Analysis of habituation and sensitization of hindlimb flexion response in acute spinal cat. A: experimental arrangement. Stimuli can be delivered to skin (S_S1 or S_S2), to cutaneous afferent nerve (S_N1), or to group Ia muscle afferents (S_N2). Responses can be recorded from dorsal root (R_S), from α‐motoneurons (R_α) or interneurons, from ventral root (R_N), or from muscle (R_M). B: comparison of changes in excitatory postsynaptic potentials (EPSPs) recorded in deep peroneal motoneuron elicited by stimulation of either cutaneous afferent nerve (1, 2, 3) or muscle afferents (4, 5, 6). Stimulation of the 2 nerves was interpolated. 1, 2, 3, Polysynaptic EPSPs elicited by single‐shock stimuli delivered to posterior femoral cutaneous nerve. 1, Control period, stimulation once every 30 s. 2, Synaptic decrement, stimulation once per second. 3, Recovery, stimulation once every 30 s. 4, 5, 6, Monosynaptic EPSPs elicited by stimulation of deep peroneal nerve at 30‐s intervals during periods corresponding to those in 1, 2, and 3. Note constancy of monosynaptic responses during period of polysynaptic EPSP decrement.

Possible role of spinal interneurons in habituation and sensitization of flexion reflex in cat. A: responses of different types of interneurons (nonplastic, type H, and type S) and muscle response during repetitive stimulation producing first sensitization and then habituation of flexion reflex. B: schematic diagram of possible neural substrates of habituation and sensitization. N, nonplastic synapses; H, habituating synapses; S, sensitizing synapses. According to this scheme, cutaneous stimuli can exert their influence on 2 separate systems: reflex system (S‐R pathway) that mediates habituation and “state” system that mediates sensitization.

Habituation and sensitization of gill‐ and siphon‐withdrawal responses in Aplysia. A: experimental arrangement for behavioral studies in intact animal showing gill and siphon in relaxed (A₁) and contracted (A₂) position. Parapodia and mantle shelf are shown retracted to reveal gill and siphon. Gill and siphon withdrawal can be elicited with tactile stimulus to siphon. Electric shock of tail (shock) or head serves as sensitizing stimulus. B: photocell recordings showing habituation and sensitization of gill‐withdrawal response. Tactile stimulus to siphon (bottom trace) elicited gill‐withdrawal response (top trace), which habituated with repeated stimulus deliveries. Application of noxious tail shock (arrow) caused marked sensitization of gill‐withdrawal response. C: neural circuit mediating gill‐ and siphon‐withdrawal responses. Siphon skin is innervated by group of ∼24 mechanoreceptor sensory cells (SN) whose soma are located in LE cluster in abdominal ganglion. Siphon sensory cells project both monosynaptically and through interneurons (INT) to gill and siphon motoneurons (MN), many of which are also located in abdominal ganglion. Effects of sensitizing stimuli in abdominal ganglion are mediated by group of facilitator interneurons (FAC INT) that terminate on sensory cells in synaptic and somatic regions. Habituation of reflex has been shown to involve a decrease in transmitter release from sensory neuron terminals [H(↓)]. Sensitization of reflex has been shown to involve an increase in transmitter release from sensory neurons [S(↑)]. Posttetanic potentiation has also been reported to occur at neuromuscular junction. [Adapted from Kandel and Schwartz 220.]

