The synaptic convergence patterns of semicircular canal and macWar afferent nerve inputs onto second-order vestibular neurons reorganize in adult frogs after a change in the activity of vestibular nerve afferent fibers. Axotomized afferent nerve fibers become silent after a vestibular nerve lesion, and second-order vestibular target neurons become disfacilitated. These changes initiate an activity-related process that was studied in detail in vitro two months after a section of the ramus anterior (RA) of N. VIII. The postlesional reaction results in an expansion of signals, preferentially from intact, remaining afferent nerve fibers, but also from excitatory commissural and spinal ascending fibers. This process of expansion takes weeks, is graded in its extent, and reversible in case of a nerve regeneration, but is not competitive, i.e., the synaptic contacts from axotomized afferent nerve fibers are maintained without a change in their efficacy.Postlesional synaptic reorganization in the brainstem is restricted to the operated side, underlies the improved responsiveness of disfacilitated second-order vestibular neurons, but also their altered spatial response tuning. The functional consequences of this reorganization were studied in vivo two months after RA nerve section by recording abducens nerve responses during linear or angular accelerations. The vector orientations of best responses of the abducens nerve of chronic RA frogs evoked by linear or angular acceleration differed from the vector orientations of controls. In chronic RA frogs, linear acceleration evoked contralesional abducens nerve responses that originated from the utricle on the intact side and from the lagena, a vertical macular organ in frogs. Such an inadequate lagenar response component was absent in controls and in the ipsilesional abducens nerve of chronic RA frogs. Similar differences were detected in the direction of abducens nerve responses of chronic RA frogs during angular acceleration. Thus, compensatory vestibulo-ocular reflexes of chronic RA frogs became more symmetric in gain, but less precise in direction.

Postlesional reorganization of vestibular afferent and commissural inputs onto second-order vestibular neurons was studied in the isolated brain after unilateral section of the N.VIII, of the ramus anterior (RA) of N.VIII, of the utricular (UT) or of the anterior vertical and horizontal canal nerves in combination. RA nerve section eliminated the inputs from utricular, anterior vertical and horizontal canal organs. In the first set of experiments we recorded field potentials on the operated side of the vestibular nuclei 2 months after RA nerve section. These responses were evoked by electrical stimulation of the RA nerve or of the posterior vertical canal nerve on the operated or on the intact side. The amplitudes of afferent field potentials evoked by stimulation of the spared posterior vertical canal nerve were increased. The amplitudes of afferent field potentials evoked by stimulation of the axotomized RA nerve remained unaltered. After N.VIII section the commissural, but not the afferent, field potentials increased significantly on the operated side following stimulation of N.VIII on the intact and on the operated side, respectively. After UT nerve section no change in commissural but an increase in the amplitude of afferent field potentials from each of the three intact canal nerves was observed on the operated side. In the context of earlier results these findings imply that second-order vestibular neurons, disfacilitated due to afferent nerve section, became receptive to additional, excitatory synaptic inputs, preferentially from intact vestibular nerve afferent fibers. The reduced excitation via afferent nerve inputs was thereby replaced by other afferent nerve inputs from spatially inadequate vestibular end-organs. The synaptic terminals of inactivated afferent nerve fibers were maintained and not repressed. The process of central reorganization after vestibular nerve lesion was activity related, the expansion of signals restricted to inputs from intact fibers, its extent graded and its onset delayed with respect to the onset of corresponding spinal changes and to the onset of postural recovery after the same type of nerve lesion. After the section of RA nerve or of an individual nerve branch the labyrinthine end-organs remained intact and were not removed as after unilateral labyrinthectomy (UL). Peripheral reinnervation of the end-organs was thus excluded after UL, but expected after one of the former types of lesion. Functional reinnervation of the utricular macula was mirrored behaviorally by the reappearance of severe postural deficits following a second RA nerve section. These lesion-induced postural deficits began to reappear if the repeated RA nerve section was delayed with respect to the first by about 3 months. We therefore studied postlesional reorganization in the brainstem 3 months after the first RA nerve section. Reinnervation of the utricular macula was accompanied by a rapid decline of the increased amplitudes of afferent and commissural vestibular field potentials towards control values, suggesting the reversibility of the lesion-induced central reorganization.

Synaptic vestibular inputs of antidromically identified motoneurons and internuclear neurons in the abducens nucleus were studied electrophysiologically and pharmacologically in the isolated brain of grass frogs (Rana temporaria). The prevailing response pattern of abducens motoneurons (AbMOT) following stimulation of the VIIIth nerve was crossed disynaptic excitation and uncrossed disynaptic inhibition. A few AbMOT (five of 46), however, exhibited uncrossed excitation instead of inhibition. Abducens internuclear neurons (AbINT), identified by antidromic activation following stimulation of the contralateral medial longitudinal fascicle, exhibited disynaptic response patterns to stimulation of the VIIIth nerve that were very similar in latency and rise time to those of AbMOT except for the absence of uncrossed disynaptic inhibition. Bath application of strychnine (50 microM), a glycine antagonist, blocked the uncrossed inhibitory vestibular input to AbMOT and AbINT completely and reversibly, whereas picrotoxin (100 microM), a GABA (gamma-aminobutyric acid) antagonist, had no detectable effect on these disynaptic potentials. These results suggest glycine as the transmitter of inhibitory vestibular projections onto AbMOT and AbINT. The pharmacology of the excitatory vestibular input of these neurons was studied by electrical stimulation of the vestibular nuclear complex. Crossed monosynaptic excitatory inputs in AbMOT and AbINT were blocked completely by CNQX (6-cyano-7-nitroquinoxaline-2,3-dione) (10 microM), an antagonist of AMPA (alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) receptors, indicating glutamatergic excitation. Comparison of these results with those in the cat suggests the presence of a basic horizontal vestibulo-ocular reflex that is very similarly organized, and corroborates the hypothesis that major behavioural differences in the performance of compensatory eye movements between species result from the properties of supplementary networks and not from differences in a common 'three-neuron' vestibulo-ocular arc.