Chapter 4 Three Dimensional Neural Probe Array
4.4 Neural Recording in Rat Brain
To demonstrate the practical function, the fabricated 3-D neural probe array was implanted into an anesthetized rat. Fig. 4.7 (A) shows the photograph of a stacked 3-D microprobe array that was inserted into the brain of an anesthetized rat by a manual 3-axis moving stage (not shown in the figure). The screw on the skull was adopted as a reference for measurement. The opening in the skull was about 2 mm by 3 mm. Fig. 4.7 (B) shows the
photomicrograph of the implantation section. The figure was modified by superimposing a lesion marker, an implantation track, and overlaying a scaled image of the microprobe array.
One lesion marker arrowhead (red) was used to identify the location of the outermost recording site, in relation to the field CA1 of the hippocampus. Fig. 4.7 (C) presents neural signals from the 16-channel microprobe array, acquired with a Multi-Channel Acquisition Processor (MAP, Plexon Inc., USA). During recordings, electrical signals were passed from the headstage to an amplifier through a band-passed filtered (spike preamp filter: 450–5 kHz, gain: 15,000–20,000) and sampled at 40 kHz per channel. Fig. 4.8 presents a 50-ms segment of spontaneous neural activity derived from Fig. 4.7 (C) for clear action potential observation.
Fig. 4.7 (A) Photograph of inserted stacked 3D microprobe in brain (B) The in situ location of probes (C) The simultaneous16-channel neural activities recording
Fig. 4.8 Recorded spontaneous neural activity in motor cortex of rat
4.5 3D Signal Acquisition in Human Brain Resection Tissue
Next, real 3D neural signal propagation observation in human cerebral resection cortex layers is achieved by practical implantation after resection surgery on the cooperation with the Buddhist Tzu-Chi Hospital in Hua-Lien City, Taiwan in this study. All the study procedures follow the IRB rule of Buddhist Tzu-Chi Hospital. The patient is suffering from epilepsy seizure needed for surgery resection treatment. After the resection surgery, the
tissue was put into a prepared ACSF solution in a water bath with 37 Celsius degree with oxygen pump to keep it alive for recording. The fabricated 3D probe array was inserted into the brain tissue by a manual 3-axis moving stage. From the recording, we can build a 3D distributed spatial signal plot. Fig. 4.9 displays the recorded raw data from the resection tissue. Fig. 4.10 (A) illustrates the 3D distributed sensing electrodes diagram in implant tissue volume. Fig. 4.10 (B) shows the normalized near-field potential intensity variation observation in 4 layers in continuous time frames. However, spontaneous recording in the resection tissue lacked for even-related response to demonstrate the propagation in three dimensional volumes. Future work will also seek for stimulated recording in implant tissue for further investigation.
Fig. 4.9 Recorded raw data from resection brain tissue
Fig. 4.10 (A) 3D distributed sensing electrodes diagram in implant tissue volume (B) Normalized LPF intensity propagation observation in 4 layers in continuous time frames
4.6 Summary
The minimal opening is the area that must be resected, including skull and dura, to fully place an implantable device onto the brain. A smaller skull opening can reduce the implantation damage such as the rise of intracranial pressure, and the probability of wound infection. In previous work, the opening area was never less than the supporting platform [74-75] to make sure all the probes were completely inserted into tissue. Therefore, the minimal surgical opening area was defined by the supporting platform in this case. Previous works requires larger the supporting platform area when ASIC chips were used for system integration [59]. In presented stacking method, the system integration will not increase the opening area because it can be accomplished by replacing spacers with ASIC chips. The opening area, depends only on the probe array dimension, is less than 1.75×1 mm2 which can be readily shrunk by using thinner and narrow shafts.
Previous work may also induce additional tissue damage in the bottom of the platform as well. The interlocking structures, including tethers and joints, can cause the protrusion and damage to the tissue underneath [59]. For the stacked 3-D array, only probe array will be in contact with the target tissue.
The strength of the assembled structure is also an important issue in implantable device.
Compared with the proposed 3D array, structures with tethers and joints used in previous work may not provide reliable strength to fix the probes on the platform during implantation [59] [74]. The thermosetting polymer in the stacked 3-D array provided an adhesive strength of 150–180 kg/cm3 after curing. Thus, sufficiently structural strength was guaranteed in the present design.
