Journal of Neuroscience Methods 168 (2008) 119126

A novel miniature telemetric system for recording EEG activity in freely moving rats
Damien Lapray 1 , J urgen Bergeler 1 , Erwan Dupont, Oliver Thews, Heiko J. Luhmann
Institute of Physiology and Pathophysiology, University of Mainz, Duesbergweg 6, D-55128 Mainz, Germany Received 20 August 2007; received in revised form 27 September 2007; accepted 28 September 2007

Abstract Telemetric recording systems offer the advantage to monitor physiological parameters in freely moving animals without any restrictions in their explorative behaviour. We present a novel, inexpensive, portable and reusable telemetric system to record the electroencephalogram (EEG) from adult freely moving rats under various experimental conditions. Our system consists of an implantable transmitter which communicates at a sampling rate of 500 Hz bi-directional with a receiver via radio transmission (in EU: 868.35 MHz; in USA: 916.5 MHz) over a distance of up to 3 m. The switching time between receiving and transmitting signals is 20 s and the data transmission rate amounts to 115.2 kbps. The receiver is connected to a laptop via an USB connection and the data are displayed and saved by a software developed by the authors. This system allows the simultaneous recording and storage of a video signal for direct comparison of the animals EEG with its behaviour. EEG recordings could be obtained over 45 weeks and under various experimental conditions (i.e. from rats swimming in water). The current system is optimized for recording electrical activity from the animals brain, but can be easily modied to record other physiological parameters. 2007 Elsevier B.V. All rights reserved.
Keywords: Telemetry; Electroencephalogram; Freely moving rat; Wireless physiology

1. Introduction Recording physiological parameters from freely moving experimental animals represents a most valuable approach to study the animals milieu interne (e.g. blood pressure, heart beat, body temperature, brain activity) under relatively natural conditions. However, a direct connection of the animal to the recording apparatus, e.g. a cable connecting the EEG electrode on the animals head with the EEG recording amplier, does not only restrict the animal in its locomotion and exploratory behaviour, but may also act as a strong stress factor thereby modifying various physiological parameters (Tang et al., 2004). Therefore, wireless, so-called telemetric recording systems which are as comfortable as possible to the awake experimental animal are in many aspects advantageous and more valuable when compared to conventional recording devices.

Corresponding author. Tel.: +49 6131 39 26070; fax: +49 6131 39 26071. E-mail address: luhmann@uni-mainz.de (H.J. Luhmann). These authors equally contributed to this work.

Telemetric recording systems should full a number of requirements in order to guarantee a long-term and powerful analysis of physiological parameters in freely moving animals under stress-free conditions: (i) All devices, which have to be implanted in the animal, should be as small and light as possible in order to minimize or better prevent stress and pain to the animal. The system should be also fully biostable and biocompatible in order to allow telemetric recordings over long periods as weeks or even months. (ii) The energy consumption of the implanted components should be as low as possible and it should be possible to control the energy source (usually a battery) from extern by switching the system on and off or by recharging the battery. (iii) The system should allow a telemetric transmission of the recorded signals over at least a few meters in order to allow free exploratory behaviour of the animals in a dened environment, e.g. a conventional open eld for rodents. (iv) The system should work reliably in water in order to allow telemetric recordings of physiological parameters from diving or swimming animals, e.g. rats in the Morris water maze (Morris, 1984). (v) Internal components should be re-useable and the external recording system should be as simple as possible to

