IOP PUBLISHING SMART MATERIALS AND STRUCTURES

Smart Mater. Struct. 17 (2008) 027001 (7pp) doi:10.1088/0964-1726/17/2/027001

TECHNICAL NOTE

The development of a PZT-basedmicrodrive for neural signal recordingSangkyu Park1, Euisung Yoon1, Sukchan Lee2, Hee-sup Shin2,Hyunjun Park3, Byungkyu Kim3,6, Daesoo Kim4, Jongoh Park5 andSukho Park1,5,6

1 Nano-Bio Research Center, Korea Institute of Science and Technology, PO Box 131,Cheongryang, Seoul 130-650, Korea2 Center for Neural Science, Korea Institute of Science and Technology, PO Box 131,Cheongryang, Seoul 130-650, Korea3 School of Aerospace and Mechanical Engineering, Korea Aerospace University, 200-1,Whajon-dong, Deokyang-gu, Koyang-city, Kyonggi-do 412-791, Korea4 Department of Biological Sciences, Korea Advanced Institute of Science and Technology,373-1, Guseong-dong, Yuseong-gu, Daejeon 305-701, Korea5 School of Mechanical Systems Engineering, Chonnam National University, 300,Yongbong-dong, Buk-gu, Gwangju 500-757, Korea

E-mail: bkim@kau.ac.kr and spark@chonnam.ac.kr

Received 4 August 2007, in final form 6 December 2007Published 5 February 2008Online at stacks.iop.org/SMS/17/027001

AbstractA hand-controlled microdrive has been used to obtain neural signals from rodents such as ratsand mice. However, it places severe physical stress on the rodents during its manipulation, andthis stress leads to alertness in the mice and low efficiency in obtaining neural signals from themice. To overcome this issue, we developed a novel microdrive, which allows one to adjust theelectrodes by a piezoelectric device (PZT) with high precision. Its mass is light enough to installon the mouse’s head. The proposed microdrive has three H-type PZT actuators and their guidingstructure. The operation principle of the microdrive is based on the well known inchwormmechanism. When the three PZT actuators are synchronized, linear motion of the electrode isproduced along the guiding structure. The electrodes used for the recording of the neuralsignals from neuron cells were fixed at one of the PZT actuators. Our proposed microdrive hasan accuracy of about 400 nm and a long stroke of about 5 mm. In response to formalin-inducedpain, single unit activities are robustly measured at the thalamus with electrodes whose verticaldepth is adjusted by the microdrive under urethane anesthesia. In addition, the microdrive wasefficient in detecting neural signals from mice that were moving freely. Thus, the present studysuggests that the PZT-based microdrive could be an alternative for the efficient detection ofneural signals from mice during behavioral states without any stress to the mice.

(Some figures in this article are in colour only in the electronic version)

1. Introduction

For research on brain activity, a neural signal, which is calledan action potential, is a representative signal activity, and

6 Authors to whom any correspondence should be addressed.

it can be recorded by electrode wires. For neural signalrecording, a microdrive with a function of precise manipulationof the electrode is installed on the head of the study animal.Therefore, many studies on neural signal measurement andanalysis using the brains of large animals such as monkeys

0964-1726/08/027001+07$30.00 © 2008 IOP Publishing Ltd Printed in the UK1

Smart Mater. Struct. 17 (2008) 027001 Technical Note

(Ludvig et al 2001), rats, and pigeons (Bilkey et al 2003) havebeen carried out since there is little restriction with regard toits size and weight of a recording device for large animals.As models of various disorders, however, gene-modified miceare generally used and a small microdrive for neural signalrecording is much required.

Since the development of the recording electrode (Adrianand Zotterman 1926), two general approaches have beenapplied for neural signal recording: acute and chronicrecording (Cham et al 2005). The manual microdrive foracute recording consists of a tube, electrodes, and a screw.The electrodes are inside the tube, which is connected to thescrew. The screw is rotated manually (Bilkey and Muir 1999)to advance and position the electrodes. The acute recordingmethod using a manual microdrive has some difficulties incontrolling and maintaining the position of the electrodes. Theelectrode system for chronic recording, the second approach,consists of many electrodes or a thin wire array. It is surgicallyimplanted to the brain region of interest (Porada et al 2000,Rousche and Normann 1998, Williams et al 1999). Neuralsignal recording using a multiple electrode array depends on aprobability of the surgical operation. To overcome the abovedrawbacks of acute recording and chronic recording and tomeasure a clear neural signal, an automatically manipulatedmicrodrive was needed.

