- Email: [email protected]
GABAergic modulation of axonal conduction in the developing rat retinotectal pathway
Developmental Brain Research 108 Ž1998. 299–302
Short communication
GABAergic modulation of axonal conduction in the developing rat retinotectal pathway Christopher H. Lim, Stephen M. Ho
)
DeÕelopmental Neurobiology Group, Research School of Biological Sciences, Australian National UniÕersity, Canberra, Australia Accepted 3 March 1998
Abstract The effect of g-aminobutyric acid ŽGABA. on nerve conduction was investigated in the developing rat optic nerves in vitro. Antidromic compound action potentials ŽCAP. were suppressed by GABA, with increasing attenuation in older preparations. Enhancement of CAP amplitude using bicuculline was observed in postnatal days 4–5 suggesting the presence of endogenous GABA activity. Our findings suggest that the role of electrical activity in the refinement of the retinotectal projection could be limited by the GABAergic action on axonal conduction. q 1998 Elsevier Science B.V. All rights reserved. Keywords: Development; Optic nerve; Retinotectal pathway; GABA; Rat
g-Aminobutyric acid ŽGABA. is the principal inhibitory neurotransmitter in the mammalian central nervous system. GABA can also impose a non-synaptic inhibitory effect by suppressing the conduction of action potentials. In the rat optic nerve, GABA A -mediated depolarization and attenuation of action potential was age dependent, with its effect largely abolished after the 3rd postnatal week w12x. Our previous studies have shown that axonal conduction could be detected by embryonic day ŽE. 17 in the rat superior colliculus ŽSC. w9x. It remains unclear whether the GABAergic effect on axonal conduction also exists in these early developmental stages. The imposition of such a conduction block between the retina and retino-recipient structures would have an influence on the transmission of electrical activity originating from spontaneously active retinal ganglion cells w3x. We have therefore extended the studies of Sakatani et al. w12x and examined the effect of GABA on axonal conduction from E18 to postnatal day ŽP. 5. Albino–Wistar rat pups and embryos were used. Embryos were obtained from timed pregnancies with the day of sperm positivity taken as E0. The day of birth was denoted as P0. Pregnant dams were anaesthetised with Nembutal Ž60 mgrkg, i.p. injection. and fetuses were
)
Corresponding [email protected]
author.
Fax:
q61-02-6249-3808;
E-mail:
0165-3806r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 3 8 0 6 Ž 9 8 . 0 0 0 5 2 - 2
delivered by caesarian section. Fetuses or neonates were then anaesthetised by hypothermia and decapitated. The brain tissue was dissected under chilled sucrose solution Žabout 108C. with the following compositions ŽmM.: sucrose, 266; KCl, 3; NaHCO 3 , 17; CaCl 2 , 2; MgCl 2 , 1; glucose, 12.2; saturated with 95% O 2 , 5% CO 2 . The brain was hemisected with the optic chiasm and contralateral optic nerve retained on one side. The preparation was then transferred to the recording chamber, which was superfused with recirculating artificial cerebrospinal fluid ŽACSF. with the following compositions ŽmM.: NaCl, 124; KCl, 5; NaHCO 3 , 26; NaH 2 PO4 , 1.25; CaCl 2 , 2; MgSO4 , 2; glucose, 11; saturated with 95% O 2 , 5% CO 2 ; bath temperature maintained at 20–228C. Antidromic compound action potentials ŽCAP. were recorded at the optic nerve head using glass a microelectrode and a conventional electrophysiological setup. The lateral SC was stimulated at supramaximal intensity Ž2–9 mA, 0.03-ms duration, square current pulse. using bipolar tungsten electrodes. A low stimulation rate Ž1r30 s. was used to avoid conduction failure. GABA and bicuculline methiodide ŽBic. were purchased from Research Biochemicals International ŽNatick, MA. or Sigma ŽSt. Louis, MO.. CAP was recorded in 25 preparations aged between E18 and P5 ŽFig. 1.. Multimodal waveform was observed in some older preparations ŽFig. 1A., probably due to different conduction velocities. A large variation in CAP amplitude Ž100–900 mV. was observed between preparations from the same age group. Bath-applied GABA suppressed
300
C.H. Lim, S.M. Ho r DeÕelopmental Brain Research 108 (1998) 299–302
Fig. 1. ŽA. Attenuation of optic nerve CAP by GABA. Each trace was averaged from 10 sweeps and arrows indicate where peak measurements were made. ŽB. Recordings obtained from a P4 preparation showing the GABAergic attenuation of CAP was abolished after co-application of bicuculline. ŽC. Histograms showing the effect of GABA, and co-application of GABA and bicuculline on the CAP amplitude averaged from postnatal preparations. Error bars represent S.E.M.. ŽD. changes in CAP amplitude vs. stimulation intensity of a P5 preparation.
