Neurotrophic Factors

First printed in R&D Systems' 1996 Catalog.


The fate of neurons in the developing and adult nervous system is controlled, in part, by members of the neurotrophin family. This consists of four proteins; nerve growth factor (NGF), brain derived neurotrophic factor (BDNF), neurotrophin 3 (NT-3), and neurotrophin 4 (NT-4). As indicated by their name (trophin), all have some survival activity on nervous tissue. For example, the number of neurons reaching a target field in the developing vertebrate nervous system exceeds the number found in the mature innervated target. The final number is attained by thinning the population of neurons through programmed neuronal death1, a process called apoptosis. Neurotrophic factors secreted by cells in the target field protect the neurons from apoptosis. Only those neurons that obtain sufficient neurotrophins survive. Thus, the final number of neurons innervating a target reflects the availability of neurotrophins. The classical "neurotrophin theory"2 does well to explain losses of sets of peripheral sympathetic neurons in NGF knockout mice3, sensory neurons in BDNF knockout mice4,5 proprioceptive neurons in NT-3 knockout mice6-8, and subsets of sensory neurons in NT-4 knockout mice9,10 However, the roles of the neurotrophic factors in the brain and in earlier development are not so well understood. Neurotrophins influence development by means other than sparing neurons from programmed death. Some neurotrophins support neuronal survival up to the time of naturally-occuring cell death and then become ineffective, an apparent result of a switch in the type of neurotrophin receptor expressed by the neuron (11-14, reviewed in 15-17). The neurotrophin NT-3 has profound effects on neuronal progenitor cells and thereby has the potential to increase the number of neurons in a population destined to have a specific phenotype.18-21 Neurotrophins can also affect electrical synapse activity and electrical synaptic activity can also increase neurotrophin expression (referenced in sections on specific neurotrophins), suggesting a role for neurotrophins in plasticity in the nervous system.

The availability of mice in which the gene encoding a specific neurotrophin is disrupted will aid in defining the roles of neurotrophins in the central nervous system. It is already clear from studies on these mutant mice that although the lack of a neurotrophin may not cause the death of a specific neuron, it may result in the inability of the neuron to reach its full differentiated potential (see BDNF knockout mice). The picture that is developing with respect to neurotrophins is that the development, maintenance, and plasticity of the nervous system involves multiple neurotrophins in a spatial and temporal manner where the timing and site of expression of neurotrophins, their receptors, as well as other factors, all contribute to the complex set of interactions that leads to the development of the mature vertebrate nervous system.23

Brain-derived Neurotrophic Factor

Brain-derived Neurotrophic Factor (BDNF) knockout mice

BDNF23-25 and its receptor, trkB (trk receptors are reviewed in 26), are widely expressed in the developing and adult nervous system.26-30 Mice in which the BDNF gene is disrupted (4, 5 and see 18 and 31 for reviews) have: 1) an 80% reduction in neurons of the vestibular ganglion that innervate the inner ear organs responsible for proper balance and locomotion; 2) have a 25-44% reduction in trigeminal neurons which connect the hindbrain with whisker pads and other facial cutaneous targets; 3) show a 55-65% reduction in neurons of the nodose and petrosal ganglia which relay information from the heart, major vessels, lung, and gut to the brain stem; 4) and, have a decreased number in a population of neurons that have been tentatively identified as mechanoreceptor neurons in the dorsal root ganglion. Knockout mice lacking functional trkB32, the receptor for BDNF, are also severely compromised in their complement of sensory neurons. Loss of this set of neurons appears to occur during innervation of the target field4 and is in accordance with the ability of BDNF to sustain viability of trigeminal neurons cultured from embryos in which target innervation is in its ealiest stages.15,33 BDNF is also required for proper differentiation of specific subsets of cortical and hippocampal GABAergic interneurons. The number of GABAergic neurons and the amount of neural GABA appear to be normal in the BDNF knockout mouse, but expression of the GABAergic neuronal markers parvalbumn, calbindin, and neuropeptide Y is decreased.4 Decreased levels of the calcium binding proteins parvalbumin and calbindin could affect synaptic inhibitory activity of these neurons. Decreased calbindin expression was also seen in the medium-sized striatal spiny neurons and in the striatal projection to the substantia nigra.4 Decreased neuropeptide Y expression was observed in 15-20 day-old mutant mice suggesting that lack of BDNF prevented, rather than delayed, normal differentiation of a set of cortical and hippocampal GABAergic neurons. These results with the BDNF knockout mouse agree with the demonstration that BDNF infusion into the adult normal rat hippocampus, cortex, and striatum increases neuropeptide Y expression in all three regions and increases expression of other neuropeptides by hippocampal and cortical neurons.34,35

