* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Intrinsic firing patterns of diverse neocortical neurons
Molecular neuroscience wikipedia , lookup
Single-unit recording wikipedia , lookup
Synaptogenesis wikipedia , lookup
Clinical neurochemistry wikipedia , lookup
Central pattern generator wikipedia , lookup
Subventricular zone wikipedia , lookup
Multielectrode array wikipedia , lookup
Electrophysiology wikipedia , lookup
Apical dendrite wikipedia , lookup
Anatomy of the cerebellum wikipedia , lookup
Chemical synapse wikipedia , lookup
Biological neuron model wikipedia , lookup
Premovement neuronal activity wikipedia , lookup
Circumventricular organs wikipedia , lookup
Neural coding wikipedia , lookup
Development of the nervous system wikipedia , lookup
Neuroanatomy wikipedia , lookup
Nervous system network models wikipedia , lookup
Stimulus (physiology) wikipedia , lookup
Neuropsychopharmacology wikipedia , lookup
Pre-Bötzinger complex wikipedia , lookup
Synaptic gating wikipedia , lookup
Optogenetics wikipedia , lookup
Intrinsic firing patterns of diverse neocortical neurons B a r r y W . C o n n o r s a n d M i c h a e l J. G u t n i c k Neurons of the neocortex differ dramatically in the patterns of action potentials they generate in response to current steps. Regular-spiking cells adapt strongly during maintained stimuli, whereas fast-spiking cells can sustain very high firing frequencies with little or no adaptation. Instrinsically bursting cells generate clusters of spikes (bursts), either singly or repetitively. These physiological distinctions have morphological correlates. RS and IB cells can be either pyramidal neurons or spiny stellate cells, and thus constitute the excitatory cells of the cortex. FS cells are smooth or sparsely spiny non-pyramidal cells, and are likely to be GABAergic inhibitory interneurons. The different firing properties of neurons in neocortex contribute significantly to its network behavior. temporal patterns of repetitive firing, and thus determine to a large extent the way individual neurons transform synaptic input into spike output. These intrinsic physiological properties constitute a reasonable basis for neuronal classification, since studies from a large number of brain areas show that the electrical fingerprint can be extremely uniform from cell to cell within a particular neuronal class (e.g. among cerebellar Purkinje cells, thalarnic relay cells, or dopaminergic cells of the substantia nigra; for a recent comprehensive review, see Ref. 4). Neurons of the neocortex are not physiologically homogeneous ~-14. Three basic types of intrinsic physiology have been recognized, and our terms for them are: regular-spiking (RS), fast-spiking (FS) and intrinsically bursting (IB). For each type, classification is based on three general variables - the characteristics of individual action potentialafterpotential complexes, the response to a justthreshold intracellular current pulse, and the repetitive response to prolonged, intracellularly applied stimuli. While these neuron classes are based entirely on physiological differences, intracellular staining experiments suggest that there are also some distinct morphological correlates. The general characteristics of the three classes of neurons, derived mostly from studies of rodent neocortex in vivo, of slices in vitro, and from tissue culture (both explants and dissociated cells), are summarized in Table I and discussed below. It is not our intention here to describe the specific ionic mechanisms underlying these neuronal behaviors, since only RS neurons have been examined in any detail. We also emphasize that these categories are not necessarily comprehensive, exclusive or definitive. On the one hand, it is very likely that subcategories exist, or that a continuum of variation better describes some cell populations; on the other, Barry W. Connorsis at the Section of Neurobiology, Division of Biology and Medicine, Box 6, Brown University, Providence, R102912, USAand MichaeIJ. Gutnick is at the Department of Physiology, Faculty of Health Sciences,BenGurion University of the Negev, Beersheva, Israel. It is axiomatic in neuroscience that the informational output of most neurons is completely defined by their temporal patterns of action potentials. This notion is implicit in all contemporary models of the functions of neocortex, making it essential to understand how each neuron