B.L. McNaughton

Hippocampal synaptic enhancement and information storage within a distributed memory system

Abstract The hypothesis that the physical substrate of memory in the mammalian brain resides in alterations of synaptic efficacy has been proposed frequently in both neuroscience 1–5 and cognitive science 6–12 , and has been widely investigated in behavioural, physiological and theoretical studies. Although this hypothesis remains unproven, considerable evidence suggests that a particular form of synaptic strengthening, induced by electrical stimulation of certain CNS fibre systems, may represent the activation of mechanisms that normally subserve associative memory. This phenomenon is known as long-term potentiation (LTP) or long-term enhancement (LTE) ∗ . It has been most intensively investigated within the hippocampal formation, a brain structure that plays a crucial role in certain forms of associative memory. Physiological investigation has revealed that LTE exhibits most of the properties implicit in Hebbs original suggestion that associative memory results from a synaptic strengthening that is contingent upon the conjunction of activity in pre- and post-synaptic elements. In this article, we outline a simple neuronal model capable of superimposing multiple memory traces within the same matrix of connections, and consider the correspondence between such models and the properties of LTE in the context of the hippocampal circuitry in which it occurs. Certain predictions are derived from this framework concerning the behavioural consequences of experimental manipulation of LTE, and we conclude by describing experimental evidence that confirms these predictions and suggests that LTE is, in fact, fundamentally involved in memory.

The contributions of position, direction, and velocity to single unit activity in the hippocampus of freely-moving rats

SummaryIsolated single units in rat dorsal hippocampus and fascia dentata were classified as ‘Theta’ or ‘Complex-Spike’ cells, and their firing characteristics were examined with respect to position, direction and velocity of movement during forced choice, food rewarded search behavior on a radial eight arm maze. Most spikes from CS cells ocurred when the animal was located within a particular place on the maze and moving in a particular direction. Theta cells had very low spatial selectivity. Both cell categories had discharge probabilities which increased somewhat as a function of running velocity but tended to asymptote well before half-maximal velocity. The place/direction specificity of CS cells was significantly higher in CA1 than in CA3 and CA3 CS cells exhibited a striking preference for the inward radial direction. The pronounced directional selectivity of CS cells, at least in the present environment, suggests that they fire in response to complex, but specific, stimulus features in the extramaze world rather than to absolute place in a non-egocentric space. An alternative possibility is that the geometrical constraints of the maze surface have a profound influence on the shapes of the response fields of CS cells.

The stereotrode: A new technique for simultaneous isolation of several single units in the central nervous system from multiple unit records

A new method is described for the recording and discrimination of extracellular action potentials in CNS regions with high cellular packing density or where there is intrinsic variation in action potential amplitude during burst discharge. The method is based on the principle that cells with different ratios of distances from two electrode tips will have different spike-amplitude ratios when recorded on two channels. The two channel amplitude ratio will remain constant regardless of intrinsic variation in the absolute amplitude of the signals. The method has been applied to the rat hippocampal formation, from which up to 5 units have been simultaneously isolated. The construction of the electrodes is simple, relatively fast, and reliable, and their low tip impedances result in excellent signal to noise characteristics.

Hippocampal granule cells are necessary for normal spatial learning but not for spatially-selective pyramidal cell discharge

SummaryThe effects of massive destruction of granule cells of the fascia dentata on the spatial and temporal firing characteristics of pyramidal cells in the CA1 and CA3 subfields of the hippocampus were examined in freely moving rats. Microinjections of the neurotoxin colchicine were made at a number of levels along the septo-temporal axis of the dentate gyri of both hemispheres, resulting in destruction of over 75% of the granule cells. By contrast there was relatively little damage to the pyramidal cell fields. As assessed by three different behavioral tests, the colchicine treatment resulted in severe spatial learning deficits. Single units were recorded from the CA1 and CA3 subfields using the stereotrode recording method while the animals performed a forced choice behavioral task on the radial 8-arm maze. Considering the extent of damage to the dentate gyrus, which has hitherto been considered to be the main source of afferent information to the CA fields, there was remarkably little effect on the spatial selectivity of “place cell” discharge on the maze, as compared to recordings from control animals. There was, however, a change in the temporal firing characteristics of these cells, which was manifested primarily as an increase in the likelihood of burst discharge. The main conclusion derived from these findings is that most of the spatial information exhibited by hippocampal pyramidal cells is likely to be transmitted from the cortex by routes other than the traditional “trisynaptic circuit”. These routes may include the direct projections from entorhinal layers II and III to CA3 and CA1, respectively.

Comparison of spatial and temporal characteristics of neuronal activity in sequential stages of hippocampal processing.

The activity of individual pyramidal cells in the CA1 and CA3 subfields of the rodent hippocampus exhibits a remarkable selectivity for specific locations and orientations of the rat within spatially-extended environments. These cells exhibit high rates of activity when the animal is present within restricted regions of space, referred to as place fields, and are extremely quiet when it is elsewhere. Although this phenomenon has been well studied in the CA fields of the hippocampus, relatively little is known about the spatial and temporal firing characteristics either of the entorhinal cortical inputs to the hippocampus, or of the subicular recipients of the output of hippocampal place cells. We report here on a comparison of spatial and temporal discharge characteristics among entorhinal cortex, CA3 and CA1, and the subiculum. CA3 complex spike cells were significantly more spatially specific than their CA1 counterparts. Neither entorhinal cortex nor subiculum exhibited the highly localized patterns of spatial firing observed in the CA fields. In addition, average discharge rates in these areas were substantially higher. However, particularly in subiculum, there was evidence for spatially consistent, but dispersed, firing in some cells, suggestive of the convergence of a number of CA1 place cells. The patterns observed are not consistent with the hypothesis that spatial selectivity is progressively refined at the various levels of hippocampal processing. Rather, hippocampal output appears to be expressed as a much more highly distributed spatial code than activity within the hippocampus proper. We suggest that the sparse coding used within the hippocampus itself represents a mechanism for increasing the storage capacity of a network whose function is to form associations rapidly.