Electrophysiological analyses of habituation and sensitization of gill‐withdrawal response. A: homosynaptic depression and presynaptic facilitation at synapses from a siphon sensory cell onto L7, a gill motoneuron. Abdominal ganglion was dissected from animal and was maintained in vitro in artificial seawater (ASW). Intracellular stimulation of sensory neuron triggered action potential (bottom trace), which elicited monosynaptic EPSP in gill motoneuron (top trace). (Sharp negative and positive deflections that precede and follow action potential in sensory neuron are stimulus artifacts from depolarizing current pulse.) With repeated stimulation, action potential elicited successively smaller EPSPs in gill motoneuron. Electrical stimulation of left connective (arrow), which carries input to abdominal ganglion from head and tail, caused a marked increase in EPSP elicited in gill motoneuron. Synaptic depression and facilitation parallel and contribute to behavioral habituation and sensitization of gill‐withdrawal reflex. B: connective stimulation (B₁) and application of 0.2 mM serotonin (B₂) increase action‐potential duration in siphon sensory cells. Abdominal ganglion was maintained in artificial seawater containing 0.1 M tetraethylammonium (TEA) (K⁺ channel blocker) to facilitate measurement of action‐potential duration and maximize effects of facilitation. Similar though smaller changes also occur in normal seawater. Because Ca²⁺ entry is voltage dependent, broadening of action potential is believed to contribute to increase in Ca²⁺ influx and subsequent transmitter release during facilitation. C: comparison of changes in action‐potential duration in siphon sensory cell (bottom trace) and amplitude of EPSPs elicited in gill motoneuron L7 (top trace) in 1 preparation. With repeated stimulation, action‐potential duration and EPSP amplitude decrease concomitantly. Weak stimulation of pleuroabdominal connective (left vertical dashed line) causes a transient increase in action‐potential duration and EPSP amplitude. Stronger stimulation (right vertical dashed line) produces larger effects. Experiment conducted in 0.1 M TEA.

Biophysical analyses of ionic changes underlying homosynaptic depression and presynaptic facilitation in siphon sensory cells. A: serotonin, a putative facilitatory transmitter, depresses an outward K⁺ current in siphon sensory cells. A₁: voltage‐clamp experiment, performed in normal seawater. Voltage step from −35 to +15 mV (bottom trace) elicits initial inward current (downward deflection, top trace) carried largely by Na⁺ and Ca²⁺, followed by slower outward current (upward deflection, top trace) carried largely by K⁺. With repeated stimulation, inward current decreases, paralleling habituation. Application of serotonin (arrow) causes marked decrease in outward current. A₂: voltage‐clamp experiment in presence of K⁺ channel blockers (TEA, tetraethylammonium; 4‐AP, 4‐aminopyridine). Same preparation as in A₁. With K⁺ currents blocked, depolarizing step (bottom trace) elicits primarily an inward current. Application of serotonin (arrow) had no effect on current elicited by voltage step, indicating that serotonin normally exerts its action by depression of outward K⁺ current. B: effect of serotonin (5‐HT) on single‐channel currents recorded from siphon sensory cells in Aplysia. Left ordinate, number of open channels (n); right ordinate, current magnitude. Individual current steps are 2.6 pA. 1, Current in absence of serotonin. Current record was well fit by binomial distribution assuming that 5 channels are active in the patch and that each channel opens with probability of 0.84. 2, Current recorded 2 min after addition of 30 μM serotonin to bath. 3, Current recorded 1 min after addition of further dose of serotonin, raising total concentration to 60 μM. The 2 traces are a continuous recording. Downward current deflection at end of bottom trace is from intracellular current pulse used to monitor cell resistance. Serotonin produced decrease in average number of open channels but had no effect on unit current. 4, Current recorded ∼5 min after superfusion with serotonin‐free artificial seawater (ASW). Note increase in number of open channels.

Biochemical analyses of presynaptic facilitation in siphon sensory cells. A: injection of cAMP‐dependent protein kinase produces presynaptic facilitation. In base‐line control period, intracellular stimulation of sensory neuron (SN, bottom trace) evoked EPSP in a follower neuron (FN, top trace). Injection of catalytic subunit of cAMP‐dependent protein kinase led to increase in duration of sensory neuron action potential and increase in amplitude of EPSP evoked in follower cell. Experiment conducted in 100 mM tetraethylammonium (TEA). B: injection of kinase inhibitor blocks serotonin‐induced spike broadening. Recordings of action potentials in 2 different sensory neurons before (left) and after (right) application of serotonin. Serotonin produced marked spike broadening in control sensory cell but relatively little spike broadening in sensory neuron that had been injected with Walsh inhibitor, which blocks activation of cAMP‐dependent protein kinase. Experiment conducted in TEA. C: proposed sequence of steps underlying presynaptic facilitation in siphon sensory cells.