transforms its input into output. Studies show that the intrinsic membrane properties of neurons in the neocortex are not homogeneous, but instead produce several categories of transform characteristics. The fine structure of the neocortex varies from region to region and from species to species; however, unifying principles underlie the assembly of individual neocorfical neurons into functional circuits 1. A fundamental unit of neocortex appears to be a group of diverse, radially organized neurons, each with its own morphological and physiological characteristics and its own pattern of synaptic connections. One classic investigational approach to this complex local circuitry is taxonomic classification. Neocortical neurons have been categorized according to various criteria such as location, morphology (i.e. size, soma shape and dendrite patterns), synaptic relationships locally and TABLE I. General classification scheme for neocortical neurons with distant cortical and subcortical Physiological class regions, and biochemical properties (especially neurotransmitters Characteristics Regular-spiking Intrinsically bursting Fast-spiking and their associated enzymes) 1'2. (RS) (IS) (FS) Nevertheless, in order to under- Single-spike: stand a particular neuron's funcrate of rise ++ ++ 4-+ tional role within a circuit, it is not ++ rate of fall + + enough to know only these charac- Single-spike afterpotential Complex, AHPs Complex, AHPs and Simple, AHP teristics. Its electrical fingerprint, and ADP ADP alone ++ Variable as determined by its intrinsic Frequency adaptation membrane properties, is also Spike bursts during injected current ++ important. Laminar location of soma II to VI IV or V II to VI It has long been known that Pyramidal or spiny Pyramidal or spiny Aspiny or neuronal membranes do not all Presumed morphology stellate stellate sparsely spiny behave similarly3. Neurons differ non-pyramidal in the types and distributions of Presumed synaptic function Excitatory Excitatory (glutamate Inhibitory (GABA) specific ion channels on their (glutamate or or aspartate) somata and dendrites. These inaspartate) trinsic membrane differences are manifest as differences in the Symbols: +, the presence and relative strength of a characteristic; -, the absence of a characteristic. Abbreviations: AHP, afterhyperpolarization; ADP, afterdepolarization. Table adapted from Ref. 15, and shapes of individual action poten- compiled from data in the following Refs: 7-12, 16, 18, Gutnick, M. J. and White, E. L., unpublished tials. They also lead to distinctive observations, and Agmon, A., PhD Thesis, Stanford University, CA, 1988. TINS, VoL 13, No. 3, 1990 © 199o.ElsevierSciencePublishersLtd.(UK) 0166-2236/90/$02.00 99 A Regular-spiking C Stimulus current I Fast-spiking 4OO 50 mV 3 nA 50 m s Fast-spiking B • ooo • :o e° .o • • • ooo °°°e eooo o. • °° oo • ooooe ° • oo • oo t~ ~ 300 ~200 Regular-spiking _J --"-I__ ,°° t 50 mV 3nA 0 0 i I i i i i i I i i I 50 100 Time during stimulus (ms) 25 m s Fig. 1. Differences in intrinsic firing patterns between regularspiking (RS) and fast-spiking (FS) neurons. (A) When stimulated with a suprathreshold step of depolarizing current, RS cells respond with an initial high-frequency spike output that rapidly declines to much lower sustained frequencies. In this and subsequent figures, intracellular voltages are displayed in the top trace, injected current steps in the bottom trace. (B) Under similar conditions F5 cells generate high frequencies that are sustained for the duration of the stimulus. ((:) Temporal patterns of typical RS and t=5 firing displayed graphically. The firing frequencies were calculated from the reciprocal of the interspike interval, and plotted as a function of time during a strong step pulse of intracellular current, as in (,4) and (B). Although each cell initially generated similar spike frequencies (about 320 Hz), the RS cell (solid line) declined within 50 ms to <100 Hz, while the F5 cell (dotted line) actually accelerated slightly to >350 Hz. [(A) taken, with permission, from Ref. 9; (B) and (C) modified from Ref. 10.] it is possible that rare cell types have not yet been described. There is also a dearth of comparative data across both species and cortical areas. Regular-spiking (RS) neurons As their name implies, the neurons most commonly encountered in electrophysiological studies generate what Mountcastle and his colleagues