INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cells in the cat visual cortex are selective for particular stimulus attributes such as the degree to which one eye provides stronger input (ocular dominance), the orientation of the optimal stimulus (orientation selectivity) and direction of motion (direction selectivity). These properties are distributed in an organized manner across the primary visual cortex of adult animals (Crair et al. 1997; Hubel and Wiesel 1974; Hubener et al. 1997). Both ocular dominance and orientation selectivity appear to be well organized in young animals (Chapman and Stryker 1993; Crair et al. 1998, 2001; Thompson et al. 1983).

It is well known, however, that the nature of an animal's early visual experience can dramatically reorganize these properties in the visual cortex. Perhaps the best known example is the effect of closure of one eye on the ocular dominance of neurons in primary visual cortex (Wiesel 1982; Wiesel and Hubel 1963). After brief periods of deprivation, the majority of cells are excited only through the nondeprived eye. Susceptibility to this type of manipulation is greatest during a critical period early in postnatal development. Ocular dominance plasticity, as measured by the effects of a 10-day period of monocular deprivation (MD), peaks around 4-6 wk of age (Olson and Freeman 1980) and declines thereafter with residual plasticity to longer periods of deprivation evident in 9-mo-old animals (Daw et al. 1992).

Critical periods are different for different visual properties (Daw 1995, 2002). For example, the critical period for direction selectivity ends earlier than that for ocular dominance plasticity. Rearing in an environment moving in one direction only has little effect on the directional selectivity of neurons in the cat visual cortex after 7 wk of age, when ocular dominance is still quite mutable (Berman and Daw 1977; Daw and Wyatt 1976). Indeed, if an animal is reared in an environment moving to the right with the right eye open until 5 wk of age, then in an environment moving left with the left eye open for another 7 wk, the majority of cells will prefer movement right and be dominated by the left eye (Daw et al. 1978). Orientation selectivity also develops earlier than ocular dominance (Chapman and Stryker 1993; Chapman et al. 1996). Because orientation selectivity in this and most other experiments is measured with a moving stimulus, orientation and direction selectivity are expected to develop together.

We have previously shown that the activity of protein kinase A (PKA) is a necessary component of the signaling pathways that lead to ocular dominance shifts in the kitten visual cortex (Beaver et al. 2001). Inhibition of PKA blocks the effect of 5 days of MD at 4-5 wk of age in the kitten. Orientation selectivity is also affected in these animals. While cells far from the osmotic minipump infusing the inhibitor of PKA showed normal orientation selectivity, the majority of cells near the pump were nonselective.

This result led to two questions: do PKA inhibitors have a general deleterious effect on receptive field properties in all developing animals, normal as well as monocularly deprived and does the effect of MD on receptive field properties decay with age before its effect on ocular dominance decays when it is combined with PKA inhibitors? Because the critical period for orientation/direction selectivity differs from that for ocular dominance, we hypothesized that the loss of orientation selectivity observed following the combination of PKA inhibition and MD would vary with the age of the kitten. That is, in older kittens that are within the critical period for ocular dominance but not for orientation plasticity, we expected to observe the blockade of ocular dominance plasticity without the accompanying loss of orientation selectivity.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Kittens (n = 23) were implanted with osmotic mini-pumps. Animals were given preanesthetic doses of acepromazine (0.1 mg/kg), atropine (0.04 mg/kg), and dexamethasone (1 mg/kg), then anesthetized with ketamine (25 mg/kg) and xylazine (1.5 mg/kg), intubated, and mounted in a stereotaxic instrument and prepared for surgery using standard aseptic techniques. Anesthesia was maintained with halothane (0-1.2%), and each animal was monitored using a thermometer, electrocardiogram (EKG), pulse oximeter, and CO₂ monitor. A small (1.5 × 1.5 mm) craniotomy was opened over the visual cortex centered approximately 5 mm posterior to the interaural zero line and approximately 1.5 mm from the midline. A small hole was made in the dura, and the tip of a 28 G cannula (Alzet Brain Infusion Kit, Durect, CA), the tip for which had been beveled, was lowered to 1.5-2.0 mm below the surface of the cortex with the bevel facing rostrally. The cannula was attached to an osmotic mini-pump (Alzet model 2001, Durect), which was placed in a pocket that had been opened underneath the skin of the neck. Pumps were filled an hour before surgery with 8-chloroadenosine-3',5'-monophosphorothioate (Rp-8-Cl-cAMPS, 20 mM) reconstituted in sterile 0.033 M phosphate-buffered saline and then incubated at 37°C in the same buffer until implantation.