0165-0270/$ see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2007.09.029

120

D. Lapray et al. / Journal of Neuroscience Methods 168 (2008) 119126

minimize the expenses. (vi) All external components of the telemetric system (e.g. the receiver) should be as small as possible and portable to allow experiments outside the conventional laboratory settings. (vii) In order to allow maximal exibility in the experimental design and in the recording protocols, the properties of the signals sent out from the transplanted transmitter should be adjustable from outside. In the present report we describe a novel telemetric recording system that fulls all these requirements. We constructed a system that allows telemetric recording of the electroencephalogram (EEG) from adult freely moving rats under different experimental conditions. In its current version the system is optimized for recording electrical activity from the animals brain, but by the utilization of other implantable devices the system can be easily used to record other physiological parameters such as blood pressure, heart beat, body temperature, etc. Telemetric recordings of the EEG in freely moving mammals is currently an important issue in neurosciences, because a number of different brain rhythms have been described in the last two decades (for review Buzs aki and Draguhn, 2004; Singer, 1999; Steriade, 2005; Uhlhaas and Singer, 2006), but the physiological function and the behavioural context of the different brain rhythms are not completely understood (for comprehensive review Buzs aki, 2006). The use of radio-telemetry to collect EEG measurements or other neurophysiological data in conscious, unrestrained animals is one valuable approach to address these questions (Guler and Ubeyli, 2002). However, the full implantable systems available so far do not transmit the signal over distances that would allow to combine recordings with behavioural test in a larger open eld (Mumford and Wetherell, 2001; Obeid et al., 2004; Williams et al., 2006). Here we describe the development of a telemetry system totally implantable for a maximum freedom from interference by the subject. The behaviour is not affected by the implantation system. This system is suitable for EEG and electrocorticogram (ECoG) in freely moving rats. Most important, the EEG signal is transmitted at a distance of up to 3 m, thus cortical electrical activity and behaviour can be synchronously recorded and then directly correlated. In addition, we investigated the question if the normal behaviour of the rats were affected by the implantation surgery and the transmitter. 2. Materials and methods 2.1. Overview Our system consists of an implantable transmitter which amplies and AD-DA converts the recorded signal and communicates bidirectional with a receiver via radio transmission over a distance of up to 3 m (Fig. 1). The receiver is connected to a computer via an USB connection and the data are displayed and saved by a software developed by the authors for this purpose. A commercially available CED and Spike2 software (Cambridge Electronic Design, Cambridge, England) can also be used to display and store the data on the host computer.

Fig. 1. Schematic illustration of the main components of the telemetric system. The recording electrodes are implanted on the cortical surface and connected to a transmitter placed in the abdominal cavity of the rat. The transmitter is switched ON/OFF by the receiver. The recorded signal is directly amplied and digitized by the transmitter, emitted in all directions and detected by the receiver at a distance of up to 3 m. The data are transferred from the receiver to a computer via an USB connection and can be visualized online with our software. The animal behaviour is synchronously recorded with a video camera and movies are stored on-line on the PC.

2.2. Implanted system The implanted part of the telemetric system (Fig. 2A, B and Table 1 ) consists of a microcomputer (PIC12F675, Microchip ), an analogue input amplier and a transceiver. An operational amplier is also present to generate the virtual ground potential and a power switch in order to reduce the power dissipation during the standby mode. The system is assembled with two stacked printed circuit board (PCB) connected by wires. The power supply is composed by two small silver-oxide batteries (Renata 399, 9.5 mm 9.6 mm, 55 mAh), connected by conductive glue (Chemtronics, USA) and connected to the PCB by a holder that is constructed by stainless steel wires. The rst stage of the analogue input path is an instrumental amplier with an amplication of 10. The input voltage is measured as a differential signal between the positive and the negative input. The reference potential (virtual Ground VGND) has a level of half of the battery voltage and is connected to the skull of the animal. The second stage is a standard noninverting operational amplier with an amplication of 500 and is dc-decoupled by a 2.2 F non-polarized capacitor. To avoid resonance effects a small capacitor with a value of 470 pF is connected in parallel to the resistor in the backward path. The
Table 1 Properties of the transmitter unit Property Weight Dimensions Power supply Carrier frequency Run mode current Standby mode current Life time battery 4 g (with batteries and wires) 40 mm 8 mm 5 mm 2 1.55 V (silveroxide) 868.35 MHz ISM (Europe) 916.5 MHz ISM (USA) 1.6 mA 2 A 14 h at 500 Hz sampling rate

The system operates at sampling rates sufcient for physiological recordings of EEG activity (2502000 Hz). The light transmitter contains a battery that on average allows 14 h of recording.