TM were the first to call 'regular' action potentials. Most published intracellular recordings from neocortex in vivo have been from RS neurons (e.g. Refs 5, 6). Individual regular spikes of RS cells are relatively long-lasting, mainly because of a slow rate of repolarization. Each spike is usually followed by a complex set of intrinsically generated afterhyperpolarizations (AHPs) and afterdepolarizations (ADPs). When stimulated at threshold, an RS neuron generates only a single spike and, in contrast to IB behavior, as the stimulus amplitude increases, the first interspike interval decreases as a strong function of current intensity. When presented with prolonged stimuli of constant amplitude, RS neurons exhibit pronounced adaptation of the spike frequency (Fig. 1A, C). spikes with an extracelular electrode. Since the development of in vitro preparations, it has been possible to study the FS neurons using intracellular methods 1°-1z'15 (Agrnon, A., PhD Thesis, Stanford University, CA, 1988). However, recordings continue to be infl'equent and technically challenging. Individual fast spikes usually last less than 0.5 ms. Although spikes of all three neocortical cell types have similar depolarization rates, those of FS cells are faster because of a more rapid rate of repolarization, and each spike is cut short by a deep, relatively brief AHP (Fig. 2B, C). The more prolonged hyperpolarizing (and depolarizing) afterpotentials that characterize RS and IB neurons are not prominent in FS cells. FS neurons are most impressively defined by their repetitive firing pattern; they undergo little or no adaptation during prolonged intracellular current pulses (Fig. 1B, C). Indeed, when strongly stimulated, they can sustain spike fl-equencies of at least 500-600 Hz for hundreds of milliseconds. FS cells are thus able to perform a relatively faithful conversion of input to output over a wide dynamic range; in strong contrast to RS and IB cells, the output of these neurons is likely to retain the precise temporal features of their input (Fig. 1C; cf. FS and RS firing patterns). Fast-spiking (FS) neurons Mountcastle et al. 13 also described rarely encountered neurons with relatively 'thin' (i.e. brief duration) extracellular spikes in monkey cortex. Nine years Intrinsically bursting (IB) neurons later, Simons TM made similar observations in rat IB neurons are distinguished by the tendency for neocortex and called these 'fast' spikes. Both groups their spikes to appear in a stereotyped, clustered remarked on how difficult it was to record thin/fast pattern, the burst 8-u. Bursts are often the minimal 100 TINS, Vo/. 13, No. 3, 1990 response to a just-threshold intracellular stimulus (Fig. 3A). The individual spikes of IB cells are quite similar to those of RS cells, although they are often followed by more prominent ADPs which can summate to form a slow, low-amplitude depolarizing wave during a burst. Within a burst, each successive spike usually declines in amplitude, presumably because the sustained depolarization inactivates sodium conductances (Fig. 3A-C). When presented with a simple, prolonged intracellular stimulus, IB neurons can respond with a complex, often periodic pattern of bursts and single spikes (Fig. 3B-D). Non-IB cells under the same conditions respond with monotonic frequency patterns (cf. Fig. 1). Specific patterns of intrinsic bursting may vary widely from one IB neuron to another. The IB neurons first described 8'1° were found either in layer IV or upper layer V of parietal and cingulate neocortex of guinea-pigs. These cells generally responded to a prolonged current pulse with one burst followed by repetitive single spikes. More recent studies in primary somatosensory cortex of mice9 and rats 16 have revealed robust IB cells restricted to layer Vb. Many of these can generate rhythmic bursts in the range of 5-15 Hz9'16 (Fig. 3C, D). The term 'burst' is commonly used in the neurophysiological literature to describe a neuron's firing behavior. However, when classifying cells, the designation 'bursting cell' is ambiguous unless modified to indicate whether the firing pattern reflects an inherent property of the cell, or a direct response to its synaptic drive. As used in this review, intrinsic bursting refers to a cell's tendency to generate dusters of high-frequency spikes solely as a manifestation of its intrinsic membrane properties, and independent of its