David G. Amaral

Neurons, numbers and the hippocampal network

Anatomists involved with studies of the hippocampal formation are being prodded by computational modelers and physiologists who demand detailed and quantitative information concerning hippocampal neurons and circuits. The beautiful camera lucida drawings of old, and the elegant descriptions of dendritic form that accompanied them are giving way to computer-reconstructed and three-dimensionally analyzed cells with rigorous determination of dendritic lengths and volumes, branching pattern and spine distribution. We will review certain quantitative aspects of hippocampal organization in the rat based on a survey of available literature and on our own intracellular labeling studies of granule cells of the dentate gyrus and pyramidal cells of the hippocampus. Some of the potential implications of these data for hippocampal information processing will be discussed.

INTRINSIC CONNECTIONS OF THE RAT AMYGDALOID COMPLEX : PROJECTIONS ORIGINATING IN THE ACCESSORY BASAL NUCLEUS

The amygdaloid complex plays an important role in the detection of emotional stimuli, the generation of emotional responses, the formation of emotional memories, and perhaps other complex associational processes. These functions depend upon the flow of information through intricate and poorly understood circuitries within the amygdala. As part of an ongoing project aimed at further elucidating these circuits, we examined the intra‐amygdaloid connections of the acessory basal nucleus in the rat. In addition, we examined connections of the anterior cortical nucleus and amygdalahippocampal area to determine whether portions of these nuclei should be included in the accessory basal nucleus (as some earlier studies suggest). Phaseolus vulgaris leucoagglutinin was injected into different rostrocaudal levels of the accessory basal nucleus (n = 12) or into the anterior cortical nucleus (n = 3) or amygdalahippocampal area (n = 2). The major intra‐amygdaloid projections from the accessory basal nucleus were directed to the medial and capsular divisions of the central nucleus, the medial division of the amygdalohippocampal area, the medial division of the lateral nucleus, the central division of the medial nucleus, and the posterior cortical nucleus. The projections originating in the anterior cortical nucleus and the lateral division of the amygdalohippocampal area differed from those originating in the accessory basal nucleus, which suggests that these areas are not part of the deep amygdaloid nuclei have different intra‐amygdaloid connections. The pattern of these various connections suggests that information entering the amygdala from different sources can be integrated only in certain amygdaloid regions.

The amygdalostriatal projections in the monkey. An anterograde tracing study

Amygdalostriatal projections have been studied in the monkey with the autoradiographic method for demonstrating axonal transport of tritiated amino acids. Amygdaloid fibers were found to project in a roughly topographical manner to widespread areas of the striatum and ventral striatum, including the nucleus accumbens, the striatal-like portions of the olfactory tubercle, ventral portions of the putamen and ventral and caudal parts of the caudate nucleus. The parvicellular part of the basal nucleus and the amygdalohippocampal area appear to be the major sources of fibers to the nucleus accumbens, whereas projections to the tail of the caudate nucleus seem to arise mainly from the magnocellular part of the basal nucleus. In many of these areas, the amygdalostriatal fibers are concentrated in patches.

The development ultrastructure and synaptic connections of the mossy cells of the dentate gyrus

SummaryOne of the most distinctive and common cell types in Golgi preparations of the hilus of the rat dentate gyrus is the mossy cell. We have used a variety of techniques including the Golgi method, the combined Golgi and electron microscopic (EM) method and the retrograde transport of horseradish peroxidase (HRP) to study the development, ultrastructure and synaptic connections of this cell type. The mossy cells identified in our light microscopic preparations are characterized by: (1) triangular or multipolar shaped somata; (2) three to four primary dendrites that arise from the soma and bifurcate once or more to produce an extensive dendritic arborization restricted, for the most part, to the hilus; (3) numerous thorny excrescences on their somata and proximal dendrites with typical spines on distal dendrites; and (4) axons that bifurcate and are directed toward the fimbria and the molecular layer of the dentate gyrus.The mossy cells have an immature appearance at birth and on subsequent days their maturation appears to lag somewhat behind that of the hippocampal pyramidal cells. On postnatal day 1, many of the dendrites bear growth cones primarily at their termini and have long, thin filipodia emanating from various points along their lengths. Many of the dendrites enter the molecular layer of the dentate gyrus, though this is rarely seen in the mature brain. Typical pedunculate spines are first commonly seen on the distal dendrites around postnatal day 7 while thorny excrescences are first commonly seen between postnatal days 11 and 14. By postnatal day 21, the dendrites have attained a mature appearance although the density of both typical spines and thorny excrescences is less than that found in adults.Two different retrograde transport methods were used to confirm that mossy cells give rise to the commissural projection to the contralateral dentate gyrus. The first method combined HRP histochemistry with a silver intensification procedure and the second method combined HRP histochemistry with Golgi staining. While the majority of commissurally projecting hilar neurons had the appearance of mossy cells, there were others that were smaller and either ovoid or fusiform.