Monocular deprivation was begun the following day. Under ketamine (25 mg/kg) and xylazine (1.5 mg/kg) anesthesia, the lid margins of the eye contralateral to the pump were removed and the eyelids sutured together using 4-0 silk. A small bead of antibiotic ointment (bacitracin-neomycin-polymixin, Pharmaderm) was placed between the lids before closing. Animals were given the antibiotic enrofloxacin (2.5 mg/kg, Baytril, Bayer) and monitored daily for the development of any openings in the sutured eye as well as for any infections around the pump cannula. Animals were given the analgesic buprenorphine hydrochloride (0.005-0.01 mg/kg, Buprenex, Reckitt and Colman Pharmaceuticals) for 48 h following surgery to relieve pain.

After 5 days of MD (6 days of drug infusion), animals were prepared for single-unit extracellular physiological recordings. Animals were sedated with acepromazine (0.1 mg/kg) and given preanesthetic doses of atropine (0.04 mg/kg) and dexamethasone (1 mg/kg). Anesthesia was induced with halothane (4%) in a mixture of 67% nitrous oxide, 33% oxygen and maintained with 0.4-1.0% halothane afterward. After a tracheotomy and insertion of a cannula into the femoral vein, the pump was removed, the skull was opened up over the lateral gyrus, and a hole made in the dura for insertion of the electrode. All wound margins were treated with lidocaine (20 mg/ml). The deprived eye was opened, and both eyes covered with contact lenses of zero power, and curvature appropriate to focus the retina on a tangent screen at 57 in. Animals were paralyzed by intravenous infusion of pancuronium bromide at 0.6-1.5 mg/h (Elkins-Sinn). Heart rate and end tidal CO₂ were monitored continuously and CO₂ was maintained at 3.5-4.3% by adjusting the respirator.

Single-unit recordings were made with single barrel carbon fiber in glass microelectrodes. Penetrations were angled at roughly 16° from vertical in an anterior to posterior direction and were spaced between 0.5 and 1 mm apart, depending on the location of surface blood vessels, to give a sample of cells that crossed a number of orientation and ocular dominance columns. This yielded approximately 14-16 cells per penetration and roughly 70 cells per animal.

After isolation of a unit, its receptive field was mapped on a tangent screen with a hand-held projector to determine the preferred orientation, optimal size, and velocity of the stimulus. Using the preferred stimulus, cells were assigned to ocular dominance categories according to the seven-category scheme of Hubel and Wiesel (1962) based on the auditory discrimination of two independent listeners. Cells in category 1 are driven exclusively by the contralateral eye, category 7 cells are driven exclusively by the ipsilateral eye, while those in category 4 are equally driven by both eyes. Five to six penetrations were made, always in the hemisphere ipsilateral to the nondeprived eye of each animal because of the normal contralateral bias.

Cells were also assigned to three categories according to their selectivity for direction of movement: Orientation selective: cells responding to a direction or axis of movement with little or no response to movement in the perpendicular direction. Orientation biased: cells responding to all directions of movement with a response to one direction or axis of movement that was noticeably stronger than the response to the perpendicular axis. And nonselective: cells responding to all directions with no noticeable preference. The specificity of cells in a particular animal for orientation was calculated by expressing the number of orientation selective cells as a percentage of the total number of cells characterized.