Analysis of habituation of crayfish escape response. A: neural circuit mediating escape response and proposed sites of plasticity. Tactile stimulation of posterior parts of crayfish produces forward somersault that is mediated by pair of neurons called lateral giants (LG). The LG have rectifying electrical synapse onto motor giants (MO G) and electrical synapse onto fast flexor motoneurons (FF MN). The LG receive direct electrical connections from sensory neurons (SN) as well as di‐ or trisynaptic input through heterogeneous group of interneurons (INT). These connections are believed to mediate the α‐ and β‐components, respectively, of complex EPSP evoked in LG by tactile stimuli. Habituation has been shown to involve a decrease in transmitter release at chemical synapses from primary sensory neurons onto interneurons. Synaptic depression can also occur at motor giant neuromuscular junction (↓) and posttetanic potentiation has been observed at neuromuscular junction of fast flexor motoneuron (↑); however, contribution of these forms of plasticity to behavioral learning is not clearly established. B: example of decrement in responses recorded from lateral giant fibers in response to repeated stimulation (1/min) of second root of abdominal nerve cord. Initial stimuli evoke action potential that fails as the β‐component of complex EPSP grows smaller with successive stimuli. Note that α‐component of EPSP remains relatively stable. C: example of decrement of monosynaptic EPSP in tactile interneuron. Stimulation of second root was adjusted to recruit single tactile afferent that elicited monosynaptic EPSP in tactile interneuron. With repeated stimulation, amplitude of EPSP decreases. C₁, C₂, C₃: first, fourth, and seventh responses to stimulation at 0.5 Hz. [A and C adapted from Zucker 450,451; B adapted from Krasne 239.]

Habituation and sensitization of crayfish defense response to tactile stimuli. A: proposed neural circuit and possible sites of plasticity. During habituation, neural responses elicited by tactile stimulus to back or tail decrease in interneurons (INT) and in excitor motoneuron (E) but do not change in inhibitor motoneuron (I) [H(0)]; these results suggest that habituation is due to changes in excitatory reflex pathway and not to buildup of peripheral inhibition. Sensitization has been proposed to involve increase in responses of interneurons and excitor motoneurons [S(↑)]; posttetanic potentiation (PTP) at neuromuscular junction of excitor motoneurons [S(↑)]; and decrease in responses of inhibitor motoneuron [S(↓)]. SN, sensory neuron; CN, command neuron. B: changes in number of excitor and inhibitor spikes accompanying sensitization. Note increase of excitor responses and decrease of inhibitor responses. C: PTP at excitor neuromuscular junction. Excitatory junction potentials recorded during test trains of excitor stimulation (30 Hz for 0.5 s) before (top trace) and 5 s after (bottom trace) conditioning train of excitor stimulation (10 Hz for 10 s). [B and C from Hawkins and Bruner 189.]

Neural circuit mediating cockroach escape response and proposed sites of plasticity during habituation and sensitization. Primary sensory neurons (SN) make monosynaptic connections onto giant and nongiant interneurons (INT), which in turn project (possibly monosynaptically) to leg motoneurons (MN). Synaptic depression occurs at sensory neuron‐interneuron synapses as well as interneuron‐motoneuron synapses [H(↓)]. Synaptic facilitation [S(↑)] occurs at interneuron‐motoneuron synapses.

Pigeon heart‐rate conditioning. A: schematic illustration of major ascending visual pathways mediating pigeon heart‐rate conditioning to visual stimuli. Acquisition of conditioned response can be prevented by lesions of retina, as indicated by L₁; by combined lesions of principle optic nucleus, nucleus rotundus, and tectofugal fibers to nucleus dorsolateralis posterior of thalamus, as indicated by L₂; and by combined lesions of striate and extrastriate cortical areas, as indicated by L₃. During conditioning, output of retinal ganglion cells appears to remain constant, as indicated by (0). Increases and decreases in neuronal responses have been recorded in principal optic nucleus and nucleus rotundus (↑, ↓). B: effects of visual‐system lesions on acquisition of conditioned response. L₁, L₂, L₃, performance of animals with combined lesions as indicated in A. Solid line, performance of control (CONT) animals. Each point, mean heart‐rate change in beats per minute (BPM) between 6‐s conditioned‐stimulus (CS) period and an immediately preceding 6‐s control period. C: training‐induced modifications of “type I” geniculate (principal optic nucleus) neurons as function of their response to training stimuli. I/I, cells that increase their discharge at CS onset and unconditioned‐stimulus (US) onset. I/II, cells that increase their discharge at CS onset but decrease their discharge at US onset. COND, units studied during associative training; SENS, units studied during nonassociative training. Curves, mean percentage change from response to light prior to training for phasic responses at CS onset. Note that type I/II cells showed marked response enhancement during associative training, while other cells showed only response decrement. [A adapted from Cohen 97; B adapted from Cohen 95; C from Cohen 96.]