Compared with previous 3-D array studies, the advantages of using the stacking method for constructing 3-D arrays include easier assembly processes, stronger structure strength, smaller opening area and less damage to the tissue surrounding the implanting region. ASIC chips can be substituted for spacers to achieve system integration without increasing device size as well. The stacking method can therefore increase the design flexibility and enhance the volume usage efficiency.
In-vivo spontaneous neural activity recording from present stacked multichip neural sensing
structure performed on anesthetized rat also illustrates the capability for practical implantable neural recording. Finally, 3D neural signal propagation observation for human cerebral cortex layers is achieved by practical implantation in excised brain tissue after resection surgery in this study. The neural signal data acquired by 3D multichip array achieves the recording and mapping of the neural signal network and interconnections among the target brain structure, which allows further studies for event-related observation.Future work will also seek for suitable solder-ball process and related biocompatible encapsulation material to realize the full concept of the stacked multichip structure.
Chapter 5
Integrated Electronics toward Microsystem
5.1 Biomedical Microsystem for Neuroprosthesis
Neuroscience engineering together with in-vivo real time biomedical information, such as neural activity, recorded by implant neural sensors and microsystems is ultimately crucial for bio-logical researches to identify and develop treatments for brain disorders i.e.
Parkinson’s disease, epilepsy, and seizures [87], as well as for developing rehabilitative prosthetic devices to help people with disabilities such as deafness [88], blindness [89], and incontinence regain at least some measure of normal function. Moreover, implantable devices have attracted the interest of the neuroscience community because of their importance in facilitating basic research on neural organization, learning, and plasticity in behaving animals. Such implantable microsystems promise to enable revolutionary advances in recording from dozens to hundreds of neurons at the cellular level [90-92]. A miniaturized long-term and reliable biomedicine implantation with versatile bio-sensors is needed for real-time prosthetics applications in free-behavior implanted subject. Current commercial products in the implants often with a heavy battery associate with the system power supply. Therefore, systems constructed by integrated circuits and bio-potential sensors with miniaturized power and volume, developed with coil inductor for RF-power receiving from an external host is highly desirable.
In this chapter, key components including inductive coupled coil for radio-frequency (RF) powering, low dropout (LDO) voltage regulator for RF-DC conversion and multi-channel analog-front-end (AFE) amplifier for biopotential conditioning are developed for the biomedical microsystem design. Section 5.2 describes the RF-powering coil and LDO regulator design, following with a 16-channel AFE chip development reported in Section 5.3. Spiral coils as a wireless power module with LDO linear regulator circuit convert RF signal into DC voltage provide a batteryless implantation for truly free-behavior monitoring.
The 16-channel AFE provides pre-amplification and signal conditioning. Also, a chip-on-broad (COB) integrated microsystem consisting of fabricated AFE chip, commercial micro control unit (MCU) and wireless transceiver module are developed for the conceptual representation of the proposed microsystem.
5.2 Wireless Powering Coil and Circuitry
Fig.5.1 illustrates the electrical system architecture of the proposed RF-powering system.
External component includes PA and RF power transmission coil, on-broad components involves receive coil, rectifier, smoothing capacitance and bias diode, and the on-chip component includes LDO regulator with band gap bias circuit and thermal protection circuitry. External RF power produced by a Class-E amplifier is coupled into the microsystem via a tuned LC network followed by a full-wave rectifier and regulators to produce stable 1.8V and 0.7V for system and reference supplies. 13.56MHz RF-powering is chosen due to lowest tissue absorbability to avoid damage on tissue [93].
Results in this section were co-worked with Sheng-Hsin Hung in Microsystem Control Laboratory, National Chiao-Tung University.
Fig. 5.1 Electrical system block diagram of the wireless powering system
5.2.1 Spiral Coil and Full-wave Rectifier
The radio frequency power is transmitted through coupled coils, which follows the electromagnetic induction of Faraday’s Law. Fig. 5.2 shows the simplified schematic diagram of an inductive power transmission link [92]. L1 is the primary coil that is attached to the skin from outside of the body, driven by an ac source Vs, which is often an efficient class-E power amplifier [94]. L2, is the secondary coil that is implanted under the skin flap along with the rest of the implant electronics. A pair of permanent magnets, one in the center of each coil, can align and hold them together to maximize their mutual inductance M.