The in vivo role of BDNF in development of motor neurons is less clear. BDNF supports motor neuron survival in lesioned animals36-39 and prevents naturally-occuring death and increases expression of neuronal markers in cultured embryonic motor neurons.40-41 A role for BDNF in motor neuron development is suggested by the TrkB knockout mice. TrkB knockout mice have reduced numbers of lumbar spinal and facial nuclear motor neurons (32). However, motor neurons in the BDNF knockout mice4,5 and in mice missing both BDNF and NT-410,11, both of which are high affinity ligands for the trkB receptor26, apparently have normal motor neuron development. Since NT-3 can also activate trkB26 there may be redundancy to ensure proper motor neuron development. Thus, BDNF may participate in motor neuron development but an absence of BDNF is compensated for by other neurotrophins. Differentiation of cholinergic neurons, which are important in cognitive functions and are involved in Alzheimer's disease, is influenced by exogenous BDNF. BDNF administration protects basal forebrain cholinergic neurons against lesion-induced death, influences neurotransmitter production, and increases expression of cholinergic markers in vivo and in vitro (41-43 and reviewed in 44). BDNF also affects differentiation of dopaminergic neurons.45-48 It is surprising that deficiencies in cholinergic neurons have not yet been reported for the BDNF knockout, BDNF/NT-4 double knockout, and trkB knockout mice. Nor have any gross deficits in the dopaminergic neurons of the midbrain been reported. A more detailed analysis of these mutant mice should clarify the role of BDNF in differentiation of these sets of neurons.

Other biological activities of BDNF

BDNF is also a survival factor for cultured rat retinal ganglion cells49, cerebellar granule neurons50, and neurons of the locus coeruleus (51, see review in ref. 44) when the neurons are obtained from embryonic or adult animals at the appropriate developmental stage. BDNF also acts in concert with other neurotrophins. Trigeminal sensory neurons15 and cerebellar neurons12 appear to switch their dependence from BDNF to nerve growth factor and NT-3, respectively. Adult rat hippocampal and cerebral cortical neurons express trkB and trkC, the receptor for NT-326, indicating that these neurons can respond to BDNF and NT-3.52 BDNF can also act via an autocrine mode. BDNF and trkB are coexpressed in many hippocampal and cortical neurons53 and antisense oligonucleotides to BDNF inhibit survival of cultured neurons from the dorsal root ganglion.54 The expression of BDNF can be modulated by activity of the GABAergic and cholinergic systems55-58 and BDNF can influence neuronal synaptic activity.59, 60 Thus, the influence of BDNF on neuronal activity and the reciprocal influence of neuronal activity on BDNF expression, combined with the stabilizing effect of BDNF on interneuronal connections, suggest that this interplay influences synaptic plasticity and anatomical reorganization in the mature nervous system.

Neurotrophin 3 (NT-3)

NT-3 knockout mice

Neurotrophin 3 (NT-3)

NT-361-64 and its receptor, trkC26, are widely expressed in the developing and adult nervous system.26,28,61,65-67 Mice in which the NT-3 gene is disrupted7-9 have a severe loss of proprioceptive neurons (convey information on limb position). Proprioceptive 1a afferents are part of the muscle-motor neuron circuit that mediates the stretch reflex. They relay information from the peripheral muscle spindles, which are the sensing organs, to motoneurons in the spinal cord, and send ascending projections to the spinocerebellar neurons. The 1a afferents and muscle spindles are greatly decreased or missing in NT-3 knockout mice and in the NT-3 receptor (trkC) knockout mice.68 Muscle spindles were present at 50% the normal number in mice heterozygous for disrupted NT-3, indicating a direct proportionality between NT-3 availability and survival of 1a afferents. Loss of proprioceptive neurons in the NT-3 knockout mice is consistent with previous findings that NT-3 is expressed by the target tissues, muscle and motoneurons. Proprioceptive neurons express trkC and NT-3 supports viability of cultured proprioceptive neurons from the trigeminal mesenencephalic nucleus.69 Golgi tendon organs, which are also involved in proprioception, were missing, suggesting that the 1b afferents that synapse with these structures will be found to be absent. Motor neurons in the ventral roots of the spinal cord were reduced by approximately 30%, whereas other examined motor neurons were normal. NT-3 has been reported to support viability of cultured embryonic spinal motor neurons.40

