Insulated, bipolar stainless steel electrodes were chronically implanted in various regions of the cat brain and the long-term structural changes in the tissue surrounding the electrodes were studied by light and electron microscopy. A sheath surrounded and separated the electrode from normal grey or white matter. A layer of foreign body giant cells of variable thickness was formed adjacent to the electrode. This layer was attenuated in some places so that it was unrecognizable by light microscopy. The bulk of the sheath structure consisted of collagen fibrils, leptomeningeal cells and hypertrophied astrocytes. Areas consisting of modified leptomeningeal cells with long thin processes we designated as spongy areas. These have not been previously reported using the electron microscope. Glycogen bodies were seen in leptomeningeal cells. Astrocytes became greatly enlarged and were more numerous in and around the sheath. Oligodendrocytes contained lamellar bodies, and direct continuity was shown between a lamellar body and an adjacent myelin sheath. Myelin was seen in abnormal sites (around oligodendrocytes and neurons) and in unusual configurations. Neuronal changes near the sheath included whorls and stacks of modified endoplasmic reticulum and the presence of cytoplasmic nucleolus-like bodies. Reactive, regenerative and degenerative axons were observed. Blood vessels were more numerous in the sheath and surrounding tissue than normal. Perivascular spaces were prominent even around capillaries and often plasma cells and monocytes were in these spaces. As compared to normal tissue the extracellular space is noticeably increased. Electrodes passing through ventricles were surrounded with a sheath covered with ependymal cells. This sheath was comparable in structure to the sheath present around the electrode in other locations.

An account is given of the responses of 557 medial reticular neurons with axons projecting down the spinal cord. All 30 experiments were on decerebrated unanesthetized cats paralyzed by Flaxedil. Recording from single neurons was by extracellular glass microelectrodes. Identification was first by location (confirmed by subsequent histology) in the medial reticular nucleus of medulla or pons, and second by antidromic activation from cord stimulation at C2 and L2 segmental levels. Axonal conduction velocities were calculated from the latency differential between L2 and C2 antidromic responses, and were usually in the range of 90-140 m/s; but about 25% were slower, ranging down to 30 m/s. Stimulation by electrodes in the ipsilateral and contralateral fastigial nuclei differentiated reticulospinal neurons into two classes according to whether they did or did not receive monosynaptic inputs, the respective populations of fully investigated neurons being 270 and 174. The fastigioreticular neurons were distinguished by a higher background frequency with mean values of 28 as against 15/s. There were also significant diffences in both the excitatory and inhibitory responses to afferent volleys from forelimb and hindlimb nerves. Comparison of the respective latency histograms showed that the responses of neurons with a fastigial input had an excess of latencies in the ranges that can be correlated with the latency histograms observed for fastigial responses. Thus, there is evidence for the effectiveness of the fastigial input and so for the pathway with monosynaptic linkage: Purkinje cells of cerebellar vermis yields fastigial neurons yields medial reticular neurons projecting down the spinal cord. Adequate stimulation of cutaneous receptors by pad taps and air-jet stimulation of hairy skin in a disppointingly small action when compared with fastigical responses. Explanations of this deficiency are suggested. Another discrpancy from the fastigial responses is that the medial reticular neurons have much wider receptive fields with little discrimination between ipsilateral and contralateral and between forelimb and hindlimb. Stimulation of the ipsilateral tegmental tract was tested on 183 reticulospinal neurons, 112 being with fastigial inputs. In about half there was a powerful monosynaptic excitation, which would identify such neurons as being on the pathway from mesencephalic and diencephalic centers to the spinal cord. There is a general discussion of transmission across successive synaptic relays, where specificity is sacrificed to integration.

The fine structure of the olfactory bulb was investigated in the cat. Five types of bulbar neurons were identified on the basis of size, location, cytoplasmic organelles and synaptic complexes with the neuron perikaryon. The smallest diameter neurons were predominantly located in the glomerular and granule cell layers. Ultrastructural characteristics for these two neuron populations, viz., periglomerular and internal granule cells, were identical. Tufted cells, smaller in diameter than mitral cells, were located in the external plexiform and glomerular layers and possessed all the morphological characteristics of the mitral cell. Mitral cells were the largest bulbar neurons and were located in a single lamina. Stellate (short-axon) neurons were found in the glomerular, external plexiform and granule cell layers.Reciprocal synapses were observed in all laminae of the bulb except the primary olfactory nerve and deeper granule cell layers. Such synapses were found in the glomeruli, on tufted and mitral cells, in the neuropil of the external plexiform layer, and on the axon hillock and initial segments of the mitral cells. The morphology in the cat olfactory bulb corroborated bulbar ultrastructure described for other mammals.

Several distinct classes of neurons have been identified in the medial reticular nucleus of the medulla and pons and in proximity thereto. Neurons projecting down the spinal cord comprised the principal class with two subclasses according as the neurons did or did not receive monosynaptic inputs from the fastigial nuclei of the cerebellum. Two other classes were recognized accordings as they projected to the cerebellum or rostrally to the mesencephalon. Topographic planar maps giving the location of these neurons have been constructed by exploring the nucleus with series of microelectrode tracks in parasagittal or in transverse planes. The different classes of neurons were not arranged in large discrete nuclei. In part they appeared to be randomly distributed, but many colonies of one or another class of neurons could be recognized with 3-11 neurons in zones with dimensions of a millimeter or so. Because of the limitations of sampling by microelectrode tracks at spacings of 0.5 mm, single colonies might have an actual population of 100 or more. Many of the class of neurons projecting to the cerebellum were in the region of the perihypoglossal nucleus. However, almost as many were located deep in the medial reticular nucleus. None was found at the pontine level. Reticulospinal neurons with fast axonal conduction velocities tended to be located dorsally to those with slow velocities. Correlation with the findings of Ito et al. leads to the conjecture that the neurons with fast axons are excitatory, while those with slow axons are primary inhibitory neurons. There is a brief reference to the problems raised by the admixture of the various neuronal classes, there being discrete colonies immersed in a scattered arrangement of all classes.

No abstract