Comparison of conventional and newly recognized mechanisms for generation of action potentials and responses to transmitters. A: conventional mechanisms. When excitatory postsynaptic potentials (EPSPs; left) are large enough, threshold for spike initiation is exceeded, triggering the voltage‐dependent conductances that elicit rising and falling phases of action potential (middle). Classic EPSP is due to opening of nonspecific ion channels that allow flow of cations (Na⁺, K⁺). Membrane potential is driven (thin arrows) toward equilibrium potential (E) for EPSP (E_EPSP), usually around −20−0 mV. Spike results from driving membrane potential toward sum of E_Na, E_K, and E_Ca. Afterhyperpolarization (AHP) is produced by opening of voltage‐dependent or Ca²⁺‐dependent K⁺ channels, driving membrane potential toward E_K. Classic inhibitory postsynaptic potential (IPSP; right) is generated by opening of Cl⁻ or K⁺ channels, driving membrane potential toward E_Cl or E_K (usually −70 to −100 mV). RMP, resting membrane potential; g, conductance. B: newly recognized synaptic mechanisms. Several alternative mechanisms have been proposed as the basis for synaptic events, including those schematized here. Some transmitter actions might be elicited by activation of electrogenic pumps (left); for example, pumping Na⁺ or Ca²⁺ out of cell would hyperpolarize membrane, whereas pumping K⁺ out would depolarize it. However, there is little strong evidence for this type of transmitter action. Generation of synaptic potentials by reducing ion conductance is on more solid experimental ground. Thus the slow EPSP in sympathetic ganglia may be partly due to closing of K⁺ channels, thus driving membrane potential away from E_K (thick arrows) and toward equilibrium potential for next most diffusible ion(s). Slow IPSP might result from closing of Na⁺ channels (however, see chapter by North in this Handbook), driving membrane potential away from E_Na and toward that of next most important ion (K⁺). In addition at least 1 class of ion channel associated with slow EPSP is a voltage‐dependent K⁺ channel: the M‐channel. M‐channel is only open over a finite range of potentials (probably −10 to −40 mV). Other voltage‐dependent loci of transmitter action are the conductances responsible for action potential or associated afterpotentials (right). Thus transmitters alter either voltage‐dependent Ca²⁺ conductance of spike, voltage‐dependent K⁺ conductance, or Ca²⁺‐dependent K⁺ conductance. Such actions would either shorten or lengthen Ca²⁺ component of spike, or decrease or augment AHP after spike.

Sample records from cultured CNS neuron illustrating types of events that can be measured with extracellular (top trace), intracellular (middle trace), and single‐channel (bottom trace) recording techniques. For extracellular recordings, relatively large‐tipped microelectrode is placed close to neuronal surface, and neuronal activity is measured from an extracellular vantage point. This technique can detect action potentials but not subthreshold changes in membrane potential. Extra‐cellular recordings provide information about firing rate and pattern and type of event generating pattern. Recorded activity reflects both synaptic and intrinsic influences, single events being generated by integration of excitatory and inhibitory synaptic input, passive and active cable properties of membrane, endogenously generated activity (if present), and any metabolic or second‐messenger influences. For intracellular recording, fine‐tipped microelectrode is inserted through membrane and used to measure potential changes at membrane level. This technique can detect action potentials, synaptic potentials, and subthreshold changes in membrane potential. For single‐channel recording, relatively large‐tipped microelectrode is placed in close contact with neuronal membrane and measures activity of individual ion channels in membrane patch. Channel activity appears as current flow through open channel. Transitions between open and closed states are instantaneous, producing boxlike events. This technique provides information about individual ion channels that underlie chemical and electrical excitability and produce the events recorded with intracellular and extracellular recording techniques.

Concentrations of ionic species responsible for electrical excitability, in extracellular and intracellular fluids of CNS. Values are expressed as mM in approximate range most often reported (see refs. 193,417). Width of channel or pore (gap in membrane) reflects relative resting permeability (p) for these ions. Note low permeability and intracellular concentration for Ca²⁺.

Simple equivalent electrical circuit analogy of neuronal membrane. Battery (E) generates unequal distribution of electrical charge (positive outside, negative inside), in analogy to action of ion pumps and selective ionic permeability. Resistor (R) symbolizes resistance (limited permeability or conductance) in membrane to flow of ions across it. Capacitor (C) signifies capacity of membrane to store a charge for finite periods of time.

Ion channels and related processes contributing to electrical activity of excitable cells. Most activity is generated by flow of ions through such conductance channels. Channels that are always open (top) are termed leak channels; they largely generate the resting membrane potential (RMP). Receptor‐mediated channels are depicted on right. These channels are activated or inactivated by chemicals (mostly hormones or neurotransmitters) and are thought to account for conventional non‐voltage‐dependent responses evoked by activation of synaptic pathways. Traditional voltage‐dependent channels (left) open only at certain membrane potentials; they contribute to the generation of an action potential, after the membrane potential is brought to threshold (trigger) level of depolarization by injected current or chemical activation of passive conductances (see Fig. 1A). One result of voltage‐sensitive conductance is shown at bottom left; entry of Ca²⁺ during action potential triggers efflux of K⁺, resulting in membrane repolarization (and sometimes hyperpolarizing afterpotential) at conclusion of action potential. Such ion‐sensitive channels are opened only when particular ions (e.g., Ca²⁺) are present. Existence of voltage‐sensitive, receptor‐mediated ion channels (bottom left) is a relatively new concept; as with conventional synapses, such channels may be opened or closed by neurotransmitters, but only at certain membrane potentials. Because many voltage‐dependent conductances are associated with action‐potential mechanisms, activation of such receptors might be expected to alter properties of the spike or its afterpotentials. Activation or inhibition of electrogenic ion pumps (bottom right) could also contribute to receptor‐ or nonreceptor‐coupled changes in membrane potential. Generation of cyclic nucleotides by nucleotide cyclases, possibly through activation of transmitter receptors, could open or close ion channels directly (or perhaps via protein phosphorylation) or alter voltage‐sensitive conductances or membrane pumps, thus significantly altering neuronal excitability. Equivalent channel effects or mechanisms could apply to presynaptic actions of transmitters (see nerve terminal, top right).