Individual differences in the cognitive and neurobiological consequences of normal aging

Defining the neural basis of age-related cognitive dysfunction is a major goal of current research on aging. Compelling evidence from laboratory animals and humans indicates that aging does not inevitably lead to cognitive decline. Conducting neurobiological investigations in subjects that have previously undergone behavioral characterization has therefore emerged as a promising strategy for identifying those alterations in brain structure and function that are specifically associated with age-related cognitive impairment.

An anatomical study of the development of the septo-hippocampal projection in the rat

Abstract The development of the septal projection to the hippocampal formation of the rat was studied in preparations stained for the demonstration of acetylcholinesterase and in experiments utilizing both orthograde and retrograde tracer techniques. Some of the cells in the medial septal nucleus and in the nucleus of the diagonal band are acetylcholinesterase-positive at least 2 days before birth (fetal day 20), and at least some of their fibers, which travel in the supracallosal stria, the fimbria, and the dorsal fornix, are also stained at this time. Within the hippocampal formation, numerous acetylcholinesterase-positive cells are observed in the hilar region of the dentate gyrus, in strata oriens and radiatum of regio inferior of the hippocampus, in the subiculum and in the entorhinal cortex on fetal day 20, but there are few stained fibers in these fields. By postnatal day 3, however, acetylcholinesterase-positive fibers are consistently observed in the rostral dentate gyrus, hippocampus and subiculum. By postnatal day 5, stained fibers are found throughout the rostro-caudal extent of the hippocampal formation and by postnatal day 14 an adult pattern of acetylcholinesterase staining is established. Especially in the adult preparations it is apparent that acetylcholinesterase-positive fibers enter the hippocampal formation by 3 distinct routes: (1) the fimbria, which appears to supply fibers to the entire rostro-caudal extent of the hippocampus and dentate gyrus; (2) the dorsal fornix, which contributes fibers to rostral levels of the subiculum, the regio superior of the hippocampus and the dentate gyrus; and (3) the supracallosal stria, in which fibers travel to mid-rostro-caudal levels of the subiculum and dentate gyrus. Cells of the medial septal nucleus and the nucleus of the diagonal band were retrogradely labeled as early as two days before birth following injections of wheat germ agglutinin-conjugated horseradish peroxidase or the fluorescent dye Fast blue, into the hippocampal formation. At this time the overall distribution of labeled cells within the septal complex appears similar to that seen in the adult. Injections of tritiated amino acids into the septal complex lead to anterograde labeling of the hippocampal formation as early as fetal day 21, the initial stage at which this procedure was attempted. While terminal labeling is diffusely distributed throughout the hippocampal formation during the first 5 postnatal days, by the end of the second postnatal week an increased density of labeling appears both above and below the granule cell layer of the dentate gyrus, within strata radiatum and lucidum of regio inferior of the hippocampus and in the inner and outer plexiform layers of the subiculum. In addition, by postnatal day 14 it is clear that the temporal half of the hippocampal formation receives a substantially greater innervation than the septal half. These studies lead to the conclusions that: (1) fibers originating in the septal complex are present within the hippocampal formation by at least fetal day 20; (2) septal terminals are diffusely distributed initially and segregate to their mature position during the second postnatal week; and (3) the time-course of development of acetylcholinesterase fiber staining within the hippocampal formation parallels the pattern of innervation by the septal nuclei as demonstrated with amino acid autoradiography, though it lags behind the latter by approximately 5 days.

Emerging principles of intrinsic hippocampal organization

The hippocampal formation has a unique and highly distributed network of intrinsic connections. What are the principles of organization that govern information flow through this system? The notion that information processing in the hippocampal formation is segregated in autonomous chips or lamellae appears to be inconsistent with the extremely divergent nature of many of the intrinsic connections. Recent neuroanatomical data suggest, however, that information may be segregated in other ways as it negotiates the links from one hippocampal region to the next.