the CAP. Fig. 1B shows that the GABAergic attenuation of CAP was abolished when GABA and bicuculline were co-applied. Fig. 1C summarises the effects of GABA and co-application of GABA and bicuculline on CAP amplitude in postnatal preparations. The amplitude of axonal CAP was reduced to about 60% with 100 m M of GABA, but remained at about 100% when the same concentration of bicuculline was co-applied with GABA. This suggests that the attenuation of CAP was mediated through a GABA A receptor-linked pathway.
To eliminate the possibility that the reduction of CAP amplitude was the result of an increase in the stimulating threshold, we recorded the CAP amplitude at various stimulation strengths below the supramaximal level in 2 postnatal preparations. Fig. 1D shows the response amplitude vs. stimulation intensity curves under control and bath-applied GABA conditions. With bath-applied GABA, CAP attenuation was similar in proportion at different stimulation intensities. The absence of lateral shifting in the response–stimulation intensity curve suggests that at-
C.H. Lim, S.M. Ho r DeÕelopmental Brain Research 108 (1998) 299–302
tenuation was not a result of a local increase in stimulation threshold. Fig. 2A summarizes the GABAergic attenuation on CAP during the E18 to P5 period. It is apparent that the GABA mediated blockade of CAP increases with development. We also addressed whether endogenous GABA activity can be detected during this developmental period. After bath-applied bicuculline Ž10 m M., an increase in CAP amplitude was observed in P4–5 preparations ŽFig. 2B.. The average increase was 124% Ž"9%. Žmean "SE, n s 10., which was significantly different from control Ž t-test, p - 0.5, n s 10., suggesting that endogenous GABA activity was present. For prenatal animals, no significant change was observed. The average amplitude after bicuculline was 97.5% Ž"15%. Ž n s 13.. We next tested the behaviour of CAP when the impulse traffic was delivered at a higher rate, as previous studies have shown that spontaneous discharge of developing retinal ganglion cells can exceed 20 Hz w3,10x. Twin-pulse stimulation at intervals between 100 to 33 ms were used
301
ŽFig. 2C.. Increasing the stimulation rate to 10–30 Hz in normal ACSF caused a graduated reduction of the normalized test-pulse amplitude by 10%–30%. Test-pulse amplitude was further reduced with bath-applied GABA ŽFig. 2D.. This indicates that optic nerve conduction at these developmental stages is subject to further suppression in the presence of GABA when the impulse traffic rate is high. In the postnatal group, a lower concentration of GABA was used since 1 mM of GABA often abolished the response of the test pulse. This also agrees with results in Fig. 2A showing an increase of GABA-mediated blockade of axonal conduction with age. The source and site of GABA activity in the developing retinal axons remain to be determined. Various reports have demonstrated GABA immunoreactivity in glial cells or axons in the optic nerve w8,11,12x. The presence of GABA transporters ŽGAT 1–3. mRNAs in neonatal rat optic nerve w5x suggests that GABA could be released by heteroexchange through GABA transporters. The location of GABA A receptors is also a contentious issue. GABA A
Fig. 2. Effects of GABA ŽA. and bicuculline ŽB. on CAP amplitude during development. Each symbol represents the normalized amplitude obtained from individual animals 30 min after bath application of the drug. ŽC. Sample traces recorded from a P5 preparation stimulated at different rates in control ACSF and 0.1 mM GABA. Test-pulse amplitudes Žarrows. are normalized to the CAP amplitude obtained at low stimulation rate Žopen arrows.. ŽD. Summary of the changes in test-pulse amplitude in pre- and postnatal preparations. Each symbol represents the averaged value pooled from 5 prenatal and 7 postnatal preparations. Error bars represent S.E.M.
