Cartesian representation of stimulus direction: Parallel processing by two sets of giant interneurons in the cockroach

The cockroachPeriplaneta americana responds to wind puffs by turning away, both on the ground and when flying. While on the ground, the ventral giant interneurons (ventrals) encode the wind direction and specify turn direction, whereas while flying the dorsal giant interneurons (dorsals) appear to do so. We report here on responses of these cells to controlled wind stimuli of different directions. Using improved methods of wind stimulation and of positioning the animal revealed important principles of organization not previously observed.All six cells of largest axonal diameter on each side respond preferentially to ipsilateral winds. One of these cells, previously thought to respond non-directionally (giant interneuron 2), was found to have a restricted directional response (Fig. 3). The organization of directional coding among the ventral giant interneurons is nearly identical to that among the dorsals (Fig. 2). Each group contains, on each side, one cell that responds primarily to wind from the ipsilateral front, another primarily in the ipsilateral rear, and a third responding more broadly to ipsilateral front and rear.These results are discussed in terms of the mechanisms of directional localization by the assembly of giant interneurons.Abbreviations GI giant interneuron - vGI ventral giant interneuron - dGI dorsal giant interneuron - CF 5-carboxyfluorescein - A6 6th abdominal ganglion - TI thoracic interneuron - BED best excitatory direction     

Motion analysis of leg joints associated with escape turns of the cockroach,Periplaneta americana

Considerable information is now available on the neural organization of the escape system of the American cockroach. To relate these data to the behavior, we need detailed information on the movements made at the principle leg joints that produce the turn. We used motion analysis of high speed video records to acquire such information. Records from both free ranging and tethered animals were analyzed. 1. We analyzed individual joint movements using a tethered preparation. Stimuli from 4 different angles around the animal were used. For all wind angles, the femur-tibia (FT) joint on the mesothoracic leg that is ipsilateral to the wind source extended while the contralateral mesothoracic FT joint flexed. This moved both of these legs laterally toward the wind source. In freely moving animals the FT movements provide forces that turn the animal away from the wind source. 2. The ipsilateral mesothoracic coxa-femur (CF) joint extended for all wind angles. The contralateral mesothoracic CF joint extended in response to most winds from the rear, but switched to flexion in response to wind from the side and front. As a result of these joint movements, rear wind resulted in rearward movements of the contralateral mesothoracic leg, while side and front wind resulted in more forward movements of that leg. 3. The CF and FT joints for both ipsilateral and contralateral metathoracic legs extended to wind from the rear and switched to flexion as the wind was placed at more anterior positions around the animal. In freely moving animals, extension of these joints would push the animal forward. Flexion would pull the animal backward. 4. Several of the joints showed correlations between rate of movement and initial joint angle. That is, joints that were already flexed at the onset of stimulation tended to move at a faster rate to a final position than joints that started at a more extended position. 5. Metathoracic FT and CF joints showed a high degree of positive correlation during the escape movements. Indeed, many curves showing movement of metathoracic FT and CF joints with time were virtually identical.     

Multivariate statistical analysis of the response of the cockroach giant interneuron system to wind puffs

The cockroach Periplaneta Americana responds to wind puffs by turning from the source of the wind. This turning movement is thought to be initiated by sensory responses to the wind puff occurring on filiform haircells on the cerci. The responses on the haircells initiate responses on 14 giant interneurons that rapidly transfer the information about the puff from the abdomen to the thorax and head. This response, in turn, is thought to initiate the escape response on the cockroach. In this paper a multivariate statistical analysis of the response of the 14 giant interneurons to wind puffs from a variety of angles is considered. A temporal-spatial integration model is proposed for the neural processing units acting as terminators for the giant interneurons and it is shown that using appropriate spatial integration, the cockroach can accurately estimate the direction of a wind puff on the basis of the response seen on the giant interneurons. It is also shown that with appropriate spatial integration, wind puffs from the left can be discriminated from wind puffs from the right. The right-left discrimination method is shown to be robust against the loss of any single giant interneuron. In some cases several giant interneurons can be lost without losing the right-left discrimination ability.     