In 23 cells, activity records were collected using computer-generated stimuli that were used to stimulate the cell with the preferred parameters. The stimulus rested at first for 1 s outside of the receptive field, then swept across the receptive field, resting for 1 s on the far side, and swept back, resting 1 s more. Sweep speed was adjusted to be close to the preferred velocity for the cell being recorded. This procedure was repeated three to five times, depending on the velocity of the stimulus, for one group of records every minute. The computer was also used to store spike discharge times and to construct a peristimulus-time histogram on-line as spikes came in. Records of the level of spontaneous activity were taken by turning off the stimuli and recording the cell's activity to a blank screen. Orientation tuning curves for individual cells were collected using computer generated bar stimuli that passed back and forth across the receptive field along 0°/180°, 30°/210°, 60°/240°, 90°/270°, 120°/300°, 150°/330° axes of movement. Data was subsequently analyzed off-line using custom-written ASYST programs (Asyst Software, Rochester, NY).

Data analysis

Ocular dominance histograms were constructed and weighted ocular dominance (WOD) scores calculated as follows
where G_i is the number of cells in ocular dominance group i and N is the total number of cells. According to this scheme, a score of 1 would mean that all cells responded only to the ipsilateral eye, whereas a score of 0 would mean that all cells responded only to the contralateral eye. Ocular dominance histograms of normal animals have an average WOD of 0.43, that is, slightly dominated by the contralateral eye. The binocularity index (BI) was calculated from the formula
Orientation selectivity scores were calculated as follows: (response to best axis  response to perpendicular axis)/(response to best axis + response to perpendicular axis). Best axis was defined as the axis of movement that had the highest spike rate when the responses to both directions of movement were summed. The spike rate included both the visual response and the spontaneous activity. Spontaneous activity is shown as a dashed circle for comparison. According to this formula, cells that are very selective for a direction or axis of movement will have a score that is close to 1, while nonselective cells will have a score that is close to 0.

Values for these parameters are given as means ± SD, and significance for the various comparisons was tested with Student's t-test, considering the number calculated from each animal as one observation.

The procedures used in this study were approved by the Animal Care and Use Committee at Yale University and conform to guidelines of National Institutes of Health, the Society for Neuroscience, and the American Physiological Society.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The PKA inhibitor, Rp-8-Cl-cAMPS, affects both ocular dominance and orientation selectivity of cells in the visual cortex of monocularly deprived cats. As we previously reported, Rp-8-Cl-cAMPS blocks the effects of a 5-day period of MD at 4-5 wk of age. The shift in ocular dominance was observed far (>5 mm) from the pump where the inhibitor was not effective but not near the pump (<4 mm) where the inhibitor was present (Fig. 1). Cells near the pump tended to be more binocular than normal and also showed a loss of orientation selectivity. A small percentage of selective cells (17.3 ± 13.8%) was recorded near the pump compared with a large percentage (82.7 ± 8.8%) further from the pump (Fig. 1), the difference being significant (t-test, P < 0.001). Nonselective cells were found in all ocular dominance categories and were not located preferentially within a particular layer. The percentage of selective cells was reduced significantly in all layers, comparing the percentage found near the pump with that found far from the pump (Fig. 2), the difference being less significant for the comparison for cells in layers 2-3 (P < 0.05) than for the comparison for cells in layer 4 and layers 5/6 (P < 0.001 in both cases). As reported previously, spontaneous activity was reduced by about 1/3, but the visual response and signal to noise ratio were not significantly affected (Beaver et al. 2001) (Fig. 2).