synaptic input. Almost any neuron might produce dusters of spikes in response to phasic synaptic input. However, this response pattern does not, of itself, justify dassifcafion as an IB neuron. The distinction between intrinsic bursts and synaptically driven bursts, which is essential for understanding a neuron's role within a circuit, cannot be made without methods that isolate the activity of a cell from its extrinsic connections. This may account, in part, for the fact that IB neurons were only recently recognized in neocortex 8,m. Moreover, IB neurons are relatively rare and are usually encountered only in certain laminae. C o r r e l a t i o n s b e t w e e n i n t r i n s i c p h y s i o l o g y and morphology By impaling neocortical neurons with micropipettes that contain an intracellular dye (such as Lucifer yellow, horseradish peroxidase or biocytin) it has been possible to correlate intrinsic physiological properties, as described above, with the traditional classifications of the Golgi anatomists, as determined by light microscopy. There is now a relatively consistent and accepted scheme for the morphological categorization of cortical neurons 1,2. Virtually all neocortical neurons can be placed into one of two groups: pyramidal cells, which have a high density of dendritic spines, prominent apical dendrites, excitatory synaptic function, and an axon that projects out of the cortex as well as locally; and non-pyramidal cells, most of which have smooth or sparsely spined TINS, VoL 13, No. 3, 1990 40 mV A 8 mV s Cell1 CeB2 3mM DGG B L 20 mV s Cell3 Cell4 20 p~ Bicuculline C 20 mV • ms i" ! Fig. 2. Neurons that elicit EPSPs have more prolonged action potentials than neurons that elicit IPSPs. Pairs of neurons were recorded intracellularly in microcultures of rat neocortical neurons. One cell was stimulated with intracellular current to generate an action potential, while the resulting synaptic response was monitored in the second cell. (,6,) In this pair, a spike in cell I yielded a monosynaptic EPSP in cell 2. Application of the excitatory amino acid antagonist 7-o-glutamylglycine (DGG) blocked the EPSP. (B) In a different pair of neurons from those shown in (A), a spike in cell 3 yielded a monosynaptic IPSP in cell 4. Application of bicuculline, a GABAA receptor antagonist, blocked the IPSP. (C) Action potentials of seven excitatory (E) and seven inhibitory (I) presynaptic neurons superimposed to show the differences in spike duration. (Figure modified from Ref. 12.) dendrites of various configurations, inhibitory (GABA-mediated) synaptic function, and axons that arborize only locally within the cortex. One major subtype, found in layer IV of many primary sensory areas, is a class of small neurons usually called spiny stellate cells 17. With their high spine densities and presumed excitatory synaptic function, spiny stellate cells are very similar to pyramidal cells. However, their axons usually do not leave the cortex. Pyramidal cells constitute the majority of neurons in the neocortex (60-70%). Their sizes range from some of the smallest to the very largest cortical cells, and their somata can be found in layers II through VI. 101 A D 40 mV 1 nA J 25 m s B 20 mV L_ 2 nA 20- J 50 m s L v ¢J =~ 1o ~3 C I. t, I _J I I I I I I 200 400 600 Time during stimulus (ms) L_ 50 m s Fig. 3. Diverse firing patterns of intrinsically bursting (IB) neurons. (A) Single intrinsic bursts evoked by threshold pulses of intracellular current in a neuron from guinea-pig sensorimotor cortex. Slightly larger currents generated a very similar burst at much shorter latency. Figure shows three superimposed traces. Voltage and current calibrations in (A) apply also to (B) and (C). (B) Response of an IB cell in mouse SI cortex to prolonged current stimulus. Sequence of bursts and single spikes terminating with a train of single spikes. Arrowhead points to a prominent ADP following a single spike. (C) Repetitive intrinsic bursting in response to prolonged stimulus. Mean interburst frequency was about 9 Hz. Recording site was near the border of layers V and Vl in mouse SI cortex. (D) Transition from rhythmic bursting to single-spiking in a deep layer V neuron from rat 51 cortex (top). Each burst consisted of three spikes firing at about 250 Hz. The graph plots the response: the neuron generated four bursts at 10-I 1 Hz (interburst frequencies plotted by broken line), then abruptly