Recording electrophysiological signals from a freely-behaving preparation is technically challenging. Unlike anesthetized or in-vitro preparations, the system used to record from a behaving animal must accomodate movement. As the preparation moves, the cable connecting the animal to the instrumentation flexes. Cable flexion results in capacitance changes which cause small currents to flow through the cable. These small currents can generate sizeable voltages when they flow through high impedances. A voltage generated by cable flexion that is large enough to interfere with the electrophysiological signals you are trying to measure is called a movement artifact. Movement artifacts can be eliminated by reducing the impedance seen by the cable. This is accomplished by mounting a special preamplifier called a headstage between the electrode and the cable as shown in Figure 1. The headstage should be as close to the electrode as possible. The connection between the electrode and the headstage must be rigid.

---

Figure 1

---

Headstage amplifiers are generally voltage-follower circuits built around an operational amplifier (op-amp) or field-effect transistor (FET). The output voltage of the headstage circuit follows the input voltage precisely. However, the output impedance of the headstage circuit is low. Thus, a headstage amplifier can be described as an impedance transformer which takes a signal from a high impedance source (the electrode) and presents it to the cable as the same signal voltage but at a lower impedance. Practical op-amp and FET realizations of voltage-follower circuits are shown below (Figure 2). Actual circuits used in particular laboratories may vary slightly from the configurations shown. However, all headstages that successfully eliminate movement artifact will act as impedance transformers and present a low impedance to the cable.

---

Figure 2

---

Both of these circuits are actively used in laboratories recording from freely-behaving animals. The selection of one circuit over another is largely based on legacy issues. Both circuits will successfully reduce movement artifacts. A comparison of the two types of headstage circuit is presented in the table below.

  Table 1:  Fet / Op-Amp Comparison

ISSUE	FET	OP-AMP
Power	Discrete FET circuits will usually draw more power than op-amp circuits and depend on the resistance used by the circuit. Thus, battery life is reduced and more heat is generated when compared to low-power op-amps. Increased circuit temperature will result in increased wideband thermal (Johnson) noise.	Micropower op-amp versions are available (< 200 microamps/channel).  Power consumption is generally unaffected by the resistor value used in the headstage circuit.
Circuit	Each circuit requires at least one resistor. Power consumption and output impedance depend on the resistance value used. There is a trade-off between power consumption and output impedance.  Increased power consumption will also generate more heat.	Some circumstances may require no resistors at all.  If the cable has sufficiently low capacitance, the 50 ohm resistor in Figure 1 can be replaced by a wire.
Size	Increased circuit complexity results in larger sized multi-channel headstages. Individually packaged FETs can result in smaller headstages for low channel-count applications. The 2N3958 that we use contains two FET devices.	Reduced circuit complexity results in smaller headstages, especially in multi-channel situations. Op-amp ICs can have as many as 4 devices in a single package.
Technology	Discrete FET technology is relatively static.	Op-amp technology is actively developed. New and improved op-amps become available several times a year.

As the table suggests, FETs have significant drawbacks when compared to op-amps, particularly in newer preparations in which activity from a large number of wires (stereotrodes, tetrodes) will be recorded.  In these cases, op-amp circuits would be far superior because of their low power consumption and reduced circuit complexity, characteristics which serve to minimize the size of the headstage mounted on the rat and also reduce the heat generated by the circuitry.

Nevertheless the lab currently uses the FET-based circuit shown in Figure 2.  Each channel is built around one half of a 2N3958 dual FET device (Arrow Advantage). This is an n-channel junction FET.  We use five devices to construct a 10 channel headstage.  The 100 k-ohm resistor limits power consumption and heat dissipation but presents an impedance to the cable that is many times larger than that of the op-amp circuit. This trade-off appears to be reasonable for extracellular single unit recordings.  However, this headstage design may not be as effective at removing movement artifact for other types of electrophysiological recording such as spontaneous EEG or evoked potentials because these signals use a different passband.

It is likely that our lab will soon adapt our headstage design to utilize an op-amp circuit.  In principle, all op-amps work similarly.  A key parameter to look for when seeking an op-amp to use in a headstage circuit is input bias current, which should be as low as possible (less than 50 picoamperes).  Size, power consumption, operating voltage and input offset voltage are other parameters to consider.  Specific examples of op-amps used by us would include the LF444 , LF442 and LF441 for four- , two- and one-channel headstages, respectively.  These could be combined for larger designs.  Smaller packages are also available.  Other IC's that have been used for unit recording include include the TI2274AC, TLC2274, LT1490, and TL084.  For more information and specs on different ICs, go to www.questlink.com or call Newark.

---

Suppliers

Newark Electronics, 7272 Park Circle Drive, Hanover MD  21076, 1-800-NEWARK  (general electronics)
Arrow Advantage, 18640 Lake Drive East, Chanhassen MN  55317, 800-833-3557  (for 2N3958 Siliconix FET)

Web Content from:  Schoenbaum, G.  Olfactory Learning and the Neurophysiological Study of Rat Prefrontal Function.  In: CRC Series: Methods and Frontiers in Neuroscience.  Edited by S.A. Simon and M.A.L. Nicolelis, CRC Press, NY, 2000.

This web page coauthored by Kevin B. Austin, Ph.D., Eclectic Engineering Studio, www.EclecticStudio.com