comfort to patients and care providers. Among the various ..... [9] A. Farajidavar, J. L. Seifert, J. E. S. Bell, Y. S. Seo, M. R.. Delgado, S. Sparagana, M. I. Romero ...
International Microwave Symposium, Baltimore MD, June 5-10 2011.
A Miniature Power-Efficient Bidirectional Telemetric Platform for in-vivo Acquisition of Electrophysiological Signals Aydin Farajidavar1, Philip McCorkle2, Timothy Wiggins2, Smitha Rao2, Christopher Hagains3, Yuan Peng3, Jennifer Seifert1, Mario Romero1, Greg O'Grady4, Leo Cheng4, Steven Sparagana5,6, Mauricio Delgado5,6, Shou-Jiang Tang7, Tom Abell7, and J.-C. Chiao1,2 1
Bioengineering, 2Electrical Engineering, 3Psychology Departments, University of Texas at Arlington, Arlington, TX, 76019, USA 4Depts. of Surgery & Auckland Bioengineering Institute, The University of Auckland, Auckland, 1010, New Zealand, 5Texas Scottish Rite Hospital for Children in Dallas, TX, 75219, USA, 6Neurology Dept., University of Texas Southwestern Medical Center, Dallas, TX, 75390, USA, 7Dept. of Digestive Diseases, University of Mississippi Medical Center, Jackson, MS, 39216, USA.
Wireless technology is transforming the modern medicine. Telemetric systems have been established for electroencephalography (EEG), neural activity recordings [1], electromyography (EMG), electrocardiography (ECG), and electro-oculography (EOG) [2]. This trend is predicted to continue rising as wireless systems afford convenience and comfort to patients and care providers. Among the various applications of medical wireless systems, recording of myoelectric signals or in vivo gastric electrical activity (GEA), electrocorticography (ECoG) and transcranical motor evoked potentials (TcMEP) have not been extensively demonstrated. The major challenges in designing a multi-channel wireless system for acquiring such signals are the device size and signal absorption by tissues [3]. The transponder should be small enough to be implanted comfortably and data communication should be reliable so that the signals can be received without distortion or interruption. The power consumption should be low for chronic uses without frequent recharging. The power consumption and bandwidth requirement however depend on the specific intended application, such as diagnosis period and sampling rates.
10 mV 1 mV
100μV 10μV
GEA
I. INTRODUCTION
Myoelectric activity of the stomach consists primarily of very low amplitude and frequency signals termed slow waves (Fig. 1) [4]. Recording slow waves provides important information for evaluating gastric motility disorders including gastroparesis. For optimal data, electrodes are placed invasively on the gastric wall with wires connecting through an abdominal incision or via the mouth of the patient to an electronic recorder. ECoG has conventionally provided valuable information on the mechanisms of brain activities [5]. ECoG acquires signals from the surface of the dura in the brain, therefore providing a better spatial resolution, broader bandwidth and higher characteristic amplitude (Fig. 1) than EEG. ECoG is more robust to motional artifacts which are the main source of noises in freely behaving/moving subjects [6]. Several systems have previously been proposed for wireless acquisition of either EEG or ECoG. However, due to bulkiness, they could not practically be used to study small animal models [7]. Recently, telemetric systems with appropriate size and weight have been developed for acquiring EEG signals in rodents [8], however they are capable of recording from only one active electrode. Therefore, a multichannel system that can acquire ECoG from multiple electrodes in small animals with high fidelity and low power consumption is needed. TcMEP consists of relatively high magnitude and frequency components evoked in the distal limb muscles as a result of electrical stimulation of the cerebral motor cortex (Fig. 1). It
Signal amplitude
Abstract — The need for in vivo wireless acquisition of biological signals is emerging in various medical fields. Electrophysiological applications including recording myoelectric signals in-vivo gastric electrical activity (GEA) to study gastric dysmotility, electrocorticography (ECoG) to study pain, and transcranical motor evoked potentials (TcMEP) for intraoperative neurophysiological monitoring of spinal cord integrity require physically miniaturized devices with low power consumption and capability of implantation. These systems should provide reliable communication in real time with sufficient data rates. We have developed three telemetric systems for GEA, ECoG and TcMEP applications based on a common transceiver platform but with different design considerations. Each has been successfully validated in appropriate animal models, to demonstrate the feasibility of wireless acquisition of key electrophysiological signals. Index Terms — Gastric electrical activity, electrocorticography, transcranical motor evoked potentials, wireless signal acquisition.