changed to single-spike firing at 15-20 Hz (plotted by solid line). Each triplet of vertical lines represents an interburst interval Each single vertical line is an interspike interval [(A) taken, with permission, from Ref. 10; (B) and (C) taken, with permission, from Ref. 9; (D) provided by Y. Chagnac-Amitai, L. R. Silva and B. W. Connors, see Ref. 17.] It is therefore not surprising that this is the cell type most frequently encountered by dye-containing microelectrodes. In recordings from neocortex both in vitro and in vivo, almost all RS cells that have been morphologically identified have been pyramidal cells (e.g. Refs 10, 18, 19). Most IB cells have also been identified as pyramidal cells; however, their somata are restricted to layers IV and V 1°' 16. A preliminary study of layer IV spiny stellate cells indicates that they show the same spectrum of physiological properties as layer IV pyramidal cells; some are RS cells while others burst intrinsically (Gutnick, M. J. and White, E. L., unpublished observations). That neurons of different morphological classes can show similar intrinsic physiological characteristics, and vice versa, implies that these parameters are not causally related. Nonetheless, there is recent evidence that they can be consistently correlated. Chagnac-Amital et al.20 have found that although RS 102 and IB cells in rat layer Vb are all pyramidal neurons, the IB cells have larger somata, more dendritic branches, and markedly different patterns of mtracortical axons. This latter difference may be an important clue to the function of these IB cells within the cortical circuit; while the RS cells all had axon collaterals that ascended and branched profusely within supragranular layers, the axons of IB cells tended to stay within infragranular layers, and often extended far in the horizontal dimension. Two primary lines of evidence imply that the elusive FS cells are GABAergic non-pyramidal cells. First, all neurons with the physiological properties of FS cells that have also been intracellularly stained have had the somatic size and shape, and the smooth or sparsely spiny dendritic morphologies that are peculiar to cells containing GABA or its associated enzyme systems 1°. Second, paired recordings in cultures of dissociated neocortical neurons have TINS, Vol. 13, No. 3, 1990 I Iv}jL II Ill V 1.~ IB t__ RS l J _1 1~ FS VI WM Fig. 4. Schematic summary of correlations between intrinsic physiology and anatomy of rodent neocortical neurons. RS neurons (open symbols) are spiny cells, either pyramidal or stellate, distributed through layers II through Vl. F5 neurons (filled symbols) are aspiny or sparsely spiny non-pyramidal cells, with presumed GABAergic inhibitory function, also distributed through layers II through Vl. IB neurons (shaded symbols) are restricted to layers IV and V, and are also spiny cells of pyramidal or stellate morphology. Neurons of layer I have not been studied physiologically. Braces on the right summarize the laminar distributions of the three neuron types, and illustrate a typical firing pattern of each. WlV1, white matter. shown that cells generating monosynaptic, GABAmediated IPSPs onto follower neurons have significantly faster spikes than those cells generating monosynaptic EPSPs 12 (Fig. 2). It is still quite possible that there exist types of neocortical GABAergic neurons that are not FS cells, as suggested for the hippocampus 21. However, the data strongly suggest that every FS neuron encountered in the neocortex is a GABAergic inhibitory cell. The data are too scant to ascertain whether classification into three types of neurons on the basis of intrinsic physiological properties is generally applicable to all cortical areas and all species. Analogous neuronal classes have been described for the dorsal cerebral cortex of turtles, where pyramidalshaped cells generate RS-like or IB-like activity and non-pyramidal interneurons generate FS-like activity22. Thus it is likely that this separation arose early in forebrain evolution, and may now be widespread. All three classes have been repeatedly observed in rodent neocortex (i.e. mice, rats and guinea-pigs), as described above. Mountcastle's original description of FS and RS neurons applied to monkey neocortex, and human neocortex also has both FS and RS cells (Ref. 23; McCormick, D. A., unpublished observations). The prevalence of IB cells across species is less