As automatic microdrives, motorized microdrives arecommercially available (Thomas Recording GmbH, FHC, andNarishige). However, the motorized microdrives are large andheavy, and thus are only appropriate for an anesthetized animal.On the other hand, micromotor-based microdrives (Fee andLeonardo 2001, Cham et al 2005) have been proposed. Dueto the length of the micromotor, the height of the microdrivebecomes too large, and thus the microdrives are not suitable forthe recording of a freely moving mouse. Previous researcheshad proposed linear piezomotors based on the inchwormmechanism (Newton et al 1997, Ni and Zhu 2000, Zhang andZhu 1997). These linear piezomotors have certain advantages,such as simple structure, precise positioning, fast response,and large moving range. However, these piezomotors also arelarge, and thus cannot be applied as a microdrive for neuralsignal recording of small animals. Recently, MEMS-basedmicrodrives for neural signal recording have been proposed(Muthuswamy et al 2005, Pang et al 2007). MEMS-basedmicrodrives can have many advantages of size and weight.However, the MEMS-based microdrives are only tested withrespect to their positioning performances, and the neural signalrecording results were not reported. Therefore, we proposea small and light-weight microdrive that can be applied forneural signal recording of an anesthetized mouse and a freelymoving mouse. Through a preliminary positioning test andin vivo experiments, the performances of the microdrive wereevaluated.

2. Construction of the PZT-based microdrive

A microdrive requires several characteristics (Bilkey et al2003), as follows: (1) small size and low weight, (2) preciseadvancement, (3) stability, (4) simple installation, and (5) low

Figure 1. Structure of the proposed microdrive.

cost. The above requirements should be considered in thedesign and fabrication of a microdrive for neural signalrecording. In this study, for the neural signal recording of asmall animal such as a mouse, a small microdrive is proposedand fabricated since gene manipulated small mice are widelyused in neuroscience research.

2.1. Design and fabrication

For the neural signal recording of a mouse, a microdrive basedon three PZT (lead zirconate titanate) actuators is proposed, asshown in figure 1. It consists of H-type PZT actuators, guidehousing, cannula and electrode. H-type PZT actuators areassembled with three stack PZT actuators. Their actuation isbased on the principle of inchworm-like movement, as shownin figure 2. Of the three PZT actuators, those at the ends (Aand C) are the actuators for clamping, and the middle PZTactuator (B) is used for elongation. By the sequential motionsof the H-type PZT actuators in figure 2(a), the microdrivecan advance stepwise. For the inchworm motions of the H-type PZT actuators, the driving voltage sequences for the threePZT actuators (A, B, and C) are illustrated in figure 2(b). Inaddition, the electrode for neural signal recording is placedinside of the cannula, and the cannula is attached to one ofthe clamp PZT actuators. Therefore, the electrode can advanceas the microdrive moves forward.

Figure 3 illustrates the fabricated microdrive. It weighs5 g and has dimensions 9 mm × 7 mm × 13 mm. In theH-type PZT actuator as shown in figure 3(a), the two clampPZT actuators (AE0203D04, TOKIN) are 5.0 mm in overalllength and yield 4.6 µm in maximum displacement. Forfine position control, an elongation PZT actuator (PL022.30,Physik Instrumente GmbH & Co.), which has dimensionsof 2 mm × 2 mm × 2 mm and maximum displacement of2.0 µm, is used. The three PZT actuators are assembled andglued by epoxy resin. The guide housing in figure 3(c) ismachined by wire-EDM (electrical discharge machining) andis polished by a hand rasp. Figure 3(b) shows the connectorfor the connection of the measured neural signal and the dataacquisition system, and the head mount in figure 3(d) is usedto fix the microdrive on the head of the mouse.

2.2. Control system for the microdrive

For the inchworm motion of the microdrive, the drivingvoltage sequences in figure 2(b) are necessary. For thegeneration of the driving sequences, we used a general purpose

2

Smart Mater. Struct. 17 (2008) 027001 Technical Note

Figure 2. Inchworm mechanism of H-type PZT actuators: (a) PZT actuator sequence and (b) driving voltage sequence.

controller (DS1103, dSPACE Inc., Germany). The drivingsequences from the controller were amplified by a PZTamplifier (PL022.30, Physik Instrumente GmbH & Co.), andthe amplified driving sequences were used as the drivingvoltage inputs of the microdrive.