Possible interactions of cAMP‐mediated synaptic receptor mechanisms and their consequent actions on neuronal physiology. Neuron receives synaptic contact from dopaminergic and noradrenergic terminals. Both types of terminals have been described on separate cell types but not on same cell. Both catecholamines activate adenylate cyclase, dopamine (DA) via DA receptor, and norepinephrine (NE) via β‐receptor. cAMP that is formed and survives phosphodiesterase catabolism can activate cAMP‐dependent protein kinases. Two such kinase actions are illustrated: 1) to phosphorylate or dephosphorylate proteins in synaptic membrane, which could then alter resistive or capacitative properties of membrane (same symbols at bottom left as in Fig. 4); and 2) to phosphorylate nuclear histones, thereby changing extent to which genetic properties are expressed in current metabolic properties of cell. Latter mechanism might also alter electrical or pharmacological properties of cell membrane through classic DNA‐RNA protein route over longer time course of action. Prostaglandins of E series inhibit responses to NE in hippocampal pyramidal neurons 685 and in cerebellar Purkinje neurons 347,711; prostaglandins potentiate DA in caudate nucleus 712,714. Antipsychotic drugs can block both NE and DA, depending on target cell.

Traditional methods for assessing changes in ionic conductance produced by neuromessengers. A: (left) brief injection (20–200 ms) of rectangular negative current (I) pulse through recording barrel is recorded as hyperpolarizing voltage (V) deflection, which, after nulling of effects of current on electrode resistances with balanced‐bridge circuit, reflects impedance properties of membrane. Slowness of falling and rising phases of this voltage pulse (electro‐tonic potential) represents time constant of cell, which is proportional to product of specific membrane resistance and membrane capacitance. Amplitude of electrotonic potential is proportional to membrane resistance; this value and that of input current allow calculation of input resistance (sum of membrane resistance for whole cell membrane and other, smaller resistances, e.g., that of cytoplasm). If such a current pulse is injected repetitively at regular intervals (1–5 s) and if recording time base is retarded, pulses appear as short downward deflections (A, right). Application of conventional inhibitory transmitter, such as GABA, during sequence reveals expected hyperpolarization, and resultant reduction in size of downward deflections suggests a reduction of input resistance (increase of membrane conductance). However, this simple method may not be adequate for analysis of drug effects on those neurons that rectify in response to injection of some current values or with drugs having voltage‐dependent effects. In both cases, nonlinear I‐ V curves would result. In these cases, construction of I‐V curves before and during messenger administration is preferable and also allows estimation of the reversal potential for messenger effect (see Fig. 8). B: multiple pulses of both polarities and various intensities (usually from 0.05 to 2 nA) are injected into cell at regular intervals, and resultant electrotonic potentials are often displayed as shown on left, using repeated fast sweeps of oscilloscope. “Sag” in the electrotonic potential seen after earlier “hump” is representative of certain central neurons such as hippocampal pyramids (see refs. 601,717,718) and probably reflects time‐dependent activation of some current. Plotting the size of these potentials (either at peak of hump or at steady state) against current used to generate each potential yields I−V curve with reproducible slope and shape (right). Repetition of this procedure during application of messenger may then reveal change in position and slope of curve: lowering of curve indicates hyperpolarization; reduction of slope (as schematized for GABA effects) indicates reduced input resistance, or increased conductance. Increased slope signifies reduced conductance. Note that control curve intercepts GABA curve at ca. −75 mV, which would be reversal potential for hyperpolarizing GABA response. Use of K‐acetate‐filled electrode is assumed.

More sophisticated methods for assessing changes in ionic conductance produced by neuromessengers. A: using current clamp, membrane potential is experimentally altered during transmitter action to determine potential at which there is no net ion flow and therefore no change in membrane potential, even though channels are open. This potential (produced by injecting steady DC into cell and assuming pipette contains no Cl⁻) is the reversal potential for response. Ideally, reversal potential occurs at the equilibrium potential for ions (here, probably Cl⁻) mediating response (ca. −75 mV for GABA, curve on right), but under certain circumstances, such as a remote location for active site, technical problems limit the agreement between equilibrium potential for ions involved and reversal potential for measured response. B: voltage clamp. In this example, voltage is held steady at −40 mV and current is measured while GABA is applied. Downward current deflections indicate inward current produced by command voltage pulses. Note that GABA application causes outward current (due to Cl⁻ influx) and increase in command current pulses (due to increased conductance). Repetition of this paradigm while clamping membrane potential at different voltage levels would indicate the reversal potential (shown on I‐V curve on right) to be the potential (−75 mV) at which GABA no longer could elicit current flow.

Theoretical explanation of the method of fluctuation or noise analysis. A: top traces show activity of 10 computer‐simulated channels. Each channel undergoes transitions between open and closed states as a result of a Poisson process. All 10 channels operate independently but have same amplitude and average lifetime in open state. Bottom trace shows summed activity of these 10 channels. Fluctuating signal is produced; it reflects moment‐to‐moment variation in number of open channels present. B: typical observations made during application of fluctuation analysis to study of synaptic channels in biological membrane. Under voltage‐clamp conditions, application of agonist during period indicated by vertical bar results in change in DC membrane current (I). In addition variance of membrane current is seen to increase from prior to agonist to during agonist. This increase is readily apparent in AC trace, which represents condenser‐coupled, amplified varion of DC trace. Additional variance is assumed to reflect moment‐to‐moment changes in number of synaptic channels opened by agonist. C: kinetic properties of these channels can be estimated from power spectral density (PSD) of agonist‐induced current fluctuations. It is assumed that simple kinetic scheme of type used to generate simulated channels in A also controls operation of synaptic channels. Under this assumption, PSD of biological current noise is expected to be of Lorentzian form (C, smooth curve). Mean open time of agonist‐induced channels can then be calculated from half‐power frequency (f_c, arrow) of Lorentz curve that affords best fit to observed spectral point. S(f), power spectral density function.

Various patch‐clamp recording configurations (top) and representative single‐channel recordings from a patch on cultured CNS neuron (bottom). Single‐channel recordings can be made in 3 configurations: cell‐attached, inside‐out, or outside‐out patch. In addition membrane patch can be broken and recording of whole‐cell voltage or current obtained (whole‐cell recording). Single‐channel activity (bottom) is recorded as current flow through open channel and appears as boxlike events. In this patch (cell‐attached) at least 2 channel types were present, as indicated by amplitudes of events. Brief upward events are channel openings that were not fully resolved because of limitations of recording equipment. Open (O) and closed (C) states for largest events are indicated. A 70‐mV depolarization was applied to membrane patch. Analysis of channel activity in this patch indicated that both channel types were K⁺ channels, 1 with a single‐channel conductance of 100 pS (•) and the other of 20 pS (•).