302
C.H. Lim, S.M. Ho r DeÕelopmental Brain Research 108 (1998) 299–302
receptors in the glial cells, when activated, could alter the environment of the retinal axons. GABA has been shown to cause depolarization of optic nerve astrocytes w1,2,6,7x. However, the effects of GABA on astrocytes do not preclude a direct effect on optic axons. A recent report shows that the GABA-induced extracelluar potassium increase in isolated optic nerve is abolished after neonatal enucleation w4x, suggesting that the receptors that are involved in this response reside on the axons rather than the astrocytes. The suppression of axonal conduction in a developing system is surprising, as neural activity is thought to be important for the establishment of projection patterns w13x. One possibility is to prevent the transmission of excessive activity from the retina to the SC when local inhibition is poorly developed during the first two postnatal weeks w14,15x. Artificially induced activity in the developing ferret can weaken the orientation selectivity of cortical neurones without affecting the overall pattern of orientation and direction domains w16x. Therefore, the GABAergic action on axonal conduction may act to filter the retinal signal with no visual coherent before eye opening. In subsequent development, this GABAergic filter is gradually switched off, as shown by Sakatani w12x, to allow activity-dependent mechanisms to take over the refinement of connections. Acknowledgements We thank Prof. R. Mark and Dr. L. Marotte for their valuable comments on this manuscript. This work was supported in part by John Crawford and Glaxo WellcomeEDB Scholarships ŽC. Lim.. References w1x A. Butt, J. Jennings, The astrocyte response to gamma-aminobutyric acid attenuates with age in the rat optic nerve, Proc. R. Soc. London B 258 Ž1994. 9–15.
w2x A. Butt, J. Jennings, Response of astrocytes to gamma-aminobutyric acid in the neonatal rat optic nerve, Neurosci. Lett. 168 Ž1994. 53–56. w3x L. Galli, L. Maffei, Spontaneous impulse activity of rat retinal ganglion cells in prenatal life, Science 242 Ž1988. 90–91. w4x A.G. Howd, S.L. Kirell, M. Rattray, D. Shewan, A.M. Butt, Action of GABA in the developing rat optic nerve, Soc. Neurosci. Abstr. 22 Ž1996. 128.14. w5x A.G. Howd, M. Rattray, A.M. Butt, Expression of GABA transporter mRNAs in the developing and adult rat optic nerve, Neurosci. Lett. 235 Ž1997. 98–100. w6x H. Kettenmann, K. Backus, M. Schachner, Gamma-aminobutyric acid opens Cly channels in cultured astrocytes, Brain Res. 404 Ž1987. 1–9. w7x H. Kettenmann, M. Schachner, Pharmacological properties of gamma-aminobutyric acid-, glutamate-, and aspartate-induced depolarizations in cultures astrocytes, J. Neurosci. 5 Ž1985. 3295–3301. w8x N. Lake, Taurine, GABA and GFAP immunoreactivity in the developing and adult rat optic nerve, Brain Res. 596 Ž1992. 124–132. w9x C.H. Lim, L. Reece, Functional development of the retino-tectal connection in embryonic rat in vitro, Proc. Aust. Neurosci. Soc. 7 Ž1996. 55. w10x R. Mooney, A.A. Penn, R. Gallego, C.J. Shatz, Thalamic relay of spontaneous retinal activity prior to vision, Neuron 17 Ž1996. 863– 874. w11x P. Rogers, D. Pow, Immunocytochemical evidence for an axonal localization of GABA in the optic nerves of rabbits, rats, and cats, Vis. Neurosci. 12 Ž1995. 1143–1149. w12x K. Sakatani, J.A. Black, J.D. Kocsis, Transient presence and functional interaction of endogenous GABA and GABA A receptors in developing rat optic nerve, Proc. R. Soc. London B 247 Ž1992. 155–161. w13x C.J. Shatz, Role for spontaneous neural activity in the patterning of connections between retina and LGN during visual development, Int. J. Dev. Neurosci. 12 Ž1994. 531–546. w14x J. Shi, S.M. Aamodt, M. Constantine-Paton, Temporal correlations between functional and molecular changes in NMDA receptors and GABA neurotransmission in the superior colliculus, J. Neurosci. 17 Ž1997. 6264–6276. w15x S.S. Warton, M. Perouansky, R. Grantyn, Development of GABAergic synaptic connections in vivo and in cultures from the rat superior colliculus, Dev. Brain Res. 52 Ž1990. 95–111. w16x M. Weliky, L.C. Katz, Disruption of orientation tuning in visual cortex by artificially correlated neuronal activity, Nature 386 Ž1997. 680–685.