View larger version (24K): [in this window] [in a new window]   Fig. 1. Monocular deprivation (MD) affects orientation selectivity at 4 wk of age in cats treated with protein kinase A (PKA) inhibitors. Ocular dominance histograms constructed from 4 kittens whose left visual cortices were infused with 8-chloroadenosine-3',5'-monophosphorothioate (Rp-8-Cl-cAMPS, 20 mM) during a 5-day period of MD. Right: histogram constructed from cells that were located far (>5 mm) from the tip of the pump cannula. Left: histogram constructed from cells near (2-4 mm) the pump cannula. Cells in category 1 are driven exclusively by the contralateral (closed) eye, category 7 cells are driven exclusively by the ipsilateral (open) eye, while those in category 4 are equally driven by both eyes. WOD, weighted ocular dominance score; BI, binocularity index; %sel, percentage of selective cells (see METHODS), uc, uncharacterized.

View larger version (22K): [in this window] [in a new window]   Fig. 2. The PKA inhibitor reduces the percentage of selective cells in all layers. Histograms from 4 animals show means ± SE, with each animal regarded as one observation.

Does the PKA inhibitor also affect orientation selectivity in normal animals? To test this, we recorded three animals that were implanted with osmotic minipumps infusing the PKA inhibitor but which did not receive a period of MD. The percentage of cells that were selective for direction of movement was high and not affected substantially by the PKA inhibitor (Fig. 3A): 86.2 ± 5.6% for cells near the pump (left) and 91.5 ± 5.2% for cells far from the pump (right). The percentage of selective cells near the pump was significantly different from that seen near the pump in animals treated with the PKA inhibitor and also monocularly deprived (P < 0.001) but not significantly different from that seen far from the pump in those animals (P = 0.58). The ocular dominance histograms of cells recorded near and far from the pump in normal animals treated with the PKA inhibitor are also very similar: WOD scores were 0.353 ± 0.051 and 0.401 ± 0.087 and BIs were 0.612 ± 0.039 and 0.707 ± 0.135 for the near and far histograms, respectively. For comparison, Fig. 3B shows a histogram of cells from two animals that were implanted with osmotic minipumps infusing only the vehicle (0.033 M phosphate buffer) and monocularly deprived combined with two that were not infused and monocularly deprived. In agreement with previous results the histogram shows a strong shift toward the nondeprived eye, but normal orientation selectivity was observed with a high percentage of selective cells (84.3 ± 10.1%). Thus neither the presence of the inhibitor nor a period of MD, by themselves, was sufficient to reduce orientation selectivity: a combination of PKA inhibition and MD was required.

View larger version (22K): [in this window] [in a new window]   Fig. 3. Inhibition of PKA, by itself, has no effect on orientation selectivity. A: ocular dominance histograms constructed from 3 kittens whose left visual cortices were infused with Rp-8-Cl-cAMPS (20 mM) for 6 days without any MD. B: for comparison, the ocular dominance histogram of cells recorded near the pump in animals (n = 4) whose left visual cortices were infused with the vehicle solution (0.033 M phosphate buffer) or not infused at all, during a 5-day period of MD, are shown on the right. Conventions as in Fig. 1.

The reduction in orientation selectivity was not found with the protein kinase G (PKG) inhibitor -phenyl-1,N²-etheno-8-bromoguanosine-3',5'-cyclic monophosphorothioate (Rp-8-Br-PET-cGMPS). Two 4-wk kittens were infused with this inhibitor during a 5-day period of monocular deprivation (Fig. 4, top), and the percentage of selective cells near the pump (80.5 ± 17.2%) was close to that for cells far from the pump (82.7 ± 1.8%), there being no significant difference (P = 0.87). As a second control, we compared the effect of the drug 8chloroadenosine-3',5'-monophosphorothioate (Sp-8-Cl-cAMPS), which is an activator of PKA and also of cyclic nucleotide-gated channels, with the effect of Rp-8-Cl-cAMPS, which is also an activator of cyclic nucleotide-gated channels at high concentrations. Two kittens were infused with Sp-8-Cl-cAMPS during a 5-day period of monocular deprivation (Fig. 4, bottom), and in this case also the percentage of selective cells close to the pump (77.5 ± 1.3%) was close to that for cells far from the pump (77.5 ± 20.5%), there being no significant difference (P = 1). Thus the reduction in orientation selectivity is specific to inhibition of PKA in combination with MD.