Simplified coronal section of spinal cord. Spinal cord neurons were 1st mammalian central neurons studied intracellularly (see refs. 97,225,226), and considerable electrophysiological data are still derived from various types of spinal cord (or associated spinal ganglionic) preparations. Spinal cord circuits are often considered to be relatively simple, yet more recent data show them to be extremely complicated. Diagram is greatly simplified to highlight neurons and pathways most often studied by synaptic physiologists and pharmacologists. Cell body and fiber locations are shown on left; laminae and other regions of gray matter are numbered on right. Gray matter of the spinal cord consists of dorsal and ventral horns; these are surrounded by fiber tracts (white matter) of ascending and descending fibers to and from brain and other spinal regions (not shown). Sensory information enters spinal cord via dorsal roots, along axons of the monopolar neurons whose cell bodies lie in dorsal root ganglia (dorsal root ganglia neurons). These axons then enter spinal cord and terminate either on neurons in ventral horn or on 1 of several cell types in the various laminae of dorsal horn. Those axons (from muscles) terminating on ventral horn neurons follow 1 of 3 projections: 1) monosynaptically, from Ia afferents onto agonistic motoneurons via excitatory terminals, with collaterals onto inhibitory interneurons that project to antagonistic motoneurons (constituting stretch reflex); 2) disynaptically, from Ib afferents that terminate on excitatory interneurons projecting to agonistic motoneurons, with collaterals terminating on inhibitory interneurons that project to antagonist motoneurons [inverse myotatic reflex, not shown; (see refs. 121,694)]; or 3) mono‐ or disynaptically from group II (muscle spindle) fibers that project either directly onto agonistic motoneurons or through inhibitory and excitatory interneurons to agonistic and antagonistic motoneurons (see refs. 121,694). Those pathways (e.g., from skin) that project onto dorsal horn neurons in turn transmit information in 1 of 2 ways: 1) fibers from dorsal horn cells cross midline and then ascend contralaterally to thalamus through spinothalamic tract or ascend in ipsilateral dorsal column to brain stem (not shown); or 2) dorsal horn cells receiving terminations of Aδ−, Aβ‐, and C‐fibers (from mechano‐, thermo‐, and pain receptors) project, probably through other interneurons, both to ipsilateral and contralateral motoneurons in ventral horns, and to spinothalamic tracts. Pathways to the motoneurons constitute part of flexor reflex, in which ipsilateral motoneurons to flexor muscles are activated by stimuli (e.g., pain), while the motoneurons innervating extensor muscles are inhibited. Reverse situation prevails contralaterally: flexors are relaxed and extensors activated. Sensory fibers entering the spinal cord through the dorsal roots probably use glutamate (Glu) (Ia, Ib, and II afferents), substance P, somatostatin (SS), vasoactive intestinal peptide (VIP), and cholecystokinin (CCK) (C‐fibers); the various interneurons in dorsal horn probably contain GABA and enkephalin, among others. In addition some descending fibers terminating in dorsal horn contain NE, DA, and serotonin [see the chapter by Zieglgänsberger in this Handbook; 481,501,694]. Several phenomena of pain perception probably involve integration in the dorsal horn substantia gelatinosa (SG). Elucidation of the interconnections of various interneurons and the histochemical delineation of presence of several putative neurotransmitters in this sensory region of spinal cord should lead to a greater understanding of nociception in general. Several ventral horn transmitters are also known. Thus motoneurons are cholinergic, Renshaw cells (RC) and some other inhibitory interneurons probably use glycine (Gly) as their transmitters, and others probably use GABA (see refs. 21,159,160,162,163,164,694). Glu or aspartate (Asp) would seem likely candidates as transmitters for some of the excitatory interneurons (see refs. 161,694). IN, interneuron; γMN, γ‐moto‐neuron; and αMN, α‐motoneuron.

Simplified coronal section of hippocampus. Hippocampal formation is a well‐defined cortical structure extensively studied with anatomical, biochemical, and electrophysiological techniques (see refs. 98,465,611,758). We include Ammon's horn and the dentate in the hippocampus proper. Neuronal elements comprising the hippocampus are organized into well‐defined lamellar structures that are maintained throughout its extent. Principal neurons of the hippocampal formation are pyramidal (Pyr) neurons that are arranged in a linear fashion in Ammon's horn. Pyr cell layer has been subdivided into areas CA1–CA4 based on anatomical and physiological differences between Pyr neurons and the synaptic organization within these regions 98,393,465,611. However, some researchers do not consider CA2 and CA4 as distinct areas. Pyr neurons receive input from local interneurons, from each other, and from other CNS regions via fornix (leftmost input pathway), from medial septum, the perforant pathway from entorhinal cortex and the alvear commissural pathways from the contralateral hippocampus. Axons of the Pyr neurons form the only efferent pathway leaving the hippocampus. Axon collaterals from CA3 Pyr neurons, called Schaffer collaterals, also provide excitatory input to other CA3 Pyr neurons and to interneurons and Pyr neurons in CA1 and CA2 region of the hippocampus. Excitatory amino acids (Glu or Asp) are putative transmitters for this pathway (see refs. 141,143,758). Other neuron types have been identified in layers adjacent to the Pyr cell layer. Those most well characterized are basket (B) cells 393,465,611 that provide inhibitory input to the Pyr neurons. B cells are activated by axon collaterals originating from nearby Pyr neurons or from distant or even contralateral Pyr neurons and are thought to use GABA as their transmitter (see refs. 10,13,19,393,758). Some neuropeptides (e.g., opioids, VIP, CCK, and SS) have been identified within other hippocampal interneurons, but their function is not clear (see Peptides, p. 61, and Other Peptides …, p. 74). Serotonergic, noradrenergic, and cholinergic fibers have also been described in hippocampus (see refs. 72,73,402,451,452,467,496,500,501,638,695,746,756,758). Targets for these fibers are not yet fully characterized, although it seems likely from light‐microscopic studies that NE projects to CA3 Pyr cells and dentate granule cells (Gr) and that cholinergic fibers contact CA1 and CA3 Pyr neurons 406,451,452,496,500,501,695,746. Principal neurons in the dentate are intrinsic excitatory neurons known as Gr cells. These cells receive afferent input from fibers in the perforant and commissural pathways. Gr cells send axons to Pyr neurons and interneurons in hilar, CA3, and CA4 regions. Several putative neurotransmitters have been suggested for Gr cells, including amino acids (see refs. 142,143), opioids 280,492,737,738, and other peptides (see refs. 737,738). Catecholaminergic and serotonergic inputs are also present in dentate 406,500,501. M, molecular layer.

