Effects of Superior Colliculus Inhibition on Visual Motion Processing in the Lateral Suprasylvian Visual Area of the Cat

Abstract

purpose. To clarify whether visual inputs of the tectothalamocortical pathway influence motion processing within the lateral suprasylvian (LS) area of the cat.

methods. This study was conducted in five cats. Tungsten microelectrodes were used for recording visual evoked potentials. The electrodes were introduced into the LS area. An array of 120 randomly located dots was projected onto the stimulus field (40° × 40°) in front of the animal by a slide projector. The dots were moved rightward and leftward alternatively with interstimulus intervals by a mirror attached to a galvanometer, the movements of which were controlled by a microcomputer. Each motion sequence consisted of an abrupt onset of motion that continued for 100 msec followed by an abrupt offset and a stationary phase of 900 msec; the total duration of each sequence was thus 1000 msec. The velocity of the motion was varied in 12 steps. The onset of motion was used as the trigger for recording evoked potentials. Single or multiple injections (two to three) of muscimol were made, mainly into the rostral superior colliculus (SC). The amplitudes of evoked potentials before and after the muscimol injection were compared.

results. A large negative wave (N1) with the peak latency of 89.80 ± 16.39 msec (mean ± SD, n = 191) was recorded consistently. The amplitude of N1 was not altered by the muscimol injection into the SC when the velocity of motion was 50 deg/sec or less. When the velocity of motion was 75 deg/sec or more, however, the amplitude of N1 was reduced to 62% to72% of that noted before the muscimol injection.

conclusions. These findings suggest that the LS area processes the visual motion inputs reaching through the two parallel pathways, the geniculostriate pathway and the tectothalamocortical pathway, when the velocity of visual motion is 75 deg/sec or more.

Recent neurophysiological studies in the monkey and human suggest the existence of two general information-processing streams in the primate visual cortex. ^{1 2 3 4 5 6 7 8} The first stream is considered to subserve form and color vision and to lie ventrally and terminate in the temporal lobe (temporal stream). The other stream is considered to be specialized for visual motion and to lie dorsally and terminate in the parietal cortex (parietal stream). These streams in the visual cortex are thought to receive visual information from two subcortical pathways, the parvocellular (P) and the magnocellular (M) pathways, which originate in the retina. ⁹ The M pathway projects through the striate cortex (V1) to subdivisions of area V2 and the middle temporal (MT) area in the superior temporal sulcus of the monkey. ^{1 2 3 4 9} Area MT is assumed to be an important neural substrate for visual motion perception in the cerebral cortex. 

In the cat, the lateral suprasylvian (LS) visual area is the region suggested to be functionally analogous to the MT area of the monkey. ^{1 10 11 12} Neurons in the LS area respond to visual motion stimuli in the preferred direction. ^{11 13 14 15 16} The LS area receives visual inputs from the geniculostriate pathway and from the extrageniculate system through the tectothalamocortical pathway. ^{17 18 19 20 21 22 23 24 25 26} The tectothalamocortical pathway consists of projections from the superior colliculus (SC) to the LS through the pulvinar and the lateral posterior nucleus of the thalamus. The functional roles of the geniculate system in visual motion perception have been investigated previously. ^{27 28 29} However, the functional roles of the tectothalamocortical pathway in visual motion processing within the cortex are still controversial. ^{30 31}  

In this study, we investigated the effects of SC inhibition by muscimol injection on the velocity profiles of motion-triggered visual evoked potentials (m-VEPs) within the LS area for various stimulus velocities. In contrast to the conventional visual evoked potentials associated with the sudden reversal of the contrast of a pattern, m-VEPs associated with the onset of visual motion are considered to represent the activity of the M pathway and the temporal stream. ^{32 33 34 35 36} This technique would allow us to estimate the activity of motion processing within a relatively large area of the LS area. Therefore, we attempted to clarify whether the visual inputs of the tectothalamocortical pathway to the LS area influence motion processing within the LS area, by investigating the effects of SC inhibition on m-VEPs. 