Coil windings have parasitic resistance and capacitance associated with them, which are represented by lumped elements Rs1 and Rs2. Capacitors Cs1 and C2 are added to form a pair of resonant LC-tank with L1 and L2, respectively.
Magnetic flux lines are formed around the primary coil as a result of the flow of current though it. They close their paths through the air and spread all around the coil. The voltage induced in the secondary coil, which is important for both power and data delivery to the
Fig. 5.2 (A) Basic schematic diagram of the inductive link (B) Equivalent circuit implant, is due to the magnetic flux passing though the secondary coil [95]. The equivalent AC load resistance Rac which will dissipate an amount of AC power equivalent to the DC power in Ro is
2
o AC
R R ( 5 - 1 ) The equivalent RL resistance can be transform into equivalent resistance
Also, the equivalent resistance Re, reflected back into the primary coil is1 where K is the coupling coefficient, which can be find by the
2
from testing probe [96]. The coupling coefficient is2 where VR is the voltage drop on the resistance load Ro. M is the mutual inductance of the coils and are the unloaded quality factor of the two coils. The optimal efficiency is given by [97]
2 The coupling coefficient is determinated by the coil dimension (diameter and distance), and the quality factor is defined by the coil structure and material [98]. In this work, low
skin-effect, low resistance single layer spiral coil structure is utilized. Then, consider of a standard 50 ohm output power amplifier, the Rs1 can be determinated as
e
s
R
R
150 ( 5 - 9 ) Also, the effective transmission distance can be estimated by using equations [99]:
2 maximized coupling coefficient can be determined through [97]. Quality factor is maximized by selecting appropriate conductor material. Calculate RAC, RS2_Optimal and RS1
from desired rectifier performance and powering specification. Therefore, L1 and L2 can be designed and obtained and realized by experimental measurements. Then, the resonance capacitance CS1 and C2 can be found according to desired transmission frequency. Fig. 5.3 (A) (B) shows the optical photographs of the fabricated receive and external coils. The receiving and external powering coil, made by 24/16 AWG cupper wire, exhibit 1.5cm and 4cm in diameter, respectively. Fig. 5.3 (C) shows the measured S22 parameter and Smith chart of the developed 13.56 MHz inductive coupled spiral coils.
Fig. 5.3 (A) Receiving coil (B) External coil (C) Measured S22 parameter (D) Smith chart of the developed SCWPM
Rectification is processed utilizing a full-wave rectifier to convert the RF AC signal, received from the spiral coil through inductive coupling, to DC voltage level for further regulation. Here, rectifier consisting of four diodes is used, as shown in Fig. 5.4.
Fig. 5.4 Full-wave bridge rectifier using 4 diodes [100]
The antenna efficiency is characterized with PDMS coating as bio-protection in implanted tissue. Measurement result under conditions in air/tissue (pork) and with/without PDMS encapsulation is illustrated in Fig. 5.5, which results no obvious influence from packaging.
Fig. 5.5 Antenna efficiency characterization with PDMS packaging as bio-protection in implant under conditions in air/tissue (pork) and with/without PDMS encapsulation
5.2.2 RF-DC Voltage Regulator
Fig. 5.6 illustrates the block diagram of the proposed low-dropout linear regulator. The architecture is modified from a typical low-dropout regulator topology [101] with power MOSFET and thermal protection unit to enhance the driving current and avoid temperature damage on tissue. The error amplifier amplifies the voltage difference between the reference voltage and divided load voltage and switches the power MOSFET (PMOS). Output voltage is described as
2 2 1
R R V R
VOUT REF ( 5 - 1 2 )
Fig. 5.6 Block diagram of the proposed low-dropout linear regulator
Another design issue in the linear regulator is the stability problem. According to the bode plot analysis, the output capacitance and its equivalent series resistance (ESR) decides the
zero point to increase the phase margin. However, the ESR must be carefully designed in appropriate range to ensure the system stability.
