| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Arizona Research Laboratories Division of Neural Systems, Memory and Aging, University of Arizona, Tucson, Arizona; 2California National Primate Research Center, Davis, California; 3Evelyn F. McKnight Brain Institute, University of Arizona, Tucson, Arizona; 4The Medical Investigation of Neurodevelopmental Disorders Institute, Department of Psychiatry and Behavioral Sciences, and Center for Neuroscience, University of California, Davis, California; and 5Departments of Psychology and Neurology, University of Arizona, Tucson, Arizona
Submitted 9 April 2007; accepted in final form 22 May 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
3 mo, while the electrodes were moved by small increments through the hippocampus and neighboring structures. After recording, the monkeys were necropsied, and the brains were sectioned and Nissl-stained, permitting identification of individual electrode tracks. The results establish that hippocampal pyramidal cells are "complex spike cells," firing at overall average rates of 0.3 Hz, with spike trains consisting of long periods of silence interspersed with bursts of activity. The results also establish that the monkey hippocampal EEG shows "sharp wave" events consisting of a high-frequency "ripple" oscillation (110 Hz) together with a large slow-wave EEG deflection lasting several hundred milliseconds. The evidence suggests that monkey sharp waves are probably generated mainly in the CA1 region and that sharp waves are associated with an inactive/drowsy-or-sleeping behavioral state, which is also associated with increased hippocampal pyramidal cell activity and increased hippocampal EEG amplitude. The results of this initial study of ensembles of primate hippocampal neurons are consistent with previous studies in rodents and consistent with the hypothesis that theories and models of hippocampal memory function developed on the basis of rat data may be applicable to a wide range of mammalian species. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
), spatial memory (O'Keefe and Nadel 1978), and to the learning of some behaviors that require forming arbitrary associations (Brasted et al. 2003; Wise and Murray 1999). Complementing this, a large body of neurophysiological data has permitted the creation of rather detailed theoretical models of the hippocampal role in memory (Burgess et al. 2001; Buzsáki 1989; Marr 1971; McNaughton and Morris 1987; Murre 1996; O'Reilly and Norman 2002; Treves et al. 1996). A crucial contribution to such theories has come from the finding that hippocampal population activity, during sleep after a period of task-related activity, shows reactivation of population patterns that occurred previously in the context of the task (Kudrimoti et al. 1999; Wilson and McNaughton 1994). This reactivation has been found (Kudrimoti et al. 1999) to be concentrated specifically during brief EEG perturbations, associated with bursts of hippocampal neuronal activity, called sharp waves (Buzsáki 1986; Chrobak and Buzsáki 1996; Ylinen et al. 1995). In rats, sharp waves are brief (50–100 ms), large deflections of the hippocampal EEG signal, associated with fast EEG oscillations (ripples; 130–200 Hz) that can be observed from electrodes situated near the CA1 or CA3 cell body layers (Buzsáki 1986; Chrobak and Buzsáki 1996; O'Keefe and Nadel 1978). They occur during states of behavioral quiessence, when the hippocampal EEG is in the large irregular activity (LIA) state (Vanderwolf 1969). Accumulating evidence suggests that sharp waves are neural events that may underlie a memory consolidation process (Battaglia et al. 2004; Hoffman and McNaughton 2002; Qin et al. 1997; Skaggs and McNaughton 1996), perhaps by providing "teaching" signals that then induce synaptic changes within the neocortex (Buzsáki 1989; Marr 1971; McClelland et al. 1995; Murre 1996).