NT-3 knockout mice have decreased numbers of sympathetic neurons in the superior cervical ganglion7, but have no obvious deficiencies in peripheral sympathetic sensory innervation as observed in the NGF knockout3 and trkA knockout70 mice. Sympathetic neurons depend on NGF for survival during innervation of targets.1 Therefore, loss in NT-3 knockout mice may reflect the absence of an NT-3-dependent event that occurs prior to innervation. NT-3 promotes in vitro survival of sympathetic neuroblasts19, which are dividing precursors to sympathetic neurons. Thus, decreased numbers of sympathetic neurons in NT-3 knockout mice may reflect a loss in an early NT-3-dependent population of neurons destined to be sympathetic. A similar scenario has been proposed for development of the nodose ganglia.13 Neuronal cell counts in the nodose ganglia indicate a partial (47%) loss of neurons in the NT-3 knockout mice.7 Neutralization of NT-3 in the developing quail embryo with antibodies against NT-3 decreases the number of neurons in the nodose ganglion.13 The decrease was proposed to indicate a decrease in the supply of neurons to the nodose ganglion rather than protection against naturally-occuring death because the timing of NT-3 neutralization preceded the stage of naturally-occuring neuronal death.29 An involvement of NT-3 in early development is also suggested by its high expression in other areas of the developing embryo before naturally-occuring cell death. NT-3 has been shown to induce differentiation of cultured enteric multipotent neural crest cells into neurons or glia20, to be mitogenic for neural crest cells cultured in the presence of somites18, to be mitogenic for cultured oligodendrocyte precursor cells71, to stimulate differentiation of motor neurons from neural tube progenitors21, and to be a neural tube-derived signal for conversion of the epithelial dermatome progenitors to mesenchyme cells of the dermis.72 The role of NT-3 during early embryogenesis should become clearer with more in depth studies of the NT-3 knockout mice.

Other biological activities of NT-3

NT-3 promotes the survival of cultured embryonic noradrenergic neurons of the locus coeruleus41 and in vivo survival of noradrenergic neurons of the locus coeruleus after 6-hydroxydopamine-induced lesions. NT-3 promotes the survival and differentiation of cultured dopaminergic and cholinergic neurons from the developing substantia nigra48, promotes the cholinergic phenotype of cultured rat motor neurons41, increases survival of Purkinje cells74, and stimulates neurite outgrowth in cultured hippocampal pyramidal neurons.75 NT-3 also reduces neuronal activity by reducing the inhibitory GABAergic synaptic transmission by cultured cortical neurons.76

Neurotrophin 4 (NT-4)

NT-4 knockout mice

NT-4 [also called NT-4/5 or NT-577,78] mRNA is present in many tissues and in major regions of the brain as determined by Northern blotting.30 Unlike all other neurotrophin or neurotrophin receptor knockout mice the NT-4 knockout mouse thrives, exhibits no overt phenotypic abnormalities, and reproduces.10,11 The only deficits thus far found in the NT-4 knockout mice are a 56-57% and a 50% deficit in sensory neurons of the nodose-petrosal and geniculate ganglions, respectively. BDNF and NT-4 use the same trkB receptor.26 The population of nodose-petrosal neurons is decreased by 59-61% and the geniculate by 48% in BDNF knockout mice.10,11 However, the nodose-petrosal neurons missing in the NT-4 and in the BDNF knockout mice appear to consist of different populations. Nodose-petrosal neurons are depleted by 79-94% when both the NT-4 gene and the BDNF gene are disrupted. The larger loss in the double knockout suggests that two subsets of nodose- petrosal neurons innervate different targets, one which supplies NT-4 and another which supplies BDNF for survival of the innervating neurons. BDNF and NT-4 have been shown to support survival of neurons from the nodose ganglion.33 Other sensory ganglia (e.g., trigeminal), which showed neuronal deficits in the BDNF mutant mouse, were unaffected in the NT-4 knockout mouse. Comparison of the 80-94% decrease in nodose neurons in the NT-4/BDNF double knockout to the 47% loss of nodose neurons in the NT-3 knockout mouse7 indicates that a population of nodose neurons has a sequential or simultaneous dependency on NT-3 and either NT-4 or BDNF. As discussed in the section on BDNF, the absence of NT-4 or NT-4 and BDNF did not decrease the number of motor neurons as was predicted by the survival and differentiation promoting activity of NT-4 on cultured motor neurons and axotomized motor neurons.39-41,79