Involvement of the suboesophageal and thoracic ganglia in the control of antennal movements in crickets

In crickets (Gryllus campestris, Gryllus bimaculatus) the contribution of the suboesophageal ganglia (SOG) and thoracic ganglia to the generation of antennal movements during visual tracking, walking and flight was investigated by the transection of connectives. Transection of one circumoesophageal connective abolished the movements and postures of the antenna ipsilateral to the lesion, while the contralateral antenna behaved normally. Simple antennal reflexes remained. Transection of one neck connective reduced fast components of antennal movements during tracking and walking. During flight the ipsilateral antenna could not be maintained in a prolonged forward position. Antennal movements during tracking and walking appeared normal after transection of one connective between pro- and mesothoracic ganglia. However, the antennal flight posture required uninterrupted connections between brain and mesothoracic ganglion. The ablation of more posterior ganglia had no effect on the antennal behaviours investigated. Recordings from an antennal motor nerve revealed a unilateral net excitation relayed via the SOG to the brain. Two ascending interneurones with activity closely correlated with antennal movements are candidates for such a relay function. The data show that the brain is not sufficient to generate antennal movements and postures as integral parts of several behaviours. The SOG and the thoracic ganglia are required in addition. Accepted: 12 March 1997     

Critical parameters of the spike trains in a cell assembly: coding of turn direction by the giant interneurons of the cockroach

Cockroaches (Periplaneta americana) respond to air displacement produced by an approaching predator by turning and running away. A set of 4 bilateral pairs of ventral giant interneurons is important in determining turn direction. Wind from a given side is known to produce more spikes, an earlier onset of the spike trains, and different fine temporal patterning, in the ipsilateral vs the contralateral set of these interneurons. Here we investigate which of these spike train parameters the cockroach actually uses to determine the direction it will turn.We delivered controlled wind puffs from the right front, together with intracellular injection of spike trains in a left ventral giant interneuron, under conditions where the animal could make normally directed turning movements of the legs and body. In trials where our stimuli caused the left side to give both the first spike and more total spikes than the right, but where our injected spike train included none of the normal fine temporal patterning, 92% of the evoked turns were to the rightopposite of normal (Figs. 4–6). In trials where the left side gave the first spike, but the right side gave more spikes, 100% of the turns were to the left-the normal direction (Figs. 8, 9). Comparable results were obtained when each of the left giant interneurons 1, 2 or 3 were electrically stimulated, and when either weak or stronger wind puffs were used. Stimulating a left giant interneuron electrically in the absence of a wind puff evoked an escape-like turn on 9% of the trials, and these were all to the right (Fig. 9).These results indicate that fine temporal patterning in the spike trains is not necessary, and information about which side gives the first spike is not sufficient, to determine turn direction. Rather, the key parameter appears to be relative numbers of action potentials in the left vs the right group of cells. These conclusions were supported by similar experiments in which extracellular stimulation of several left giant interneurons was paired with right wind (Figs. 11, 12).Abbreviations GI giant interneuron - vGI ventral giant interneuron - dGI dorsal giant interneuron - LY Lucifer yellow - CF carboxyfluorescein     

Attenuation of phrenic motor discharge by phrenic nerve afferents

Short latency phrenic motor responses to phrenic nerve stimulation were studied in anesthetized, paralyzed cats. Electrical stimulation (0.2 ms, 0.01-10 mA, 2 Hz) of the right C5 phrenic rootlet during inspiration consistently elicited a transient reduction in the phrenic motor discharge. This attenuation occurred bilaterally with an onset latency of 8-12 ms and a duration of 8-30 ms. Section of the ipsilateral C4-C6 dorsal roots abolished the response to stimulation, thereby confirming the involvement of phrenic nerve afferent activity. Stimulation of the left C5 phrenic rootlet or the right thoracic phrenic nerve usually elicited similar inhibitory responses. The difference in onset latency of responses to cervical vs. thoracic phrenic nerve stimulation indicates activation of group III afferents with a peripheral conduction velocity of approximately 10 m/s. A much shorter latency response (5 ms) was evoked ipsilaterally by thoracic phrenic nerve stimulation. Section of either the C5 or C6 dorsal root altered the ipsilateral response so that it resembled the longer latency contralateral response. The low-stimulus threshold and short latency for the ipsilateral response to thoracic phrenic nerve stimulation suggest that it involves larger diameter fibers. Decerebration, decerebellation, and transection of the dorsal columns at C2 do not abolish the inhibitory phrenic-to-phrenic reflex.     