Fig. 5.7 shows the circuit schematic of the error amplifier. MI27 is the power MOSFET illustrated in Fig. 5.7. Also, stability is one of the design key points. In this design, ceramic capacitance with low ESR is used as output capacitance, which can cause low phase margin and lead into un-stable condition. Therefore, extra RC-compensation is introduced to enhance the system stability. In Fig. 5.7, current limitation control by two Poly resistances which is inverse proportional to current. PIP capacitance I48 and resistor RI31 are used for Miller compensation. PIP capacitance MI44 and MI39 are used as internal smoothing capacitances for spike rejection. MOS MI3 acts as output switch, which controlled by thermal protection circuit.
Fig. 5.7 Error amplifier circuit
Fig. 5.8 (A) shows the detailed schematic of a temperature-independent bandgap reference voltage generator (RVG) [102]. In Fig. 5.8 (A), the PNP BJT QI16 and QI17 act as negative- and positive-TC voltage devices produce a temperature independent voltage difference to the operational amplifier. A resistor RI18 is used to adjust the output VREF. PIP capacitance MI31 and resistor RI2 are used for Miller compensation. MI5 and MI6 act as a startup for quicker response. MI4 is an extra gain stage for higher PSRR. MI20 is used for extra ESD protection, while the PIP capacitance MI3 used as a filter and bias. Fig. 5.8 (B) illustrates the thermal protection circuit schematic. The over-temperature protection (OTP) design is used to avoid thermal damage on tissue due to the high temperature (> 40°C) caused by the operating circuit. In Fig. 5.8 (B), RI3 is a ROND resistor in tsmc 0.35um process, which has high temperature coefficient (PTC1=1.51e-3). When T raise, voltage drop on RI3 increase, then turn-on MI14, increase current flow passing MI14, turn-on MI15, pull down OTP. Hysteresis functions when temperature drop down lower than 40°C, MI14 will not off immediately due to the current provided by BJT QI2. Until the temperature is lower enough, say <37°C, voltage drop on RI3 decrease, current decrease, MI14 turn-off, MI15 off, OTP pull high.
Fig. 5.8 (A) Bandgap voltage reference generator and (B) Over-thermal protection circuit The proposed LDO regulator chip is fabricated via TSMC 0.35um 2P4M process. The die size is 1.42 x 0.95 mm2. Fig. 5.9 (A) shows the optical microphotograph. Fig. 5.9 (B) is the test result of the Power supply rejection ratio. The measured PSRR is around 70.883 dB and 0.165 dB at 10k Hz and 13.56M Hz, respectively. Comparing to the simulation result of PSRR (~81.44 dB), the main difference is that simulation consists of only pure capacitance, but in practice, parasitic inductance does exist in the output ceramic capacitance, which decreases the high frequency performance in PSRR.
Fig. 5.9 (A) Optical microphotograph of the fabricated chip (B) PSRR of the LDO The stability test of the LDO regulator displays the ripple and noise on the output voltage level under different current loading. Fig. 5.10 (A) Output stability observation under IOUT=200 mA. Result shows that the maximal noise level is lower than 3-4mV. Low ripple and noise performance is one of the great property of linear regulator. Fig. 5.10 (B) present the output ripple caused by a current loading from 0 mA to 200mA. Output transient behaves a 308-264 mV spike, say 17%-14% variance in output voltage, under the condition VIN from 3V to 6V, respectively. The output transient performance can be easily improved by increase the Quiescent Current but also increase the power consumption. Fig. 5.13 (C)
illustrates the output ripple observation under a sudden maximal current loading. The maximal output current varies under different VIN condition. Result shows that the peak of the spike is less than 300mV. Line Transient is the output current ripple caused by the input voltage variance. Fig. 5.13 (D) shows the spike observation under loading current 200 mA with VIN variance from 2.6V to 6V. The overall Line Regulation performance is between 1.5-4.5mV, cause 0.0833~0.25% variance in output current.
