- Email: [email protected]
Identification of the midbrain locomotor nuclei and their descending pathways in the teleost carp, Cyprinus carpio
Brain Research 773 Ž1997. 1–7
Research report
Identification of the midbrain locomotor nuclei and their descending pathways in the teleost carp, Cyprinus carpio K. Uematsu ) , T. Todo Faculty of Applied Biological Science, Hiroshima UniÕersity, Higashi-Hiroshima, Hiroshima 739, Japan Accepted 30 April 1997
Abstract In order to identify the mesencephalic spinal pathways for initiation of swimming in the carp, we employed electrical and chemical microstimulation of the mesencephalic tegmentum. Electrical stimulation of the midbrain in decerebrate carp produced bilateral or unilateral rhythmic movements of the tail. Bilateral alternating movements were induced by stimulation with the lowest threshold currents to the brain region just beneath the third ventricle at the level of the mid mesencephalon. The region included the nucleus of medial longitudinal fasciculus ŽNflm., the medial longitudinal fasciculus Žflm., the red nucleus ŽNrb.. To specify the nuclei of the origin of the descending pathway, we microinjected 0.1 M L-glutamic acid to the region. Both bilateral and unilateral tail movements were induced, the majority being the latter. The unilateral movements were accompanied with tail flips toward the ipsilateral side of stimulation sites. The smallest injection volume required for initiation of the movement was recorded at the Nflm. Bilateral tail movements were produced only by injections into the medial region between the nucleus of the both sides. The present results imply a crucial role of Nflm neurons in the initiation of swimming Nflm neurons on one side project through flm to the ipsilateral spinal cord along its entire length and regulate activities of the individual central pattern generators. q 1997 Elsevier Science B.V. Keywords: Midbrain locomotor region; Swimming; Teleost; Electrical stimulation; Chemical stimulation
1. Introduction In animals that swim by an undulatory body movement such as many fish and amphibian tadpoles, the behavior is produced by an alternating activation of motoneurons on the left and the right side of each segment along the body w6,7,17x. When decerebrated, most teleost fishes, except for the eel, cannot swim at all w18x, suggesting that initiation of their swimming requires descending pathways activating the central pattern generator ŽCPG. located within the spinal cord w8x. Since the first success of induction of locomotion in cat by electrical stimulation of the midbrain locomotor region ŽMLR. w19x, similar experiments have been carried out in a variety of lower vertebrates including lamprey, stingray, and teleost fishes w2,3,10–12,20x. In
) Corresponding author. Fax: q81 Ž824. 240790; E-mail: [email protected]
0006-8993r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S 0 0 0 6 - 8 9 9 3 Ž 9 7 . 0 0 6 1 9 - 7
teleosts, the MLR is assumed to be localized beneath the ventricle on the midline in the midbrain tegmentum w3,10,20x. However, definite nuclei or nerve tracts corresponding to the MLR and its descending pathways were not yet identified, as the purpose of their study was originally to be able to elicit locomotion under experimental control and they also utilized relatively high stimulating currents to induce locomotion. Oka et al. retrogradely labeled brainstem nuclei projecting to the spinal cord of a salmon, and suggested that the nuclei concerned were the nucleus fasciculus longitudinalis medialis andror the nucleus ruber, because of their localities and morphology; neurons in both these nuclei project to the spinal cord w14x. In the present study, we applied electrical and chemical microstimulation to the midbrain tegmentum to identify the origins of the locomotor pathways to the spinal cord. To localize the nuclei involved, we adopted a method of microinjection of an excitatory amino acid to stimulate cell bodies w5x.
2
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
2. Materials and methods
plotted onto a standard brain map made from a fish by a conversion according to differences in brain size.
2.1. Electrical microstimulation 2.2. Chemical microstimulation Carp Ž Cyprinus carpio, body weight 50–102 g, n s 8. were anesthetized by immersion in 100 ppm tricaine methanesulfonate ŽCrescent Research Chemicals.. Fish were mounted in a holding apparatus, reconstructed from a stereotaxic apparatus for rat ŽNarishige ST-7. with some modification to stabilize fish with a mouthpiece, a pair of eye bars and a body clamp. Fish were fixed to the apparatus with a particular angle so that the head was rather higher than the tail. The ventral portion of the body and almost all the tail, was immersed in water in a tray situated below the fish. During surgery and the experiment, aerated water was run into the buccal cavity through the mouthpiece to irrigate the gills. While under anesthesia, the skull overlying the brain and the telencephalon were removed by dissection to eliminate spontaneous movements of the fish. After recovery from anesthesia, electrical stimulation was applied to the whole midbrain region by a monopolar stainless electrode ŽFHC a20-05-1, 9–12 M V . with a cathodal rectangular pulse Ž100 mA, 8 ms duration, 100 Hz.. The electrode was moved horizontally in 500-mm steps along the longitudinal and lateral axes of the fish, and in each track it was advanced in a 10-mm step perpendicularly. Induced movement of the fish body was recorded with a video camera ŽSony VX-1. during experiments. Two patterns of rhythmic bending of the body and tail resembling swimming were produced by the stimulation; symmetrical bilateral alternating movements and asymmetrical unilateral ones. We call them bilateral and unilateral movements, respectively. At sites where symmetrical movements were observed, stimulating currents were lowered in a 10-mA step to obtain thresholds. Finally, particular sites that gave lowest thresholds were marked by anodal DC currents of 10–20 mA for 10–20 s to deposit iron ions around the sites. After completion of experiments, the brain was removed and processed for visualization of iron deposits on sections by a Prussian Blue method. Each effective stimulation site was recorded as a 3-dimensional coordinate, where the origin of the coordinate axes was placed on an intersecting point made by a transversal plane through the rostral end of the optic tectum, a horizontal plane through the dorsal surface of the cerebellum and the medial sagittal plane. All data were
Chemical stimulation was carried out by using a glass micropipette ŽNarishige GD-1, tip diameter 16–24 mm., which was filled with 0.1 M L-glutamic acid and 2% Pontamine Sky Blue ŽNacalai Tesque. in a Ringer solution for fresh water fish, and connected to a microdispenser ŽDrummont 500. via a polyethylene tubing ŽHibiki No. 6. filled with silicon oil ŽToshiba Silicone TSF451-500.. Carp Ž Cyprinus carpio, body weight 60–100 g, n s 11. were attached to the recirculating fish holder and decerebrated. After recovery from anesthesia, the micropipette was inserted toward the midbrain locomotor region determined in the previous experiment. Injection was done with the solution of up to 500 nl at a rate 100 nlrmin until rhythmic bending of the body and tail was produced. If no movement occurred 10 min after 500 nl injection, the site was judged to be negative for induction of the movement. When the same site was injected twice, the second injection was done after an interval longer than 30 min. When more than two sites were injected in a fish, every electrode track was set at an interval of more than 1 mm horizontally. The movements induced were classified into bilateral and unilateral ones. We injected into a total of 119 sites in 11 fish, and among them 57 sites were effective Ž1–12 sites per fish.. At these effective sites, we measured injection volume, frequency and duration of the induced movements. As negative controls, two fish were injected with a Ringer solution into the midbrain locomotor region; no movement occurred with this treatment. After experiments, the brain was fixed with formalin, sectioned and viewed under a microscope to determine injection sites and the diffusion range of Pontamine Sky Blue.