Materials and Methods

Surgical Preparations

This study was conducted in five cats, weighing 2.5 to 3.5 kg. Each cat was deeply anesthetized with 2% to 4% halothane. After the trachea and the saphenous vein were cannulated, ketamine hydrochloride (initial dose, 25 mg/kg, intramuscular [IM]) and α-chloralose (25 mg/kg, intravenous [IV]) were substituted for halothane anesthesia. The animal was immobilized with pancuronium bromide (initial dose, 0.1 mg/kg, IV) and artificially ventilated. Pancuronium bromide (0.05 mg/kg IV) was administered every 60 minutes. The animal was attached to a stereotaxic head-holder frame. Two small holes were made in the parietal skull for later insertions of microelectrodes and glass micropipettes into the LS area and the SC. All incisions and pressure points were infiltrated with 2% lidocaine hydrochloride. Rectal temperature was maintained at 37.5°C using a feedback-controlled heating pad. During the experiment, supplemental doses of ketamine hydrochloride (15 mg/kg, IM) and α-chloralose (10 mg/kg, IV) were administered every 30 minutes. All experimental protocols were approved by the Sapporo Medical University Animal Care and Use Committee and complied with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Visual Stimuli

An array of 120 randomly located dots (1.0° in diameter, 20 candela [cd]/m², 99.8% contrast) was projected by a slide projector onto the stimulus field (40° × 40°) positioned 57 cm in front of the animal (Fig. 1) . The stimulus was centered on the receptive field in each cat. Random dot displays have several virtues for use in motion psychophysics. ³⁷ They stimulate the visual motion system, while minimizing familiar position cues. ^{38 39 40 41} The dots were moved right and left alternatively with interstimulus intervals by a mirror attached to a galvanometer, the movements of which were controlled by a microcomputer. Each motion sequence consisted of the abrupt onset of motion which lasted for 100 msec followed by an abrupt offset and a stationary phase of 900 msec; the total duration of each sequence was thus 1000 msec (Fig. 1) . The velocity of motion was varied in 12 steps (5, 10, 15, 20, 25, 35, 50, 75, 100, 125, 150, and 200 deg/sec). All recordings were performed in a sound-attenuated shield chamber under a background luminance of 1 cd/m². 

Eye positions were monitored by tapetal reflections before the experimental period. ⁴² A fiber optic system was used to introduce light into the eye. The optics then produced an image of the fundus on a tangent screen in front of the animal, and the alignment of the two eyes was corrected using prisms. The random dot pattern was then introduced into the same area of the retina in the two eyes. 

Recording and Muscimol Injection

Tungsten microelectrodes insulated with Isonel 31 (Nisshoku, Osaka, Japan) were used for recording the visual evoked potentials. The electrodes were introduced into the medial bank of the LS area at an angle of 30° to 35° from the vertical axis in a coronal plane, and were positioned at A0 to A1 of stereotaxic coordinates, which is the area corresponding to the posteromedial LS cortex. ⁴³ Single neuronal activities were initially recorded, and the receptive field and the directional selectivity of each neuron was identified. In each cat, a single neuronal activity that was selective for the horizontal stimulus of motion was found, and the electrode was positioned at this point. Visual evoked potentials were recorded from the area consisting of visual cells with horizontal directional selectivity. The onset of motion was used as the trigger for recording the evoked potentials. Evoked potentials to either rightward or leftward motion were averaged, depending on the motion selectivity of the neuron in each cat. After amplification with a band-pass filter of 0.5 to 100 Hz, 128 epochs of 1-second duration were averaged and digitized by a computer at a sampling rate of 10 kHz. 