D. Lapray et al. / Journal of Neuroscience Methods 168 (2008) 119126

121

Fig. 2. Properties and function of the implanted part of the telemetry system. (A) Electronic circuit diagram of the implanted unit. (B) Photograph of the telemetric device. (C) Illustration of the protocol, which starts with initiating pulses to adjust the automatic-gain-control of the receiver and subsequently transmits the data. After 300 s the transceiver switches to the reception mode and can also detect other transmitters using the free ISM band. During this period instructions can be received to modify the systems status. (D) Flowchart of the machine program.

analogue signal is internally digitized in the microcomputer with a resolution of 10 bits and a conversion time of 2 s. The microcomputer controls the system status running and standby, the digitalization and the receiving and transmitting of radio signals. It contains a non-volatile EEPROM memory of 128 bytes where the parametric data are stored. The implanted system acts as a receiver and a transmitter following the protocol illustrated in Fig. 2C. The hybrid transceiver system (TR1001, RF Monolithics ) is available for European (868.35 MHz) and other ISM-band frequencies (e.g.: USA 916.5 MHz). This part has a unique rapid switching time between receiving and transmitting of 20 s in combination with a transmission data rate of 115.2 kbps. Fig. 2D shows the owchart of the machine program. After inserting the battery the system remains in a standby status with a very

low power dissipation of less than 2 A. The implanted system wakes up every 2.5 s, switches the receiver ON and scans for an instruction to start. This operation takes less than 10 ms. If no signal is received, the system falls back to standby status. After receiving an ON signal the unit starts to measure and transmit the data until it receives an OFF instruction. The acquisition rate can be pre-selected by software from 250 Hz up to 2 kHz. In order to minimize the size of the system most of the parts are Surface Mount Devices with a size of 0402. Before implantation, electrode wires (0.3 mm stainless steel wires with HiFlex nylon coating, Grifn, Schramberg, Germany) with small connectors at their tip are plugged on stainless steal pins on the transmitter and a small layer of epoxy is applied on the connectors to protect them. These wires can easily be replaced to reuse the system for new implantations. The

122

D. Lapray et al. / Journal of Neuroscience Methods 168 (2008) 119126

transmitter is then covered with two layers of a biocompatible silicone (Elastosil N2010, Wacker Chemie AG, Munich, Germany). Between the two layers of silicone and the system one layer of water-soluble separating agent (PVA, R&G, Waldenbuch, Germany) is rst applied in order to facilitate the cleaning of the transmitter after extraction from the animal at the end of the experiments. The shape of the silicone is made as round and smooth as possible to avoid discomfort and distress of the animals. 2.3. Control system The receiving and control system is composed of a personal computer and an interface box (Fig. 3A, B) connected via standard USB. The receiver of the interface is the same as in the implanted units except some modications to adapt the high frequency power and the modulation depth. According to the received data rate the implemented microcontroller (PIC16F876A, Microchip ) is connected by a high speed serial interface (2.5 mbps) to a special interface chip FT232R (IC1, FTDI ) which converts the serial data stream into the complex USB protocol. The program is written in machine language

to optimize the usage of memory space and processing speed. Because the microcontroller has the ability to write to its own program memory the software can be updated by the PC. 2.4. PC software The PC software consists of three independent modules for (1) capturing data, (2) replaying/analyzing data, and (3) software servicing of the control system and the implanted devices. In the capture module the user has to dene the parameters of the capture session (e.g. sampling rate) and whether parallel to the EEG data acquisition a video recording will be performed. The user can also dene a detailed time schedule describing the points of time and the duration of each recording session. After initialization of the control device the software starts the data transmission of the implanted system. All data will be sent as a continuous stream in the predened sample rate. For quality assurance each data block contains a unique identication and a successive numbering so that a loss of single data packages during wireless transmission can be identied. The data will be written to the PC hard disk in a binary format together with a timestamp. In parallel the received data will be displayed on

Fig. 3. Electronic circuit diagram of the control interface (A) and photograph of the USB unit (B).