3. Results 3.1. Midbrain locomotor regions and their descending pathways Both bi- and unilateral rhythmical body and tail movements were frequently produced by electrical stimulation of the reticular formation beneath the caudal diencephalon, the mesencephalon and the cerebellum. This wide region includes many nuclei and nerve tracts, such as the nucleus
Fig. 1. Locomotion evoked by electric microstimulation of the midbrain of the carp. Each pair of the brain map represents the results obtained at brain areas indicated by lines on the dorsal view of the brain at the top. The right drawing of each pair shows names of brain regions Žright half., and positive sites to evoke bilateral movement of the fish body with symbols indicating stimulation thresholds obtained on each site Žleft half.. The left drawing shows positive sites where unilateral body movement toward ipsilateral or contralateral side of the stimulation were recorded. Bar s 1 mm. Abbreviations: see Section 5.
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
3
4
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
fasciculus longitudinalis medialis ŽNflm., the nucleus ruber ŽNrb., the nucleus originis nervi oculomotoris ŽN3., the nucleus originis nervi trochlearis, the fasciculus longitudinalis medialis Žflm., the fasciculus retroflexus, the brachium conjunctivum Žr conj., the tractus mesencephalocerebellaris anterior, and the tractus rotundus Žtr.. For convenience of description, we divided the brain region concerned into eight transversal slices; namely five parts beneath the optic tectum and three parts beneath the corpus cerebellum level. On some occasions, positive stimulation sites were plotted outside the brain structure demarcated in the standard brain maps. This may be attributed to distortion of the brain tissue by tracking with electrodes. 3.1.1. Bilateral moÕements Stimulation of a small region ventrolateral to the ventricle at the most rostral midbrain level, only induced bilateral movements, when the current thresholds were very high, nearly 100 mA ŽFig. 1A.. At a slightly proximal level, some regions included in the diencephalon and the ventral optic tectum were positive for the stimulation-induced movements, in addition to the periventricular region ŽFig. 1B.. Positive sites in the ventral optic tectum were assumed to correspond to the tractus opticus and the commissura transversa. The sites in the periventricular area were located ventromedially to the ventricle, and the thresholds in this area were much lower than those in other regions ŽFig. 1B.. At the middle level of the tectum, there were many positive sites with extremely low thresholds Žsometimes lower than 20 mA. to induce the bilateral movements, in a relatively broad area in the medioventral region of the ventricle ŽFig. 1C.. This area included flm, Nflm, nervus oculomotorius Žn3. and decussatio brachium conjunctivum Ždr conj.. The red nucleus was located beside the area. No positive sites were found in the diencephalon. In the caudal midbrain, positive stimulation sites were located around the ventricle, especially in the dorsal and lateral aspects ŽFig. 1D and E.; flm, nucleus originis nervi trigemini ŽN5. and r conj might correspond to these sites. Among them, sites with low current thresholds Žlower than 20 mA. were restricted to the vicinity of the ventricle. At the rostral cerebellar level, the effective sites were located separately in the reticular formation away from the third ventricle ŽFig. 1F.. The current thresholds of these sites were relatively high, however sites with lower thresholds Žlower than 60 mA. might be located near flm and the tractus gustatorius secundarius ŽST.. At the middle and caudal cerebellar levels, every positive stimulation site could correspond to flm, while their current thresholds were relatively high Žhigher than 61 mA. ŽFig. 1G and H.. 3.1.2. Unilateral moÕements By stimulating a wide area in the rostral midbrain reticular formation, tail flips toward the one side ipsilateral to the stimulation sites were usually evoked ŽFig. 1A.. On the other hand, stimulation to hypothalamus in the same
levels caused movements toward the contralateral side ŽFig. 1A.. At the more proximal level, positive sites were centered on the reticular formation. However, ipsilateral and contralateral movement sites were restricted to dorsal and ventral aspects of the region, respectively ŽFig. 1B.. In the middle midbrain level, most unilateral movements induced were toward the ipsilateral side and were elicited by stimulation of a wide area ventral to the third ventricle, while movements toward the contralateral side were also induced when the medial ventral region were stimulated ŽFig. 1C.. In the caudal midbrain level, positive sites for both types of movement were segregated separately or scattered in the dorsal part of the reticular formation ŽFig. 1D and E.. At the cerebellar level, positive sites became fewer caudally and were scattered in the area. However, some were located beside flm and in the ventral aspect of the corpus cerebellum.
Fig. 2. Locomotion evoked by microinjection of L-glutamic acid to the midbrain tegmentum of the carp. Injections were carried out at the levels around the nucleus of medial longitudinal fasciculus. Maps A and C correspond to maps B and C in Fig. 1. Map B is the midst of maps B and C in Fig. 1. Both bilateral and unilateral body movement were induced by the injection. However, most are unilateral ones with tail beating toward ipsilateral side of the injection. Bilateral movement often induced by injection to the medial portion of the region. Bar s1 mm. Abbreviations: see Section 5.
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
3.2. Chemical microstimulation In this experiment, injection of glutamic acid solution around Nflm, Nrb, tr and flm caused both bi- and unilateral movements. Most movements occurred within 5 min following the onset of injection and the rest did so within 10 min after Pontamine Sky Blue was injected with glutamic acid diffused approximately 150 mm longitudinally and 100 mm laterally from the injected sites when 500 nl solution was injected. Bilateral movements of the body and tail were induced mostly when the solution was injected into the medioventral region of the ventricle, namely loci between flm and Nflm of the both sides ŽFig. 2.. The movements could be induced by injections of less than 50 nl solution, especially at the middle midbrain level ŽFig. 2A.. Unilateral movements occurred more frequently and the positive sites that required injection volumes less than 100 nl were located in the vicinity of the Nflm ŽFig. 2.. Injections into the Nflm required only 30 nl of the solution to cause the movement ŽFig. 2A.. All unilateral movement induced were tail flips toward the ipsilateral side of the injection sites. Injections to sites apart from Nflm or at the caudal midbrain level where Nflm was not found, needed injections of 100–500 nl solution to cause the movements ŽFig. 2.. The movements with frequencies of 72–126 timesrmin were induced by injection of 30–200 nl of glutamic acid solution and lasted for 6.5–15 min. At some positive sites, the effect of repetitive injection was examined with a interval of 30 min. The second injections required 100–500 nl of solution for inducing the movements, that lasted for 1–10 min and exhibited oscillations of 60–82 timesrmin.
