INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Binocular visual responsiveness of neurons in visual cortex (VC) of cats, monkeys and other mammals including humans can be changed by monocular visual deprivation in the critical period of postnatal development, one of the best-known examples of experience-dependent modification of brain function (Awaya et al. 1973; Hubel et al. 1977; Wiesel and Hubel 1963). It is hypothesized that afferents from each eye compete with one another for synaptic connections with cortical neurons so that less active afferents from the deprived eye fail to maintain the connections, i.e., the balance between activities of both afferents is critical to determine the fate of each eye-specific pathway (Katz and Shatz 1996; Sherman and Spear 1982; Singer 1995; Sur and Learney 2001; Wiesel 1982). This synaptic competition hypothesis has been supported by various types of deprivation experiments in kittens (Blakemore and Van Sluyters 1974; Chapman et al. 1986; Daw 1995). If such a competition hypothesis is valid, then an increase in inputs from one eye instead of decrease due to deprivation should also change binocular responsiveness of cortical neurons. Otherwise the deprivation-induced change might be due to "disuse atrophy" of the pathway. It is rather surprising that the hypothesis has not successfully been tested with such a monocular activation paradigm except for a recent paper reporting an effect of pharmacological activation of retinal activity on the development of retinogeniculate pathways (Stellwagen and Shatz 2002). In the present study, we activated one of the optic nerves (ONs) by electrical stimulation in behaving kittens for 2 days and found that such a chronic monocular activation induced a shift in ocular dominance of visual responses of cortical neurons.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All of the kittens (n = 12) used in the present study were born in the breeding colony of Osaka University Graduate School of Medicine, and experimental procedures met the regulation of the Animal Care Committee of the School. For chronic activation of afferents from the eyes, both ONs were exposed by a retrobulbar approach (Tsumoto 1978) under anesthesia with 3-4% isoflurane in 50% N₂O-50% O₂. Two silver-wires coated with Teflon except for sharpened tips with a separation of 1.0-1.5 mm were glued side by side. The tips were plated with chlorine to minimize polarization of the electrode in the tissue. This bipolar electrode was placed on the ONs, and its bulging part near tips were tightly glued to the sheath of the ONs with a surgical adhesive agent (-cyanoacryllate monomer, Sankyo). Thus the ONs were stimulated in a constant condition during the experiments. Also we implanted bipolar electrodes into the dorsal lateral geniculate nucleus (LGN) to record field potentials evoked by ON stimulation and strobe flashes. An antimicrobial agent (enrofloxacin, 5 mg/kg) was injected subcutaneously before surgery and then once every day until the day when single-unit recordings were begun.

Single-unit recordings were carried out with elgyloy microelectrodes from the primary visual cortex and LGN under anesthesia with intravenous administration of pentobarbital sodium (2-4 mg · kg¹ · h¹) with inhalation of 50% N₂O-50% O₂. To prevent eye movements the animals were paralyzed by intravenous infusion of gallamine triethiodide (10 mg · kg¹ · h¹) in 5% dextrose Ringer solution and maintained under artificial respiration. The respiratory rate and volume were adjusted to result in an end-tidal expired CO₂ concentration of approximately 4%. Body temperature was maintained at 38°C. Level of anesthesia was monitored by observing whether there was an increase in heart rate following a paw pinch.

Properties of receptive fields of cortical and LGN neurons were explored initially with light slits or spots projected onto the tangent screen by means of a hand-held projector. For more quantitative analyses, peristimulus time histograms (PSTHs) of unitary spikes to moving slits or ON-OFF of stationary spots were constructed with a computer. These visual stimuli were presented on a cathode ray tube located 30 cm in front of the animal. In most of the cortical neurons, their eye dominance was determined by observing responses elicited through each eye by manual stimulation.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

