A detailed description of the training paradigm and experimental procedures is found in a preceding paper (Populin and Yin, 1998
). Briefly, using the magnetic search coil technique (Fuchs and Robinson, 1966) to measure eye movements, we trained cats to look at the location of sound sources with their heads fixed. Auditory and visual stimuli were presented from an array of speaker and light-emitting diode (LED) assemblies located in front of the cat. Acoustic stimuli were broadband (0.1-25 kHz) noise bursts. The animals were food-deprived and rewarded when their eyes were within electronic windows surrounding the targets.
A typical session included a random mixture of various experimental tasks: visual and auditory fixations, standard and delayed saccades, and sensory probes. Not all tasks were used in every experimental session, but a variety was always presented to prevent the cat from anticipating upcoming trials.
Pinna movements recorded during standard and delayed saccades to visual and auditory targets (Populin and Yin, 1998) and auditory sensory probe trials are included in this report. In the standard saccade task the cat was required to first fixate an LED at the primary position (0°,0°) and then saccade to a spatially disparate visual or auditory target presented at the time the fixation LED was turned off. In the delayed saccade task the target was presented some time (500-700 msec) before the offset of the fixation LED, which constituted the signal for the cat to saccade. In sensory probe trials the fixation LED remained on for the entire duration, and the cat was expected to fixate it, even when an acoustic probe was presented during the fixation period. In saccade trials the cat was rewarded for making a saccade to the target within specified spatial and temporal windows (Populin and Yin, 1998), whereas in sensory probe trials the cat was rewarded for not breaking fixation. No behavioral contingencies were placed on pinna movements in any task.
Coil implant. Pinna movements were routinely recorded with coils anchored to the pinna. We considered two options for attaching the coil to the ear: taping the coil to the external aspect of the pinna before each session (Jay and Sparks, 1987; Hartline et al., 1995) and implanting it under the skin. We chose the latter because it permitted more stable recordings than a removable coil, which was subject to slippage during an experiment; it maximized reproducibility, because it would be impossible to attach the coil in the same position day after day; it minimized noise from loose leads; and it minimized discomfort for the cat, which is very particular about its pinnae. The disadvantages of implanting the coil were the potential for damaging the musculature, innervation, or blood supply and restricting pinna movements with taut coil leads. None of the cats implanted with this procedure showed any signs of discomfort or impaired pinna function. Furthermore, postmortem inspection indicated that the coil was firmly held in place. We conclude that the disadvantages were kept to a minimum.
The search coils were implanted subcutaneously under sterile surgical conditions at the same time that the head restraint was attached. In Cat06 we implanted the coil on the flat, medial aspect of the right pinna, because it would help preserve the geometry of the coil (Fig. 1, right ear) (see Figs. 7, 10). However, the resulting position of the coil was far from perpendicular to the magnetic fields and complicated the calibration procedure. The phase detectors of our magnetic search coil system (CNC Engineering) produce an output signal that is a sinusoidal function of angle; thus the output is highly nonlinear when the coil is oriented 
30° from the coronal plane. In subsequent cats we implanted the coils more caudally in the pinna, as close to the coronal plane as possible within the limitations imposed by its geometry (Fig. 1, left ear; see Figs. 4-6, 8, 9).
A 6-7 mm incision was made in the skin, and a pocket large enough to house the coil was dissected. Each coil was 10 mm in diameter and made of three turns of AS633 fine wire (Cooner Wire Co.). Precautions were taken to avoid damaging the delicate muscles, their innervation, or blood supply. The leads from the coils were routed loosely, so as not to tether the pinna, underneath the skin to the top of the head where they were soldered to brass connectors (Microtech Inc., Boothwyn, PA) embedded in dental cement near the head post.
To calibrate the pinna coils we made use of the observation that the cats would generally orient their pinnae to a more or less standard position when they fixated a visual stimulus at the primary position (0°,0°). The calibrations were then made around this position. The orientation of each pinna coil when the cat fixated at (0°,0°) was estimated by eye during behavioral testing with a calibrating ring (Fig. 2, cr). This ring, made of malleable copper wire, was carefully manipulated to overlie the implanted coil, the contour of which could be seen while the cat was fixating an LED at the primary position. Once the orientation of the ring was judged to match that of the pinna coil, the cat was removed from the experimental setup, and the angular deviation of the ring from the coronal plane was measured. A dummy coil, physically matched to those implanted, was placed at the same orientation in the center of the field coils and rotated along the x- and y-axes separately, over a range of ±20°. The voltage output of the coil system was sampled at 10° intervals, and the data points relating angular deviation and voltage output were fit with linear equations. The coefficients were used by the data analysis programs to transform the voltage output of the actual pinna coil into angular deviation.
Data analysis. The onset of pinna movement was determined with a criterion similar to the one used previously for eye movements (Populin and Yin, 1998). Pinna movement onset was defined as the time at which the velocity trace was >2 SD of the mean velocity baseline, from 125 msec before to 5 msec after the onset of the stimulus. The horizontal and vertical components were computed separately, and the component with the shorter latency was chosen for analyses. Because of the small amplitude of some pinna movements and large first derivatives of the noise in the traces of some trials, we had to override the objective criterion and pick the onset of movement by eye in ~15-20% of the trials. A 2 msec acoustic delay was subtracted from all latency measurements. Statistical analyses are presented as means and confidence intervals.
Answers to all your Biology Questions
