Cerebellar influence on the time structure of movement in the electric fish Eigenmannia

Abstract

To determine how cerebellar activity influences motor co-ordination, a simple motor act of an electric fish, Eigenmannia spp, was studied. Whenever the fish encounters an object of conductivity different from water, it bends the caudal two-thirds of its body towards the object. This behavior, called “probing”, was analysed in the normal fish and in fish with a small cerebellar lesion. Probing could be described as a wave travelling along the body very similar to what one observes when shaking a rope of moderate elasticity. In normal fish, the wave starts at the tip of the tail and proceeds towards the head, dying out after about two-thirds of the body length. A lesion in the corpus cerebelli causes this wave to start at the head and move towards the tail, though it is otherwise the same as in normal fish, i.e. it has the same frequency components as revealed by Fourier analysis. In addition to the change in probing, the lesioned fish also undulates its body while swimming whereas it is normally kept straight. This undulating wave turns out to be the same as that during probing where it is executed whilst the fish keeps itself on the spot with its locomotory apparatus, the anal fin, which is driven by muscles independent of the trunk muscles used for probing. Since there is evidence [McClellan and Grillner (1982) Soc. Neurosci. Abstr. 8, 250–258] that the locomotory pattern is governed centrally, the above results suggest the following scheme: whenever the fish intends movement of the trunk, the pattern generating network is started and produces a wave which travels from head to tail with a fixed stereotyped phase relation between the segments. The cerebellar activity co-ordinates this phase relation at the central level according to the intended movement and the electrosensory needs to keep the body (at least the rostral third) straight enough so that the fish's own movements do not disturb its dipole field. A model calculation was used to estimate the amount of phase shift between the segments needed to change the pattern of movement of a lesioned animal to one of a normal animal. It turns out to be in the range of 0.1 ms, which suggests that this type of motor activity has to be designed in advance.