well described. They have not been observed in extensive investigations of layer V neurons in cat sensorimotor cortex in vitro18; however, these studies targeted only the largest (presumed Betz) cells by using microelectrodes with large tip diameters. An earlier study of cat pyramidal tract cells in vivo described some features of the rhythmic IB cells seen in rodents (see Fig. 6C in Ref. 6). There are at least two preliminary reports of IB TINS, VoL 13, No. 3, 1990 neurons in human neocortex 23'24. Figure 4 schematically depicts the general distribution of neuron classes in neocortex, based largely upon studies in rodents. There are several morphological classes of neocortical neurons whose intrinsic physiological properties have not yet been examined 2. These include the small population of non-GABAergic, non-pyramidal neurons (notably peptidergic bipolar cells) and the assorted enigmatic neurons of layer I, many of which are GABAergic. Also, neurons of layer VI have been only sparsely studied. Finally, it would be of great interest to know whether the diversity of anatomy and biochemistry among GABAergic neurons 2 is paralleled by a diversity of intrinsic firing patterns 21. Significance of diverse intrinsic firing patterns in neocortex The intrinsic physiological properties of a neuron's membrane play a central role in determining (1) how it transforms the information it receives into an output pattern, (2) how these transformations are modulated by humoral or environmental factors, and (3) whether (and with what pattern) the neuron generates spontaneous activity. Since these properties can vary widely from neuron to neuron, knowledge of the quirks of each cell type is an essential step in unraveling the functions of a neural circuit 25. In the neocortex, it is evident that RS cells will attenuate prolonged excitatory stimuli while favoring the transmission of phasic ones; by contrast, FS cells offer a wide-band responsiveness and, if necessary, sustained high-frequency output. The complexities of IB cell behavior suggest more varied possibilities. Near threshold for firing they have very high gains, 103 Acknowledgements We thank our colleaguesAric Agmon, Yael Chagnac-Amitai, Alon Friedman, David McCormick, David Prince, LaurieSilva and Ed White for their invaluable contributions to the work describedhere. We also thank Larry Cauller for comments on the manuscript, and J. E. Huettner and R. W. Baughman for permission to reproduce their data in Fig. 2. Theauthors' studies were supported by grants from the NIH and the Klingenstein Fund (BWC) and the DFG (MJ6). 7 Stafstrom, C. E., Schwindt, P. C. and Crill, W. E. (1984) generating either a very large output or none at all in J. Neurophysiol. 52, 264-277 response to small increments of input amplitude. 8 Connors, B. W., Gutnick, M. J. and Prince, D. A. (1982) When coupled to each other synaptically IB cells may J. Neurophysiol. 48, 1302-1320 serve as initiators of synchronized cortical activities 15. 9 Agmon, A. and Connors, B. W. (1989) Neurosci. Lett. 99, Those IB cells with a tendency to oscillate may play a 137-141 pivotal role in cortical rhythmogenesis. During dif- 10 McCormick, D. A., Connors, B. W., Lighthall, J. W. and Prince, D. A. (1985) J. Neurophysiol. 54, 782-806 ferent behavioral states, various neurotransmitters 11 Wolfson, B., Gutnick, M. J. and Baldino, F. (1989) Exp. Brain would be expected to modulate the intrinsic firing Res. 76, 122-130 patterns of RS, and presumably IB and FS, neurons 12 Huettner, J. E. and Baughman, R. W. (1988) J. Neurosci. 8, 160-175 by altering their intrinsic membrane properties 26'27. Many important questions remain. We are just 13 Mountcastle, V. B., Talbot, W. H., Sakata, H. and Hyvarinen, (1969) J. NeurophysioL 32,452-484 beginning to understand the varied functional prop- 14 J.Simons, D. J. (1978) J. Neurophysiol. 41,798 erties of neocortical neurons and how they correlate 15 Chagnac-Amitai, Y. and Connors, B. W. (1989) J. Neurowith other cellular features. The ionic mechanisms of physiol. 62, 1149-1162 intrinsic firing patterns 7'8'18 and their sensitivity to 16 Silva, L. R., Chagnac-Amitai, Y. and Connors, B. W. (1989) Soc. Neurosci. Abstr. 