Short communication
GABAergic modulation of axonal conduction in the developing rat retinotectal pathway Christopher H. Lim, Stephen M. Ho
)
DeÕelopmental Neurobiology Group, Research School of Biological Sciences, Australian National UniÕersity, Canberra, Australia Accepted 3 March 1998
Abstract The effect of g-aminobutyric acid ŽGABA. on nerve conduction was investigated in the developing rat optic nerves in vitro. Antidromic compound action potentials ŽCAP. were suppressed by GABA, with increasing attenuation in older preparations. Enhancement of CAP amplitude using bicuculline was observed in postnatal days 4–5 suggesting the presence of endogenous GABA activity. Our findings suggest that the role of electrical activity in the refinement of the retinotectal projection could be limited by the GABAergic action on axonal conduction. q 1998 Elsevier Science B.V. All rights reserved. Keywords: Development; Optic nerve; Retinotectal pathway; GABA; Rat
g-Aminobutyric acid ŽGABA. is the principal inhibitory neurotransmitter in the mammalian central nervous system. GABA can also impose a non-synaptic inhibitory effect by suppressing the conduction of action potentials. In the rat optic nerve, GABA A -mediated depolarization and attenuation of action potential was age dependent, with its effect largely abolished after the 3rd postnatal week w12x. Our previous studies have shown that axonal conduction could be detected by embryonic day ŽE. 17 in the rat superior colliculus ŽSC. w9x. It remains unclear whether the GABAergic effect on axonal conduction also exists in these early developmental stages. The imposition of such a conduction block between the retina and retino-recipient structures would have an influence on the transmission of electrical activity originating from spontaneously active retinal ganglion cells w3x. We have therefore extended the studies of Sakatani et al. w12x and examined the effect of GABA on axonal conduction from E18 to postnatal day ŽP. 5. Albino–Wistar rat pups and embryos were used. Embryos were obtained from timed pregnancies with the day of sperm positivity taken as E0. The day of birth was denoted as P0. Pregnant dams were anaesthetised with Nembutal Ž60 mgrkg, i.p. injection. and fetuses were
)
Corresponding [email protected]
author.
Fax:
q61-02-6249-3808;
E-mail:
0165-3806r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 3 8 0 6 Ž 9 8 . 0 0 0 5 2 - 2
delivered by caesarian section. Fetuses or neonates were then anaesthetised by hypothermia and decapitated. The brain tissue was dissected under chilled sucrose solution Žabout 108C. with the following compositions ŽmM.: sucrose, 266; KCl, 3; NaHCO 3 , 17; CaCl 2 , 2; MgCl 2 , 1; glucose, 12.2; saturated with 95% O 2 , 5% CO 2 . The brain was hemisected with the optic chiasm and contralateral optic nerve retained on one side. The preparation was then transferred to the recording chamber, which was superfused with recirculating artificial cerebrospinal fluid ŽACSF. with the following compositions ŽmM.: NaCl, 124; KCl, 5; NaHCO 3 , 26; NaH 2 PO4 , 1.25; CaCl 2 , 2; MgSO4 , 2; glucose, 11; saturated with 95% O 2 , 5% CO 2 ; bath temperature maintained at 20–228C. Antidromic compound action potentials ŽCAP. were recorded at the optic nerve head using glass a microelectrode and a conventional electrophysiological setup. The lateral SC was stimulated at supramaximal intensity Ž2–9 mA, 0.03-ms duration, square current pulse. using bipolar tungsten electrodes. A low stimulation rate Ž1r30 s. was used to avoid conduction failure. GABA and bicuculline methiodide ŽBic. were purchased from Research Biochemicals International ŽNatick, MA. or Sigma ŽSt. Louis, MO.. CAP was recorded in 25 preparations aged between E18 and P5 ŽFig. 1.. Multimodal waveform was observed in some older preparations ŽFig. 1A., probably due to different conduction velocities. A large variation in CAP amplitude Ž100–900 mV. was observed between preparations from the same age group. Bath-applied GABA suppressed
300
C.H. Lim, S.M. Ho r DeÕelopmental Brain Research 108 (1998) 299–302
Fig. 1. ŽA. Attenuation of optic nerve CAP by GABA. Each trace was averaged from 10 sweeps and arrows indicate where peak measurements were made. ŽB. Recordings obtained from a P4 preparation showing the GABAergic attenuation of CAP was abolished after co-application of bicuculline. ŽC. Histograms showing the effect of GABA, and co-application of GABA and bicuculline on the CAP amplitude averaged from postnatal preparations. Error bars represent S.E.M.. ŽD. changes in CAP amplitude vs. stimulation intensity of a P5 preparation.