Eye‐blink conditioning in cat. A: proposed neural circuit. Unconditioned eye‐blink response (UR) to glabella tap unconditioned stimulus (US) is mediated by facial nucleus (VII), which receives input directly and/or indirectly from trigeminal nucleus (V). Conditioned response (CR) pathway has been proposed to involve coronal pericruciate sensorimotor cortex (SENS‐MOTOR CX). Lesions (L) of sensorimotor cortex block acquisition of CR. Increases in neural excitability occurring in sensorimotor cortex and facial nucleus (↑) have been proposed to be intrinsic to these regions and to contribute to response specificity during learning. Increases in neural responses to conditioned stimulus (CS) have been recorded in auditory association cortex (AUD ASSOC CX) and have been proposed to be due to changes in afferent or presynaptic elements and to contribute to stimulus specificity during conditioning. Lesions of caudal cortical areas, including auditory association cortex, do not prevent acquisition, and essential afferent pathways for click CS are not known. B: effects of bilateral lesions of cortical motor areas on acquisition of conditioned eye blink. Average percent conditioned responses (± SD) of normal cats (N) and cats with lesions of rostral cortex (L) during training of conditioned eye blink, n = 5 and n = 5, respectively. One additional animal (not included) whose lesion was confined to coronal pericruciate region reached 35% performance levels. C: mean ± SD threshold currents required to discharge “both” projective neurons of coronalpericruciate cortex with intracellular stimulation for 6 behavioral groups. Cortical cells were classified as “both” cells if their intracellular or extracellular stimulation evoked electromyogram responses in nose and eye musculature. Stippled bars, groups of cats that received paired CS‐US presentations, including extinction (US‐CS) group for whom order of stimuli was reversed during testing phase. Open bars, groups that received only CS‐alone trials or US‐alone trials. Delay groups were tested 3–28 days (DEL CS‐US) or 25–100 days (DEL US) after training and received no stimuli during testing phase. Asterisks, significant decreases (P < 0.05) in threshold compared with CS‐only group. US‐only group exhibited only moderate decreases in thresholds (P < 0.10). D: average changes in membrane resistance in cells of coronal pericruciate cortex given either iontophoresis of acetylcholine (ACh) plus depolarizing current sufficient to repeatedly discharge cell (ACh + DISCHARGE), iontophoresis of ACh alone (ACh ONLY), or current‐induced discharge only (DISCHARGE ONLY). Iontophoresis of ACh occurred from period of Ach to 0 as indicated on abscissa.

Behavioral and electrophysiological changes during classical conditioning of rabbit nictitating membrane response for animals that received paired training (open circles) or explicitly unpaired training (closed circles). Each point, ∼23 training trials. Top: changes in magnitude of behavioral conditioned nictitating membrane response. Middle: changes in number of hippocampal multiple‐unit discharges occurring between conditioned stimulus onset and unconditioned stimulus onset. Bottom: changes in amplitude of population spike elicited in dentate granule cell layer by stimulation of perforant path. LTP, long‐term potentiation.