Principal neurons and afferents in cerebellar cortex. The cerebellum is another laminated region that has been extensively studied anatomically, biochemically, and physiologically. Its neuronal organization and synaptic connections have been well characterized 98,227,574 and are similar throughout the structure. In the cortical region, 5 neuronal types are present: Purkinje neurons (P), basket cells (Ba), stellate cells (St), Golgi cells (Gg), and granule cells (Gr). Gr cells are the only excitatory neurons; all others are inhibitory. Parallel fibers, formed by axons of Gr cells, provide excitatory input to the other 4 types of neurons in cerebellar cortex. Glu and Asp are putative transmitters for the Gr cells [see EXCITATORY AMINO ACIDS …, p. 34; 141,142,741]. Ba and St cells provide inhibitory input to the P neurons, whereas Gg cells provide feedback inhibition to Gr cells (not shown). P cell axons are the only efferent pathway from cerebellar cortex and provide inhibitory input to neurons of deep cerebellar nuclei (DCN), whose axons in turn exit the cerebellum for other regions of CNS. GABA is strong transmitter candidate for all of the inhibitory neurons in cerebellum, including P cell (see refs. 70,826). At least 3 afferent pathways transmit information from other regions of CNS to cerebellar cortex. Climbing fibers (CFs), which are axons originating from cell bodies in inferior olive, provide powerful, bursting excitatory input to P cells. CF axon collaterals also innervate the other 4 types of neurons in the cerebellar cortex and DCN 574. Mossy fibers (MFs), whose axons originate from neurons in several CNS regions (including pons), provide excitatory afferents to Gr cells and Gg cells of cerebellar cortex and to DCN (not shown). Glu, Asp, and serotonin (latter from dorsal raphe nucleus) are putative excitatory transmitters for these MF pathways (see refs. 89,141,142). A 3rd afferent input arises from neurons of nucleus locus coeruleus (LC) and provides inhibitory but enabling 80 input to P neurons; NE is the transmitter for this pathway [see NOREPINEPHRINE, p. 40; 345,347,700].

Membrane current fluctuations to iontophoretically applied GABA in cultured mouse spinal neuron voltage clamped (V_c) at −70 mV. Membrane current is displayed unfiltered on DC trace (B) and at 10 × gain on AC trace (C) filtered at 0.2–200 Hz. Variance associated with filtered signal, updated at 1‐s intervals, is displayed in D. Increasing iontophoretic currents cause inward current responses of increasing amplitude, each of which is associated with a thickening of the DC and AC traces and increases in membrane current variance. Largest changes in variance at beginning and end of current responses reflect relatively rapid changes in membrane current occurring at these times due to AC coupling. Arrowheads mark spontaneous inward current events, which have a fast rise time and exponential decay, suggesting they are synaptic in origin.

Gly‐ and GABA‐receptor channel currents recorded from outside‐out membrane patches isolated from soma of 3 different spinal cord neurons in culture. Patches were isolated from neurons bathed in normal bath solution [in mM: 140 NaCl, 1 MgCl₂, 1 CaCl₂, 1 KCl, and 10 Na‐HEPES (pH 7.2)], which was then exchanged for a solution containing Tris⁺ as major cation. For most patches this procedure removed background current activity. Tris⁺ substitution did not alter either chemo‐sensitivity of membrane to Gly or to GABA or conductance properties of the activated Cl⁻ channels. Pipette solution facing intracellular side was (in mM): 140 KCl, 3 NaCl, 1 MgCl₂, 11 K‐EGTA, and 10 K‐HEPES (pH 7.2). All recordings were done at room temperature (22°C–25°C). A: top trace shows absence of current activity before and after rapid application (indicated by bar) of 50 μM GABA to bath solution. GABA was then washed out, and bottom trace shows increase in current caused by addition of 50 μM Gly (day 35 neuron). B: outside‐out patch in which application of 20 μM Gly did not increase current, but after washout of Gly and application of 20 μM GABA, there was a large increase in current, which decreased with time so that current steps of ∼2 pA were discernible. Note much noisier appearance of GABA‐activated current compared with the Gly‐activated current in A, which reflects higher frequency of brief interruptions evident in single GABA‐receptor currents compared with Gly‐receptor currents. C: outside‐out patch in which application of 5 μM GABA (not shown) activated a relatively low frequency of current steps (2‐pA amplitude), but when concentrations of GABA was increased (bar) to 50 μM, additional channels were activated. At this higher concentration current response quickly and completely desensitized so that after 90 s, no current steps were evident. At this point, addition of 20 μM Gly to bath solution activated current steps of 3‐pA amplitude.

Multiphasic synaptic responses of CA1 Pyr neuron in hippocampal slice preparation to stimulation of stratum radiatum (SR). Top traces were recorded at resting membrane potential (RMP) and lower traces while membrane was artificially hyperpolarized by negative current injections in amounts (mV) shown. Vertical arrow indicates time of SR stimulation. Stimulus artifact is followed by short EPSP (seen best in left recording hyperpolarized by 5 mV); in recording at RMP, the EPSP is followed by an early (angled arrow) and a late (•) hyperpolarization. Note that early hyperpolarization, probably GABAergic IPSP (see refs. 9,10,19,195,529), is reduced and then inverted by increasing hyperpolarizations, whereas late hyperpolarization is nullified only at greater hyperpolarizations (−11 to −16 mV below RMP). These and other data (see refs. 195,529,766) suggest that the early IPSP has a reversal potential near the expected equilibrium potential for Cl⁻, whereas late hyperpolarization [possibly late or slow IPSP 529,766] has a reversal potential more like that for K⁺. Spikes (spontaneous or those driven by strong SR stimulation) are attenuated by slow rise time of polygraph.

Inhibitory effectiveness and Cl⁻ dependence of depolarizing IPSP (in rat hippocampal slice preparation). A: train (2 Hz) of depolarizing current pulses initiated full‐sized somatic action potentials.

Responses of a single cultured brain neuron to L‐Glu (A, 100 μM), D,L‐kainate (B, 100 μM), D,L‐homocysteic acid (D,L‐HCA; C, 100 μM), and L‐Asp (D, 100 μM) demonstrating decreased conductance. All amino acids were applied (C2: 150 ms; D2: 450 ms; all others: 350 ms) in the presence of 3 μM tetrodotoxin (TTX) (no spikes were evoked with brief cathodal current pulses). Trace below each voltage recording indicates period of drug application (upward deflections). Constant current anodal pulses (50 ms; not shown) were used to assay Gm in (A1–D1). Recordings were at neuron's resting level, −60 mV. Calibrations: A2, C2, and D2: 4 s; D1 and A1: 8 s; and B1, B2, and C1: 20 s. C1: 40 mV; all others: 20 mV. A: responses to L‐Glu. Initial phase is large depolarization (∼18 mV) accompanied by rapid rise in conductance (Gm; reduction in voltage deflections). Later, Gm drops dramatically. Note small, irregular potentials on falling phase of A2. Longer applications of L‐Glu lead to full spikelike potentials (not shown). B: responses to D,L‐kainate. Entire response to D,L‐kainate is increase in Gm. Gm remains elevated even when membrane falls close to preapplication values. Note lack of spikelike potentials in B2. Shape of offset of D,L‐kainate response differs from other amino acids because of absence of decreased Gm. In addition the duration is significantly greater than those to L‐Glu and L‐Asp. C: responses to DL‐HCA. Responses were particularly large and lasted longer than those to L‐Glu and L‐Asp. Although initial depolarization is associated with increase in Gm, response quickly reverts to surprisingly large decrease in Gm. Gain of C1 was reduced because of the size of voltage deflections. C2: generation of large spikelike potentials occurring during falling phase of response. D: responses to L‐Asp. This amino acid induces a depolarization comparable to that evoked by L‐Glu, but in this case only a decrease in Gm was observed. Spikelike potentials are also seen (D2). Higher concentrations of L‐Asp (500 μM) also increased Gm (not shown).