Most of what is currently known about hippocampal physiology comes from rats and mice. Only a small fraction of published studies have used other species, such as cats, dogs, monkeys, or humans, and most recently, bats (Ulanovsky and Moss 2007). Many investigators who work with rodents make the assumption that the majority of findings in rodents will be applicable to other mammalian species, including humans. The evidence for the validity of this assumption is limited, however. Lesion studies have established that the hippocampal system plays a role in memory in several species, but showing that it is the same role in all species has been challenging (Eichenbaum 2000; Squire 1992; Squire et al. 2004). The most striking feature of rodent hippocampal EEG, the theta rhythm, has been observed in rabbits, cats, and dogs, but has been difficult to observe in primates and humans (Ekstrom et al. 2005; Green and Arduini 1954; Halgren et al. 1978; Kahana et al. 2001), except in a study of urethane-anesthetized squirrel monkeys where it was clearly present (Stewart and Fox 1991). Also, it has not been clearly established that hippocampal neurons, in particular pyramidal cells, have similar spike train properties across the full range of species (Fox and Ranck 1975; Wirth et al. 2003). Furthermore, it has not been shown that the primate hippocampus shows distinct behavior-related physiological states corresponding to the rhythmic slow activity (RSA), LIA, and small irregular activity (SIA) states observed in rats (Buzsáki 1989; Jarosiewicz et al. 2002; Skaggs and McNaughton 1998; Vanderwolf 1969). This study was designed to address these issues by using multitetrode recordings to examine whether hippocampal EEG and unit activity properties in rhesus macaques show these features that have been observed in rodents.
A major factor contributing to the difficulty of verifying similarities between rodent and primate hippocampal physiology has been technical: the methods used for recording hippocampal activity from behaving rats are quite different from those most commonly used with primates. For rats, the most productive approach has been to implant bundles of blunt-tipped electrodes chronically, advancing them into the cell layers over the course of days. This approach can yield stable recordings for hours or days and, particularly when tetrodes are used, permits high-quality isolation of units with very low firing rates (Thompson and Best 1990; Wilson and McNaughton 1993). In primates, however, chronic implantation of electrodes for hippocampal unit recording has been more challenging, for at least four reasons: 1) primates require electrodes and devices that advance them that are well protected; 2) primates are more susceptible to infection than rodents; 3) the hippocampal target area in primates lies at the bottom of the forebrain, >30 mm below the brain surface, compared with the CA1 layer at a depth of 2 mm in rats; and 4) the number of subjects available for a primate experiment is generally limited by resource constraints and ethical concerns, making it essential to acquire a large amount of data from each subject. Because of these factors, recording from the primate temporal lobe has almost always been done using acutely implanted electrodes, which can be inserted and removed hundreds of times. The requirement for a large amount of data from each subject, in particular, has militated against the use of chronically implanted electrodes. To overcome this problem, this study used 12-tetrode "hyperdrives" (in which multiple cells can be isolated on each probe), whose tips were advanced through the hippocampus over the course of several weeks, yielding several hundred isolated hippocampal units per subject.
The data from these experiments permit three significant conclusions. First, primate hippocampal pyramidal cells are complex spike cells, generally firing at rates well below 1 Hz when averaged across a session, but often firing in bursts with interspike intervals (ISIs) <20 ms. Second, the primate hippocampal EEG shows sharp wave/ripple events resembling those observed in rats (Buzsáki 1986), which occur during quiet/resting behavioral states. Third, it is confirmed that, as found in previous studies, there is little evidence for robust theta-frequency rhythmicity, at least under the behavioral conditions of this study.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
600 ms). For both training and recording, the behavioral tasks were managed by a computer running Cortex software (http://www.cortex.salk.edu). Monkey 1 was a 31-yr-old (geriatric) female; monkey 2 was a 13-yr-old (middle-aged adult) male; and monkey 3 was a 14-yr-old female. These monkeys were part of a long-term study of the effects of aging on the primate temporal lobe; however, the number of subjects in this report is insufficient to support conclusions about age-related differences between subjects. All three monkeys were born and housed continuously at the California National Primate Research Center, Davis, CA. All procedures followed an Animal Care and Use protocol approved by the IACUC at the University of California, Davis, CA.