3. Preliminary positioning test of the PZT-basedmicrodrive

In order to confirm the positioning performance of ourmicrodrive, we executed preliminary positioning tests ofthe microdrive. For displacement measurement, a laserposition sensor (laser doppler vibrometer, Polytec, Inc.) with20 nm resolution was used. The overall schematic diagramof the positioning tests is shown as figure 4. The stepdisplacement of the proposed microdrive depended on thedriving voltage of the middle elongation PZT actuator (B).Therefore, the positioning tests were executed under thevariation of the driving voltages of the PZT actuator (B). Theexperimental results are presented in figure 5. The microdriveshowed stepwise forward movements. The step displacementsincreased as the applied voltages of the elongation PZTactuator (B) increased. In particular, if the applied voltagewas 100 V, the step displacement was about 2 µm. From theexperimental results, it was confirmed that the proposed andfabricated microdrive had sufficient positioning accuracy forneural signal recording.

Figure 3. Fabricated microdrive: (a) H-type PZT actuators,(b) connector, (c) guide housing, and (d) head mount.

4. Experiment for neural signal recording

4.1. Experimental setup for neural signal recording

The overall neural signal recording system for gene-manipulated mice is illustrated in figure 6. The neuralrecording system consists of a mouse, the proposed microdrive,controller, pre-amplifier, main amplifier, and a data acquisition

3

Smart Mater. Struct. 17 (2008) 027001 Technical Note

Figure 4. The overall schematic diagram of the positioning tests.

0 1 2 3 4 5 6 7

2

4

6

8

10

Steps

20V40V60V80V100V

0

12

Dis

pla

cem

ent(

um)

Figure 5. Displacement of the microdrive for actuating voltages.

system. For neural signal recording, the electrode is preciselypositioned by the proposed microdrive, and the measuredneural signal is amplified by the pre-amplifier and mainamplifier. The neural signal is a very weak signal and itis very sensitive to external noises. For noise reduction ofthe measured signal, therefore, the pre-amplifier (a currentamplifier with a unit gain) is connected to the microdriveand mounted on the head of the mouse. The main amplifierconnected with the data acquisition system is a voltageamplifier, and the gain of the main amplifier is about 3000–10 000. The proposed microdrive is surgically mounted onthe head of the mouse by the mount in figure 3(d) and gluingcement. The microdrive is controlled by the dSPACE controllerand PZT amplifier.

Figure 7 shows a surgically implanted microdrive on thehead of a mouse. Our target region in the brain of the mousewas the thalamus, which is a well known sensory junction.For the initial positioning of the electrode, the mouse brainin stereotaxic coordinates (Paxinos and Franklin 2001), whichcan give information about the location of the interest region inmouse brain, is employed. From the mouse atlas, the thalamus

Figure 6. Microdrive and neural signal recording system.

Figure 7. Microdrive mounted on the head of a mouse.

region includes the VPM (ventral posterior medial nucleusof thalamus) and the VPL (ventral posterior lateral nucleusof thalamus) which is located at about 1.7 mm from bregma(the junction of the sagittal and coronal sutures at the top ofthe skull), laterally 1.7 mm from the center of the brain, andvertically 3.0 mm from the cortex of the brain. In the surgicaloperation, therefore, the electrode of the proposed microdrivewas inserted close to the thalamus region. Then, the proposedmicrodrive advanced the electrode stepwise and searched forthe action potentials from neuron cells.

4.2. Neural signal recording experiments

First, we carried out an experiment of neural signal recordingfor a mouse under general anesthesia. Figure 8 shows themeasurement results of the neural signals from the anesthetizedmouse, and the mark (�) implies the action potentials of theneural cell signals. The magnitudes of the action potentialswere about 20 mV, and the base noise was about 2 mV. The

4

Smart Mater. Struct. 17 (2008) 027001 Technical Note

Figure 8. Neural signal from a mouse under anesthesia before formalin stimulation.

Figure 9. Neural signal of a mouse under anesthesia after formalin stimulation.

5

Smart Mater. Struct. 17 (2008) 027001 Technical Note

Figure 10. Neural signal in a freely moving mouse.

signal to noise ratio was about 10:1, and the neural signal wasappropriate for analysis.

In another experiment, in order to stimulate the neuroncells, a pain inducing drug (formalin) was injected to theabdomen of the mouse, and the change of the neural signalwas monitored. Figure 9 shows the neural signal afterthe injection of the drug. From the result, we could findthat the action potentials were very vigorous, implying thatthe neural cells actively responded to the pain induced byformalin.