Retrograde transport of D-[3H]-aspartate injected into the monkey amygdaloid complex

SummaryThe possibility that certain of the afferents of the primate amygdaloid complex use an excitatory amino acid transmitter was evaluated by injecting D-[3H]-aspartate into the amygdala of twoMacaca fascicularis monkeys. The distribution of D-[3H]-aspartate labeled neurons was compared with those labeled with the nonselective retrograde tracer WGA-HRP injected at the same location as the isotope. Retrogradely labeled cells of both types were observed in a variety of cortical and subcortical structures and in discrete regions within the amygdala. D-[3H]-aspartate labeled neurons were observed in layers III and V of the frontal, cingulate, insular and temporal cortices. In the hippocampal formation, heavily labeled cells were observed in the CA1 region and in the deep layers of the entorhinal cortex. Of the subcortical afferents, the claustrum and the midbrain peripeduncular nucleus contained the greatest number of D-[3H]-aspartate labeled cells. Subcortical afferents that are not thought to use excitatory amino acids, such as the cholinergic neurons of the basal nucleus of Meynert, did not retrogradely transport the isotope. Within the amygdala, the most conspicuous labeling was in the paralaminar nucleus which forms the rostral and ventral limits of the amygdala. When the D-[3H]-aspartate injection involved the basal nucleus, many labeled cells were also observed in the lateral nucleus. Retrograde transport of D-[3H]-aspartate injected into the amygdala, therefore, appears to demonstrate a subpopulation of inputs that may use an excitatory amino acid transmitter.

Evidence for a direct projection from the superior temporal gyrus to the entorhinal cortex in the monkey.

During the course of a larger study of the afferent and efferent connections of the entorhinal cortex in the macaque monkey we have found evidence for a hitherto undescribed projection to the entorhinal cortex from the superior temporal gyrus. The evidence is derived principally from experiments in which small volumes of wheat germ agglutinin-conjugated horseradish peroxidase (WGA-HRP) were injected into different parts of the entorhinal cortex, but has been confirmed by 3H-amino acid autoradiography. After WGA-HRP injections into the entorhinal cortex, retrogradely labeled neurons have been seen mainly in layer III, but also to some extent in layer VI, throughout much of the superior temporal gyrus. The projection appears to be topographically organized in the sense that the ventral insular cortex and the adjoining temporal operculum have been found to project to the periamygdaloid cortex and the lateral division of the entorhinal cortex; the convexity of the superior temporal gyrus and the cortex along the dorsal bank of the superior temporal gyrus project further caudally to the medial division of the entorhinal cortex; and the cortex surrounding the fundus of the superior temporal sulcus projects to the perirhinal cortex. Following an injection of 3H-amino acids into the convexity of the superior temporal gyrus, terminal labeling has been seen over layers I and II of the entorhinal cortex and over layer I in the most lateral portion of the presubiculum. While the distribution of retrogradely labeled cells in our WGA-HRP experiments encompasses several cytoarchitectonically distinguishable areas in the superior temporal gyrus, the most heavily labeled field appears to coincide with what Gross and his colleagues have termed the superior temporal polysensory area on the dorsal bank of the superior temporal sulcus.

Recognition memory deficits in a subpopulation of aged monkeys resemble the effects of medial temporal lobe damage

The present study examined individual differences in recognition memory function in a group of Old World monkeys (Macaca mulatta). Four young (9-11 years) and 10 aged (22-33 years) monkeys were tested in the same delayed-nonmatching-to-sample (DNMS) recognition memory procedure that has been widely used to study the effects of experimental hippocampal lesions in young subjects. Animals were first trained to a 90% correct learning criterion in the DNMS task using a 10-second delay between the sample and recognition phase of each trial. The memory demands of the task were then increased by gradually extending the retention interval from 15 seconds to 10 minutes. Three of the aged monkeys performed as accurately as young subjects at all delays. The remaining aged monkeys performed well at the shortest delays (15 and 30 seconds), but progressively greater impairments emerged across delays of 60 seconds, 2 minutes, and 10 minutes. These results suggest that recognition memory is only compromised in a subpopulation of aged monkeys. Moreover, aged monkeys that are impaired in the DNMS task exhibit the same delay-dependent pattern of deficits that is the hallmark of memory dysfunction resulting from medial temporal lobe damage.