15, 660 neuromodulators 26'27 have only been extensively 17 Lund, J. S. (1984) in Cerebral Cortex, Vol. I, Cellular examined in RS neurons, and there is an urgent need Components of the Cerebral Cortex (Peters, A. and Jones, to extend these analyses to IB and FS cells. We also E. G., eds), pp. 255-308, Plenum Press need more extensive comparative data, across both 18 Stafstrom, C. E., Schwindt, P. C., Flatman, J. A. and Crill, W. E. (1984) J. Neurophysiol. 52,244-263 species and neocortical areas, to establish the generLandry, P., Wilson, C. J. and Kitai, S. J. (1984) Exp. Brain Res. ality of the principles outlined here. As computational 19 57, 177-190 approaches to the functions of neocortex advance 28, 20 Chagnac-Amitai, Y., Luhmann, H. and Prince, D. A. J. Comp. many realistic models will need to incorporate not only Neurol. (in press) the morphological and biochemical aspects of cortical 21 Lacaille, J-C. and Schwartzkroin, P. A. (1988) J. NeuroscL 8, 1400-1410 connectivity, but also the specific biophysical charac22 Connors, B. W. and Kriegstein, A. R. (1986) J. Neurosci. 6, teristics of each class of neuron. Selected references 1 White, E. L. (1989) Cortical Circuits: 5ynaptic Organization of the Cerebral Cortex - Structure, Function and Theory Birkh~user 2 Peters, A. and Jones, E. G. (1984) Cerebral Cortex, Vol. 1, Cellular Components of the Cerebral Cortex Plenum Press 3 Hodgkin, A. (1948) J. Physiol. (London) 107, 165-181 4 Llin~s, R. (1988) Science 242, 1654-1664 5 Takahashi, K. (1965) J. Neurophysiol. 28, 908-924 6 Calvin, W. H. and Sypert, G. W. (1976) J. Neurophysiol. 39, 420-434 164-177 23 Masukawa, L. M., Strowbridge, B. W., Kim, J., Spencer, D. D. and Shepherd, G. M. (1987) Biophys. J. 51, 67a 24 Frederick, D., Wilson, C. J., Wyler, A. J. and Foehring, R. C. (1989) Soc. Neurosci. Abstr. 15, 1309 25 Getting, P. (1989) Annu. Rev. NeuroscL 12, 185-204 26 McCormick, D. A. (1989) Trends NeuroscL 12,215-221 27 Foehring, R. C., Schwindt, P. C. and Crill, W. E. (1989) J. NeurophysioL 61,245--256 28 Koch, C. and Segev, I. (1989) Methods in Neuronal Modeling. From Synapses to Networks The MIT Press Hormonalcontrola neuropeptidegene expressionin sexuallydimorphicolfactorypathways R i c h a r d B. S i m e r l y RichardB. Simedyis with the Howard HughesMedical Institute at TheSalk Institute for Biological Studies, LaJolla, CA 92037, USA. An abundance of experimental literature has established that gonadal steroid hormones are responsiblefor the sexual differentiation of neural circuitry, mediating a rarity of reproductive behaviors and physiological mechanisms. These same hormones regulate the expression of reproduetivefunction in the adult and may influence the responsiveness of the brain to specific olfactory cues. The recent demonstration that the expression of the neuropeptide cholecystokinin is activationally regulated by estrogen at the mRNA level, within a sexually dimorphic p@ulation of neurons in the medial amygdala, suggests a possible cellular mechanism for the hormonal modulation of olfactory information relayed along the vomeronasal pathway to the h39othalamus. Gonadal steroid hormones influence manlmalian reproductive function in two fundamental ways. First, during the perinatal period these hormones can permanently alter the pattern of copulatory behavior and gonadotropin secretion expressed during adulthood, and second, the expression of these sexually 104 dimorphic functions in the adult is dependent on adequate levels of circulating gonadal hormones. In the rat, perinatal exposure to differential levels of gonadal steroids determines, at least in part, whether the mounting behavior that is typical of male rats, or the lordosis response typical of females, will be displayed by adult animals 1. In a similar way, perinatal steroids determine whether the adult pattern of gonadotropin secretion will be cyclic (with the periodic surges of luteinizing hormone that lead to ovulation in female rats), or relatively constant as it is in male animals 2. Furthermore, gonadectomy reduces copulatory behavior in both mature male and female animals, and subsequent hormone treatment restores, or 'activates' these behaviors in response to appropriate sensory cues. The discovery that the morphological organization of certain brain regions thought to mediate reproductive function is sexually dimorphic led to the now widely accepted idea that functional sexual differentiation has an anatomical basis 3'4. Thus, gonadal steroids permanently alter the morphology and connections of certain groups of © 1990,ElsevierSciencePublishersLtd,(UK) 0166-2236/90/$02.00 TINS, Vol. 13, No. 3, 1990