4. Discussion In our electric microstimulation study, the positive sites for inducing the bilateral body and tail movement with the lowest stimulating current Žless than 20 mA. were localized around the ventral aspect of the ventricle at the middle and caudal midbrain level. Our data are consistent with those obtained by Kashin et al. w10x, Fetcho and Svoboda w3x and Uematsu and Ikeda w20x. They have demonstrated that such a region can produce swimming in fish. In these studies, the major aim was to artificially induce swimming, and they stimulated the brain with stronger currents than those used in this study. In contrast, we conducted this study in order to identify the neural structures, such as neuronal nuclei and nerve bundles, which are intimately involved with the initiation of fish swimming. For this, we stimulated the entire midbrain region and rostral medulla with relatively small currents Žless than 100 mA.. Around the positive region in our data there are some nuclei and nerve bundles, namely the nucleus fasciculus
5
longitudinalis medialis ŽNflm., the nucleus ruber ŽNrb., the nucleus originis nervi oculomotoris ŽN3., the nucleus originis nervi trochlearis ŽN4., the fasciculus longitudinalis medialis Žflm., the fasciculus retroflexus Žfrf., the brachium conjunctivum Žr conj., the decussatio brachium conjunctivum Ždr conj., and nervus oculomotorius Žn3.. They would be candidates for neuronal structures concerned with initiation of swimming. However, N3, N4 and n3 might be eliminated as candidates, since they govern eye movements. Also, frf that connects the habenula and the interpeduncular nucleus w13x, and r conj and dr conj, both of which send fibers to the optic tectum w4x, could also be deleted from the list of candidates. Hence, the structures closely associated with the initiation of swimming are supposed to include Nflm, Nrb and flm. Oka et al. w14x speculated from their retrograde labeling data in salmon that Kashin et al. w10x might have stimulated Nflm andror Nrb to induce carp swimming. Both nuclei in the zebrafish and the goldfish also project to the spinal cord w1,15,16x. Fetcho and Svoboda w3x succeeded in inducting fictive swimming in the goldfish and concluded that they had stimulated parts of flm as well as surrounding structures. In fact, in the present study, flm was the structure around which many positive sites were localized at the medulla level. From this, we are inclined to conclude here that flm is the major descending pathway from the hypothetical swimming center in the midbrain. In addition to the possible neural structure related to swimming initiation, there are many positive sites for inducing the bilateral movements in other structures with relatively high threshold currents. They are distributed in the optic tectum, hypothalamus and midbrain reticular formation, the latter of which included r conj, the nucleus originis nervi trigemini and the tractus gustatorius secundarius. Although the relationship of these structures with swimming initiation is obscure, fibers from these areas might make connections with the swimming center to contribute to the modification of the swimming mode. In any case, the higher current threshold found on positive sites in these areas may imply that they are not the core of the swimming center but may be able to influence it. Unilateral movements were caused by stimulation of broad areas including the fasciculus longitudinalis lateralis, and tractus mesencephalocerebellaris anterior and posterior. Since these tracts are involved in either afferent pathways mediating lateral line sensation or connections between the cerebellum and midbrain areas w13x, they might transmit modulatory signals to the midbrain swimming region. They might normally send signals intermittently to the region on an occasion when the fish decides to make a turn or to change swimming direction, but continual stimulation given in this study might cause repetitive tail flips toward one side. In midbrain and mesencephalic regions, small differences in depth of stimulation sites often produced changes in the direction of unilateral tail flips, suggesting that there is a complicated arrange-
6
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
ment of nerve tracts connected directly or indirectly with the swimming center. Our aim in the present study is to identify the nucleus of origin of the descending pathway for initiating fish swimming. For this, we made a microinjection of Lglutamic acid into the focused site, the ventral aspect of the third ventricle at the midbrain level, to eliminate the possibility of stimulating nerve tracts of passage. The concentration of the amino acid we used was much higher than those used in the stingray for a similar purpose w11x; however, our injection volume Ž20–500 nl. was much less than they adopted Ž100–1600 nl.. In other words, although approximately comparable quantities of the amino acid were injected in both studies, our stimulation sites were much more discretely focused. Minimal spread of the Pontamine Sky Blue injections with the amino acid also supports the reliability of our method. As a control, we injected Ringer’s solution into the locus concerned and observed no movements, indicating that the fish movements in this study were induced by the injection of the excitatory amino acid. Most chemically induced movements began during injection, suggesting that the stimulated cells were localized in the vicinity of stimulation sites. The nuclei situated around the sites in the present study were Nflm and Nrb. Both nuclei have neurons projecting to the spinal cord and are considered to be strong candidates for the midbrain swimming center w14x. Injections of relatively small volume of the amino acid into Nflm evoked tail flips, whereas a larger amount of injections to Nrb did not. Interestingly, the movements induced by injections to Nflm of the one side were always unilateral tail flips toward the ipsilateral side. This result suggest that neurons in Nflm of the one side send axons into the spinal cord and activate central pattern generators ŽCPG. located in that side. Bilateral, symmetrical tail movements were induced only when the amino acid was injected into the ventral aspect of the third ventricle along the midline. In this case, the injected amino acid could diffuse to the Nflm on both sides and give excitatory inputs to neurons in the nucleus. In mammals, locomotion is initiated by stimulation to the mesencephalic locomotor region and the signal is relayed by reticulospinal neurons to the spinal cord w9x. Reticular formation has also been shown to be the origin of descending locomotor pathways in stingrays w2x. These results appear to be inconsistent with our conclusion stated above that Nflm neurons directly project to the spinal cord. We interpret this discrepancy as follows. Most mammals and stingrays walk and swim by movements of their appendages, limb or the pectoral fins, while typical fish swimming is achieved by body undulations, suggesting that different motor systems might be subject to different brain mechanisms for the initiation of locomotion. Hence, it is probable that there might be two exclusive brain mechanisms for initiating fish swimming by axial muscles and paired fins. The former system would project to the
spinal cord directly and the latter would do so via reticulospinal neurons. 4.1. Conclusion The present stimulation experiments suggest that Nflm neurons may provide the excitatory drive onto the spinal CPGs to directly activate them, and that the neurons in Nflm of one side connect with the CPG array in the ipsilateral spinal hemisegment. Supposing that each set of Nflm and the spinal CPGs works independently, in a fish swimming straight both the Nflm would be activated simultaneously and equally by higher centers. In contrast, the strength or timing of the excitatory drive onto the nuclei could possibly differ when a fish makes a turn.