Glass micropipettes, which were filled with 1 μg/μl saline solution of muscimol (Sigma, St. Louis, MO), were introduced stereotaxically into the SC along the vertical axis on the same side as the recording side of evoked potentials. The projection from the SC to the LS area is exclusively ipsilateral. ^{21 44 45 46} Therefore, only the ipsilateral SC was inhibited in this study. The solution was stained with fast green for later identification of the injection sites. Single or multiple injections ^{2 3} of muscimol, spaced 0.25 to 0.5 mm apart, were made into the region of the SC which corresponds to the area of the representation of the visual field compatible with the receptive field location at the recording site. The total amount of muscimol injected ranged from 0.2 to 0.5 μl. In the control, 0.2 to 0.5 μl saline was injected into the SC to indicate no effects. 

Data Analysis and Histologic Processing

We evaluated the amplitudes and peak latencies of the positive and negative peaks of the evoked potentials for each stimulus velocity before and after the muscimol injection into the SC. Evoked potentials were recorded six to eight times for each stimulus velocity in each animal. The amplitudes of the evoked potentials before and after the muscimol injection were compared by two-way repeated-measures analysis of variance. 

After the experiments, the animals were deeply anesthetized with pentobarbital sodium and perfused transcardially. Two liters physiological saline was introduced, followed by 2 l fixative solution containing 10% paraformaldehyde and 1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). After perfusion, the brains were exposed and blocked in the stereotaxic plane, placed in 0.1 M phosphate buffer containing 30% sucrose, and kept in a refrigerator overnight. The brains were then sectioned into 100-μm serial coronal sections on a freezing microtome and collected in compartmentalized trays. The sections were then mounted on gelatin-coated slides and stained with neutral red. 

Each section was examined using both low- and high-magnification lenses under bright-field illumination. The distributions of the injection sites of muscimol were plotted on sheets of paper with the aid of a drawing tube attached to a microscope. 

Results

The activity of single neurons was initially recorded, and this activity, which exhibited horizontal directional selectivity, was finally identified in each cat. The receptive fields of five neurons in the five cats corresponded to the lower quadrant of the visual field of the side contralateral to the recording side with sizes ranging from 15° to 30° in width (Fig. 2) . Evoked potentials were recorded from the point where the single neuronal activity was recorded in each cat. The patterns of evoked potentials recorded were essentially the same for the five cats and were dependent on the velocity of motion. Figure 3 shows examples of m-VEPs for various stimulus velocities in one cat. A large negative wave designated as N1 was recorded consistently. A positive wave after N1 was also observed, but was not recorded consistently in every trial. The mean peak latency of N1 was 89.80 ± 16.39 msec (mean ± SD, n = 191). Figure 4 show the relationship between the velocity of motion and the mean amplitude of N1 in the five cats for each stimulus velocity. The amplitude of N1 changed with stimulus velocity. It increased as the velocity increased up to 100 deg/sec, but for velocities more than 100 deg/sec the amplitude of N1 was fairly constant. 

Muscimol (1 μg/μl saline solution) was injected stereotaxically into the SC. Figure 5 indicates the distributions of the injection sites in the five cats. The injection sites included the SC in all cats. The effect of the muscimol injection on the amplitude of N1 was essentially the same for all cats. Figure 3B shows examples of evoked potentials after the muscimol injection. Waveforms were similar before and after the injection. Figure 4 shows the mean amplitude of N1 for each stimulus velocity before and after the muscimol injection. The amplitude decreased significantly after the muscimol injection (62%–72% of that before the injection) at velocities of 75 deg/sec or more. However, N1 was not completely abolished by the muscimol injection, and its amplitude remained unchanged after the muscimol injection when the velocity of motion was 50 deg/sec or less. The mean peak latency of N1 after the muscimol injection was 89.15 ± 23.26 msec (mean ± SD, n = 183). It was not significantly different from the value before the injection. 

Discussion

The present study indicated that the amplitude of N1 in the LS area is independent from visual inputs through the tectothalamocortical pathway for velocities 50 deg/sec or less. However, visual inputs through the tectothalamic pathway to the LS area contribute partially to the N1 amplitude at velocities of 75 deg/sec or more. 