After training, the monkeys were chronically implanted with a hyperdrive holding 12 individually movable tetrodes (McNaughton et al. 1983), plus a movable reference electrode and a fixed ground connection (Fig. 1A). The design of the hyperdrive was based on models previously used to record from rat brains (Gothard et al. 1996; Qin et al. 1997; Wilson and McNaughton 1993), but modified to permit the electrodes to reach structures located 30–50 mm below the top of the brain and to handle the mechanical stresses that could be produced by primates. The tetrodes were glued into 160-µm-diam silica tubing, beveled at the tip, which was 5 mm back from the tip of the tetrode. These silica-encased tetrodes were directed into the temporal lobe through sharpened 28-gauge stainless steel cannulae. At maximum extension, the tetrodes could reach 12 mm below the tips of the stainless steel cannulae. The placement of the hyperdrive was determined on the basis of MRI scans conducted before surgery. Locations varied slightly from monkey to monkey, but on average, the drives were aligned vertically in the standard stereotaxic frame of reference, centered 5–10 mm anterior to the interaural line and 10–15 mm lateral from the midline. The cannulae were lowered to a depth of 30 mm below the brain surface, which was intended to yield an initial electrode position slightly above the top of the hippocampus.
|
23 µm in diamater. The outer diameter of the twisted tetrode probe was 71 µm, and the impedances were reduced to 100–200 KOhm by gold plating. The tetrode probes were arranged in a hexagonal array with mean spacing of 1.2 mm at the top of the cannula tracks; however, the cannulae did not travel perfectly straight, and the spacing varied considerably at the temporal levels targeted in the study, as can be seen in the example CT scan section in Fig. 1B. Electrophysiological data were acquired and recorded using a computer running Neuralynx Cheetah software (Neuralynx, Tucson, AZ). The Cheetah system was linked to the Cortex behavior control system through a custom-built electronic device that permitted each action taken by the Cortex system to produce a time-stamped event flag in the Cheetah data file. Also, analog eye position signals from the ISCAN system were fed to the Cheetah system and recorded continuously.
Recording started after recovery from the implantation surgery. For each recording session, the monkey was chaired, headposted, and placed in the isolation box. A headstage, with cables leading to the recording system, was plugged into a connector board at the top of the hyperdrive, and the data acquisition system was started. Most sessions began with a 10- to 20-min period of inactivity, intended to create the type of resting/sleeping state that has been associated with sharp waves in rodents (Kudrimoti et al. 1999), one or more periods of task performance, lasting typically from 30 min to 2 h depending on the specific task requirements and the monkey's performance, and a second period of inactivity lasting 15–30 min. In the second inactive period, and to a lesser degree in the first inactive period, the monkeys often closed their eyes for several minutes at a time, although long periods of full sleep were rare. (Monkey 3, intolerant of more than a few seconds of inactivity at any time, was an exception.) The monkeys also occasionally stopped working for a while during the task-performance periods, sometimes becoming drowsy.
The reference for EEG and unit recordings was a tetrode channel located in the temporal lobe above the level of the hippocampus. EEG signals were amplified 1,000 times, filtered with a band-pass of 1–475 Hz, digitized at a rate of 1,000 samples/s, and recorded continuously. Signals from spike channels were amplified 2,000 times, filtered with a band-pass of 600-6,000 Hz, and sampled at a rate of 32 KHz. Whenever the voltage on any of the four channels of a tetrode exceeded threshold, a 1-ms sample of "spike" data from all four channels was recorded. The spike threshold was set by the investigator using the Neuralynx control software. Different thresholds could be set for each channel of a tetrode, in cases where some channels were noisier than others. The aim, in choosing thresholds, was to record as many neuron-generated spikes as possible while excluding the majority of events generated by noise-driven voltage fluctuations. Unit isolation was performed off-line by manual cluster-cutting using the program xclust (M. A. Wilson). Further analyses of spike and EEG were done using custom-written software.