N‐methyl, D‐aspartate (NMA) produces dose‐related biphasic conductance change. A: responses of cell to increasing ionophoretic currents of NMA. Small NMA depolarizations are accompanied by apparent rise in input resistance, whereas large responses are dominated by fall in input resistance. B: NMA pulses (150‐ms wide), with and without constant current negative pulses to measure input resistance. C: in another cell, a 40‐ms, 180‐nA NMA pulse produces an apparent rise in input resistance, as measured with positive current pulses, with no underlying depolarization. A−C from different cells.

I−V curve for cultured spinal cord neuron under voltage clamp (2‐electrode clamp). RMP was −38 mV and both depolarizing and hyperpolarizing command steps were employed to construct curve. Control curve (•) was performed in bathing solution supplemented with TTX (2 μM) and repeated during constant microperfusion with L‐aspartic acid (500 μM; •). Net inward current (negative by convention) was evoked by L‐Asp. Slope of this steady‐state relationship was reduced in range from −70 to −30 mV.

Intracellular recordings from rat cerebellar Purkinje cells in vivo. A: recording and electrophoretic setup: 3‐barreled micropipette with a Purkinje cell. Intracellular electrode protrudes beyond orifices of the 2 extracellular microelectrophoretic barrels. B: multispiked spontaneous climbing fiber discharge obtained during intracellular recording from a Purkinje cell. RMP (in mV) is given in parentheses. Calibration: 20 ms and 25 mV. C: changes in membrane potential and membrane resistance of 4 Purkinje cells in response to drugs. All specimens in each row of records are from same cell. Bar above each record indicates extracellular electrophoresis of indicated drug (100–150 nA). RMP (in mV) is given in parentheses below each recording. Calibration: 10 s and 20 mV for NE, dibutyryl cAMP (DB), and cAMP; 5 s and 10 mV for GABA. Right: effective input resistance was judged by size of pulses resulting from passage of a brief constant current (1‐nA) pulse across the membrane before, during, and after electrophoresis of respective drugs (1 mV = 1 MΩ). Discontinuities in fast transients of pulses (and loss of spikes) result from loss of high frequencies (>10 kHz) and from chopped nature of the frequency‐modulated magnetic tape recording used. All pulse records were graphically normalized to same base‐line level. Calibration: 80 ms and 15 mV for all pulse records.

Intracellular recording from a CA1 pyramidal neuron in hippocampal slice in vitro preparation. Isoproterenol (IP) reduced amplitude and duration of afterhyperpolarization (AHP) evoked by depolarizing current pulses that generated action potentials (attenuated by polygraph). AHP is a hyperpolarization and inhibition of spontaneous activity after termination of depolarizing current pulse (arrow). Amplitude and duration of AHP increases as amplitude of depolarizing pulse increases. Note also activity increase produced by IP.

Intracellular recordings show effect of adrenergic agonists on membrane potential and spontaneous activity of CA1 pyramidal neurons in hippocampal slice preparation of rat. A: typical response of CA1 pyramidal neuron to superfusion of IP. In all neurons tested, IP evoked an increase in subthreshold activity (indicated by thickness of base line) and action‐potential generation (large upward deflections). Superfusion of IP began at downward arrow and stopped at end of bracket. Delay to onset of increase in activity is partially due to “dead time” (1–2 min) of perfusion system. Increase in activity evoked by IP was prolonged in duration and far outlasted estimated washout time (2–4 min). Note depolarization evoked by IP. B: typical response of CA1 pyramidal neurons to the α‐adrenergic agonist clonidine. NE decreased spontaneous activity and produced a small hyperpolarization. Concentrations of 2–10 μM were usually required to produce this effect. This is same neuron as in A, where superfusion with β‐agonist IP increased spontaneous activity. Clonidine mimicked action of NE in this neuron, but in other neurons NE was excitatory whereas clonidine was inhibitory. These data suggest that CA1 pyramidal neurons have both α− and β‐adrenergic receptors.

Suggested locus of NE action at membrane of hippocampal pyramidal cell. NE, in binding with β‐adrenergic receptor (triangular notches) on neuronal membrane, could block either of 2 ion channels: 1) the Ca²⁺ channels that are voltage‐sensitive channels opened by a depolarization such as occurs during an action potential; or 2) the Ca²⁺‐dependent K⁺ channels that are opened by influx of Ca²⁺. Data showing no NE effect on Ca²⁺ spike during TTX treatment (see refs. 302,477) suggest that 2nd possibility is more likely. Because NE‐induced inactivation of the Ca²⁺‐dependent K⁺ channels would retard repolarization of membrane during an action potential, more Ca²⁺ might then enter the cell during this state of depolarization by virtue of voltage‐dependence of Ca²⁺ channel.

NE activates K⁺ conductance in the locus coeruleus in vitro. A: pressure application of NE [50 ms, 10 psi (68 kPa); ▴] produced hyperpolarization and increased membrane conductance. Top traces (in 1 and 2) are of membrane potential and bottom traces are of membrane current. Parts 1 and 2 are from same neurons. NE was applied twice in each trace. Full amplitude of NE‐induced hyperpolarization is seen after 1st application. After 2nd application, membrane potential was maintained at rest with manual‐clamp technique. Twice as much NE was applied in 2 as in 1. B: determination of NE reversal potential in solutions of varying K⁺ content. A single pressure‐ejection pulse [50 ms, 10 psi (68 kPa)] was applied at each arrowhead. Left: normal, 2.5 mM K⁺ solution; as membrane was hyperpolarized, NE potential decreased and reversed at −100 mV. Middle: 4.5 mM K⁺ solution; reversal potential decreased to −85 mV. Right: 10.5 mM K⁺ solution; reversal potential was −65 mV.

Reversal potential of IPSP in locus coeruleus in vitro. A: synaptic potentials were evoked by single pulse (30 V, 0.5 ms; arrow) at different RMPs. At −67 mV the IPSP was hyperpolarizing, at −105 mV it was nullified, and at −127 mV it was depolarizing. Durations of EPSP and IPSP were reduced at more negative potentials, probably because of membrane rectification at these potentials. B: amplitude of IPSP in another cell is plotted as function of membrane potential in solutions with different K⁺ content. •, Normal K⁺ concentration (2.5 mM); •, 4.5 mM K⁺; and ▴, 10.5 mM K⁺. Estimated reversal potentials were −121, −105, and −92 mV. (These reversal potentials were the most negative of cells included in Table 1 of ref. 229.)