Fig. 5.10 (A) Stability test@IOUT=200 mA (B) Load transient test IOUT=0-200 mA@VIN =3V (C) Ripple observation@load 0-252mA (D) Output spike@VIN =2.6-6V
Table 1 shows the measured Load/Line Regulation result. Note that when the IOUT=200mA under VIN from 5V to 6V, the system is thermal shutdown due to the over thermal protection (OTP). Generally speaking, the Load Regulation is between 0.5-5mV, which is around 0.0278-0.278% variance in output voltage.Table 5 summarizes the measured performance of the presented LDO regulator comparing to other works.
Table 5.1 Load/Line Regulation Test Result
VIN (V) 2.6 3.3 4.2 5.0 6.0 Line Regulation IOUT=0mA
VOUT (V)
1.828 1.8285 1.829 1.8295 1.8285 1.5mV
IOUT=50mA 1.828 1.829 1.83 1.8305 1.8305 2.5mV
IOUT=100mA 1.828 1.83 1.831 1.8325 1.8325 4.5mV
IOUT=200mA 1.8285 1.8315 1.834 OTP OTP NA
Load Regulation 0.5mV 3mV 5mV NA NA
Presented chip shows low drop out voltage, wide input range with reasonable Quiescent Current. Additionally, thermal-protection (<40°C) design is also included in the present linear regulator to avoid damage in implanted target for implant device applications. Some equations used for specification comparison are listed below:
MAX Table 5.2 Summary of LDO Regulator performance
[103] [104] [105] This work utilized for system performance characterization. To generate the RF powering, a power amplifier module using commercial chips is used to amplify the 13.56 MHz signal from a signal generator. Fig. 5.11 shows (A) Fabricated wireless powering system (B) Schematic of the class-E amplifier (C) Receive module. Fig. 5.11 (D) illustrates the temperature raise test of the SCWPM under 37°C environmental temperature, which is used to simulate the practical implant environmental condition with maximal output current (200mA) for 1hr.
Result shows that less than 2°C raise is observed thus meet the implantation requirement [56] [106].
Finally, Fig. 5.12 shows observation on the output performance of receive coil, rectifier and LDO chip by applying 13.56M Hz signal to the. Fig. 5.12 (A) and (B) display the received power on receiving coil, the rectifier output and the LDO output when input signal on
E-class amplifier is 2.5V and 6.8V, while related output current is 0 mA and 200 mA, respectively. The maximal output ripple is around 50mV due to the low PSRR performance at 13.56M Hz. However, the ripple can be reduced down to less than 20mV is extra 0.1μF ceramic capacitance is added at the LDO output. Fig. 5.12 (C) shows the load regulation under current loading varies from 100 mA to 200 mA. The Transient voltage is about 400mV in peak. Table 5.3 shows the result of measured Load regulation of the full system.
To summarize, the spiral coils as a wireless power module is presented in power management for batteryless medical instrumentation applications. The coil and circuit design, fabrication and system implementation are exhaustively discussed in this section.
Finally, practical measurement result provides the detailed performance characterization of the presented wireless power module.
Fig. 5.11 (A) Fabricated system (B) Schematic of the class-E amplifier (C) Receive module (D) Temperature raise test under 37°C with maximal output current for 1hr
Fig. 5.12 (A) Output ripple observation under VIN = 2.5V, IOUT = 0 mA (B) VIN = 6.8V, IOUT
= 200 mA (C) Load regulation under IOUT = 100-200 mA
Table 5.3 Measured Load regulation of the system
Current VOUT
IOUT=0mA 1.828V
IOUT=50mA 1.827V
IOUT=100mA 1.826V IOUT=150mA 1.823V IOUT=200mA 1.820V Load Regulation 8mV
5.3 Multichannel Analog Front End
The studies of biomedical signals, or say biopotential, play an important role in the researches of biomedical science. Generally, biopotentials used in diagnoses are monitored with electrodes to acquire these electrical signals. Generally speaking, EEG and ECG are most widespread used in clinical diagnoses, and they span over a relatively low frequency
The studies of biomedical signals, or say biopotential, play an important role in the researches of biomedical science. Generally, biopotentials used in diagnoses are monitored with electrodes to acquire these electrical signals. Generally speaking, EEG and ECG are most widespread used in clinical diagnoses, and they span over a relatively low frequency