Previous studies have indicated the effects of SC removal on the visual responses of neurons in the LS area of the cat and the MT area of the monkey. ^{30 31} According to these results, the tectothalamic inputs to the LS area are primarily inhibitory in nature and are not necessary for most of the properties of LS neurons. These inputs include inhibition of responses to slow stimulus movement, symmetrical internal inhibition, and surround spatial inhibition, but do not include inhibition of responses to fast stimulus movement. ³⁰ Previous studies have also indicated that neurons in the LS area and the MT area retain visual responsiveness in the absence of inputs from the striate cortex and have suggested that the residual responses found in the LS area and the MT area after striate cortex removal derive inputs from the tectothalamic pathway. ^{27 28} These findings suggest that the tectothalamic inputs to the LS area are facilitatory at least for fast stimulus movement. The results of the present study indicate that tectothalamic inputs to the LS area are facilitatory when the target velocity was 75 deg/sec or more, although the facilitatory inputs were relatively weak. N1 amplitude was reduced to 70% of the control value after SC inhibition. In contrast, SC inhibition had no effect on N1 amplitude when the target velocity was 50 deg/sec or less. In this study, disinhibition of the tectothalamic inputs to the LS area was not demonstrated. Therefore, the LS area receives converging inputs from both the geniculostriate and the tectothalamic pathways, and visual inputs through the tectothalamic pathway to the LS area may be involved in motion processing for fast stimulus movement, although the tectothalamic inputs do not relate to motion processing for slow stimulus movement. 

Visual cells in the superficial layers of the SC respond to motion stimuli. ^{47 48 49} The directional selectivity of collicular cells is dependent on the velocity of motion. ⁴⁹ There is an optimal velocity at which the directional selectivity is most marked, above and below which both the discharge rate and directional selectivity decrease. With few exceptions, the optimal velocity has been found to be more than 50 deg/sec. ⁴⁹ The superficial layers of the SC project to the LS area through the pulvinar and the lateral posterior nucleus of the thalamus. ^{17 18 19 20 21 22 23 24 25 26 50 51} It is probable that the visual inputs of the tectothalamic pathway to the LS area convey mainly motion signals of high velocities more than 50 deg/sec. Based on these findings we conclude that the LS area processes visual motion inputs from the two parallel pathways, the geniculostriate pathway and the tectothalamocortical pathway, but primarily at velocities of 75 deg/sec or more. 

 Supported by a Grant-in-Aid for Scientific Research 04771362 from the Japanese Ministry of Education, Science, and Culture.

 Submitted for publication August 4, 1999; revised October 21, 1999; accepted November 8, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

System for recording m-VEPs (A). An array of 120 randomly located dots was projected by a slide projector onto the stimulus field (40° × 40°) positioned 57 cm in front of the animal. The electrodes were introduced into the LS area at an angle of 30° to 35° from the vertical axis in the coronal plane, and positioned at A0 to A1 of stereotaxic coordinates. Saline solution of muscimol was injected stereotaxically into the SC along the vertical axis on the same side as the recording side of evoked potentials. Visual stimulus paradigm (B). Each motion sequence consisted of abrupt onset of motion, lasting for 100 msec followed by abrupt offset and a stationary phase of 900 msec.

Figure 2.

 

Receptive fields of the five neurons at the recording sites in five cats. These corresponded to the lower quadrant visual field of the side contralateral to the recording side, with sizes ranging from 15° to 30° in width.

Figure 3.

 

Examples of motion-triggered visual evoked potentials for various stimulus velocities in one cat (cat 203) before (A) and after muscimol injection (B). A large negative wave was recorded consistently and was designated as N1 (arrow).

Figure 4.

 

Relationships between the velocity of motion and the mean amplitude of N1 of five cats before (○) and after muscimol injection (•) for each stimulus velocity. Error bars, SD.

Figure 5.

 

Drawings of serial coronal sections through the rostral SC in the stereotaxic plane showing the distributions of the muscimol injection sites in five cats. The solid black region represents the extent of spread of fast green stain.