Effects of iontophoretic DA and protons (H⁺) on cat caudate neurons in vivo. A: iontophoretically applied DA depolarizes membrane of cat caudate neuron and decreases its firing rate, whereas Na⁺ ejected at twice the current intensity and same pH used for DA had no comparable effect. B: H⁺ ejected from barrel containing HCl at pH of 0.5 produces long‐lasting increases of firing rate of another caudate neuron (ratemeter output). Excitatory effect of Glu is much faster in onset and offset, whereas DA had no effect on firing rate of this cell displaying no spontaneous action potentials. 2/1 and 69Q2, Experiment and cell numbers.

DA‐induced effects with short (50 μm) distances between tips of recording and iontophoretic electrodes on 2 different cells of cat caudate in vivo. A: DA induces hyperpolarizations. B: if DA is applied for longer period on another cell, initial hyperpolarization is followed by slow depolarization. Single arrow, cortically evoked EPSP; double arrows, cortically evoked EPSP with superimposed action potential. Glu, glutamate.

Effect of carbachol on current relaxations generated by hippocampal CA1 neuron (in slice preparation) at rest and at depolarized holding potential. Traces are currents initiated by hyperpolarizing voltage‐clamp step (ΔV) of 14 mV from holding level of −42 mV (top row) and from close to resting potential, −62 mV (bottom row), before, during, and 10 min after brief (2 min) bath application of carbachol (48 μM). During drug administration an inward current developed at the more positive holding potential, as indicated by positioning of traces, but not when cell was held at −62 mV. Note approximately ohmic behavior of cell at latter potential, whereas a similar current response was only observed in presence of carbachol at depolarized level, i.e., when time‐dependent inward relaxation had been abolished. TTX (0.5 μM) was included in bathing medium to abolish Na⁺‐dependent action potentials; bath temperature was 23°C.

Effects of ACh on rat CA1 pyramidal cells in hippocampal slice preparation. All responses are from same cell. A: chart record of control AHP after 60‐ms direct depolarizing current pulse (trace 1) and film record of response to 600‐ms pulse (trace 2). Current record is positioned below voltage record. B: ACh (200 μM) superfusion depolarized membrane and increased cell's input resistance. C: blockade of AHP (trace 1) and accommodation (trace 2) in presence of ACh. D: addition of atropine (0.5 μM) in presence of ACh reversed effects of ACh. E: atropine also reversed effect of ACh on AHP (trace 1) and on accommodation (traces 2 and 3). Current pulse in trace 3 was identical to those of trace 2 in A and C. Current pulse was increased in trace 2 to match depolarization evoked in presence of ACh (trace 2) in C. Gain in trace 1 in A applies to all chart records. Time calibration for trace 1 in A also applies to trace 1 in C and E, and time calibration in D and B are the same. Calibration (time, volts, and current) for trace 2 in A applies to all film records. RMP = −57 mV.

Hippocampal slice preparation. Effect of exogenous ACh mimicked by stimulating stratum oriens. Records from 4 different cells are shown (A‐D). A: immediately after train of stimuli (S), cell hyperpolarizes (condensed on time scale) and then depolarizes (arrow in trace 1). In trace 2, slow depolarizaton was manually voltage clamped, which clearly shows increase in size of hyperpolarizing current pulses. RMP = −53 mV. B: slow depolarization (trace 1) was completely blocked by 1 μM atropine (trace 2). RMP= −59 mV. Calibration in A also applies to B. C: AHPs before pathway stimulation and 4 s after stimulus. In control solution (trace 1) AHP was slightly depressed after stimulus. After addition of 1 μM eserine (trace 2), AHP was greatly reduced by stimulus (arrow); AHP was restored to control amplitude by 1 μM atropine (trace 3). RMP = −56 mV. D: identical protocol as in C, except showing accommodation during 600‐ms pulse before and after pathway stimulation. RMP = −64 mV.

Effect of adenosine on membrane potential and input resistance of pyramidal neuron in rat hippocampal slice. A: low concentration (5 μM) of adenosine (•) has no detectable effect on input resistance measured from electrotonic potentials (bottom inset oscillographs in 1 and 2) generated by current injection (top inset oscillographs in 1 and 2) through a bridge circuit and recording electrode. 1: Control; 2, adenosine superfusion; I‐V curve during adenosine superfusion is superimposable with that of control period (x). B: higher adenosine concentration (20 μM) reduces input resistance during hyperpolarization. Right, current and electrotonic potentials before (1), during (2), and after (3) adenosine perfusion. Calibration: 20 mV, 50 ms. Spiking after offset of current pulse was reduced by adenosine. All voltage measurements were taken from later, flat portions of electrotonic potentials. Curves were normalized to RMP. C: 20 μM adenosine hyperpolarizes and inhibits spontaneous discharge of pyramidal neuron via postsynaptic action; slice was continuously perfused with 10 mM MgCl₂ to prevent transmitter release. Spikes were attenuated by slow frequency response of polygraph. Brackets show duration of adenosine (plus 10 mM MgCl₂) perfusion. Note rebound depolarization often seen after withdrawal of adenosine.

Effects of adenosine on Ca²⁺ spikes from TTX‐treated hippocampal neurons in slice preparation. All traces have been digitized (0.2‐ to 0.4‐ms sample time). A: individual responses before (A) and after (B‐D) 20 μM adenosine. Traces B‐D are successive responses at 15−, 30−, and 60‐s intervals after addition of adenosine to recording chamber. Spike seen in traces A and B is largely abolished in trace C; trace D shows no active response to current injection. Depolarizing current pulse was held constant (1.0 nA) for all traces. B: superimposed voltage traces from cell before (A), during (B), and after (C) perfusion with 20 μM adenosine. Depolarizing current was 0.8 nA (top traces) and 1.1 nA (bottom). C: superimposed records showing response of cell (top traces) to depolarizing current injection before (C₁), and during (C₂) treatment with 20 μM adenosine. Amount of current injected is illustrated by current monitor (bottom trace). Note that calcium spike can be elicited in presence of adenosine, but that considerably higher currents are required (top 2 traces in C₂). Calibration: 10 mV (2 nA for current traces) and 20 ms for each panel.

Possible sites of action of nucleosides or nucleotides in nervous system. XYP is a molecule in which X could be any purine or pyramidine base and Y is number of phosphate groups, from 0 to 3. These substances could have several roles: 1) as local modulators released from glia or neurons and acting either pre‐ or postsynaptically; 2) classic transmitters; 3) presynaptic modulators of transmitter release; 4) cotransmitters with some other transmitter such as ACh; 5) 2nd messengers acting at several possible sites, including at some nucleotide cyclase, which would then generate a 3rd messenger such as cAMP or cGMP; and 6) unconventional transmitters acting on voltage‐dependent conductance such as M‐current (I_M).