TcMEP
ECoG/EEG
1μV
1 Hz
10 Hz 100 Hz Signal f requency
1 kHz
Fig. 1. The amplitude and frequency ranges of GEA, ECoG, and TcMEP signals.
Page 1 of 4
Back-end
Front-end 2
1
(a) AIN0 Radio
…..
Labview
…..
ADC
Radio
AIN1 Amplifier & Filter
AIN6
μC nRF24Le1
(b)
μC nRF24Le1
(c)
Fig. 2. (a) Block diagram of the wireless system, including the front- and back-ends. Dashed line #1 shows the transmitted packets from the front-end to back-end, and #2 shows acknowledgement in the opposite direction. (b) The graphical user interface (GUI). (c) Top-view of the assembled front-end device.
has been used in intraoperative monitoring (IONM) during spine surgery, and often requires up to 40 lead wires for stimulating and recording [9]. The wire connections limit the maneuverability of the surgeon and are susceptible to noise and interference from power lines and electronic equipment in the operating room. We have designed, fabricated and examined three different systems for acquiring in vivo GEA, ECoG and TcMEP signals using a commercially-available wireless transceiver nRF24Le1 that combines radio, microcontroller (μC) and analog to digital (ADC) converter all in a single chip. The design considerations for each transponder are addressed.
Gaussian frequency-shift keying (GFSK) modulation. Packets were received in the back-end and sent to a computer via serial communication where a graphical user interface (GUI) designed in Labview (National Instrument) displayed the data in parallel panels (Fig. 2). Figure 3 shows the flow chart in both μCs of front- and back-ends. In the front-end, after the μC acquired data from ADC (step 1) and loaded it into the radio (step 2), the radio turned into the transmitting mode, which took 130 μs, and transmitted the data packets (step 3). Then the radio took another 130 μs returning back to the receiving mode (step 4). At the same time, the radio in the back-end which was in the receiving mode looked for the data (step 5). If it received the packet, it turned into the transmitting mode and sent back an acknowledgement to the front-end (step 6) while loading the data on the UART (universal asynchronous receiver/transmitter, at 500 kBaud) to be sent to the computer (step 7) as the μC on the front-end loaded the next set of data into the radio. If the data packet was not received by the back-end, without the acknowledgement packet the radio on the front-end retransmitted the data packet to the back-end (back to step 3) until attainment was verified. The re-transmitting procedure mainly depended on the sampling rate of the μC and the time for each packet to travel in air, which varies between the three systems. Each packet is composed of 1 preamble byte, 3 to 5 address bytes, up to 32 bytes of payload and 1 to 2 bytes for cyclic redundancy check (CRC). The time on air TOA can be calculated from TOA = (Packet length)/(Air data rate). The air data rate was chosen to be 1 Mbps and the packet length differed in each application. The following explains the detailed specifications on each system. GEA: The analog board for the GEA system was designed to amplify signals at 65 dB and filter the undesired signals out Step 1
μC acquire data (n) from ADC
II. SYSTEM CONFIGURATION
Step 3
Radio turns into Tx mode (130 μs) and transmits packet (n)
Step 4
Radio turns into Rx mode (130 μs)
No
Packet received?
Yes Step 6
Step 7
Radio turns into Tx mode (130 μs) and sends an acknowledgement to Front-end
Back-end
Step 5
Front-end
μC loads the data (n) into the radio
Step 2 n= n+1
The three different systems share a common platform in digitization and transmission of signals. However, there are different design considerations for each system. The common platform is discussed prior to discussion of the specific considerations for each application. It should be noted that all three systems are designed to acquire signals from four independent analog channels. Platform: Each system constituted a front-end and a backend. The front-end consisted of an analog board, an ADC, a μC and a 2.4-GHz transceiver. The analog board conditioned the signals through a two-stage amplification and band-pass filtration. Signals passed through an instrumentation amplifier INA333 (Texas Instruments) then to a second-order band-pass filter to eliminate noise before the next stage of amplification. The gains and filter passbands depend on respective applications. The amplified signals were sampled, digitized, put into packets and sent by the transceiver, which utilized
μC loads the data (n) into the UART
End
Fig. 3. The flow chart for the μCs in the front and back-ends to ensure reliable communication. Page 2 of 4
(a)
(b)
Fig. 4. Two minutes of slow GEA wave signals recorded by (a) wired and (b) wireless systems.