5. Abbreviations CC cp DI dr conj fll flm fr frf Nflm n3 N5 N6 n7 Nri Nrb Nr n8 r conj st tmca tmcp TO TS tr trop ttbc ttbr VC V3 V4
corpus cerebelli commissura posterior diencephalon decussatio brachium conjunctivum fasciculus longitudinalis lateralis fasciculus longitudinalis medialis formatio reticularis fasciculus retroflexus nucleus fasciculi longitudinalis medialis nervus oculomotorius nucleus originis nervi trigemini nucleus originis nervi abducentis nervus facialis nucleus reticularis inferior nucleus ruber nucleus raphes nervus octavi brachium conjunctivum tractus gustatorius secundarius tractus mesencephalocerebellaris anterior tractus mesencephalocerebellaris posterior tectum opticum torus semicircularis tractus rotundus tractus optics tractus tectobulbaris cruciatus tractus tectobulbaris rectus valvula cerebelli third ventricle fourth ventricle
Acknowledgements The authors acknowledge Dr. Keith Sillar for improving the manuscript. The investigation was supported partly by
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
a research grant to K.U. from the Ministry of Education, Science and Culture, Japan Ž05660209.. References w1x T. Becker, M.F. Wullimann, C.G. Becker, R.R. Bernhardt, M. Schachner, Axonal regrowth after spinal cord transection in adult zebrafish, J. Comp. Neurol. 377 Ž1997. 577–595. w2x N.A. Bernau, R.L. Purdrowski, R.B. Leonard, Identification of the midbrain locomotor region and its relation to descending locomotor pathways in the Atlantic stingray Dasyatis sabina, Brain Res. 557 Ž1991. 83–94. w3x J.R. Fetcho, K.R. Svoboda, Fictive swimming elicited by electrical stimulation of the midbrain in goldfish, J. Neurophysiol. 70 Ž1993. 764–780. w4x T.E. Finger, Organization of the teleost cerebellum, in: R.G. Northcutt, R.E. Davis ŽEds.., Fish Neurobiology, Vol. 1, University of Michigan Press, Ann Arbor, 1983, pp. 261–284. w5x A.K. Goodchild, R.A.L. Dampney, R. Bandler, A method for evoking physiological responses by stimulation of cell bodies, but not axons of passage, within localized regions of the central nervous system, J. Neurosci. Methods 6 Ž1982. 351–363. w6x S. Grillner, T. Matsushima, The neural network underlying locomotion in lamprey synaptic and cellular mechanisms, Neuron 7 Ž1991. 1–15. w7x S. Grillner, J.T. Buchanan, P. Wallen, ´ L. Brodin Ž1988. Neural control of locomotion in lower vertebrates: from behavior to ionic mechanisms, in: A.H. Cohen, S. Rossignol, S. Grillner ŽEds.., Neural Control of Rhythmic Movements in Vertebrates, John Wiley and Sons, New York, pp. 1–40. w8x S. Grillner, T. Deliagina, A Ekeberg, A. El Manira, R.H. Hill, A. Lansner, G.N. Orlovsky, P. Wallen, ´ Neural networks that co-ordinate locomotion and body orientation in lamprey, Trends Neurosci. 18 Ž1995. 270–279. w9x L.M. Jordan, Initiation of locomotion from the mammalian brainstem, in: S. Grillner, P.S.G. Stein, D.G. Stuart, H. Forssberg, R.M.
w10x
w11x
w12x
w13x
w14x
w15x
w16x
w17x
w18x
w19x
w20x
7
Herman ŽEds.., Neurobiology of Vertebrates Locomotion, Macmillan, London, 1986, pp. 21–37. S.M. Kashin, A.G. Feldman, G.N. Orlovsky, Locomotion of fish evoked by electrical stimulation of the brain, Brain Res. 82 Ž1974. 41–47. C.A. Livingston, R.B. Leonard, Locomotion evoked by stimulation of the brain stem in the Atlantic stingray Dasyatis sabina, J. Neurosci. 10 Ž1990. 194–204. A.D. McClellan, S. Grillner, Activation of ‘fictive swimming’ by electrical microstimulation of brainstem locomotor regions in an in vitro preparation of the lamprey central nervous system, Brain Res. 300 Ž1984. 357–361. R. Nieuwenhuys, E. Pouwels, The brain stem of actinopterygian fishes, in: R.G. Northcutt, R.E. Davis ŽEds.., Fish Neurobiology, vol. 1, University of Michigan Press, Ann Arbor, 1983, pp. 25–88. Y. Oka, M. Satou, K. Ueda, Descending pathways to the spinal cord in the hime´ salmon Žlandlocked red salmon Oncorhynchus nerca., J. Comp. Neurol. 254 Ž1986. 91–103. P.D. Prasada Rao, Descending projection neurons to the spinal cord of the goldfish Carassius auratus, J. Comp. Neurol. 265 Ž1987. 96–108. P.D. Prasada Rao, A.G. Jadhao, S.C. Sharma, Topographic organization of descending projection neurons to the spinal cord of the goldfish Carassius auratus, Brain Res. 620 Ž1993. 211–220. A. Roberts, S.R. Soffe, N. Dale, Spinal interneurones and swimming in frog embryos, in: S. Grillner, P.S.G. Stein, D.G. Stuart, H. Forssberg, R.M. Herman ŽEds.., Neurobiology of Vertebrates Locomotion, Macmillan, London, 1986, pp. 279–306. B.L. Roberts, The organization of the nervous system of fishes in relation to locomotion, in: M.H. Day ŽEd.., Vertebrate Locomotion, Academic Press, London, 1981, pp. 115–136. M.L. Shik, F.V. Severin, G.N. Orlovsky, Control of walking by means of electrical stimulation of the midbrain, Biophysics 11 Ž1966. 756–765. K. Uematsu, T. Ikeda, The midbrain locomotor region and induced swimming in the carp Cyprinus carpio, Nippon Suisan Gakkaishi 59 Ž1993. 783–788.