After having implanted electrodes for electrical stimulation and field potential recordings at both ONs and in the left LGN, the animals were allowed to recover completely from the surgery without any electrical stimulation (see Fig. 1A, horizontal bar). Experimental schedules for each of the animals are summarized in Table 1. After the recovery period, a battery of tests was carried out under isoflurane anesthesia to check the possibility for the electrodes to have impaired the ONs. Initially, mass fiber responses evoked by strobe flash stimuli to the eyes were recorded from each ON (Fig. 1B, left). Also field potentials evoked by flash stimuli to each eye were recorded from the LGN (Fig. 1C, left). In three kittens, the responses to flashes were not clearly recognizable in these structures. Therefore they were omitted from the further analysis. Then, field responses to electrical stimulation of each ON given at every 2 or 3 s were recorded from the LGN to confirm that ON stimulation could effectively activate LGN neurons.

View larger version (30K): [in this window] [in a new window]   Fig. 1. Schematic depiction showing experimental procedures and field potentials indicating unimpaired function of the optic nerves (ONs). A, top: sites for stimulation and recordings. Extracellular single-unit recordings were performed from the right lateral geniculate nucleus (LGN) and visual cortex (VC) after chronic stimulation of the left or right ON. Bottom: bar shows a time table for experimental procedures. B: field potentials of ONs evoked by strobe flashes to each eye. Top and bottom: potentials evoked by flash stimulation given to the eyes at the chronically stimulated and nonstimulated side, respectively. Left and right: indicate potentials recorded before and after the chronic stimulation, respectively. Arrow heads indicate the time of flash stimulation. Each trace is an average of 10 sweeps. Time and amplitude calibration at the right apply to records in the same row. Negativity, upward. C: field potentials of LGN evoked by strobe flashes to each eye. Time and amplitude calibration at the bottom right apply to all the records. Other conventions are the same as in B.

After having recorded these responses, one of the ONs was stimulated with high-frequency, biphasic current pulses (4 pulses at 100 Hz) given at 5 Hz for 2 s. Biphasic pulses were used to minimize a possible polarization of the electrodes. This stimulation was repeated at an interval of 10 s for 48 h. Such chronic stimulation was done without anesthesia and motor paralysis, which were suggested to disturb ocular dominance plasticity of cortical neurons (Freeman and Bonds 1979). During this monocular activation period, the kittens behaved quite normally without any sign of discomfort: they slept from time to time, walked in a breeding cage, ate foods, and drank water as normal kittens do. Immediately after cessation of the chronic stimulation, the same set of tests as that before stimulation was carried out to confirm that the ONs were not impaired (Fig. 1, B and C, right). In two kittens, these field potentials became almost undetectable. So the data from these animals were omitted from the analysis.

To confirm further that both ONs were functionally unimpaired after the chronic stimulation for 48 h, visual responses of single neurons were recorded extracellularly from layers A and A1 (layers receiving inputs from the contra- and ipsilateral eyes, respectively) of the right LGN. Neurons recorded from each layer had clear on or off responses to a small light spot projected onto their receptive fields (Fig. 2A). Altogether 43 (23 ON and 20 OFF) and 52 (29 ON and 23 OFF) cells were recorded from the layers that received inputs from the activated and nonactivated ON, respectively. As exemplified in Fig. 2A, they responded quite normally to spot lights projected onto their receptive fields, indicating that visual inputs from both eyes arrived at the LGN without noticeable disturbance through the ONs. Furthermore, there was no recognizable bias in the response magnitude of LGN neurons toward the eye of the activated side.

View larger version (26K): [in this window] [in a new window]   Fig. 2. Examples of visual responses of LGN and VC neurons. A: example of peristimulus time histograms (PSTHs) of on (top) and off responses (bottom) of LGN cells of the on and off types, respectively. Left and right: responses to visual stimulation given to the eyes of the nonactivated and activated side, respectively. Number of sweeps, 10. The time when the stimulus light is turned on is shown at the bottom. Bin width, 16 ms. B: example of responses of a cortical neuron to motion of the slit. Orientations of the slit and directions of the motion are indicated at top of each PSTH. The velocity of slit motion was 5.8°/s. Number of sweeps, 5. Bin width, 32 ms.