View larger version (29K): [in this window] [in a new window]   Fig. 4. Neither inhibition of protein kinase G (PKG) nor activation of PKA affects orientation selectivity in MD animals. Top: results from 2 MD animals infused with 1 mM -phenyl-1,N2-etheno-8-bromoguanosine-3',5' cyclic monophosphorothioate (Rp-8-Br-PET-cGMPS). Bottom: results from 2 MD animals infused with 5 mM and 10 mM 8-chloroadenosine-3',5'-monophosphorothioate (Sp-8-Cl-cAMPS). Conventions as in Fig. 1.

Next we investigated the effect of age on the loss of orientation selectivity from MD in animals infused with PKA inhibitors. At 4 wk of age, inhibition of PKA interferes with ocular dominance plasticity from MD and normal orientation selectivity is not maintained. At older ages, because the critical period for plasticity of orientation/direction selectivity ends earlier than the critical period for ocular dominance plasticity, PKA inhibition should block the ocular dominance shifts following 5 days of MD but the orientation selectivity should be affected to a progressively smaller degree. To test this hypothesis, we examined the effect of 5 days of MD combined with Rp-8-Cl-cAMPS in two 6-wk and three 9-wk-old animals and compared it to the results from the 4-wk-old animals. Figure 5A (top) shows the results from the 6-wk-old animals. Similar to the animals at 4 wk of age the population of cells recorded far from the pump shows a characteristic ocular dominance shift toward the open eye (WOD = 0.90 ± 0.01; BI = 0.19 ± 0.02). Near to the pump infusing the inhibitor, however, there are more binocular cells (WOD = 0.59 ± 0.08; BI = 0.58 ± 0.06), the difference being significant for both indices (P < 0.05). The percentage of orientation-selective cells near the pump was larger in the 6-wk-old (52 ± 16.3%) than in the 4-wk-old animals (17.4 ± 13.8%), but it did not reach the percentage seen far from the pump (80.7 ± 3.7). A further change was observed in the 9-wk-old animals (Fig. 5A, middle). In this case, the WOD and BI near the pump were 0.70 ± 0.09 and 0.55 ± 0.13, respectively, and the percentage of selective cells was increased to 68 ± 18%. In summary, the WOD for cells far from the pump was consistently near 0.9 and for cells near the pump was consistently considerably less than this (Fig. 5B, left). The percentage of selective cells for cells far from the pump was consistently near 80%, but for cells near the pump, it increased from 17.4% at 4 wk of age to 68% at 9 wk of age (Fig. 5B, right). At these older ages, Rp-8-Cl-cAMPS still reduces the ocular dominance shifts, but the MD has little effect on orientation selectivity in the presence of the inhibitor.

View larger version (25K): [in this window] [in a new window]   Fig. 5. The effect of PKA inhibition on orientation selectivity is reduced in 6- and 9-wk-old animals. A: ocular dominance histograms constructed from 2 6-wk-old (top) and 3 9-wk-old (bottom) kittens whose left visual cortices were infused with Rp-8-Cl-cAMPS (20 mM) during a 5-day period of MD. Conventions as in Fig. 1. B: summary graphs showing the WOD scores (left) and percent nonselective cells (right) for cells near () and far () from the pump as a function of the age at which the experiment was performed. Histograms give means ± SE.