Rabbit nictitating membrane conditioning. A: proposed neural circuitry underlying rabbit nictitating membrane and eyelid responses. Unconditioned‐response (UR) pathway is a di‐ or trisynaptic arc involving fifth sensory ganglion (VG), fifth sensory nuclei (VS), and motoneurons mediating response in accessory abducens, abducens, and facial nuclei (VIA, VI, and VIII). Secondary pathway through reticular formation (RF) has also been proposed. Lesions (L) of middle cerebellar peduncle (MCP), dentate‐interpositus nuclei (DENT INT), superior cerebellar peduncle (SCP) before and after point of decussation, and red nucleus (RED N) all abolish conditioned responses (CR) to tone or light but have little or no effect on unconditioned responses to corneal air puff, indicating that these areas are essential components of CR pathway (shaded structures). At present, role of cerebellar cortex is controversial. Conditioned‐stimulus (CS) pathway has been proposed to involve auditory input from cochlear nucleus (VIII) projecting ultimately to dorsolateral pontine nuclei (PONT N) and possibly lateral reticular nucleus, and their mossy fiber projections to cerebellum via middle cerebellar peduncle. Unconditioned‐stimulus (US) pathway has been proposed to involve somatosensory projections to inferior olive, which in turn projects to cerebellum via climbing fibers in inferior cerebellar peduncle (ICP). Training‐dependent increases in neuronal activity evoked by tone CS have been recorded in deep cerebellar nuclei (↑). Similar responses have also been recorded in cerebellar cortex and red nucleus. B: effects of unilateral lesions of dentate‐interpositus cerebellar nuclei on mean peak amplitude of CR and UR nictitating membrane (NM) responses. CR were measured in 250‐ms period between tone CS onset and air‐puff US onset; UR were measured in the 250‐ms period following US onset. Data presented in 4 periods of training trials per session, ∼27 trials per period. Animals received 3 days of training (P1–3) with US delivered to left eye, while movement of left NM was monitored (n = 14). Following lesions of left cerebellar nuclei, animals were trained for 4 more days (L1–4) to test for retention and recovery of CR. CR were almost completely abolished by lesion, but UR amplitude was unaffected. On fifth postlesion day (L5), US was switched to right (nonlesioned) side, and movement of right NM was monitored; training was then returned to left eye (n = 13). Right (nonlesioned) side learned quickly, controlling for nonspecific lesion effects, while conditioned responding on left side showed essentially no recovery. C: example of unusually robust change in neuronal unit activity recorded from medial dentate lateral interpositus nuclei during explicitly unpaired training (C₁) and paired training (C₂). Top trace, NM extension; bottom trace, peristimulus histogram (9‐ms bins) of multiple‐unit activity recorded from dentateinterpositus (DI) nuclei. Left vertical line, CS tone onset; right vertical line, US corneal air‐puff onset. C₁: average of behavioral and neural responses to tone CS on first day of training in which tone CS and air‐puff US were presented in explicitly unpaired fashion. C₂: responses to tone CS on second day of paired training. Onset of unit response preceded behavioral NM response within a trial by 36–58 ms. Note that pattern of unit discharge appears to parallel topography of conditioned behavioral response more closely than unconditioned behavioral response evoked by air puff. Stimulation through recording electrode elicited ipsilateral eyelid closure and NM extension. D: comparison of changes in conditioned NM response and neuronal activity recorded from dentate‐interpositus in second half of CS period over course of training (n = 7). Magnitude of conditioned NM response (dashed line) was measured as area under curve described by amplitude time course of NM response in millimeter milliseconds. Standard scores of unit activity were calculated by comparing neuronal response in CS period to pre‐CS background period according to following equation: (CS half‐block – pre‐CS half‐block)/(SD pre‐CS session). Note that neuronal responses of dentate‐interpositus nuclei increase in close relation to size of CR (r = 0.90).

