Retinal ganglion cells transmit visual information from the eye to the brain with sequences of action potentials. Here, we demonstrate how to record the action potentials of single ganglion cells in vivo from anesthetized rats.
Part I: Fabrication of Tungsten-in-Glass Electrodes
Part II: Stereotaxic Setup and Spike Train Recordings
Materials
Adult Brown-Norway rats (250-400g) were purchased from a commercial vendor and housed under a regulated (12/12) light/dark cycle. Anesthesia was induced with an intraperitoneal injection of ketamine hydrochloride and xylazine (70 and 2 mg/kg, respectively, Henry Schein Inc) and maintained for the duration of the experiment with an intravenous infusion of ketamine and xylazine (30 and 1 mg/kg/hr) mixed with dextrose, saline, and gallamine triethiodide (40 mg/kg/hr, Fischer Inc) through a catheter (0.13mm OD, Small Parts Inc) in the right femoral vein. Gallamine is included in the mixture to paralyze eye movements. The infusion rate of the pump (WPI Inc) was adjusted as needed to maintain a stable plane of anesthesia as assessed by heart rate and blood pressure variability. After paralysis, the animal was mechanically ventilated (Harvard Apparatus, Model 683) through a tracheal cannula at 60-80 breaths/min (2cc volume), with the rate adjusted as necessary to maintain end-tidal CO2 measured with an in-line capnometer (Novametrix Inc, Model 710Sp) at 30%. Body temperature was regulated via a homeothermic blanket control system (Harvard Apparatus Inc). Body temperature, end-tidal CO2, heart rate, and blood pressure were continuously monitored throughout the experiment by a LABVIEW program. Visual stimuli were presented on a Sony Multiscan 17e CRT monitor (mean luminance of 30 cd/m2) running at 100 Hz with a resolution of 800×600 pixels. Data acquisition and monitor output were controlled with custom software written in Matlab and LabView in conjunction with a video image processor (Cambridge Research Systems Inc, Bits++) and the Psychophysics Toolbox4. The animal viewed the stimulus display (40.4 x 30.2 cm) at a distance of 16.5cm through contact lenses (Ocular Instruments Inc), and ophthalmic solution was applied periodically during the experiment to keep the eyes moist. The visual field accessible to stimulation was maximized by reverse mounting the animal in a stereotaxic apparatus (Stoelting Co) elevated 14cm above the surface of a floating table (TMC Inc) and setting the monitor on a custom-designed sled that moved on the table along a ±100 deg arc about the nose. This provided a repositionable display field that extended 60-deg above and 35-deg below eye level and ±42-deg laterally from the center of monitor. Tungsten electodes were fabricated with wire obtained from Small Parts Inc and custom borosilicate glass obtained from Friedrich and Dimmock Inc. A standard micropippete puller (Sutter Instruments, Model P-97) was used to shape the glass tip. The electrode impedance tester was purchased from Bak-Electronics. Electrodes were advanced with a motorized microdrive system (Newport Inc, StepperMike). Electrocardiogram and optic nerve signals were amplified and filtered with a high input impedance multielectrode amplifier (FHC Inc, X-Cell 3×4). Heart beats were detected with a window discriminator (WPI Inc, Model 121). Nerve action potentials were detected and time stamped with 0.1ms resolution by a digital spike discriminator (FHC Inc, APM). The spike discriminator was gated by a trigger signal from the video image processor so that spike times were locked to the stimulus.
Optic fiber recordings are an attractive approach for addressing experimental questions about retinal information encoding and transmission that require an intact eye. Moreover, the signaling properties of both eyes can be studied in virtually the same physiological state if the electrode is positioned in the optic chiasm or tract where the activity of crossed and uncrossed optic nerve fibers can be recorded with a single electrode penetration. Optic fiber recordings are common in cat but not in other popular animals models used in vision research, such as rodents, perhaps owing to their small size. We have tried a variety of commercial electrodes of similar impedance and material, none of which were successful at picking up single fiber activity in rat, let alone recording it for several hours like our electrodes can. This implies that the particular geometry of our electrodes, which have a long, sharp tip as opposed to the wide blunt tip of typical commercial ones, is important for reliable and stable isolation of ganglion cell axon spike trains. In addition to showing how to fabricate these electrodes, we illustrate our custom-designed stereotaxic system for in vivo visual neurophysiological research. The system is constructed to protect the high impedance microelectrode from environmental vibration and from electromagnetic noise. This is critical for recording the small action potentials produced by axons (as compared to cell bodies) for a prolonged period of time, especially with a computer monitor positioned nearby. It does so with exceptional stability and signal-to-noise ratio, with typical recording times of an hour or more and noise levels on the order of tens of microvolts. These features make the setup particularly useful for vision researchers aiming to record in vivo from nerve fibers of the retina or other brain regions.
The authors have nothing to disclose.
We thank Dr Dan Green for providing technical input during the development of these experimental techniques. This work was supported by NIH Grant R01-EY016849A and the Smith Family New Investigator Award.