Enkephalin opens K⁺ channels in locus coeruleus neurons in vitro. A: D‐Ala, D‐Leu‐enkephalin (DADL) was applied by pressure [arrow; 50‐ms pulse, 70 kN/m² (10 psi)] at various membrane potentials. The K⁺ concentration was as indicated (in mM). At elevated K⁺ concentrations, enkephalin response reversed polarity with membrane hyperpolarization. Marked rectification occurred with hyperpolarization and elevated K⁺ concentration. B: relationship between hyperpolarization amplitude and membrane potential. C: enkephalin reversal potential (E_rev) calculated from many experiments such as those in A and B. Dashed line, E_rev = RT/F ln([K]_o/[K]_i), where [K]_i, intracellular K⁺ concentration, was 145 mM. [In normal K⁺ solutions (2.5 mM), it was usually not possible to reverse enkephalin hyperpolarization; therefore extrapolated values were used.] Good agreement between observed values of E_rev and those predicted from Nernst equation (Eq. 4) suggests a predominantly somatic site of action; large steady‐state conductance increase that accompanied membrane hyperpolarization would effectively remove dendritic contribution.

Stratum radiatum (Rad) stimulation produces EPSP/IPSP sequence in CA1 pyramidal cells (A1) of hippocampal slice. Application of 100 μM D‐Ala, D‐Leu‐enkephalin (DADL) reduces IPSP (A2) and leads to depolarization and burst of spike activity after EPSP (A3). Alvear (Alv) stimulation produces an IPSP (B1) that is blocked by 100 μM DADL (B2). Recovery can occur within 10 min (B3). However, in presence of 2 μM naloxone, alvear‐induced IPSP (C1) is not blocked by repeated applications of 100 μM enkephalin (C2). •, Time of stimulation. Calibration: A, 10 mV and 40 ms; B and C, 5 mV and 40 ms.

COOH‐terminal portion of porcine prodynorphin sequence (residues 171–256) deduced by Kakidana et al. 390. Three [Leu⁵]enkephalin‐containing segments are bracketed by putative peptide‐processing signal “Lys‐Arg”. α‐Neo‐ and β‐neoendorphin overlap as prodynorphin residues 175–184 and 175–183, respectively. Dynorphin A(1–17) corresponds to full sequence of 2nd segment. Dynorphin A(1–8) is produced by unusual cleavage to left of Arg²¹⁷. Similar cleavage to left of Arg²⁴¹ produces dynorphin B from 3rd [Leu⁵]enkephalin‐containing segment. Presence of full sequence of dynorphin B(1–29) in brain tissue has not been reported.

Substance P (SP) responses of cultured spinal cord neuron. Clear reversal of SP‐response polarity was not obtained. SP and eledoisin‐related peptide response amplitude varied as function of membrane potential, but responses did not reverse polarity. SP‐response amplitude increased when membrane was depolarized and decreased when it was hyperpolarized by 1 of 2 intracellular micropipettes. Response amplitude did not vary as linear function of membrane potential at potentials <‐50 mV or >‐95 mV in this neuron. Although activation of voltage‐dependent conductances might reduce voltage response recorded at potentials <‐50 mV, it was not clear why no reversal of response polarity was seen. Even when this neuron was hyperpolarized to very negative potentials (−106 and −114 mV), SP responses did not reverse polarity. RP^e, extrapolated reversal potential.

Structure of the presumed somatostatin (SS) prepro‐hormone SS‐28 and 2 fragments, SS‐28(1–12) and original SS‐14, or SS‐28(15–28), derived from SS‐28. Heavy brackets with arrows denote the 2 fragments; light bracket indicates ring structure of SS‐28 and SS‐14. Note that SS‐28 is NH₂‐terminally extended version of SS‐14. Cleavage of SS‐28 to SS‐14 would also appear to liberate SS‐28(1–12) and an Arg‐Lys dipeptide (see refs. 63,64).

Effect of SS on CA1 pyramidal neuron, recorded with intracellular microelectrode in rat hippocampal slice preparation. Slices were 250 μm thick. Recording microelectrodes were filled with 1 M K‐acetate (tip resistance = 90–150 MΩ). Center of a 4‐barreled microelectrode was filled with 3 mM solution of SS in 5 mM Na‐acetate (pH 7.0), and the peripheral barrels were filled with 1 M NaCl, 1 M Na acetate, 0.5 M ACh chloride (pH 4.5), or 1 M monosodium l‐Glu (pH 6.7). A backing current of 25 nA and appropriate polarity was applied to each barrel throughout experiment. In some experiments synthetic SS was used with similar results. SS was tested only on cells from which stable intracellular RMP recordings > −60 mV were obtained. A: SS [somatomedin‐release inhibitory factor (SRIF)], applied by passage of 196 nA negative current (−ve), caused depolarization of 21 mV (RMP of −75 mV). Termination of application resulted in release of some voltage coupling between recording and iontophoretic electrodes, which masked an even greater depolarization. Depolarizing pulses and both the evoked and spontaneously occurring action potentials disappeared when photographic conditions suitable for reproduction of changes in DC level at high gain were selected. B: same record at a faster sweep speed shows effect of SS on excitability of cell at times a‐d indicated in A. Bottom trace shows depolarizing current pulse injected through recording electrode, adjusted so as to evoke 2 action potentials in resting conditions. C: similar records show excitation of cell by l‐Glu (GLUT).

Effect of SS analogue in medium with high Mg²⁺. A: membrane potential response of pyramidal cell in hippocampal slice. Cell displays reversible hyperpolarization in response to addition of SS nonapeptide agonist Ac‐Cys³‐Des AA^1,2,4,5,12 (D‐Trp⁸, D‐Cys¹⁴)‐SS to perfusate. Artificial cerebrospinal fluid contains 12 mM MgCl₂, which eliminated stratum radiatum‐induced action potentials. B: voltage deflections in response to depolarizing and hyperpolarizing current before, during, and after SS‐analogue perfusion. C: I‐V curves constructed from hyperpolarizing potentials. Decrease in slope of the I‐ V curve indicates that SS analogue increased conductance in this cell.

SS‐28 hyperpolarizes and inhibits activity of hippocampal pyramidal cell in vitro with a slight decrease in input resistance. SS‐28 was superfused onto hippocampal slice preparation during intracellular recording. Deflections below the membrane potential line are electrotonic potentials resulting from intracellular current injection through a bridge circuit. Note small reduction in these potentials during SS‐28‐induced hyperpolarizations. This apparent reduction in input resistance is also indicated by slight diminution of slope of I‐V curve generated during SS‐28 superfusion [membrane potential normalized to RMP (Vm) during SS‐28]. Note slow onset and long duration of action of SS‐28 and reduction in subthreshold (thin vertical lines above RMP) and spike activity (thick vertical lines). Spikes were attenuated by slow rise time of polygraph.