Research report
Identification of the midbrain locomotor nuclei and their descending pathways in the teleost carp, Cyprinus carpio K. Uematsu ) , T. Todo Faculty of Applied Biological Science, Hiroshima UniÕersity, Higashi-Hiroshima, Hiroshima 739, Japan Accepted 30 April 1997
Abstract In order to identify the mesencephalic spinal pathways for initiation of swimming in the carp, we employed electrical and chemical microstimulation of the mesencephalic tegmentum. Electrical stimulation of the midbrain in decerebrate carp produced bilateral or unilateral rhythmic movements of the tail. Bilateral alternating movements were induced by stimulation with the lowest threshold currents to the brain region just beneath the third ventricle at the level of the mid mesencephalon. The region included the nucleus of medial longitudinal fasciculus ŽNflm., the medial longitudinal fasciculus Žflm., the red nucleus ŽNrb.. To specify the nuclei of the origin of the descending pathway, we microinjected 0.1 M L-glutamic acid to the region. Both bilateral and unilateral tail movements were induced, the majority being the latter. The unilateral movements were accompanied with tail flips toward the ipsilateral side of stimulation sites. The smallest injection volume required for initiation of the movement was recorded at the Nflm. Bilateral tail movements were produced only by injections into the medial region between the nucleus of the both sides. The present results imply a crucial role of Nflm neurons in the initiation of swimming Nflm neurons on one side project through flm to the ipsilateral spinal cord along its entire length and regulate activities of the individual central pattern generators. q 1997 Elsevier Science B.V. Keywords: Midbrain locomotor region; Swimming; Teleost; Electrical stimulation; Chemical stimulation
1. Introduction In animals that swim by an undulatory body movement such as many fish and amphibian tadpoles, the behavior is produced by an alternating activation of motoneurons on the left and the right side of each segment along the body w6,7,17x. When decerebrated, most teleost fishes, except for the eel, cannot swim at all w18x, suggesting that initiation of their swimming requires descending pathways activating the central pattern generator ŽCPG. located within the spinal cord w8x. Since the first success of induction of locomotion in cat by electrical stimulation of the midbrain locomotor region ŽMLR. w19x, similar experiments have been carried out in a variety of lower vertebrates including lamprey, stingray, and teleost fishes w2,3,10–12,20x. In
) Corresponding author. Fax: q81 Ž824. 240790; E-mail: [email protected]
0006-8993r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S 0 0 0 6 - 8 9 9 3 Ž 9 7 . 0 0 6 1 9 - 7
teleosts, the MLR is assumed to be localized beneath the ventricle on the midline in the midbrain tegmentum w3,10,20x. However, definite nuclei or nerve tracts corresponding to the MLR and its descending pathways were not yet identified, as the purpose of their study was originally to be able to elicit locomotion under experimental control and they also utilized relatively high stimulating currents to induce locomotion. Oka et al. retrogradely labeled brainstem nuclei projecting to the spinal cord of a salmon, and suggested that the nuclei concerned were the nucleus fasciculus longitudinalis medialis andror the nucleus ruber, because of their localities and morphology; neurons in both these nuclei project to the spinal cord w14x. In the present study, we applied electrical and chemical microstimulation to the midbrain tegmentum to identify the origins of the locomotor pathways to the spinal cord. To localize the nuclei involved, we adopted a method of microinjection of an excitatory amino acid to stimulate cell bodies w5x.
2
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
2. Materials and methods
plotted onto a standard brain map made from a fish by a conversion according to differences in brain size.
2.1. Electrical microstimulation 2.2. Chemical microstimulation Carp Ž Cyprinus carpio, body weight 50–102 g, n s 8. were anesthetized by immersion in 100 ppm tricaine methanesulfonate ŽCrescent Research Chemicals.. Fish were mounted in a holding apparatus, reconstructed from a stereotaxic apparatus for rat ŽNarishige ST-7. with some modification to stabilize fish with a mouthpiece, a pair of eye bars and a body clamp. Fish were fixed to the apparatus with a particular angle so that the head was rather higher than the tail. The ventral portion of the body and almost all the tail, was immersed in water in a tray situated below the fish. During surgery and the experiment, aerated water was run into the buccal cavity through the mouthpiece to irrigate the gills. While under anesthesia, the skull overlying the brain and the telencephalon were removed by dissection to eliminate spontaneous movements of the fish. After recovery from anesthesia, electrical stimulation was applied to the whole midbrain region by a monopolar stainless electrode ŽFHC a20-05-1, 9–12 M V . with a cathodal rectangular pulse Ž100 mA, 8 ms duration, 100 Hz.. The electrode was moved horizontally in 500-mm steps along the longitudinal and lateral axes of the fish, and in each track it was advanced in a 10-mm step perpendicularly. Induced movement of the fish body was recorded with a video camera ŽSony VX-1. during experiments. Two patterns of rhythmic bending of the body and tail resembling swimming were produced by the stimulation; symmetrical bilateral alternating movements and asymmetrical unilateral ones. We call them bilateral and unilateral movements, respectively. At sites where symmetrical movements were observed, stimulating currents were lowered in a 10-mA step to obtain thresholds. Finally, particular sites that gave lowest thresholds were marked by anodal DC currents of 10–20 mA for 10–20 s to deposit iron ions around the sites. After completion of experiments, the brain was removed and processed for visualization of iron deposits on sections by a Prussian Blue method. Each effective stimulation site was recorded as a 3-dimensional coordinate, where the origin of the coordinate axes was placed on an intersecting point made by a transversal plane through the rostral end of the optic tectum, a horizontal plane through the dorsal surface of the cerebellum and the medial sagittal plane. All data were
Chemical stimulation was carried out by using a glass micropipette ŽNarishige GD-1, tip diameter 16–24 mm., which was filled with 0.1 M L-glutamic acid and 2% Pontamine Sky Blue ŽNacalai Tesque. in a Ringer solution for fresh water fish, and connected to a microdispenser ŽDrummont 500. via a polyethylene tubing ŽHibiki No. 6. filled with silicon oil ŽToshiba Silicone TSF451-500.. Carp Ž Cyprinus carpio, body weight 60–100 g, n s 11. were attached to the recirculating fish holder and decerebrated. After recovery from anesthesia, the micropipette was inserted toward the midbrain locomotor region determined in the previous experiment. Injection was done with the solution of up to 500 nl at a rate 100 nlrmin until rhythmic bending of the body and tail was produced. If no movement occurred 10 min after 500 nl injection, the site was judged to be negative for induction of the movement. When the same site was injected twice, the second injection was done after an interval longer than 30 min. When more than two sites were injected in a fish, every electrode track was set at an interval of more than 1 mm horizontally. The movements induced were classified into bilateral and unilateral ones. We injected into a total of 119 sites in 11 fish, and among them 57 sites were effective Ž1–12 sites per fish.. At these effective sites, we measured injection volume, frequency and duration of the induced movements. As negative controls, two fish were injected with a Ringer solution into the midbrain locomotor region; no movement occurred with this treatment. After experiments, the brain was fixed with formalin, sectioned and viewed under a microscope to determine injection sites and the diffusion range of Pontamine Sky Blue.