Cat leg‐flexion conditioning. A: experimental arrangement and proposed neural circuit. Conditioned stimulus (CS) is train of pulses to cerebral peduncle (CP); unconditioned stimulus (US) is shock to forelimb. Cerebral peduncle is cut caudal to red nucleus to restrict corticofugal output to red nucleus. Red nucleus neurons receive monosynaptic excitatory synaptic input onto distal dendritic regions from cerebral cortex and from nucleus interpositus of cerebellum (IP) onto proximal dendritic and somatic regions. Red nucleus projects via interneurons (INT) to flexor motoneurons innervating biceps brachii muscle controlling leg flexion. Conditioning in this paradigm has been proposed to involve sprouting of new corticorubral synapses onto proximal dendritic and somatic regions of cells in red nucleus (dashed connection, ↑). B: example of acquisition and extinction of conditioned leg flexion in 1 cat. During acquisition, subject received paired CS‐US training for 10 sessions; during extinction, animal received backward US‐CS trails. C: physiological evidence for sprouting of new corticorubral synapses following conditioning. EPSPs elicited in red nucleus neurons by stimulation of cerebral peduncle in conditioned cat (C₁) exhibit component with relatively fast rise time, whereas EPSPs elicited by stimulation of cerebral peduncle in control cat that had received no training (C₂) exhibit relatively slow rise time. Frequency histograms of time to peak for EPSP elicited by stimulation of cerebral peduncle in conditioned cat (C₃) and nontrained control cat (C₄). Ordinate, number of cells; abscissa, time to peak for EPSPs.

A: experimental arrangement for leg‐position learning in locust. Animal given paired training (P) receives electric shock from stimulator (STIM) when leg is lowered below a defined position. Since the 2 animals are arranged in series, yoked control (R) receives same shock irrespective of where it holds its leg. B: effects of up‐training on firing rate of AAdC, a motoneuron controlling leg position. Demand level was set at 15 Hz, and 10 shocks were delivered. 1, Before up‐training, firing rate in high Mg²⁺‐zero Ca²⁺ solution was 11.4 Hz. 2, In normal saline, firing rate was similar (11.0 Hz). 3, After up‐training, firing rate of AAdC had increased to 28.5 Hz. 4, After second infusion of high Mg²⁺‐zero Ca²⁺, firing rate was 22.1 Hz. The fact that firing rate remained elevated in high Mg²⁺‐zero Ca²⁺ solution (which inhibits chemical synaptic transmission) suggests that some of the increase was due to intrinsic changes in AAdC motoneuron. C: effect of up‐training on input resistance of AAdC motoneurons. Intracellular constant current pulse (bottom trace) elicits hyperpolarization of AAdC motoneuron (top trace). Note increase in size of voltage step (sharp downward deflection) after up‐training (C₂) relative to voltage step in naive preparation (C₁). Note also change in action potentials (small up‐ward deflections) after training.

Partial neural circuit mediating Pleurobranchaea feeding response and sites of altered neural response following aversive conditioning. Chemosensory neurons (SN) in oral veil produce both indirect excitation and inhibition of command neurons (CN) for feeding. Conditioned food stimuli produce less excitation and more inhibition (↓) in command neurons in animals that have had paired training than in animals that have had unpaired training or naive animals. This effect is partially accounted for by greater inhibition (↓) of interneurons (INT) that excite feeding command cells and greater excitation (↑) of interneurons that inhibit feeding command cells. MN, motoneuron. Triangles, excitatory synapses; circles, inhibitory synapses.

Classical conditioning in Hermissenda. A: acquisition, retention, and reacquisition of changes in response latencies to enter a test light following different training paradigms (RR, random rotation; RL, random light; ULR, unpaired light and rotation; RLR, random light and rotation; NLR, no light or rotation; PLR, paired light and rotation). Median response ratio compared latency during test (A) with pretraining base‐line response latency (B); values below 0.5 indicate increase in response latency. Group receiving paired light and rotation showed significantly greater suppression of phototactic response during both acquisition and retention phases than did control groups, indicating conditioning had occurred. During reacquisition, the 3 groups shown were subject to paired training. B: proposed neural circuit and sites of plasticity mediating conditioning in Hermissenda. Open circles, inhibitory synapses; open triangles, excitatory synapses. Following pairing of a light CS and a rotation US with caudal orientation [which excites caudal hair cells (CH)], there is an increase in excitability of type B photoreceptors (↑) and a decrease in excitability of type A photoreceptors (↓). Electrically coupled S and E cells of optic ganglion are shown as single cell. [A from Crow and Alkon 108, copyright 1978 by the American Association for the Advancement of Science; B adapted from Alkon 8.]