D. Lapray et al. / Journal of Neuroscience Methods 168 (2008) 119126

123

the PC screen in scope-like fashion. In the case that a video camera is attached to the PC via an USB port the software is capable to capture a video stream synchronously to the EEG data recording on the PC hard disk. Due to the large number of data in this mode the rate of the real-time EEG data display on the scope is reduced. However, due to extensive buffering a loss of data can be ruled out. Terminating the capture session can be initiated either manually or automatically by the predened time schedule. For this reason repetitive recordings over a long period of several days without the presence of a human operator are possible which may be advantageous in the case of long-term behavioural studies. The replay software module reads the EEG data together with the stored video stream and displays them synchronously on the screen. With a time mark the user can indicate time intervals of interest. The accompanying EEG data of these time frames can either be stored separately on the hard disk or be exported in a format compatible for use with other analysis software (e.g., Spike2, MatLab). In this module the incorporation of individual specic analysis routines is also possible. The service module of the software can be used to parameterize (e.g. setting the sample rate) the implanted devices. The whole software package is written in Delphi (Cupertino, CA, USA) using modules from the JEDI Visual Component Library (JVCL, http://jvcl.sourceforge.net). Besides the large number of EEG data, video streaming produces extensive le sizes. For this reason, a suitable size of the hard disk is necessary. 2.5. Surgery All experiments were conducted in accordance with the national and European (86/609/EEC) laws for the use of animals in research and were approved by the local ethical committee (Landesuntersuchungsamt Koblenz, 23 177-07/G071-001). Male Wistar rats weighing 250400 g were used. The animals were housed individually in standard plastic cages (42 cm 26 cm 20 cm) under a 12 h lightdark cycle (lights on at 7 am). The room temperature was maintained at 21 2 C and relative humidity at 50 5%. Standard rodent food and tap water were available ad libitum. The surgery was performed after deeply anaesthetising the animals with an intra-peritoneal (IP) injection of chloral hydrate (400 mg/kg, chloral hydrate 99%, SigmaAldrich, Steinheim, Germany). This was followed after 10 min by the IP injection of ketamine (1 ml/kg, ketamine 500 mg/10 ml, Ratiopharm, Ulm, Germany) and by the subcutaneous (SC) injection of atropine (0.8 mg/kg, SigmaAldrich). The anaesthesia was maintained by ketamine injections during the whole surgery. For a better recovery all rats were given an antibiotic (IP) (0.1 ml Baytril 2.5 %, Bayer Vital, Leverkusen, Germany) and an analgesic treatment (SC) (0.05 ml Rimadyl , Pzer, Karlsruhe, Germany). The area from the part between the eyes, back between the ears and across the neck as well as the abdominal skin were shaved and cleaned with iodine. The animal was then placed on a heating pad and an incision was made in the left part of the abdomens mid-line, 1 cm caudal to the xyphoid cartilage, and the transmitter was placed in the peritoneal cavity (Fig. 4A).

The unit could be sutured to the body wall with non-absorbable nylon. After secured in the stereotaxic apparatus a 4 cm midsagittal incision was made on the scalp and the skin reected with hemostats to expose the entire skull. A way for the leads was made between the skin and muscles by moving aside connective tissue and transmitter leads were slipped subcutaneously from the abdomen to the incision made on the head (Fig. 4B and C). The skull was then cleaned and dried. The Bregma was marked according to the coordinates (Paxinos and Watson, 1998) and three holes were made through the skull with a driller (High Speed Micro Drill, Fine Science Tools inc., Heidelberg, Germany) xed on the stereotaxic apparatus in order to control the descent and avoid any brain damage or bleeding. The

Fig. 4. Implantation procedure. (A) Placement of the transmitter in the abdominal cavity. (B) The leads were slipped under the skin from the belly to the heads incision and holes were drilled in the skull to place the recording electrodes, which were xed to the skull with grip cement. (C) The ground and the reference electrodes were placed above the cerebellum. In this study, the recording electrode was placed above the somatosensory cortex.

124

D. Lapray et al. / Journal of Neuroscience Methods 168 (2008) 119126

recording electrode was placed above the somatosensory cortex (L = + 5.5 mm, AP = 2.3 mm), the reference and the ground electrodes were placed above the cerebellum. Three stainlesssteal screws (0.5 mm diameter) were used as electrodes and soldered to the leads. The assembling was anchored in place with grip cement (Dentsply Caulk International, Milford, USA). Both incision sites were closed using 4-0 Resolon (Resorba, N urnberg, Germany). Surgery lasted a maximum of 3 h from induction of anaesthesia and the success rate was 100% for recovery from anaesthesia and surgery. The animals were weighed daily throughout the experiment as an indicator of general health. 2.6. Data recordings After surgery, a recovery period of 5 days was given to the animals before starting the rst recording session, corresponding to the time needed to gain the pre-surgical weight again. The rats were placed in the recording eld that could be a cage, an open eld, a labyrinth or a water maze. The recording system was switched on through the receiver and the signal directly observed on a computers screen. The signal was recorded with our software at a sampling rate of 500 Hz and saved on the hard-disk. Data were then imported to MatLab (MatLab 7, The MathWorks Inc., Natick, USA) for further analysis. Frequency analysis was performed by the use of Fast Fourier Transform (one epoch of 100 s) with a band-pass lter between 1 and 80 Hz. 2.7. Statistics Statistical analyses were performed with Systat Version 10 (Systat Software, Erkrath, Germany). Values throughout this