3. Results 3.1. Midbrain locomotor regions and their descending pathways Both bi- and unilateral rhythmical body and tail movements were frequently produced by electrical stimulation of the reticular formation beneath the caudal diencephalon, the mesencephalon and the cerebellum. This wide region includes many nuclei and nerve tracts, such as the nucleus
Fig. 1. Locomotion evoked by electric microstimulation of the midbrain of the carp. Each pair of the brain map represents the results obtained at brain areas indicated by lines on the dorsal view of the brain at the top. The right drawing of each pair shows names of brain regions Žright half., and positive sites to evoke bilateral movement of the fish body with symbols indicating stimulation thresholds obtained on each site Žleft half.. The left drawing shows positive sites where unilateral body movement toward ipsilateral or contralateral side of the stimulation were recorded. Bar s 1 mm. Abbreviations: see Section 5.
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
3
4
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
fasciculus longitudinalis medialis ŽNflm., the nucleus ruber ŽNrb., the nucleus originis nervi oculomotoris ŽN3., the nucleus originis nervi trochlearis, the fasciculus longitudinalis medialis Žflm., the fasciculus retroflexus, the brachium conjunctivum Žr conj., the tractus mesencephalocerebellaris anterior, and the tractus rotundus Žtr.. For convenience of description, we divided the brain region concerned into eight transversal slices; namely five parts beneath the optic tectum and three parts beneath the corpus cerebellum level. On some occasions, positive stimulation sites were plotted outside the brain structure demarcated in the standard brain maps. This may be attributed to distortion of the brain tissue by tracking with electrodes. 3.1.1. Bilateral moÕements Stimulation of a small region ventrolateral to the ventricle at the most rostral midbrain level, only induced bilateral movements, when the current thresholds were very high, nearly 100 mA ŽFig. 1A.. At a slightly proximal level, some regions included in the diencephalon and the ventral optic tectum were positive for the stimulation-induced movements, in addition to the periventricular region ŽFig. 1B.. Positive sites in the ventral optic tectum were assumed to correspond to the tractus opticus and the commissura transversa. The sites in the periventricular area were located ventromedially to the ventricle, and the thresholds in this area were much lower than those in other regions ŽFig. 1B.. At the middle level of the tectum, there were many positive sites with extremely low thresholds Žsometimes lower than 20 mA. to induce the bilateral movements, in a relatively broad area in the medioventral region of the ventricle ŽFig. 1C.. This area included flm, Nflm, nervus oculomotorius Žn3. and decussatio brachium conjunctivum Ždr conj.. The red nucleus was located beside the area. No positive sites were found in the diencephalon. In the caudal midbrain, positive stimulation sites were located around the ventricle, especially in the dorsal and lateral aspects ŽFig. 1D and E.; flm, nucleus originis nervi trigemini ŽN5. and r conj might correspond to these sites. Among them, sites with low current thresholds Žlower than 20 mA. were restricted to the vicinity of the ventricle. At the rostral cerebellar level, the effective sites were located separately in the reticular formation away from the third ventricle ŽFig. 1F.. The current thresholds of these sites were relatively high, however sites with lower thresholds Žlower than 60 mA. might be located near flm and the tractus gustatorius secundarius ŽST.. At the middle and caudal cerebellar levels, every positive stimulation site could correspond to flm, while their current thresholds were relatively high Žhigher than 61 mA. ŽFig. 1G and H.. 3.1.2. Unilateral moÕements By stimulating a wide area in the rostral midbrain reticular formation, tail flips toward the one side ipsilateral to the stimulation sites were usually evoked ŽFig. 1A.. On the other hand, stimulation to hypothalamus in the same
levels caused movements toward the contralateral side ŽFig. 1A.. At the more proximal level, positive sites were centered on the reticular formation. However, ipsilateral and contralateral movement sites were restricted to dorsal and ventral aspects of the region, respectively ŽFig. 1B.. In the middle midbrain level, most unilateral movements induced were toward the ipsilateral side and were elicited by stimulation of a wide area ventral to the third ventricle, while movements toward the contralateral side were also induced when the medial ventral region were stimulated ŽFig. 1C.. In the caudal midbrain level, positive sites for both types of movement were segregated separately or scattered in the dorsal part of the reticular formation ŽFig. 1D and E.. At the cerebellar level, positive sites became fewer caudally and were scattered in the area. However, some were located beside flm and in the ventral aspect of the corpus cerebellum.
Fig. 2. Locomotion evoked by microinjection of L-glutamic acid to the midbrain tegmentum of the carp. Injections were carried out at the levels around the nucleus of medial longitudinal fasciculus. Maps A and C correspond to maps B and C in Fig. 1. Map B is the midst of maps B and C in Fig. 1. Both bilateral and unilateral body movement were induced by the injection. However, most are unilateral ones with tail beating toward ipsilateral side of the injection. Bilateral movement often induced by injection to the medial portion of the region. Bar s1 mm. Abbreviations: see Section 5.
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
3.2. Chemical microstimulation In this experiment, injection of glutamic acid solution around Nflm, Nrb, tr and flm caused both bi- and unilateral movements. Most movements occurred within 5 min following the onset of injection and the rest did so within 10 min after Pontamine Sky Blue was injected with glutamic acid diffused approximately 150 mm longitudinally and 100 mm laterally from the injected sites when 500 nl solution was injected. Bilateral movements of the body and tail were induced mostly when the solution was injected into the medioventral region of the ventricle, namely loci between flm and Nflm of the both sides ŽFig. 2.. The movements could be induced by injections of less than 50 nl solution, especially at the middle midbrain level ŽFig. 2A.. Unilateral movements occurred more frequently and the positive sites that required injection volumes less than 100 nl were located in the vicinity of the Nflm ŽFig. 2.. Injections into the Nflm required only 30 nl of the solution to cause the movement ŽFig. 2A.. All unilateral movement induced were tail flips toward the ipsilateral side of the injection sites. Injections to sites apart from Nflm or at the caudal midbrain level where Nflm was not found, needed injections of 100–500 nl solution to cause the movements ŽFig. 2.. The movements with frequencies of 72–126 timesrmin were induced by injection of 30–200 nl of glutamic acid solution and lasted for 6.5–15 min. At some positive sites, the effect of repetitive injection was examined with a interval of 30 min. The second injections required 100–500 nl of solution for inducing the movements, that lasted for 1–10 min and exhibited oscillations of 60–82 timesrmin.