Fig. 5. ECoG signals recorded by a wired (top) and our wireless (bottom) systems. Only a section of 7.3 seconds was shown.
of the range of 0.05 – 0.3 Hz. Since the slow wave signal occurs at very slow frequencies, it was sampled at a rate of 8 Sps and transmitted via the payload with a size of 4 bytes. Therefore, the TOA was calculated as 0.096 ms at its maximum, which is much shorter than the sampling period. ECoG: The analog board on the ECoG amplified the signals at 74 dB and filtered the undesired signals out of the range of 1 – 150 Hz. The sampling rate for this system was 1 kSps for each channel. The number of payloads was chosen as 4 bytes; hence, the TOA is the same as the one in the EGG system. TcMEP: The analog board for the TcMEP was intended to amplify the signals at 54 dB and filter the undesired signals out of the range of 100 – 1000 Hz. Since this system required a higher sampling rate and there was no need for acknowledgement as the TcMEP signals were continuously sent for signal integration, the re-transmission function was deactivated for this system. The signals were sampled at 6.2 kSps for each analog channel. The number of payloads in this case was chosen as 32 bytes and the TOA was calculated as 0.32 ms at its maximum.
Animal preparation: Three different animal models were utilized for examining each system in vivo. (1) For acquiring GEA signals, the anesthetized canine model was used in which signals were acquired from serosal membrane of the stomach using flexible PCB electrodes [4]. (2) For examining the ECoG signals, a rat was implanted under approved surgery protocols with stainless steel screws in the somatosensory cortex as active electrodes and two over the cerebellum as the ground and reference electrodes. (3) For conducting TcMEP recording, a rat was anesthetized with pentobarbital and subdermal needle electrodes were placed over the motor cortex and in the gluteal muscle for stimulation and recording respectively. The reference electrode was located in the paw and a ground electrode was inserted in the ventral thigh. In bench top experiments, sinusoidal waves with different amplitudes and frequencies were fed into the transmitter and recorded at the receiver to determine the spectral characteristics of the communication channels. All of the passbands reside in the acceptable ranges of the respective signals shown in Fig. 1. The reliable transmission range for each device was measured at 30 meters for GEA and ECoG transponders and 15 meters for TcMEP. The power consumptions for GEA, ECoG and TcMEP were 13.5, 21, and 30 mW, respectively, at 3V bias. For in vivo examinations of each system, a relevant commercially-available system (subsequently referred to as the wired system) was used as the “gold standard” to compare the signals obtained from the developed wireless system. For GEA, slow wave signals were acquired for 80 minutes and the numbers of events recorded by both systems were identical. Figure 4 shows 2 minutes of slow wave signals recorded by (a) the wired Biosemi™ system and (b) our wireless system. The signals recorded by the wireless system preserved the critical steep peaks essential for tracking slow (a)
(b)
III. IN VIVO TESTING AND RESULTS Each system was first examined via bench-top settings and then in vivo with appropriate animal models. All the procedures were approved by the relevant institution.
Fig. 6. Evoked potentials recorded in the rat gluteus muscle and results from electrical stimulation of the motor cortex. (a) and (b) represent signals recorded by the Cadwell Cascade™ and our wireless systems, respectively.
Page 3 of 4
wave propagation across the stomach [4]. The ECoG signals were successfully recorded for several hours with a 1401Plus unit (CED, Inc) wired system and our wireless system. 7.3 seconds of the data from both systems are plotted in Fig. 5. The signals were identical as both frequency and amplitude components were properly preserved. The TcMEP signals were recorded for several hours using the Cadwell Cascade™ wired and our wireless systems. Figure 6 (a) and (b) show the evoked potentials recorded by Cadwell Cascade™ and wireless systems, respectively. The signature peaks, indicated by the arrows, and times between peaks were identical. There was a slight shift in time (latency) between the two waveforms. This difference is deemed to be unimportant and is due to a slight delay in triggering signals between in the wired system. IV. DISCUSSION AND CONCLUSION In this study, we developed three bidirectional telemetric systems based on a common platform with various system considerations. The systems were made using a commerciallyavailable chip, nRF24Le1, and discrete components to acquire EGG, ECoG and TcMEP signals, which range from