report are given as mean S.E.M. For statistical comparisons, an unpaired samples t-test was performed. 3. Results 3.1. Recovery of the animals after implantation All the implantations were followed by a lost of weight with a maximum peak 2 days after the surgery (94.6 0.4% of pre-surgery body weight). Five days after the implantation the animals started again to take weight and reached the surgerys one after 10 days. No infections or mortality due to the implantation were observed on the six rats used for this study. To test the impact of the system on the animal movements, rats were placed for 10 min in an empty open eld (60 cm 70 cm) and a behavioural recording was made with EthoVision (Noldus Information Technology, Berlin, Germany). The total distance of moving and the velocity of the animals were then compared to non-implanted rats in the same conditions. No signicant differences were observed in these two parameters (30.3 1.7 m and 5.1 0.3 cm/s for the control animals versus 29.2 1.7 m and 4.9 0.3 cm/s for the implanted animals, each group n = 6). The device did not restrict limb movements during locomotion. The telemetry device did not cause any discomfort to the animals which were still able to curl up to sleep, a prominent sleeping posture in rodents (Morton et al., 2003; Tang et al., 2007). 3.2. The telemetric system gives recordings that are comparable with known studies Animals were recorded during three distinct brain states: wakefulness (Fig. 5A), slow-wave sleep (Fig. 5B) and under

Fig. 5. Examples of EEG recordings from the somatosensory cortex of adult rats under different experimental conditions. Characteristic EEG recordings from individual rats are shown. Lower EEG traces are higher magnications of the corresponding upper EEG trace. Each picture is a snapshot from the video recorded during the EEG recording time. FFT spectra were calculated from 100 s traces. (A) The animals were recorded in their cages during wakefulness and (B) slow-wave sleep. (C) Example of EEG recording during treatment with the volatile anaesthetic isourane. (D) Telemetric EEG recordings from animals in an empty open eld and (E) in a two arms maze.

D. Lapray et al. / Journal of Neuroscience Methods 168 (2008) 119126

125

isourane anaesthesia (Fig. 5C). EEG recordings were measured with a sampling rate of 500 Hz in a normal laboratory setting with no precautions taken to limit any kind of external electromagnetic interferences. Wakefulness was characterized by a desynchronized low-amplitude EEG with dominant theta activity (48 Hz, supplementary movie 1). Slow-wave sleep was clearly distinguished by a synchronized EEG with high-voltage slow-waves in the delta frequency range (14 Hz) and some spindles (1014 Hz) in an immobile animal with closed eyes (supplementary movie 2). Under isourane anaesthesia burst-suppression pattern characterized by a depressed background activity alternating with high voltage activity were observed (supplementary movie 3). Our EEG recordings obtained with the telemetric system are in good agreement with previous reports on EEG activity in sleep states (Franken et al., 1998; Gervasoni et al., 2000; Timo-Iaria et al., 1970) as well as under isourane anaesthesia (Hudetz, 2002). 3.3. The telemetric system allows the simultaneous recording of brain activity and animal behaviour We obtained telemetric recordings from rats in different environmental conditions, like an open eld (60 cm diameter, 70 cm height) (Fig. 5D, supplementary movie 1), in a two arms maze (Fig. 5E, supplementary movie 4) and in water (supplementary movie 5). These recordings were performed with a distance between the transmitter and the receiver of 0.23 m. The signal was devoid of movement artefacts during the recording sessions except for the grooming behaviour where some electrical spikes could be observed especially when the animal was cleaning its belly. These artefacts can be easily identied by the video monitoring and removed out-line. For the water test we performed recordings of the animals in an openeld full of water (around 150 l). The signal recorded was devoid of distortions and stable in an area of 30 cm around the antenna when this one was placed beside the openeld at a level below the waters surface. 4. Conclusions We present a novel telemetric recording system that offers several advantages compared to the currently available telemetric systems. The relatively small size and weight of our fully implantable device is very well accepted by adult rats which do not show any kind of behavioural impairments or sign of discomfort. The incorporation of an AD-DA converter in the transmitter, the homemade software to display and store the data simultaneously with the video signal on the host computer and the possibility to change easily wires and battery make these system inexpensive, portable and easy to use and reuse, allowing experiments in a wide variety of environments. Another advantage of our system is the transmitting range of the signal, up to 3 m, without any lost of data which allows any kind of open eld experiments and for EEG recordings in water (e.g. could be used in Morris water maze). Due to this large transmitting range and the impossibility to switch ON