Synaptic specificity in the first instar cockroach: patterns of monosynaptic input from filiform hair afferents to giant interneurons

Summary Direct evidence for monosynaptic connections between filiform hair sensory axons and giant interneurons (GIs) in the first instar cockroach, Periplaneta americana, was obtained using intracellular recording and HRP injection followed by electron microscopy. GIs 1–6 all receive monosynaptic input from at least one filiform afferent axon. GI1, GI2 and GI5 receive input only from the medial (M) axon, while GI3, GI4 and GI6 receive input from both M and lateral (L) axons. The dendrites of GI3 and GI6 which are contralateral to the cell bodies receive input from both axons whereas the smaller ipsilateral dendritic fields have synapses only from the L axon. GI5 has M axon input only onto its contralateral dendrites. In 50% of preparations GI7 receives weak input from the ipsilateral L axon. There is no obvious relationship between the morphology of the giant interneurons and the pattern of input they receive from the filiform afferents.Abbreviations GI giant interneuron - HRP horseradish peroxidase - L lateral axon - M medial axon     

Sensorimotor pathways involved in interjoint reflex action of an insect leg

Coordination of motor output between leg joints is crucial for the generation of posture and active movements in multijointed appendages of legged organisms. We investigated in the stick insect the information flow between the middle leg femoral chordotonal organ (fCO), which measures position and movement in the femur-tibia (FT) joint and the motoneuron pools supplying the next proximal leg joint, the coxa-trochanteral (CT) joint. In the inactive animal, elongation of the fCO (by flexing the FT joint) induced a depolarization in eight of nine levator trochanteris motoneurons, with a suprathreshold activation of one to three motoneurons. Motoneurons of the depressor trochanteris muscle were inhibited by fCO elongation. Relaxation signals, i.e., extension of the FT joint, activated both levator and depressor motoneurons; i.e., both antagonistic muscles were coactivated. Monosynaptic as well as polysynaptic pathways contribute to interjoint reflex actions in the stick insect leg. fCO afferents were found to induce short latency EPSPs in levator motoneurons, providing evidence for direct connections between fCO afferents and levator motoneurons. In addition, neuronal pathways via intercalated interneurons were identified that transmit sensory information from the fCO onto levator and/or depressor motoneurons. Finally, we describe two kinds of alterations in interjoint reflex action: (a) With repetitive sensory stimulation, this interjoint reflex action shows a habituation-like decrease in strength. (b) In the actively moving animal, interjoint reflex action in response to fCO elongation, mimicking joint flexion, qualitatively remained the same sign, but with a marked increase in strength, indicating an increased influence of sensory signals from the FT joint onto the adjacent CT joint in the active animal. © 1997 John Wiley & Sons, Inc. J Neurobiol 33: 891–913, 1997     

Identification of thoracic interneurons that mediate giant interneuron-to-motor pathways in the cockroach

Paired intracellular recordings were made to identify thoracic interneurons that receive stable short latency excitation from giant interneurons (GIs). Eight metathoracic interneurons were identified in which EPSPs were correlated with GI activity which was evoked either by wind or intracellular electrical stimulation or occurred spontaneously. In all cases EPSPs in the thoracic interneurons followed GI action potentials faithfully at short latencies. EPSPs associated with GI action potentials consistently represented the upper range of amplitudes of a large sample of EPSPs recorded in the thoracic interneurons. Seven of the interneurons were correlated with activity in ventral GIs but were not correlated with activity in dorsal GIs. Four of these interneurons were part of a discrete population of interneurons whose somata are located in the dorsal posterior region of the ganglion. The eighth interneuron (designated the T cell) was positively correlated with activity in dorsal GIs. The four dorsal posterior group interneurons and the T cell were depolarized intracellularly to establish their potential for generating motor activity. In all cases evoked activity was stronger in leg motor neurons (primarily Ds and the common inhibitor) located on the side contralateral to the interneuron's soma. The results indicate that significant polysynaptic pathways exist by which GI activity can evoke motor activity. The implications of this conclusion to investigations on the cockroach escape system are discussed.     