Other biological activities of NT-4

Because NT-4 and BDNF exert their effects through the trkB receptor it is not surprising that, in most cases, addition of exogenous NT-4 to cultured neurons or administration of NT-4 in vivo has the same effect as administration of BDNF. NT-4 induces differentiation of cultured spinal41 and basal forebrain cholinergic neurons51, hippocampal neurons80, and cerebellar granule cells81 and survival of cultured embryonic dopaminergic neurons of the mesencephalon49,82, and noradrenergic neurons of the locus coeruleus.51 Administration of NT-4 induces in vivo differentiation of dopaminergic, GABAergic, and serotonergic neurons in and near the substantia nigra.45 NT-4 also appears to transiently support survival of trigeminal and jugular neurons during embryonic development in a fashion analogous to BDNF.33

Structural Information

The neurotrophins are synthesized as approximately 250 amino acid residue precursors and are secreted as basic non-covalently linked dimers of two identical subunits of approximately 120 amino acid residues.1,2,21 Mature forms of specific neurotrophins show extremely high (90-100%) homology between species. The mature forms of BDNF24,25, NT-361-64, and NT-478 have approximately 50% amino acid identity to NGF. The crystal structure of the NGF subunit is known83 and the regions of homology between BDNF, NT-3, NT-4 and NGF are found in the structural segments that include three pairs of antiparallel beta strands that form a hydrophobic core that is supported by three intramolecular cysteine disulfides. Core formation exposes a flat surface by which two subunits associate to give a twofold axis of symmetry. The three beta hairpin loops (amino acids 29-35, 43-48, and 92-98; numbers refer to NGF), the reverse turn (amino acid 59-66), and the carboxyl and amino termini contain the major regions of amino acid variation between the neurotrophins. Based on chimeric molecules containing portions of the various neurotrophins, basic residues in the variable regions are critical for binding to the appropriate trk.84 It appears that a different but overlapping set of basic residues is important for binding the low-affinity neurotrophin receptor, p75LNGFR.85


The neurotrophins bind to one or more of a family of cell surface receptor tyrosine protein kinases which are glycosylated proteins of approximately 825 amino acid residues.87,88,26 Binding of neurotrophin induces receptor dimerization at the cell surface followed by autophosphorylation of receptor tyrosine residues. The phosphorylated tyrosines then serve to recruit intracellular proteins involved in signal transduction. The binding constants for the high-affinity ligands are in the range of 10-11. TrkA is the receptor for NGF, trkB is the receptor for BDNF and NT-4, and trkC is the receptor for NT-3 (see 26). However, NT-3 can also bind to trkA and trkB, but with lower affinity than to trkC, and with lower affinity than the primary ligands for these receptors. NT-4 also binds to trkA but with lower affinity. To complicate the issue, truncated isoforms of trkB89 and trkC90 also exist, which lack the cytoplasmic tyrosine kinase catalytic region and are unable to elicit a response to neurotrophin binding. Like the intact receptors, these forms are widely expressed throughout the developing and adult central and peripheral nervous systems. The kinase-deficient trkB appears to be preferentially expressed in non-neuronal cells whereas full length trkB is preferentially expressed in neurons. Whether the noncatalytic forms of the receptors act as agonists or antagonists is not clear and may depend on the context in which they are found, i.e., agonist by concentrating a neurotrophin after neuron injury or antagonist by decreasing the availability of a specific neurotrophin to competing neurons during innervation. All four neurotrophins also bind to the low affinity nerve growth factor receptor, p75LNGFR. The role of this receptor is controversial.26,86 P75LNGFR may increase the affinity of the trk receptors for their respective neurotrophic ligands and/or be important in retrograde transport of the bound neurotrophins. Mutated NT-4 which is unable to bind p75LNGFR has been shown to be much less potent as a trkB inducer of phosphorylation85, suggesting that p75LNGFR takes part in NT-4 signalling.