very low (cycles per minute) to high (kHz) frequencies. The functionalities of the systems were examined on bench-top settings and in vivo with appropriate animal models. In the GEA system, the reliability of wireless communication and power efficiency of the system was pronounced as the front-end will be implanted inside the body. The system did not show signal distortion, which may happen due to tissue absorption [10]. The re-transmission function plays a vital role to ensure the signal integrity. The low sampling rate of 8 Sps provides 125-ms quiet time for the transceiver operation which includes a 260-μs delay in the radio mode switching and 96-μs time on air (TOA), so it can retransmit each packet up to 351 times to guarantee the reception. For ECoG, the maximum number for retransmission is two, which was sufficient since the implant was placed under the skull where the signals do not experience significant distortion or attenuation. The transponder front-end could work consistently for 124, 80 and 56 hours in the GEA, ECoG and TcMEP designs, respectively, based on a 560-mAh coin cell rechargeable battery, which is sufficient to provide valuable information for diagnostic applications. The differences between the trace shapes of the GEA and TcMEP signals recorded with wired and wireless systems (Figs. 4 and 6) are due to the unnecessary higher sampling rates in the wired systems and difference in pass-bands of the wired and wireless systems. This issue did not present a problem for diagnosis since the waveform characteristics (peaks) of the two methods did not show notable difference during 80 minutes of GEA recording or 100 repetitions of 13.5-ms evoked events for TcMEP recording in the animal models. No significant difference was observed between the
ECoG signals acquired through the commercial wired instrument and our developed wireless system. We have demonstrated the feasibility and expected performance of the wireless recording systems for EGG, ECoG, and TcMEP signals with in vivo animal models. More animal experiments need to be performed prior to human trials. ACKNOWLEDGEMENTS The authors greatly appreciate the valuable support of Dr. Nancy Clegg, Patricia Rampy, Dr. Christopher Lahr, Belinda Allen, Cynthia Luby and the staff at Texas Scottish Rite Hospital for Children in Dallas and University of Mississippi Medical Center. REFERENCES [1] C. T. Lin, L. W. Ko, J. C. Chiou, J. R. Duann, R. S. Huang, S. F. Liang, T. W. Chiu and T. P. Jung, “Noninvasive neural prostheses using mobile and wireless EEG,” Proc IEEE, vol. 96, no. 7, pp. 1167-1183 , July 2008. [2] E. Jovanov, A. Milenkovic, C. Otto and P. C. de Groen, “A wireless body area network of intelligent motion sensors for computer assisted physical rehabilitation,” J. Neuroeng Rehabil., vol. 2, pp. 6, March 2005. [3] K. Hachisuka, A. Nakata, T. Takeda, Y. Terauchi, K. Shiba, K. Sasaki, H. Hosaka, and K. Itao, “Development and performance analysis of an intra-body communication device,” International Conference on Transducers, Solid-State Sensors, Actuators and Microsystems, Vol. 2, pp. 1722–1725, 2003. [4] P. Du, G. O'Grady, J. U. Egbuji, W. J. Lammers, D. Budgett, P. Nielsen, J. A. Windsor, A. J. Pullan, L. K. Cheng, “Highresolution mapping of in vivo gastrointestinal slow wave activity using flexible printed circuit board electrodes: methodology and validation,” Ann Biomed Eng. vol. 37, no. 4, pp. 839-846, 2009. [5] A. Ritaccio, P. Brunner, M. C. Cervenka, N. Crone, C. Guger, E. Leuthardt, R. Oostenveld, W. Stacey, G. Schalk, “Proceedings of the first international workshop on advances in electrocorticography,” Epilepsy & Behavior, Octrober 2010. [6] T. Ball, M. Kern, I. Mutschler, A. Aertsen, A. Schulze-Bonhage, “Signal quality of simultaneously recorded invasive and noninvasive EEG,” NeuroImage, vol. 46, no. 3, pp. 708-716, July 2009. [7] R. Lin, R. G. Lee, C. L. Tseng, Y. F. Wu and J. A. Jiang, “Design and implementation of wireless multi-channel EEG recording system and study of EEG clustering method,” Biomed. Eng. App. Basis Communications, vol. 18, no. 6, pp. 276-283, 2006. [8] D. Lapray, J. Bergeler, E. Dupont, O. Thews and H. J. Luhmann, “A novel miniature telemetric system for recording EEG activity in freely moving rats,” J. Neurosci. Methods, vol. 168, no. 1, pp. 119-126, February 2008. [9] A. Farajidavar, J. L. Seifert, J. E. S. Bell, Y. S. Seo, M. R. Delgado, S. Sparagana, M. I. Romero and J. C. Chiao, “A wireless system for monitoring transcranial motor evoked potentials,” Ann. Biomed. Eng. pp. 1-7, 2010. [10] S. Gabriel, R. Lau and C. Gabriel, “The dielectric properties of biological tissues: II. measurements in the frequency range 10 Hz to 20 GHz,” Phys. Med. Biol., vol. 41, no. 11, pp. 22512269, November 1996.
Page 4 of 4