Effect of corticotropin‐releasing factor (CRF) on current‐evoked bursts of action potentials and associated AHPs in CA1 pyramidal cell of hippocampal slice. A: 0.5 μM CRF reduces AHPs that follow trains of spikes generated by passing depolarizing current pulses of 3 different intensities through active bridge circuit and recording electrode. Late components of AHPs are completely lost during CRF superfusion, despite greater number of spikes generated by equivalent current strengths. CRF also increases spontaneous firing and depolarizes the cell (membrane potentials are indicated for each trace). B: effect of CRF in presence of TTX (100 nM). TTX alone (left) abolishes early, fast‐action potentials, leaving slower Ca²⁺ spike. CRF (400 nM) added to TTX solution (middle) nearly abolishes AHP (arrows), but did not significantly alter Ca²⁺ spike. Recovery is shown on right. V, voltage trace; I, current trace. Current pulse = 0.2 nA; calibration: 10 mV and 0.2 s.

Suggested locus of CRF action on hippocampal pyramidal cell membrane. CRF molecule, in binding with a CRF receptor (triangular notches) on neuronal membrane, could block either of 2 ion channels: 1) Ca²⁺‐channels that are voltage‐sensitive channels opened by depolarization, e.g., during action potential, or 2) Ca²⁺‐dependent K⁺ channels, which are opened by Ca²⁺ influx. Data showing lack of a CRF effect on Ca²⁺ spike during TTX treatment 8,707 suggests that the 2nd possibility is more likely. Because CRF inactivation of Ca²⁺‐dependent K⁺ channels would retard repolarization of membrane during an action potential, more Ca²⁺ might enter the cell during this state of depolarization by virtue of voltage dependence of Ca²⁺ channel.

Comparison of oxytocin/vasopressin related neuropeptide structures.

Two major brain slice recording methodologies currently used. In both cases slice is usually placed on nylon net (open circles), on an electron microscope grid, or on filter paper. Recording chamber (shown in cross‐section) may be circular or rectangular. A: static or perifusion (microdrop) method involves filling recording chamber with warm artificial CSF (ACSF; gassed with carbogen) only up to edges of the slice. Recording may be done without perfusion because space above slice is in contact with warm carbogen gas saturated with H₂O; H₂O‐gas interface at top of slice allows passage of O₂ into tissue. However, “perifusion” may also be performed, wherein the inflow and outflow of ACSF are perfectly balanced so that the level of ACSF never changes and top of slice remains exposed to carbogen. Microdrop involves pressure application of drugs from pipettes to top, exposed surface of slice. B: superfusion method involves covering the entire slice with warm ACSF saturated with carbogen; fast flow rate (2–5 ml/min) is used to assure adequate oxygenation of tissue. Outflow must balance inflow so fluid level does not change; otherwise, tissue movement and changes in recording characteristics could occur. Drug pipettes in both types of chambers could also be used to pass drugs by iontophoresis as well as by pressure. Several laboratories now use a slice recording method that combines some properties of both methods (see refs. 312,317,318).

Hypothetical integrative interactions between conventional and newly described synaptic messages. Input T: a conventional test EPSP produced by conductance increase; inputs A‐D: conditioning postsynaptic potentials (PSPs) that can modify the test EPSP, as well as each other. Inputs A, C, and E are EPSPs and B and D are IPSPs; inputs A and B are conventional (i.e., associated with increased conductance or decreased resistance), whereas inputs C‐E are associated with decreased conductance. Results of synaptic interactions, recorded in the soma, are shown as possible summations of various PSP types, and partially result from effects of changes in input resistance and consequently the space constant. “A, then T” shows decremental effects of conduction of A EPSP down dendrites to cell body; the initial EPSP is greatly attenuated compared with potential generated at dendritic sites. Also note shunting of secondary T EPSP by reduced resistance and space constant brought about by A input, even though both PSPs are depolarizing and should summate to some extent. “B, then T” shows similar shunting effect, combined with summation of opposite values. However, in “C, then T”, the increased input resistance (R_i) results in a larger T EPSP, overcoming to some extent effects of decremental conduction down the dendrite. Similar augmentation of T might occur even with hyperpolarization, if IPSP is associated with an increase in R_i (“D, then T”). When EPSPs with increased R_i occur nearly simultaneously (“E, then C”), the 1st may nullify 2nd, especially if same conductances are involved (see the chapter by North in this Handbook). In “E, then D” it is not clear what would result, although logic suggests that summation of opposite signs would nullify the responses. See Fig. 48 for interactive effects of other newly described messages. Vm, membrane potential.

Different functional types of interactive messages. Column A: responses displayed over a long time course (0.5–600 s); column B: typical electrotonic potential shape produced by intracellular current injection before (left) and during (right) application of messenger. Curves above long recording traces in A represent sizes and shapes of EPSPs (1–3) and changes in spikes (top trace) and AHPs (bottom trace) (4) before and after application of representative messenger. Sharp downward deflections in traces in A indicate repetitively generated electrotonic potentials, of which those in column B (expanded in time) are representative. Note that shunting messenger in A1 reduces size of another (conductance increase) EPSP, even when both are depolarizing. In fact, both depolarization (by driving membrane potential toward E_epsp) and the reduced input resistance would tend to reduce the conventional EPSP. Furthermore, electrotonic potentials in B1 indicate that decreased input resistance would also shorten membrane rise time (note increased “squareness” of potential with messenger), effectively shortening EPSP. Messengers that increase input resistance in A2 (larger downward deflections) would enlarge EPSPs with conductance increases (and may constitute an enabling mechanism) by increasing voltage drop across membrane and by increasing membrane time constant (note longer time required for electrotonic potential in B2 to reach steady state). However, a depolarizing enabling messenger might still reduce the EPSP by driving membrane potential substantially nearer E_epsp, thereby reducing driving force for ion flow. Note long time course (see calibrations of enabling responses compared with those of shunting messengers). Modulating messengers in A3 might disenable (reduce) or enhance (not shown) other synaptic inputs, without altering membrane properties (B3). This enhancement could also be a 2nd mechanism for enabling. Messengers that alter voltage‐dependent conductances in A4 might prolong (or shorten) spike (above trace) by enhancing the voltage‐dependent Ca²⁺ conductance, or by decreasing K⁺ conductances, but without altering resting membrane properties (B4). AHPs (below trace) might also be reduced or augmented by similar effects on voltage‐dependent Ca²⁺ or K⁺ conductances, again without any effect on resting properties.