The central morphology of the giant interneurons and their spatial relationship with the thoracic motorneurons in the cockroach, Periplaneta americana (Insecta)

Three groups of giant fibers are found in the cockroach ventral nerve cord. A latero-dorsal group (dorsal GIs), a latero-ventral group (ventral GIs) and a medio-ventral group. The morphology of all three groups of fibers within the thoracic ganglia is described. The morphology of the dorsal and ventral GI pathways in the abdominal and suboesophageal ganglia is also described. The projection patterns of the neurons in each ganglion are remarkably similar which suggests a common function. When motorneurons 5rl (depressor) and 6Br4 (levator) are stained simultaneously with the dorsal and ventral GI groups, some branches from both motor and giant neurons converge. The branching of the remaining medio-ventral group of fibers and their proximity to areas receiving motorneuronal input suggests that these are the small diameter axons described by Dagan and Parnas (1970).     

Real-time imaging of neural activity during binaural interaction in the guinea pig auditory cortex

Spatio-temporal patterns of binaural interaction in the guinea pig auditory cortex (AC) were observed using optical recording with a 12 × 12 photodiode array and a voltage-sensitive dye. The amplitudes of the sound-induced light signals from the cortex were transformed into sequential two-dimensional images every 0.58 ms. Binaural sound stimuli evoked an excitatory response followed by a strong inhibition, and contralateral stimuli evoked a strong excitatory response followed by a weak inhibition. Ipsilateral sound stimuli evoked a weak response. Binaural stimulation induced two types of ipsilateral inhibition: a fast binaural inhibition which was detected only after the contralateral and ipsilateral responses were subtracted from the binaural responses, and which appeared 12–25 ms after the onset of stimulation, and a slow binaural inhibitory effect which was clearly observed in the binaural responses themselves, appearing 70–95 ms after the onset of stimulation. The fast binaural inhibition was observed in the same area as the contralateral excitatory response. The inhibited area became stronger and more widespread with increasing intensity of ipsilateral stimulation. We did not observe the specialized organization of binaural neurons as electrophysiologically found in the cat AC, in which binaural neurons of the same binaural response type are clustered together and alternate with clusters of other response types. Accepted: 14 August 1997     

Mechanisms underlying the cardioinhibitory and pressor responses elicited from the medullary neurons in the gigantocellular tegmental field of cats

A stimulation of the gigantocellular tegmental field (FTG) in the medulla oblongata often increases systemic arterial blood pressure (SAP) and decreases heart rate (HR). We investigated if the cardioinhibitory/depressor areas, including the nucleus ambiguus (NA), the dorsal motor nucleus of vagus (DMV) and the caudal ventrolateral medulla (CVLM), underlied the functional expression of FTG neurons in regulating cardiovascular responses. In 73 chloralose-urethane anesthetized cats, the HR, SAP and vertebral nerve activity (VNA) were recorded. Neurons in the FTG, NA, DMV and CVLM were stimulated by microinjection of sodium glutamate (25 mM Glu, 70 nl). To study if the NA, DMV, and CVLM relayed the cardioinhibitory messages from the FTG, 24 mM kainic acid (KA, 100 nl) was used as an excitotoxic agent to lesion neurons in the NA, DMV or CVLM. We found that the cardioinhibition induced by FTG stimulation was significantly reduced by KA lesioning of the ipsilateral NA or DMV. Subsequently, a bilateral KA lesion of NA or DMV abolished the cardioinhibitory responses of FTG. Compared to the consequence of KA lesion of the DMV, only a smaller bradycardia was induced by FTG stimulation after KA lesion of the NA. The pressor response induced by Glu stimulation of the FTG was reduced by the KA lesion of the CVLM. Such an effect was dominant ipsilaterally. Our findings suggested that both NA and DMV mediated the cardioinhibitory responses of FTG. The pressor message from the FTG neurons might be partly working via a disinhibitory mechanism through the depressor neurons located in the CVLM.     

Imaging the spatio-temporal dynamics of supragranular activity in the rat somatosensory cortex in response to stimulation of the paws