Clinical Interest (see references 44, 91-94)

The ability of neurotrophins to spare neuronal death in lesioned animals and to increase expression of cholinergic and dopaminergic neuronal phenotypes have evoked interest in the therapeutic potential for neurotrophins. Neurotrophins are candidates for treatment of the degenerative motor neuronal disorders: amyotrophic lateral sclerosis (ALS), spinal muscular atrophy, and post-polio syndrome. Current clinical trials using BDNF and ciliary neurotrophic factory in patients with ALS are ongoing. Neurotrophins also have therapeutic potential for epileptic seizures and Alzheimer's and Parkinson's disease. However, delivery of proteins to the brain remains a major obstacle in any treatment proposing to affect brain function.

For more information on Neurotrophins

Neurosciences on the Internet contains a searchable index of neuroscience resources available on the World Wide Web and other parts of the Internet. Neurobiology, neurology, neurosurgery, psychiatry, psychology and cognitive science sites are covered, including basic science and clinical fields - by Neil A. Busis, M.D.


  1. Oppenheim, R.W. (1991) Annu. Rev. Neurosci. 14:453.
  2. Levi-Montalcini, R. (1987) Science 237:1154.
  3. Crowley, C. et al. (1994) Cell 76:1001.
  4. Jones, K.R. et al. (1994) Cell 76:989.
  5. Enfors, P. et al. (1994) Nature 368:147.
  6. Enfors, P. et al. (1994) Cell 77:503.
  7. Tessarollo, L. et al. (1994) Proc. Natl. Acad. Sci. USA 91:11844.
  8. Farinas, I. et al. (1994) Nature 369:658.
  9. Liu, X. et al. (1995) Nature 375:238.
  10. Conover, J.C. et al. (1995) Nature 375:235.
  11. Segal, R.A. et al. (1992) Cell 9:1041.
  12. Birren, S.J. et al. (1993) Development. 119:597.
  13. Gaese, F. et al. (1994) Development. 120:1613.
  14. Buchman, V.L. and A.M. Davies (1993) Develop. 118:989.
  15. Davies, A.M. (1994) Current Biol. 4:273.
  16. Davies, A.M. (1994) Prog. Growth Factor Res. 5:263.
  17. Davies, A.M. (1994) J. Neurobiol. 25:1334.
  18. Kalcheim, C. et al. (1992) Proc. Natl. Acad. Sci. USA 89:1661.
  19. Di-Cicco-Blum, E. et al. (1993) Neuron 11:1101.
  20. Chalazonitis, A. et al. (1994) J. Neurosci. 14:6571.
  21. Averbach-Heller, L. (1994) Proc. Natl. Acad. Sci. USA 91:3247.
  22. Korsching, S. (1993) J. Neurosci. 13:2739.
  23. Barde, Y.-E. et al. (1982) EMBO J. 1:549.
  24. Leibrock, J. et al. (1982) Nature 341:149.
  25. Rosenthal, A. et al. (1991) Endocrinology. 129:1289.
  26. Barbacid, M. (1994) J. Neurobiol. 25:1386.
  27. Wetmore, C. et al. (1990) Exp. Biol. 109:141.
  28. Hofer, M. et al. (1990) EMBO J. 9:2459.
  29. Maisonpierre, P.C. et al. (1990) Neuron 5:501.
  30. Timmusk, T. et al. (1993) Eur. J. Neurosci. 5:605.
  31. Snider, W.D. (1994) Cell 77:627.
  32. Klein, R. et al. (1993) Cell 74:113.
  33. Davies, A.M. et al. (1993) J. Neurosci. 13:4961.
  34. Croll, S.D. et al. (1994) Eur. J Neurosci. 6:1343.
  35. Mizuno, K. et al. (1994) Dev. Biol. 165:243.
  36. Yan, Q. et al. (1994) J Neurosci. 14:5261.
  37. Clatterbuck, R.E. et al. (1994) J. Comp. Neurol. 342:45.
