Potentials in the tectum of large (12--20 cm) goldfish, evoked by stimulation of the optic nerve, were recorded extracellularly with double-barrelled electrodes (d.c., saline and a.c., Woods metal--Pt). Four fibre groups (E, M1, M2, M3) were recorded at latencies of approximately 2, 3, 5 and 8 ms after stimulation (conduction velocities of approximately 7, 5, 3 and 2 m/s). The same four groups were recorded from the optic nerve when the tectum was stimulated. The fastest fibre groups (E) did not give rise to a postsynaptic wave. Fibre groups M1, M2 and M3 gave rise to postsynaptic potentials which, following computation of their second spatial derivatives with depth, were found to have current sinks at depths of approximately 100-150 micrometers, 150--200 micrometers and 250--350 micrometers respectively. Thus the fastest conducting retinotectal fibres make their synapses most superficially, the opposite of the arrangement in the frog tectum. These postsynaptic waves fatigued at repetitive stimulus rates of 20--50 per second, and in twin pulses at interstimulus intervals of 10--15 ms; and they were reversibly blocked by topical application of pentobarbitol. The fibre potentials, however, were virtually undecremented under these conditions. To compare these electrophysiological findings with the anatomy, the cobalt procedure was used to visualize the profiles of the optic fibres in the various tectal laminae. A thick dense projection filled the superficial grey and white (s.g.w.) layer, and there was a thin satellite band just superficial to it. In addition, there were two deeper bands of sparse innervation, in the middle of the central grey zone (c.g.) and in the deep white (d.w.) layer. These bands were associated with the field potential sinks through lesions made with recording electrodes. The two deep bands correspond to the M3 fibre group. The dense s.g.w. innervation contains both the M1 and M2 fibre groups, the M1 just superficial to the M2. The fastest fibre group, E, which had no postsynaptic wave associated with it, persisted at least six weeks after retinal removal, and probably represents efferent cells with fibres projecting back through the optic nerve to the retina. Filled cell profiles could not be positively identified with the cobalt technique, but could be seen with the HRP technique, when the optic afferents were first allowed to degenerate. The filled cells were the pyramidals of the s.g.w. layer.
Read full abstract