the CAP. Fig. 1B shows that the GABAergic attenuation of CAP was abolished when GABA and bicuculline were co-applied. Fig. 1C summarises the effects of GABA and co-application of GABA and bicuculline on CAP amplitude in postnatal preparations. The amplitude of axonal CAP was reduced to about 60% with 100 m M of GABA, but remained at about 100% when the same concentration of bicuculline was co-applied with GABA. This suggests that the attenuation of CAP was mediated through a GABA A receptor-linked pathway.
To eliminate the possibility that the reduction of CAP amplitude was the result of an increase in the stimulating threshold, we recorded the CAP amplitude at various stimulation strengths below the supramaximal level in 2 postnatal preparations. Fig. 1D shows the response amplitude vs. stimulation intensity curves under control and bath-applied GABA conditions. With bath-applied GABA, CAP attenuation was similar in proportion at different stimulation intensities. The absence of lateral shifting in the response–stimulation intensity curve suggests that at-
C.H. Lim, S.M. Ho r DeÕelopmental Brain Research 108 (1998) 299–302
tenuation was not a result of a local increase in stimulation threshold. Fig. 2A summarizes the GABAergic attenuation on CAP during the E18 to P5 period. It is apparent that the GABA mediated blockade of CAP increases with development. We also addressed whether endogenous GABA activity can be detected during this developmental period. After bath-applied bicuculline Ž10 m M., an increase in CAP amplitude was observed in P4–5 preparations ŽFig. 2B.. The average increase was 124% Ž"9%. Žmean "SE, n s 10., which was significantly different from control Ž t-test, p - 0.5, n s 10., suggesting that endogenous GABA activity was present. For prenatal animals, no significant change was observed. The average amplitude after bicuculline was 97.5% Ž"15%. Ž n s 13.. We next tested the behaviour of CAP when the impulse traffic was delivered at a higher rate, as previous studies have shown that spontaneous discharge of developing retinal ganglion cells can exceed 20 Hz w3,10x. Twin-pulse stimulation at intervals between 100 to 33 ms were used
301
ŽFig. 2C.. Increasing the stimulation rate to 10–30 Hz in normal ACSF caused a graduated reduction of the normalized test-pulse amplitude by 10%–30%. Test-pulse amplitude was further reduced with bath-applied GABA ŽFig. 2D.. This indicates that optic nerve conduction at these developmental stages is subject to further suppression in the presence of GABA when the impulse traffic rate is high. In the postnatal group, a lower concentration of GABA was used since 1 mM of GABA often abolished the response of the test pulse. This also agrees with results in Fig. 2A showing an increase of GABA-mediated blockade of axonal conduction with age. The source and site of GABA activity in the developing retinal axons remain to be determined. Various reports have demonstrated GABA immunoreactivity in glial cells or axons in the optic nerve w8,11,12x. The presence of GABA transporters ŽGAT 1–3. mRNAs in neonatal rat optic nerve w5x suggests that GABA could be released by heteroexchange through GABA transporters. The location of GABA A receptors is also a contentious issue. GABA A
Fig. 2. Effects of GABA ŽA. and bicuculline ŽB. on CAP amplitude during development. Each symbol represents the normalized amplitude obtained from individual animals 30 min after bath application of the drug. ŽC. Sample traces recorded from a P5 preparation stimulated at different rates in control ACSF and 0.1 mM GABA. Test-pulse amplitudes Žarrows. are normalized to the CAP amplitude obtained at low stimulation rate Žopen arrows.. ŽD. Summary of the changes in test-pulse amplitude in pre- and postnatal preparations. Each symbol represents the averaged value pooled from 5 prenatal and 7 postnatal preparations. Error bars represent S.E.M.