We employed voltage-sensitive dye (VSD) imaging to investigate the spatio-temporal dynamics of the responses of the supragranular somatosensory cortex to stimulation of the four paws in urethane-anesthetized rats. We obtained the following main results. (1) Stimulation of the contralateral forepaw evoked VSD responses with greater amplitude and smaller latency than stimulation of the contralateral hindpaw, and ipsilateral VSD responses had a lower amplitude and greater latency than contralateral responses. (2) While the contralateral stimulation initially activated only one focus, the ipsilateral stimulation initially activated two foci: one focus was typically medial to the focus activated by contralateral stimulation and was stereotaxically localized in the motor cortex; the other focus was typically posterior to the focus activated by contralateral stimulation and was stereotaxically localized in the somatosensory cortex. (3) Forepaw and hindpaw somatosensory stimuli activated large areas of the sensorimotor cortex, well beyond the forepaw and hindpaw somatosensory areas of classical somatotopic maps, and forepaw stimuli activated larger cortical areas with greater activation velocity than hindpaw stimuli. (4) Stimulation of the forepaw and hindpaw evoked different cortical activation dynamics: forepaw responses displayed a clear medial directionality, whereas hindpaw responses were much more uniform in all directions. In conclusion, this work offers a complete spatio-temporal map of the supragranular VSD cortical activation in response to stimulation of the paws, showing important somatotopic differences between contralateral and ipsilateral maps as well as differences in the spatio-temporal activation dynamics in response to forepaw and hindpaw stimuli.     

Electrophysiological responses from the peripheral olfactory system of the American eel,Anguilla rostrata

Summary A method is described for recording with microelectrodes from central neurones in locusts,Schistocerca gregaria americana, that are free to perform a large fraction of their behavioural repertoire. This tethered preparation has been used to examine the individual responses of large neurones in the neck connectives to a range of sensory stimuli.From differences in the responses of the units examined and from their positions in the connective, as determined by dye iontophoresis, 31 separate neurones have been identified. The axons of these cells had relatively constant diameters and cord positions in different animals and appeared in both right and left connectives but with their positions mirror reversed. The majority of these 31 cells carried descending information from the head ganglia and under our experimental conditions, 7 were found to have wind stimulation as their strongest sensory input, 17 had visual stimulation, 4 had sound stimulation and 3 had proprioceptive input.Abbreviations DCMD descending contralateral movement detector (neurone) - DIMD descending ipsilateral movement detector (neurone)     

The code for stimulus direction in a cell assembly in the cockroach

Summary The cockroachPeriplaneta americana responds to the approach of a predator by turning away. A gentle wind gust, caused by the predator's approach, excites cereal wind receptors, which encode both the presence and the direction of the stimulus. These cells in turn excite a group of giant interneurons (GI's) whose axons convey the directional information to thoracic motor centers. A given wind direction is coded not by a single GI functioning as a labeled line, but rather by some relationship among the spike trains in an assembly of GI's. This paper analyzes the code in this assembly.It is shown that all three pairs of GI's with the largest axonal diameters respond differentially to wind from left front vs. right front (Figs. 3, 4; Table 2). Each GI encodes these angles by both the time of its first action potential, and the number of action potentials, relative to its contralateral homolog. It is shown that the behavioral discrimination cannot rely solely upon the leftright differences in the time of the first action potential.A model of the assembly code is developed that involves a comparison of the numbers of action potentials in the left vs. the right group of giant interneurons. The model is shown to account for a large number of pre-existing experimental data on direction discrimination. The model requires, however, the involvement of additional cells in the left and right groups, besides the specific GI's whose role had been tested in prior experiments. The model is then tested by further experiments designed to verify the involvement of these added cells. These experiments support the model.Abbreviations A abdominal ganglion - GI giant interneuron     

Motor patterns during kicking movements in the locust

Locusts (Schistocerca gregaria) use a distinctive motor pattern to extend the tibia of a hind leg rapidly in a kick. The necessary force is generated by an almost isometric contraction of the extensor tibiae muscle restrained by the co-contraction of the flexor tibiae (co-contraction phase) and aided by the mechanics of the femoro-tibial joint. The stored energy is delivered suddenly when the flexor muscle is inhibited. This paper analyses the activity of motor neurons to the major hind leg muscles during kicking, and relates it to tibial movements and the resultant forces.During the co-contraction phase flexor tibiae motor neurons are driven by apparently common sources of synaptic inputs to depolarized plateaus at which they spike. The two excitatory extensor motor neurons are also depolarized by similar patterns of synaptic inputs, but with the slow producing more spikes at higher frequencies than the fast. Trochanteral depressors spike at high frequency, the single levator tarsi at low frequency, and common inhibitors 2 and 3 spike sporadically. Trochanteral levators, depressor tarsi, and a retractor unguis motor neuron are hyperpolarized.Before the tibia extends all flexor motor neurons are hyperpolarized simultaneously, two common inhibitors, and the levator trochanter and depressor tarsi motor neurons are depolarized. Later, but still before the tibial movement starts, the extensor tibiae and levator tarsi motor neurons are hyperpolarized. After the movement has started, the extensor motor neurons are hyperpolarized further and the depressor trochanteris motor neurons are also hyperpolarized, indicating a contribution of both central and sensory feedback pathways.Variations in the duration of the co-contraction of almost twenty-fold, and in the number of spikes in the fast extensor tibiae motor neuron from 2–50 produce a spectrum of tibial extensions ranging from slow and weak, to rapid and powerful. Flexibility in the networks producing the motor pattern therefore results in a range of movements suited to the fluctuating requirements of the animal.     