4. Discussion In our electric microstimulation study, the positive sites for inducing the bilateral body and tail movement with the lowest stimulating current Žless than 20 mA. were localized around the ventral aspect of the ventricle at the middle and caudal midbrain level. Our data are consistent with those obtained by Kashin et al. w10x, Fetcho and Svoboda w3x and Uematsu and Ikeda w20x. They have demonstrated that such a region can produce swimming in fish. In these studies, the major aim was to artificially induce swimming, and they stimulated the brain with stronger currents than those used in this study. In contrast, we conducted this study in order to identify the neural structures, such as neuronal nuclei and nerve bundles, which are intimately involved with the initiation of fish swimming. For this, we stimulated the entire midbrain region and rostral medulla with relatively small currents Žless than 100 mA.. Around the positive region in our data there are some nuclei and nerve bundles, namely the nucleus fasciculus
5
longitudinalis medialis ŽNflm., the nucleus ruber ŽNrb., the nucleus originis nervi oculomotoris ŽN3., the nucleus originis nervi trochlearis ŽN4., the fasciculus longitudinalis medialis Žflm., the fasciculus retroflexus Žfrf., the brachium conjunctivum Žr conj., the decussatio brachium conjunctivum Ždr conj., and nervus oculomotorius Žn3.. They would be candidates for neuronal structures concerned with initiation of swimming. However, N3, N4 and n3 might be eliminated as candidates, since they govern eye movements. Also, frf that connects the habenula and the interpeduncular nucleus w13x, and r conj and dr conj, both of which send fibers to the optic tectum w4x, could also be deleted from the list of candidates. Hence, the structures closely associated with the initiation of swimming are supposed to include Nflm, Nrb and flm. Oka et al. w14x speculated from their retrograde labeling data in salmon that Kashin et al. w10x might have stimulated Nflm andror Nrb to induce carp swimming. Both nuclei in the zebrafish and the goldfish also project to the spinal cord w1,15,16x. Fetcho and Svoboda w3x succeeded in inducting fictive swimming in the goldfish and concluded that they had stimulated parts of flm as well as surrounding structures. In fact, in the present study, flm was the structure around which many positive sites were localized at the medulla level. From this, we are inclined to conclude here that flm is the major descending pathway from the hypothetical swimming center in the midbrain. In addition to the possible neural structure related to swimming initiation, there are many positive sites for inducing the bilateral movements in other structures with relatively high threshold currents. They are distributed in the optic tectum, hypothalamus and midbrain reticular formation, the latter of which included r conj, the nucleus originis nervi trigemini and the tractus gustatorius secundarius. Although the relationship of these structures with swimming initiation is obscure, fibers from these areas might make connections with the swimming center to contribute to the modification of the swimming mode. In any case, the higher current threshold found on positive sites in these areas may imply that they are not the core of the swimming center but may be able to influence it. Unilateral movements were caused by stimulation of broad areas including the fasciculus longitudinalis lateralis, and tractus mesencephalocerebellaris anterior and posterior. Since these tracts are involved in either afferent pathways mediating lateral line sensation or connections between the cerebellum and midbrain areas w13x, they might transmit modulatory signals to the midbrain swimming region. They might normally send signals intermittently to the region on an occasion when the fish decides to make a turn or to change swimming direction, but continual stimulation given in this study might cause repetitive tail flips toward one side. In midbrain and mesencephalic regions, small differences in depth of stimulation sites often produced changes in the direction of unilateral tail flips, suggesting that there is a complicated arrange-
6
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
ment of nerve tracts connected directly or indirectly with the swimming center. Our aim in the present study is to identify the nucleus of origin of the descending pathway for initiating fish swimming. For this, we made a microinjection of Lglutamic acid into the focused site, the ventral aspect of the third ventricle at the midbrain level, to eliminate the possibility of stimulating nerve tracts of passage. The concentration of the amino acid we used was much higher than those used in the stingray for a similar purpose w11x; however, our injection volume Ž20–500 nl. was much less than they adopted Ž100–1600 nl.. In other words, although approximately comparable quantities of the amino acid were injected in both studies, our stimulation sites were much more discretely focused. Minimal spread of the Pontamine Sky Blue injections with the amino acid also supports the reliability of our method. As a control, we injected Ringer’s solution into the locus concerned and observed no movements, indicating that the fish movements in this study were induced by the injection of the excitatory amino acid. Most chemically induced movements began during injection, suggesting that the stimulated cells were localized in the vicinity of stimulation sites. The nuclei situated around the sites in the present study were Nflm and Nrb. Both nuclei have neurons projecting to the spinal cord and are considered to be strong candidates for the midbrain swimming center w14x. Injections of relatively small volume of the amino acid into Nflm evoked tail flips, whereas a larger amount of injections to Nrb did not. Interestingly, the movements induced by injections to Nflm of the one side were always unilateral tail flips toward the ipsilateral side. This result suggest that neurons in Nflm of the one side send axons into the spinal cord and activate central pattern generators ŽCPG. located in that side. Bilateral, symmetrical tail movements were induced only when the amino acid was injected into the ventral aspect of the third ventricle along the midline. In this case, the injected amino acid could diffuse to the Nflm on both sides and give excitatory inputs to neurons in the nucleus. In mammals, locomotion is initiated by stimulation to the mesencephalic locomotor region and the signal is relayed by reticulospinal neurons to the spinal cord w9x. Reticular formation has also been shown to be the origin of descending locomotor pathways in stingrays w2x. These results appear to be inconsistent with our conclusion stated above that Nflm neurons directly project to the spinal cord. We interpret this discrepancy as follows. Most mammals and stingrays walk and swim by movements of their appendages, limb or the pectoral fins, while typical fish swimming is achieved by body undulations, suggesting that different motor systems might be subject to different brain mechanisms for the initiation of locomotion. Hence, it is probable that there might be two exclusive brain mechanisms for initiating fish swimming by axial muscles and paired fins. The former system would project to the
spinal cord directly and the latter would do so via reticulospinal neurons. 4.1. Conclusion The present stimulation experiments suggest that Nflm neurons may provide the excitatory drive onto the spinal CPGs to directly activate them, and that the neurons in Nflm of one side connect with the CPG array in the ipsilateral spinal hemisegment. Supposing that each set of Nflm and the spinal CPGs works independently, in a fish swimming straight both the Nflm would be activated simultaneously and equally by higher centers. In contrast, the strength or timing of the excitatory drive onto the nuclei could possibly differ when a fish makes a turn.