302
C.H. Lim, S.M. Ho r DeÕelopmental Brain Research 108 (1998) 299–302
receptors in the glial cells, when activated, could alter the environment of the retinal axons. GABA has been shown to cause depolarization of optic nerve astrocytes w1,2,6,7x. However, the effects of GABA on astrocytes do not preclude a direct effect on optic axons. A recent report shows that the GABA-induced extracelluar potassium increase in isolated optic nerve is abolished after neonatal enucleation w4x, suggesting that the receptors that are involved in this response reside on the axons rather than the astrocytes. The suppression of axonal conduction in a developing system is surprising, as neural activity is thought to be important for the establishment of projection patterns w13x. One possibility is to prevent the transmission of excessive activity from the retina to the SC when local inhibition is poorly developed during the first two postnatal weeks w14,15x. Artificially induced activity in the developing ferret can weaken the orientation selectivity of cortical neurones without affecting the overall pattern of orientation and direction domains w16x. Therefore, the GABAergic action on axonal conduction may act to filter the retinal signal with no visual coherent before eye opening. In subsequent development, this GABAergic filter is gradually switched off, as shown by Sakatani w12x, to allow activity-dependent mechanisms to take over the refinement of connections. Acknowledgements We thank Prof. R. Mark and Dr. L. Marotte for their valuable comments on this manuscript. This work was supported in part by John Crawford and Glaxo WellcomeEDB Scholarships ŽC. Lim.. References w1x A. Butt, J. Jennings, The astrocyte response to gamma-aminobutyric acid attenuates with age in the rat optic nerve, Proc. R. Soc. London B 258 Ž1994. 9–15.
w2x A. Butt, J. Jennings, Response of astrocytes to gamma-aminobutyric acid in the neonatal rat optic nerve, Neurosci. Lett. 168 Ž1994. 53–56. w3x L. Galli, L. Maffei, Spontaneous impulse activity of rat retinal ganglion cells in prenatal life, Science 242 Ž1988. 90–91. w4x A.G. Howd, S.L. Kirell, M. Rattray, D. Shewan, A.M. Butt, Action of GABA in the developing rat optic nerve, Soc. Neurosci. Abstr. 22 Ž1996. 128.14. w5x A.G. Howd, M. Rattray, A.M. Butt, Expression of GABA transporter mRNAs in the developing and adult rat optic nerve, Neurosci. Lett. 235 Ž1997. 98–100. w6x H. Kettenmann, K. Backus, M. Schachner, Gamma-aminobutyric acid opens Cly channels in cultured astrocytes, Brain Res. 404 Ž1987. 1–9. w7x H. Kettenmann, M. Schachner, Pharmacological properties of gamma-aminobutyric acid-, glutamate-, and aspartate-induced depolarizations in cultures astrocytes, J. Neurosci. 5 Ž1985. 3295–3301. w8x N. Lake, Taurine, GABA and GFAP immunoreactivity in the developing and adult rat optic nerve, Brain Res. 596 Ž1992. 124–132. w9x C.H. Lim, L. Reece, Functional development of the retino-tectal connection in embryonic rat in vitro, Proc. Aust. Neurosci. Soc. 7 Ž1996. 55. w10x R. Mooney, A.A. Penn, R. Gallego, C.J. Shatz, Thalamic relay of spontaneous retinal activity prior to vision, Neuron 17 Ž1996. 863– 874. w11x P. Rogers, D. Pow, Immunocytochemical evidence for an axonal localization of GABA in the optic nerves of rabbits, rats, and cats, Vis. Neurosci. 12 Ž1995. 1143–1149. w12x K. Sakatani, J.A. Black, J.D. Kocsis, Transient presence and functional interaction of endogenous GABA and GABA A receptors in developing rat optic nerve, Proc. R. Soc. London B 247 Ž1992. 155–161. w13x C.J. Shatz, Role for spontaneous neural activity in the patterning of connections between retina and LGN during visual development, Int. J. Dev. Neurosci. 12 Ž1994. 531–546. w14x J. Shi, S.M. Aamodt, M. Constantine-Paton, Temporal correlations between functional and molecular changes in NMDA receptors and GABA neurotransmission in the superior colliculus, J. Neurosci. 17 Ž1997. 6264–6276. w15x S.S. Warton, M. Perouansky, R. Grantyn, Development of GABAergic synaptic connections in vivo and in cultures from the rat superior colliculus, Dev. Brain Res. 52 Ž1990. 95–111. w16x M. Weliky, L.C. Katz, Disruption of orientation tuning in visual cortex by artificially correlated neuronal activity, Nature 386 Ž1997. 680–685.