  38. Sendter, M. et al. (1992) Nature 360:757.
  39. Friedman, B. et al. (1995) J Neurosci. 15:1044.
  40. Henderson, C.E. et al. (1993) Nature 363:266.
  41. Wong, V. et al. (1993) Eur. J. Neurosci. 5:466.
  42. Knusel, B. et al. (1992) J. Neurosci. 12:4391.
  43. Alderson, R.F. et al. (1990) Neuron 5:297.
  44. Lindsay, R.M. et al. (1994) Trends Neurosci. 17:182.
  45. Altar, C.A. et al. (1994) Exp. Neurol. 130:31.
  46. Frim, D.M. et al. (1994) Proc. Natl. Acad. Sci. USA 91:5104.
  47. Knusel, B. et al. (1991) Proc. Natl. Acad. Sci. USA 88:961.
  48. Hyman, C. et al. (1994) J. Neurosci 14:335.
  49. Johnson, J.E. et al. (1986) J. Neurosci. 6:3031.
  50. Lindholm, D. et al. (1993) Eur. J. Neurosci. 5:1455.
  51. Friedman,W.J. et al. (1993) Exp. Neurol. 119:72.
  52. Kokaia, Z. et al. (1995) NeuroReport. 6:769.
  53. Kokaia, Z. et al. (1993) Proc. Natl. Acad. Sci. USA 90:6711.
  54. Acheson, A. et al. (1995) Nature 374:450.
  55. Wetmore, C. et al. (1994) J. Neurosci. 14:1688.
  56. Berzaghi, M.P. et al. (1993) J. Neurosci 13:3818.
  57. Gwag, B.Y. and J.E. Springer (1993) NeuroReport. 5:125.
  58. Zafra, F. et al. (1991) Proc. Natl. Acad. Sci. USA 88:10037.
  59. Shen, R.-Y. et al. (1994) Proc. Natl. Acad. Sci. USA 91:8290.
  60. Knipper, M. et al. (1994) Eur. J. Neurosci. 6:668.
  61. Enfors, P. et al. (1990) Proc. Natl. Acad. Sci. USA 87:5454.
  62. Rosenthal, A. et al. (1990) Neuron 4:767.
  63. Maissonpierre, P.C. et al. (1990) Science 247:1446.
  64. Hohn, A. et al. (1990) Nature 344:339.
  65. Lamballe, F. et al. (1994) J. Neurosci. 14:14.
  66. Merlio, J.-P. et al. (1992) Neuroscience. 51:513.
  67. Tessarollo, L. et al. (1993) Develop. 118:463.
  68. Klein, R. et al. (1994) Nature 368:249.
  69. Hory-Lee, F. et al. (1993) Proc. Natl. Acad. Sci. USA 90:2613.
  70. Smeyne, R.J. et al. (1994) Nature 368:246.
  71. Barres, B.A. et al. (1994) Nature 367:371.
  72. Brill, G. et al. (1995) Development. 121:2583.
  73. Arenas, E. and H. Persson (1994) Nature 367:368.
  74. Mount, H.T.J. et al. (1994) NeuroReport. 5:2497.
  75. Morfini, G. et al. (1994) J. Neurosci. Res. 39:219.
  76. Kim, H.G. (1994) Proc. Natl. Acad. Sci. 91:12341.
  77. Hallbook, F. et al. (1991) Neuron 6:845.
  78. Berkemeier, L.R. et al. (1991) Neuron 7:857.
  79. Koliatsos, V.E. (1994) Proc. Natl. Acad. Sci. USA 91:3303.
  80. Ip, N.Y. et al. (1993) J. Neurosci. 13:3394.
  81. Gao, W.-Q. et al. (1995) J. Neurosci. 15:2656.
  82. Hynes, M.A. et al. (1994) J. Neurosci. 37:144.
  83. McDonald, N.Q. et al. (1991) Nature 354:411.
  84. Ibanez, C.F. et al. (1993) EMBO J. 12:2281.
  85. Ryden, M. et al. (1995) EMBO J. 14:1979.
  86. Bothwell, M. (1995) Annu. Rev. Neurosci. 18:223.
  87. Squinto, S.P. et al. (1991) Cell 65:885.
  88. Lamballe, F. et al. (1991) Cell 66:967.
  89. Klein, R. et al. (1990) Cell 61:647.
  90. Velenzuela, D.M. et al. (1993) Neuron 10:963.
  91. Lindvall, O. et al. (1994) Trends Neurosci. 17:490.
  92. Maness, L.M. et al. (1994) Neurosci. Behav. Rev. 18:143.
  93. Brewster, W.J. et al. (1994) Trends Neurosci. 17:321.
  94. Barinaga, M. (1994) Science 264:772.