one system independently, the implanted animal under recording session has to be separated from the other implanted ones. Except this last point no particular precautions have to be taken in order to limit any kind of external electromagnetic interferences allowing recordings directly in the animal facility or any other places. In addition the relatively long lifetime of the batteries, due to the low power consumption of the system and possibility to switch on and off by external control, permits experiments lasting a few weeks. Finally after sacricing the implanted animal and extracting the transmitter, the system can be reused several times after changing batteries and wires. A further miniaturization of the transmitter, the extension of the system to at least two recording channels, a higher transmission and acquisition rate and the addition of a bi-directional signal transfer (e.g. for electrical stimulation of brain regions) will make this system in the future even more powerful. Acknowledgments We are grateful to Pascal Ravassard and Dr. Romain Goutagny for help with the surgical experiments. We thank Dr. Nicolas Heck for helpful comments on the manuscript, Hans Reiner Polder for technical discussions and Florian Lebayle for his illustrations abilities. DL is member of the neuroscience graduate school at the University of Mainz (DFG GRK 1044). This work was funded by the Stiftung Rheinland-Pfalz f ur Innovation and by the EC (LSH-CT-2006-037315, EPICURE). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jneumeth.2007.09. 029. References
Buzs aki G. Rhythms of the brain. Oxford University Press; 2006. Buzs aki G, Draguhn A. Neuronal oscillations in cortical networks. Science 2004;304:19269. Franken P, Malafosse A, Tafti M. Genetic variation in EEG activity during sleep in inbred mice. Am J Physiol 1998;275:R112737. Gervasoni D, Peyron C, Rampon C, Barbagli B, Chouvet G, Urbain N, et al. Role and origin of the GABAergic innervation of dorsal raphe serotonergic neurons. J Neurosci 2000;20:421725. Guler NF, Ubeyli ED. Theory and applications of biotelemetry. J Med Syst 2002;26:15978. Hudetz AG. Effect of volatile anesthetics on interhemispheric EEG crossapproximate entropy in the rat. Brain Res 2002;954:12331. Morris R. Developments of a water-maze procedure for studying spatial learning in the rat. J Neurosci Methods 1984;11:4760. Morton DB, Hawkins P, Bevan R, Heath K, Kirkwood J, Pearce P, et al. Renements in telemetry procedures. Seventh report of the BVAAWF/FRAME/RSPCA/UFAW Joint Working Group on Renement, Part A. Lab Anim 2003;37:26199. Mumford H, Wetherell JR. A simple method for measuring EEG in freely moving guinea pigs. J Neurosci Methods 2001;107:12530. Obeid I, Nicolelis MA, Wolf PD. A multichannel telemetry system for single unit neural recordings. J Neurosci Methods 2004;133:338. Paxinos G, Watson C. The rat brain in stereotaxic coordinates. 4th ed. San Diego: Academic Press; 1998.

126

D. Lapray et al. / Journal of Neuroscience Methods 168 (2008) 119126 Timo-Iaria C, Negrao N, Schmidek WR, Hoshino K, Lobato de Menezes CE, Leme dR. Phases and states of sleep in the rat. Physiol Behav 1970;5:1057 62. Uhlhaas PJ, Singer W. Neural synchrony in brain disorders: relevance for cognitive dysfunctions and pathophysiology. Neuron 2006;52:15568. Williams P, White A, Ferraro D, Clark S, Staley K, Dudek FE. The use of radiotelemetry to evaluate electrographic seizures in rats with kainateinduced epilepsy. J Neurosci Methods 2006;155:3948.

Singer W. Neuronal synchrony: a versatile code for the denition of relations? Neuron 1999;24:4965. Steriade M. Sleep, epilepsy and thalamic reticular inhibitory neurons. Trends Neurosci 2005;28:31724. Tang X, Orchard SM, Liu X, Sanford LD. Effect of varying recording cable weight and exibility on activity and sleep in mice. Sleep 2004;27:80310. Tang X, Yang L, Sanford LD. Sleep and EEG spectra in rats recorded via telemetry during surgical recovery. Sleep 2007;30:105761.