Long-term enhancement of hippocampal synaptic transmission and the acquisition of spatial information

The hypothesis that memories are stored as a specific distribution of strengths in a population of modifiable synapses was examined by the bilateral induction of long-term enhancement in synapses of the main afferent fiber system to the hippocampal formation in rats. Brief, high- frequency activation of the perforant pathway in chronically prepared animals resulted in a persistent increase in the field EPSP and population spike, measured extracellularly in fascia dentata. This treatment resulted in a profound and persistent deficit in the acquisition of new spatial information in a task requiring spatial “reference” memory, and disruption of recently acquired spatial information. Well-established spatial memory was completely unaffected, however, as was the acquisition of spatial information into short-term “working” memory. These results support the hypothesis that, during the formation of “cognitive maps,” spatial information must be temporarily stored at modifiable synapses at the input stage to the hippocampal formation, but that this information is not needed once the representation of the environment is well established. Spatial working memory, in a familiar environment, appears not to depend on the distribution of synaptic strengths in this system at all.

A selective increase in phosphorylation of protein F1, a protein kinase C substrate, directly related to three day growth of long term synaptic enhancement

Increased in vitro phosphorylation of the 47 kdalton, 4.5 pI protein F1 was observed in dorsal hippocampal tissue from animals exhibiting long term enhancement (LTE) three days after high frequency stimulation of the perforant pathway, as compared to tissue from low frequency stimulated controls or from unoperated animals. The increase in protein F1 phosphorylation was related to LTE rather than simple activation of perforant path-dentate gyrus synapses. This is the first report of a change in brain protein phosphorylation accompanying synaptic enhancement lasting days. The extent of growth of LTE over the three days following stimulation was directly related (r = +0.66, P less than 0.05) to protein F1 phosphorylation. Among the phosphoproteins studied this relationship between LTE and phosphorylation was selective for protein F1. This suggests that protein F1 may regulate growth of synaptic plasticity for at least a three day period. The mechanism for the LTE-related increase in protein F1 phosphorylation has not been established. However, recent evidence from this laboratory indicates: that protein F1 is phosphorylated by the calcium/phospholipid-dependent protein kinase C; and that kinase C is activated 1 h after LTE. Therefore, the increase in protein F1 phosphorylation following LTE may result from long term activation of protein C kinase.

Cortical-hippocampal interactions and cognitive mapping: A hypothesis based on reintegration of the parietal and inferotemporal pathways for visual processing

Spatial navigation and the firing of hippocampal place cells can be driven as much by what an animal knows about its spatial world as by what it immediately experiences at a given location. If presented first with a set of spatially orienting cues, which are then removed during a test, navigation to a place of reward is accurate and place cells still fire in their correct locations despite the absence of the controlling cues (O’Keefe & Speakman, 1987). Similarly, as reported here, place fields are disrupted in a familiar room in darkness if the animal is not shown its starting location, but remain intact in darkness if the starting location is known to the animal. A minimal computational model is presented to account for these results. The hypothesis proposes that conditional associations between places and movements are established during learning about an environment. Given a known starting location, these compound movement/place representations can be used to recall a sequence of target locations on the basis of the corresponding movement sequences alone. Recordings of posterior parietal neurons in rats performing a radial maze task reveal that this cortical region contains cells that are selective for specific combinations of environmental spatial features and motion states. The proposed model suggests how these compound movement/place representations could be combined with hippocampal spatial representations to account for the “blind” navigation phenomena described above. The model may also help us to understand the reasons for segregation of mammalian visual information processing into parietal and inf erotemporal streams as described by Ungerleider and Mishkin (1982).

Reversible inactivation of the medial septum differentially affects two forms of learning in rats

The contribution of the medial septum to different aspects of spatial information processing was assessed by examining the effects of reversible septal inactivation on radial maze performance of rats. In addition, the selectivity with which the medial septum affects learning was studied by testing the effects of septal inactivation on the acquisition of non-spatial information. Rats were first trained according to a spatial working memory procedure that included a 30-min delay between the first 4 (forced) choices and subsequent test (free) choices. The forced choices comprised the sample phase of the experiment while the free choices comprised the test phase. Saline or tetracaine (a local anesthetic) was injected into the medial septal area either before the sample phase, after the sample phase (i.e. at the beginning of the delay period), or just before the test phase. In contrast to the saline injections, tetracaine injected just before the sample or test phases produced a significant increase in errors at test. Tetracaine injection at the beginning of the delay period did not affect test choice accuracy. EEG records showed that septal inactivation drastically, yet temporarily, reduced the hippocampal theta rhythm. Thus, when septal inactivation occurred either before the sample phase or at the beginning of the delay period, hippocampal theta recovered by the time of the test phase. Septal inactivation also produced a significant retardation of learning on a non-spatial reference memory task, although clear improvement over trials did occur. Moreover, the results of subsequent saline injections suggest that at least some of the performance deficit was due to variables other than learning per se.(ABSTRACT TRUNCATED AT 250 WORDS)

Enhancement of hippocampal field potentials in rats exposed to a novel, complex environment

The hippocampus plays a crucial role in place learning in rodents and also exhibits a long-term enhancement of synaptic strength and postsynaptic excitability following electrical stimulation of its principal afferents. In the present report we suggest that these two observations may be related, by demonstrating an increase in synaptic and postsynaptic field potential amplitudes resulting from exposure to a spatially complex environment.