Then we observed responses of single cortical neurons to various visual stimuli given to each eye and rated their ocular dominance, according to the categorization scheme of ocular dominance into five groups (Noda et al. 1971) modified from the original seven group categorization by Hubel and Wiesel (1962). In four kittens, this test was carried out in the double-blind condition, i.e., experimenters who rated the ocular dominance of neurons did not know which side of the ONs had been activated by the chronic stimulation, and those who stimulated one of the ONs did not know which side they were stimulating. For this double-blind test, the wire leads from the optic nerve electrodes were hidden in a socket fixed on the animal's skull. In the other three kittens, the tests were done in the nonblind condition. In both groups of kittens, we found that most of the cortical neurons responded preferentially to visual stimuli applied to the eye of the activated side. An example of responses of a cortical neuron to moving bar stimulus is shown in Fig. 2B. This neuron responded vigorously to motion of the slit oriented as shown in the second row of the histograms through the eye of the activated side (Fig. 2B, right). The same set of stimuli to the other eye elicited much weaker responses even if the orientation of the slit was optimal (left). Thus the ocular dominance of this neuron was classified as group 4. This neuron did not well respond to light slits at the nonoptimal orientation even through the activated eye, indicating that it had selectivity for orientation of stimuli. The orientation selectivity of most of the cortical neurons recorded in the present study seemed to be almost normal, although we did not quantitatively test the orientation tuning of all the cortical neurons. The ocular dominance histogram of all the cortical neurons recorded from this kitten (kitten 7) is shown in the bottom left of Fig. 3. It is seen that most of the neurons had a bias toward the activated side.

View larger version (30K): [in this window] [in a new window]   Fig. 3. Ocular dominance histograms of cortical neurons of 7 kittens after chronic stimulation of ON.  and , data obtained in the double-blind and nonblind conditions, respectively. , the summed histogram. , the side of activated ON. Ocular dominances 1 and 5 represent cells driven exclusively by the contra- and ipsilateral eye, respectively. Ocular dominance 3 represents cells driven nearly equally from both eyes. Ocular dominances 2 and 4 represent binocular cells dominated by the contra- and ipsilateral eye, respectively. Category U represents cells with undetectable visual responses. In the summed histogram at the bottom right, the columns indicated by act and nonact represent cells that responded exclusively to the eye at the chronically stimulated and nonstimulated side, respectively. The 2nd to 4th columns represent binocular cells that responded dominantly to the activated side, equally to both sides and dominantly to the nonactivated side, respectively.

Ocular dominance histograms obtained from all the kittens are shown in Fig. 3. All the histograms show the strong bias toward the activated side indicated by arrows. To examine the ocular dominance shift quantitatively, the bias index for the activated side (activated-eye bias index, ABI), modified from the contralateral bias index (Reiter and Stryker 1988), was calculated for each kitten (see Table 1, 2nd column from the right). This index corresponds to the contralateral bias index when recordings are made from the contralateral cortex to the activated eye and is expected to be slightly larger than 50, reflecting a weak bias toward the contralateral eye in the kittens at the age of recording session in the present experiments (P30-56) (Crair et al. 1998). In the present study, single-unit recordings were made from the ipsilateral cortex to the activated eye in five kittens and from the contralateral cortex in two kittens. As mentioned in the preceding text, the ABI should be slightly smaller and larger than 50 in the former and latter groups of kittens, respectively, if there is no effect of activation. However, the ABI was clearly larger than 50 even in the former group of kittens (73.0 ± 10.1; mean ± SD).