Electrophysiological analyses of changes in type B photoreceptors as function of conditioning in Hermissenda. A: comparison of input resistance of isolated (cut‐nerve), dark‐adapted type B photoreceptors from animals given paired or random presentations of light and rotation. 1, Representative linear current‐voltage relationship from experimental and random control groups. 2, Example of changes in membrane potential (top trace) to hyperpolarizing current pulses (bottom trace) in paired experimental and random control animals. Change in membrane potential to a given current pulse is larger following paired training relative to random control. B: example of voltage‐dependent outward currents in type B cells isolated from paired, random, and naive animals. Command pulses to 0 mV (bottom trace) elicit an early, inactivating K⁺ current (I_A) and a more slowly inactivating Ca²⁺‐dependent K⁺ current () (top trace). Records were chosen to illustrate reduction of I_A and for paired animals as compared with random and naive animals. C: intracellular voltage recordings of type B photoreceptors during and after presentation of unpaired light and rotation (1), light alone (2), or paired light and rotation (3). Responses to second of 2 succeeding 30‐s light steps are shown. Square pulse (solid line), light step. Ramped pulse (connected circles), rotational stimulus. Dashed lines, cell's initial resting potential preceding first light step; shaded areas, depolarization above resting level after second light step.

Differential classical conditioning of siphon‐withdrawal response in Aplysia. A: experimental preparation. Either siphon or mantle shelf could serve as conditioned stimulus (CS). Unconditioned stimulus (US) was an electric shock to tail. For illustrative purposes, parapodia and mantle shelf are shown retracted; however, all experiments were conducted using freely moving animals whose parapodia had been surgically removed. B: experimental paradigm. One group (SIPHON+) received siphon CS (CS⁺) followed by tail‐shock US and specifically unpaired mantle CS (CS⁻). Other group (MANTLE+) received mantle CS (CS⁺) followed by tail‐shock US and specifically unpaired siphon CS (CS⁻). Intertrial interval was 5 min. C: results from differential conditioning experiment using paradigm shown in B (n = 12). Testing was carried out 30 min after 15 training trials. SIPHON+ animals showed significantly longer siphon‐withdrawal responses to siphon CS than mantle CS (P < 0.05), while MANTLE+ animals showed significantly longer siphon‐withdrawal responses to mantle CS than to siphon CS (P < 0.01). Data in C and D expressed as means ± SE. D: pooled data from C. Test scores from CS⁻ and CS⁺ pathways following training are compared with their respective test scores obtained before training. Responses to stimuli delivered to CS⁺ pathway were significantly longer than responses to stimuli to CS⁻ pathway (P < 0.001), indicating differential conditioning had occurred.

Electrophysiological analysis of classical conditioning of siphon‐withdrawal reflex in Aplysia. A: experimental arrangement (A₁) and training protocol (A₂). Intracellular recordings were obtained in vitro from 2 different sensory neurons (SN) and siphon motoneuron (MN) to which they both projected. EPSP elicited in motoneuron by each sensory neuron was tested before and after training. Training consisted of trials in which train of spikes in 1 SN (paired SN) was followed by electric tail‐shock US; train of spikes in other SN (unpaired SN) was explicitly unpaired. Intertrial interval was 5 min. B: examples of EPSPs produced in siphon MN by paired SN and unpaired SN before (PRE) and 1 h after (POST) training protocol shown in A. Note that there is greater facilitation of EPSP elicited by paired SN than by unpaired SN. C: example of differential broadening of action potential in paired SN relative to an unpaired SN in the same preparation. Action potentials recorded before (PRE) and 3 h after (POST) 15 training trials using protocol shown in A. Experiment conducted in 50 mM TEA. D: removal of extracellular Ca²⁺ diminishes spike broadening produced by serotonin (5‐HT) paired with spike activity. Trains of action potentials in single SN were paired with brief (‐1 s) applications of serotonin alternately in normal seawater (10⁻² M Ca²⁺) and in Ca²⁺‐free seawater. Action‐potential duration was measured in normal seawater before and after pairing in each solution. Note that after action potentials were paired with serotonin application in normal seawater, increase in action‐potential duration was significantly greater than after pairing in absence of Ca²⁺ influx (n = 8, P < 0.001).