The ultimate objective of our experiments was theneural recording of a freely moving mouse under automaticpositioning of the electrode. Figure 10 shows a neuralsignal from a freely moving mouse. Compared with theresult of the anesthetized mouse, the action potential signalsfrequently occurred, and the magnitude of the action potentialwas about 10 mV and reduced by half compared with theaction potential under anesthesia condition. Through the aboveneural recording experiments with the anesthetized mouse andthe freely moving mouse, we can confirm that the proposedmicrodrive can precisely advance the electrode for neuralsignal recording and has sufficient positioning performanceand compatibility to find the action potentials of neuron cells.

5. Conclusions

This study proposed and fabricated a miniaturized microdrivefor the neural signal measurement of a mouse. The microdriveconsisted of H-type PZT actuators, guide housing, andelectrode, and advanced the electrode based on inchworm-like

locomotion. Through the positioning tests of the microdrive,we could achieve an accuracy of about 400 nm and a longstroke of about 5 mm. By using the developed microdrive,the neural signals from an anesthetized mouse with or withoutformalin stimulation were recorded. Action potential signalsof 20 mV from an anesthetized mouse were measured, andthrough the stimulation, the action potential signals appearedvery vigorously. In order to understand the difference ofthe action potential between an anesthetized mouse and afreely moving mouse, neural signal recording tests wereperformed on a freely moving mouse, and the action potentialsignals were also recorded. Through the neural signalrecording experiments with the anesthetized mouse and thefreely moving mouse, therefore, the proposed microdrive isexpected to be a useful tool for recording signals from in vivoextracellular single-unit recording.

Acknowledgments

This research has been supported by the 21st Century FrontierR&D Projects (the Intelligent Microsystem Center) and theNext Generation New Technology Development Programs(No. 10030037) from the Korea Ministry of Commerce,Industry and Energy.

References

Adrian E and Zotterman Y 1926 The impulses produced by sensorynerve endings J. Physiol. 61 49–72

6

Smart Mater. Struct. 17 (2008) 027001 Technical Note

Bilkey D and Muir G 1999 A low cost, high precision subminiaturemicrodrive for extracellular unit recording in behaving animalsJ. Neurosci. Methods 92 87–90

Bilkey D, Russell N and Colombo M 2003 A lightweight microdrivefor single-unit recording in freely moving rats and pigeonsMethods 30 152–8

Cham J, Branchaud E, Nenadic Z, Greger B, Andersen R andBurdick1 J 2005 Semi-chronic motorized microdrive andcontrol algorithm for autonomously isolating and maintainingoptimal extracellular action potentials J. Neurophysiol.93 570–9

Fee M and Leonardo A 2001 Miniature motorized microdrive andcommutator system for chronic neural recording in smallanimals J. Neurosci. Methods 112 83–94

Ludvig N, Botero J, Tang H, Gohil B and Kral J 2001 Single-cellrecording from the brain of freely moving monkeys J. Neurosci.Methods 106 179–87

Muthuswamy J, Okandan M, Jain T and Gilletti A 2005 Electrostaticmicroactuators for precise positioning of neural microelectrodesIEEE Trans. Biomed. Eng. 52 1748–55

Newton D, Gracia E and Horner G 1997 A linear piezoelectric motorSmart Mater. Struct. 6 295–304

Ni J and Zhu Z 2000 Design of a linear piezomotor with ultra-highstiffness and nanoprecision IEEE/ASME Trans. Mechatron.5 441–3

Pang C, Tai Y, Burdick J and Anderson R 2007 Electrolysis-basedparylene balloon actuators for neural probes Proc. 2nd IEEE Int.Conf. on Nano/Micro Engineered and Molecular Systems(Bankook, Thailand)

Paxinos G and Franklin K B J 2001 The Mouse Brain in StereotaxicCoordinates 2nd edn (New York: Academic)

Porada I, Bondar I, Spatz W and Kruger J 2000 Rabbit and monkeyvisual cortex: more than a year of recording with up to 64microelectrodes J. Neurosci. Methods 95 13–28

Rousche P and Normann A 1998 Chronic recording capability of theUtah intracortical electrode array in cat sensory cortexJ. Neurosci. Methods 82 1–15

Williams J, Rennaker R and Kipke D 1999 Long-term neuralrecording characteristics of wire microelectrode arraysimplanted in cerebral cortex Brain Res. Prot. 4 303–13

Zhang B and Zhu Z 1997 Developing a linear piezomotor withnanometer resolution and high stiffness IEEE/ASME Trans.Mechatron. 2 22–9

7