Responses of the rat cardiovascular system to substance P,neurotensin and bombesin

The carotid arterial blood pressure and heart rate responses to intravenous injections of substance P, neurotensin and bombesin were compared in anaesthetized rats. In rats anaesthetized with urethane neurotensin produced only a fall in blood pressure but in rats anaesthetized with sodium thiobutabarbitone, the fall was preceded by a transient rise in blood pressure. The reason for the different responses to neurotensin with the two anaesthetics was not investigated. The hypotensive effect of neurotensin observed with both anaesthetics was abolished by mepyramine and therefore appeared to be mediated by action on H₁ receptors either of neurotensin directly or of histamine released. On the other hand, catecholamines might be implicated in the pressor response to neurotensin observed in rats anaesthetized with sodium thiobutabarbitone since it was reduced by phentolamine and hexamethonium. Low doses of substance P produced a depressor response which was not inhibited by the antagonists tested. At higher doses marked tachycardia occurred and the depressor response was less and was often followed by a pressor response. The tachycardia was abolished by propranolol but not by cervical cord section or by hexamethonium. Bombesin produced a pressor response which was unaffected by hexamethonium but was reversed to depressor by phentolamine. This depressor response to bombesin was abolished by propranolol. It was concluded that substance P produced a depressor response by action on its own specific receptors and tachycardia by catecholamine release whereas neurotensin and bombesin produced cardiovascular actions which were mediated entirely by amine release.     

Development of cortical somatosensory evoked potentials in rats.

The development of the contra- and ipsilateral cortical potential evoked by electrical sciatic nerve stimulation was studied in 77 male albino rats aged 5 to 45 days. A contralateral response was already recorded, as double negativity, in the youngest animals, while an ipsilateral evoked potential was not reliably present until the 10th day. At this time, however, both responses started with an inconstant positive wave and their shape was practically the same. During subsequent development the responses differed only in respect to their dominant component: in the contralateral response, the N1 wave had the highest amplitude for most of the time, while in the ipsilateral response the delayed N2 wave was the largest component. The latent periods of contralateral responses were somewhat shorter than those of ipsilateral evoked potentials. During development we noticed a phase of abrupt shortening of the latent period, which took place before the 15th day in the contralateral response and before the 20th day in the ipsilateral response. We also found a difference in the fatigability of the responses, which was greater in immature rats than in adult animals; in the ipsilateral evoked potential it approached adult values more slowly. The development of the ipsilateral response is thus delayed compared with the development of the contralateral response.     

Behavioral recovery associated with central nervous system regeneration in the snail Melampus

The pulmonate snail Melampus bidentatus regenerates central nervous tracts following commissurotomy, connective transection, and cerebral ganglion ablation. Our goal was to determine whether or not neural regrowth within the central nervous system restored behaviors disrupted by lesions. One behavior that is disrupted by commissurotomy is retraction of facial structures that are contralateral to a stimulated facial region, a response that normally accompanies the ipsilateral retraction. Tentacle withdrawal on the side contralateral to stimulation reappeared on a timescale that was correlated with growth of a commissural link (8-19 days post-lesion). Electrophysiological recordings from a labial nerve pathway that has a contralateral component similar to the contralateral tentacle response showed that development or strengthening of an alternative pathway could also mediate contralateral responses. Thus, a major conclusion of this study was that both tract regeneration and changes in existing CNS pathways can underlie recovery. The percentage (approx. 75%) of snails that regenerate the cerebral commissure and show behavioral recovery is established early in the period following commissure transection. Behavioral recovery and anatomical evidence of regeneration were also correlated in the other two operations: single cerebral ganglion removal and unilateral cerebropleural and cerebropedal connective transection. We conclude that Melampus is able to regenerate neuronal connectivity that can restore normal behavior.