The point that loss of orientation selectivity was seen with a combination of MD and PKA inhibition, but not with MD alone or with PKA inhibition alone, was a surprise. It implies a complex interaction between ocular dominance and orientation selectivity. To interpret this result, it is necessary to determine whether the receptive field properties are the same in deprived and nondeprived eyes or different. Receptive field properties may be determined separately for the two eyes at the input level to the visual cortex with orientation selectivity lost for the deprived eye but not for the nondeprived eye. Alternatively, receptive field properties may be determined at the binocular level within the visual cortex by connections that are accessed by both eyes. In this case, changes in receptive field properties that are due to input from one eye can affect the receptive field properties found when stimulating the other eye. We therefore made quantitative measurements of receptive field properties for both eyes separately in two animals with PKA inhibition and MD at 4 wk of age.

Cells recorded far from the pump were generally monocular and showed high orientation selectivity as expected (Fig. 6A). The cell shown gave a robust response to a bar of light moving along the 90°/270° axis of movement with little response to movement along the perpendicular axis. Its ocular dominance score was 7 and its orientation selectivity score was 0.91. Twenty-three cells with binocular input were recorded close to the pump, and the receptive field properties were generally similar in both eyes. Some cells, an example of which is shown in Fig. 6B, had somewhat sharper tuning in the nondeprived eye (0.54) than in the deprived eye (0.21). In others, the tuning curves were broad in both eyes: for example 0.19 in the nondeprived eye compared with 0.08 in the deprived eye (Fig. 6C) and 0.04 in the nondeprived eye compared with 0.01 in the deprived eye (Fig. 6D). Orientation-selectivity scores for both the deprived and nondeprived eye inputs were always smaller than for cells recorded further from the pump.

View larger version (19K): [in this window] [in a new window]   Fig. 6. Examples of orientation tuning curves for cells in 4-wk-old animals whose left visual cortices were infused with Rp-8-Cl-cAMPS (20 mM) during a 5-day period of MD. Top: orientation tuning curves for the nondeprived eye. Bottom: orientation tuning curves for the deprived eye. Orientation selectivity scores (see METHODS) for each eye are shown above the tuning curve. A: an example of an orientation-selective neuron located far (>5 mm) from the tip of the cannula. B-D: representative examples of tuning for cells located near (<4 mm) from the tip of the cannula.

There was a tendency for the tuning curve for the nondeprived eye to have a higher orientation-selectivity score than the tuning curve for the deprived eye. When the orientation score for the deprived eye is plotted against the orientation score for the nondeprived eye, most of the points fall below the "line of similarity" (Fig. 7). There was also a tendency for cells dominated by the deprived eye (categories 1-3) to have a lower orientation score than cells dominated by the nondeprived eye (categories 5-7) (Fig. 8). Nevertheless, the overall conclusion is that orientation scores for both deprived and nondeprived eyes are considerably less than orientation scores for cells not affected by PKA inhibitors.

View larger version (20K): [in this window] [in a new window]   Fig. 7. Orientation selectivity in the deprived eye plotted as a function of the orientation selectivity of the nondeprived eye. Points along the dashed line would have the same score in both eyes.

View larger version (19K): [in this window] [in a new window]   Fig. 8. Orientation-selectivity scores (see METHODS) for the nondeprived (left) and deprived (right) eye as a function of the ocular dominance score.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results from the present experiments can be summarized as follows: inhibition of PKA blocks or reduces ocular dominance shifts at 4, 6, and 9 wk of age; MD in combination with inhibition of PKA at 4 wk of age reduces orientation selectivity when either one alone has no effect; the sensitivity of the effect on orientation selectivity declines with the critical period for orientation/direction selectivity; and the reduction of orientation selectivity affects both the deprived eye and nondeprived eye inputs.