5. Abbreviations CC cp DI dr conj fll flm fr frf Nflm n3 N5 N6 n7 Nri Nrb Nr n8 r conj st tmca tmcp TO TS tr trop ttbc ttbr VC V3 V4
corpus cerebelli commissura posterior diencephalon decussatio brachium conjunctivum fasciculus longitudinalis lateralis fasciculus longitudinalis medialis formatio reticularis fasciculus retroflexus nucleus fasciculi longitudinalis medialis nervus oculomotorius nucleus originis nervi trigemini nucleus originis nervi abducentis nervus facialis nucleus reticularis inferior nucleus ruber nucleus raphes nervus octavi brachium conjunctivum tractus gustatorius secundarius tractus mesencephalocerebellaris anterior tractus mesencephalocerebellaris posterior tectum opticum torus semicircularis tractus rotundus tractus optics tractus tectobulbaris cruciatus tractus tectobulbaris rectus valvula cerebelli third ventricle fourth ventricle
Acknowledgements The authors acknowledge Dr. Keith Sillar for improving the manuscript. The investigation was supported partly by
K. Uematsu, T. Todor Brain Research 773 (1997) 1–7
a research grant to K.U. from the Ministry of Education, Science and Culture, Japan Ž05660209.. References w1x T. Becker, M.F. Wullimann, C.G. Becker, R.R. Bernhardt, M. Schachner, Axonal regrowth after spinal cord transection in adult zebrafish, J. Comp. Neurol. 377 Ž1997. 577–595. w2x N.A. Bernau, R.L. Purdrowski, R.B. Leonard, Identification of the midbrain locomotor region and its relation to descending locomotor pathways in the Atlantic stingray Dasyatis sabina, Brain Res. 557 Ž1991. 83–94. w3x J.R. Fetcho, K.R. Svoboda, Fictive swimming elicited by electrical stimulation of the midbrain in goldfish, J. Neurophysiol. 70 Ž1993. 764–780. w4x T.E. Finger, Organization of the teleost cerebellum, in: R.G. Northcutt, R.E. Davis ŽEds.., Fish Neurobiology, Vol. 1, University of Michigan Press, Ann Arbor, 1983, pp. 261–284. w5x A.K. Goodchild, R.A.L. Dampney, R. Bandler, A method for evoking physiological responses by stimulation of cell bodies, but not axons of passage, within localized regions of the central nervous system, J. Neurosci. Methods 6 Ž1982. 351–363. w6x S. Grillner, T. Matsushima, The neural network underlying locomotion in lamprey synaptic and cellular mechanisms, Neuron 7 Ž1991. 1–15. w7x S. Grillner, J.T. Buchanan, P. Wallen, ´ L. Brodin Ž1988. Neural control of locomotion in lower vertebrates: from behavior to ionic mechanisms, in: A.H. Cohen, S. Rossignol, S. Grillner ŽEds.., Neural Control of Rhythmic Movements in Vertebrates, John Wiley and Sons, New York, pp. 1–40. w8x S. Grillner, T. Deliagina, A Ekeberg, A. El Manira, R.H. Hill, A. Lansner, G.N. Orlovsky, P. Wallen, ´ Neural networks that co-ordinate locomotion and body orientation in lamprey, Trends Neurosci. 18 Ž1995. 270–279. w9x L.M. Jordan, Initiation of locomotion from the mammalian brainstem, in: S. Grillner, P.S.G. Stein, D.G. Stuart, H. Forssberg, R.M.
w10x
w11x
w12x
w13x
w14x
w15x
w16x
w17x
w18x
w19x
w20x
7
Herman ŽEds.., Neurobiology of Vertebrates Locomotion, Macmillan, London, 1986, pp. 21–37. S.M. Kashin, A.G. Feldman, G.N. Orlovsky, Locomotion of fish evoked by electrical stimulation of the brain, Brain Res. 82 Ž1974. 41–47. C.A. Livingston, R.B. Leonard, Locomotion evoked by stimulation of the brain stem in the Atlantic stingray Dasyatis sabina, J. Neurosci. 10 Ž1990. 194–204. A.D. McClellan, S. Grillner, Activation of ‘fictive swimming’ by electrical microstimulation of brainstem locomotor regions in an in vitro preparation of the lamprey central nervous system, Brain Res. 300 Ž1984. 357–361. R. Nieuwenhuys, E. Pouwels, The brain stem of actinopterygian fishes, in: R.G. Northcutt, R.E. Davis ŽEds.., Fish Neurobiology, vol. 1, University of Michigan Press, Ann Arbor, 1983, pp. 25–88. Y. Oka, M. Satou, K. Ueda, Descending pathways to the spinal cord in the hime´ salmon Žlandlocked red salmon Oncorhynchus nerca., J. Comp. Neurol. 254 Ž1986. 91–103. P.D. Prasada Rao, Descending projection neurons to the spinal cord of the goldfish Carassius auratus, J. Comp. Neurol. 265 Ž1987. 96–108. P.D. Prasada Rao, A.G. Jadhao, S.C. Sharma, Topographic organization of descending projection neurons to the spinal cord of the goldfish Carassius auratus, Brain Res. 620 Ž1993. 211–220. A. Roberts, S.R. Soffe, N. Dale, Spinal interneurones and swimming in frog embryos, in: S. Grillner, P.S.G. Stein, D.G. Stuart, H. Forssberg, R.M. Herman ŽEds.., Neurobiology of Vertebrates Locomotion, Macmillan, London, 1986, pp. 279–306. B.L. Roberts, The organization of the nervous system of fishes in relation to locomotion, in: M.H. Day ŽEd.., Vertebrate Locomotion, Academic Press, London, 1981, pp. 115–136. M.L. Shik, F.V. Severin, G.N. Orlovsky, Control of walking by means of electrical stimulation of the midbrain, Biophysics 11 Ž1966. 756–765. K. Uematsu, T. Ikeda, The midbrain locomotor region and induced swimming in the carp Cyprinus carpio, Nippon Suisan Gakkaishi 59 Ž1993. 783–788.