The index obtained in the double-blind test ranged from 60 to 73 (66.5 ± 5.3) and was not significantly different from that in the nonblind test (74.3 ± 8.1, unpaired t-test, P > 0.05). Therefore the two groups are combined into the summed histogram of ocular dominance (Fig. 3, bottom right). This summed histogram showed the change of ocular dominance toward the activated side. A statistical analysis between the number of neurons which responded exclusively or dominantly to the eye of the activated side and that of neurons that responded exclusively or dominantly to the other eye indicated that the bias toward the activated side was highly significant (paired t-test, P < 0.0001). The same analysis was made in the five kittens in which recordings were done from the ipsilateral cortex to the activated eye to exclude the factor of natural contralateral bias. Again the bias toward the activated side was significant (paired t-test, P < 0.005).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study demonstrates that the chronic activation of afferents from one eye can change binocular visual responsiveness of cortical neurons. Previously it was reported that changes in ocular dominance of cortical neurons during on-line recording of their activities were induced by monocular visual activation associated with eye-movement-inducing stimuli (Tsumoto and Freeman 1981) or with depolarization of postsynaptic neurons (Frégnac et al. 1988). However, the changes in these studies were observed in the limited number of cells and were largely not persistent, i.e., the degree of the changes decayed with time. In the present study, we have observed that the chronic, electrical stimulation of one of the ONs for 2 days could induce the significant change in ocular dominance of cortical neurons in all the kittens tested, and the change seemed to persist longer than 24 h of the recording session.

Within our knowledge, there were two previous studies that reported that chronic electrical stimulation of ONs affected postnatal development of visual responsiveness of cortical neurons. In a study in which the two ONs were asynchronously stimulated in kittens aged from 2 to 6-8 postnatal weeks, 72% of visual cortical neurons were monocular, suggesting that the asynchronous stimulation prevented normal development of binocular visual responses (Stryker and Strickland 1984). In the present study, visual cortical neurons showed stronger response to the activated eye, but most of them still retained binocular responsiveness. This may be due to the much shorter duration of activation in the present study or the later onset of activation after visual cortical neurons had acquired binocular responsiveness (Freeman and Ohzawa 1992) or may imply that the bilateral, asynchronous stimulation and the unilateral activation of ONs have different actions on binocular responsiveness of visual cortical neurons. In another study in which stimulation of the ON (30 Hz for 1.8 s) was repeated every 20 s for 23 days in young ferrets, it was reported that normal development of orientation selectivity of visual cortical neurons was disturbed (Weliky and Katz 1997). In the present study, however, we found that most of the visual cortical neurons seem to have almost normal orientation selectivity, although we did not quantitatively test all the cases. This difference also may be due to the shorter duration of the chronic stimulation in the present study or due to the relatively late onset of the stimulation. In kittens, the majority of cortical neurons are known to be already orientation-selective at the age of postnatal days 28-54 (Frégnac and Imbert 1984) when the activation was initiated in the present study, while in ferrets most neurons are not so at the age of postnatal days 27-29 (Chapman and Stryker 1993) when the effective stimulus was begun.

The present results that the ocular dominance of most cortical neurons was changed but the orientation selectivity was not after ON stimulation for 2 days seem to be consistent with the previous observations that binocularity of cortical neurons is disrupted by monocular deprivation for as short as 1 or 2 days (Olson and Freeman 1975; Trachtenberg et al. 2000), while orientation selectivity is more resistant to abnormal visual inputs (Kim and Bonhoeffer 1994; Sherk and Stryker 1976). The rapid change in binocular responsiveness of cortical neurons in the present study may reflect morphological and/or functional changes in geniculocortical afferents in the cortex. In previous studies, such a rapid change was not observed in synaptic density (Silver and Stryker 1999) or was very minor if any in axonal structures (Antonini and Stryker 1996) of LGN afferents after the brief monocular deprivation for 2-4 days. Therefore it seems possible that functional but not such morphological changes in geniculocortical synapses may underlie the rapid plasticity that we observed, although we cannot exclude any change in intracortical circuits.

In sum, it seems possible to conclude that the unilateral activation of afferents from the eyes can rapidly change binocular responsiveness without disturbing other response properties of visual cortical neurons as far as the duration of stimulation is limited within two days. These results support the synaptic competition hypothesis and further implicate that chronic electrical stimulation of nerve fibers can be used to change physiological properties of cortical neurons.