Cellular basis of vestibular compensation: analysis and modelling of the role of the commissural inhibitory system.

In this study we used a cellular network model of the brainstem vestibulo-ocular reflex (VOR) pathways to investigate the role of the vestibular commissural system in "vestibular compensation", the behavioural recovery that takes place after unilateral labyrinthectomy (UL). The network was initialized on the basis of mathematical analysis and trial simulations to generate a VOR response with a physiologically realistic gain and time constant. The effects of a selective decrease in the strength of commissural inhibitory input to the ipsi-lesional medial vestibular nucleus (MVN) neurones, without changes in other parts of the network, were investigated. Thus we simulated the marked down-regulation of GABA receptor efficacy that our recent experimental results have demonstrated in these cells after UL. The main outcome of this study is the delineation, for the first time, of a specific region of parameter space within which an adaptive change in commissural inhibitory gain is appropriate and sufficient to bring about a re-balancing of bilateral vestibular nucleus activity after UL. For this to be achieved, the relative contribution of the intrinsic, pacemaker-like membrane properties of the ipsi-lesional MVN cells must be equal to or greater than the synaptic input from the primary vestibular afferents in determining the in vivo resting discharge rate of these cells. Recent experimental evidence supports the view that the intrinsic properties of the MVN cells do contribute substantially to their resting discharge in vivo. Previous modelling studies that have excluded a role for the commissural system in vestibular compensation have arrived at this conclusion, because their models operated outside this region of parameter space. A second finding of this study is that, in a network that compensates through a selective change in commissural gain, the time constant of the VOR response is significantly reduced, mimicking the loss of velocity storage after UL in vivo. By contrast, the time constant is unchanged in a network that compensates through changes involving other nonvestibular inputs. These findings indicate that adaptive changes in commissural gain, through the dynamic regulation of GABA receptor efficacy in the vestibular nucleus neurones, may play an important role in vestibular plasticity.