the Ph.D. degree at KAIST. He has worked on developing a high speed and low power transceiver for body channel communication. His current research interests include lowenergy transceiver design for body area networks and analysis of human body channel characteristics. Unsoo Ha (S 12) received the B.S. degree in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in He is currently working toward the M.S. degree at KAIST. His current research interests include lowenergy transceiver design for wireless body area networks. Hoi-Jun Yoo (M 95 SM 04 F 08) received the M.S. and Ph.D. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea He was the VCSEL Pioneer with Bell Communications Research, Red Bank, NJ, and Manager of the DRAM Design Group, Hyundai Electronics, designing from 1M dynamic random-access memory (DRAM) to 256M SDRAM. Currently, he is a Full Professor in the Department of Electrical Engineering at KAIST and the Director of the System Design Innovation and Application Research Center (SDIA). From 2003 to 2005, he served as the Full Time Advisor to the Minister of Korean Ministry of Information and Communication for SoC and Next Generation Computing. He has authored more than 250 papers, and wrote or edited five books: DRAM Design (Hongneung, 1997), High Performance DRAM (Hongneung, 1999), Low Power NoC for High Performance SoC Design (CRC, 2008), Mobile 3D Graphics SoC (Wiley, 2010), and BioMedical CMOS ICs (co-edited with Chris Van Hoof, 2010, Springer), and many chapters of books. His current research interests are bioinspired IC Design, network on a chip, multimedia SoC design, wearable healthcare systems, and high speed and low power memory. Dr. Yoo received the Korean National Medal for his contribution to Korean DRAM Industry in 2011, the Electronic Industrial Association of Korea Award for his contribution to DRAM technology in 1994, the Hynix Development Award in 1995, the Korea Semiconductor Industry Association Award in 2002, the Best Research of KAIST Award in 2007, the Design Award of 2001 ASP- DAC, Outstanding Design Awards of 2005, 2006, 2007, 2010, 2011 A-SSCC, and the Korean Scientist of the Month Award (December 2010). He is a member of the executive committee of the Symposium on VLSI and A-SSCC. He was the TPC Chair of the A-SSCC 2008, a guest editor of IEEE JSSC and IEEE TBioCAS. He was the TPC Chair of the International Symposium on Wearable Computer (ISWC) 2010, IEEE Distinguished Lecturer ( ), Far East Chair of ISSCC ( ), and currently ISSCC Technology Direction Subcommittee Chair and an Associate Editor of IEEE TCAS-II.