Previous experiments that examined the effects of N-methyl-D-aspartate (NMDA) receptor antagonists on ocular dominance plasticity have also shown that MD produces changes in orientation selectivity that vary with age. Bear et al. (1990), Rauschecker et al. (1990), and Daw et al. (1999) all reported that antagonism of NMDA receptors blocks ocular dominance shifts in the visual cortex of kittens. Bear et al. (1990) and Rauschecker et al. (1990) also reported that the orientation selectivity of neurons was altered but Daw et al. (1999) reported that it was not. This can be explained by the difference in the ages of the kittens used: 3 wk of age for Rauschecker et al. (1990), 3-5 wk of age for Bear et al. (1990), and 6 wk of age for Daw et al. (1999). Similarly Ramoa et al. (2001) found that ferrets treated with MD and antisense oligonucleotides for NMDA receptors starting at P21 had disrupted orientation selectivity, while Roberts et al. (1998) found no loss of orientation selectivity with the same treatment starting at P50, when orientation selectivity has matured, but ocular dominance is still mutable (Chapman and Stryker 1993; Chapman et al. 1996). Thus antagonism of NMDA receptors acts like antagonism of PKA in that they both reveal the effect of MD on orientation selectivity early in development, but later in the critical period ocular dominance is affected and orientation selectivity is not.

The finding that orientation selectivity is disrupted early in the critical period in the nondeprived eye as well as in the deprived eye is surprising. Neither Bear et al. (1990) nor Daw et al. (1999) compared the receptive field properties in deprived and nondeprived eyes in their experiments with MD and NMDA receptor blockers. Rauschecker et al. (1990) did and found reduced orientation selectivity in both eyes. We made careful quantitative measurements of orientation selectivity in animals treated with both PKA antagonists and MD and found that the difference between the two eyes was small compared with the difference between these animals and those treated with PKA antagonists alone, MD alone, or normal animals. The implication of this finding is that orientation/direction selectivity is organized within the visual cortex at a level with binocular input, not separately within left and right eye columns, and can be modified by deprived eye input as well as nondeprived eye input. This is consistent with the result that the pattern of orientation columns that develops with only one eye open remains the same when the nondeprived eye is closed and the previously deprived eye is opened (Godecke and Bonhoeffer 1996; Kim and Bonhoeffer 1994).

Treatment with MD and neurotrophins for the trkB receptors also affects both ocular dominance shifts and orientation selectivity, but the results are complicated. NT4/5 infused into the visual cortex at 4 wk of age in MD kittens abolishes the ocular dominance shift, and orientation selectivity is also abolished in both deprived and nondeprived eyes, but the effect of the neurotrophin on orientation selectivity in normal animals was not reported (Gillespie et al. 2000). BDNF infused at 4-5 wk in MD kittens produced a reverse ocular dominance shift (dominance by the deprived eye) and orientation selectivity was abolished in both deprived and nondeprived eyes (Galuske et al. 1996, 2000). BDNF infused at 4-5 wk in normal kittens produced an ocular dominance shift toward the ipsilateral eye and also abolished orientation selectivity (Galuske et al. 2000). BDNF produces a general sprouting of geniculocortical terminals in layer IV of visual cortex (Hata et al. 2000), but there must be effects on intracortical connections as well to explain these results. In any case, the difference in results between trkB neurotrophins and PKA inhibitors suggests that different mechanisms are involved.

Our results show that changes in orientation selectivity occur in the presence of the PKA inhibitor. This implies that PKA does not play a role in plasticity of orientation selectivity. This would lead to the conclusion that plasticity of ocular dominance and plasticity of orientation selectivity either occur through different mechanisms at the same synapses or that they occur at different synaptic networks in the visual cortex. One possibility is that plasticity of ocular dominance reflects changes primarily at excitatory synapses, while orientation and direction selectivity may involve changes in patterns of inhibitory connections as well.

In summary, the major result of these experiments is that when ocular dominance shifts are prevented by infusion of a PKA inhibitor in MD animals, the deprivation causes a loss of orientation selectivity which is noticeable in both eyes. This result was also seen in previous experiments with antagonists to the other major factor necessary for ocular dominance plasticity, the NMDA receptor, but the conclusion that this implies, that the deprived eye instructs the nondeprived eye, was not brought out because it is counterintuitive.