Sharp waves were identified using a method applied previously to data from rats (Jarosiewicz et al. 2002; Kudrimoti et al. 1999), which searches for brief periods of high-frequency oscillation, in the 100+ Hz (ripple) range. As shown in Fig. 2, the algorithm involves high-pass filtering the EEG signals with a cut-off frequency of 100 Hz, rectifying the result, smoothing by low-pass filtering with a cut-off of 20 Hz, and searching for peaks. The high-pass/rectify/low-pass procedure yields a positive-valued curve that approximately follows the envelope of the high-frequency component of the signal. Peaks with amplitude exceeding an observer-specified threshold were designated as ripples; the ripple amplitude was defined as the height of the peak. The threshold was chosen to yield a good match between events that appeared salient to a human observer and events that were detected by the algorithm.
|
Monkeys 1 and 3 performed, during each recording session, a passive paired associate task (PPAT), which was a go/no-go task with predictors (Erickson and Desimone 1999). The monkey initiated each trial by holding a bar while watching a computer monitor. The beginning of the trial was signaled by the appearance of a fixation spot. Once the monkey achieved fixation, a "predictor" image appeared on the monitor and stayed visible for 500 ms. The monitor then went blank for 800–900 ms, after which a second choice image appeared. Depending on whether the choice image signaled a go or a no-go response, the monkey was required either to release the bar within 2 s or to continue holding the bar for 2 s to obtain a juice reward. Trials occurred at a maximum rate of approximately one every 5 s.
Monkey 2 performed a visual paired comparison (VPC) task (Zola et al. 2000). In this task, the only behavioral requirement, to receive juice reward, was for the monkey to maintain fixation while images were being presented. The beginning of a trial was signaled by the appearance of a fixation spot. Once the monkey achieved fixation, a sample image appeared on the monitor and stayed visible for 2 s. The monitor then went blank for 1–20 s, after which a pair of images were shown for 2 s, one matching the sample image and the other a different image. The left-right relationship between the sample and novel images changed randomly from trial to trial.
For monkeys 1 and 2, each session began and ended with a 10- to 20-min period of inactivity, during which the monkeys sometimes became drowsy, and occasionally slept, but at other times simply waited patiently. These time periods were used to observe neural activity in the absence of a specific behavioral task, because the rat hippocampus shows enhanced activity under such conditions. Monkey 3 would only rarely relax enough to become drowsy, and at most times was intolerant of any interruption or delay of a behavioral task. Consequently, for this monkey, it was not possible to begin and end sessions with planned periods of inactivity.
Histology
At necropsy, the subjects were transcardially perfused for 2 min with 1% paraformaldehyde in 0.1 M phosphate buffer at a flow rate of 250 ml/min, followed by an 8-min perfusion with 4% paraformaldehyde in 0.1 M phosphate buffer at a rate of 250 ml/min and a 50-min perfusion with 4% paraformaldehyde in 0.1 M phosphate buffer at a rate of 100 ml/min. Ice was packed around the head of the animals during perfusion. After removal from the skull, the brains were postfixed in 4% paraformaldehyde in 0.1 M phosphate buffer for 6 h and cut into three blocks. The blocks were cryoprotected for 24 h by placement in 10% glycerol in 0.1 M sodium phosphate buffer + 2% DMSO at 4°C and again for 72 h in 20% glycerol in 0.1 M sodium phosphate buffer at 4°C. Next, the brains were frozen in isopentane chilled with dry ice and ethanol bath 40 min for large blocks and 25 min for small blocks. To allow for evaporation of isopentane from the surface of the brain blocks, they were stored for 24 h in a –72°C freezer before sectioning.
The brains were coronally sectioned at 30 µm on a freezing sledge or sliding microtome. Sections were collected in four series for processing and storage. Two series were stored in formalin for later processing using cell and myelin stains. One of these series was Nissl-stained for the entire extent of the brain. For monkey 1, a portion of a second series, surrounding the sites at which tetrodes were implanted, was stained with Fluoro-Jade (Schmued et al. 1997), to reveal tissue damage produced by the lesions that were made on the day before necropsy.
For each subject, a series of Nissl-stained sections encompassing the recording sites were photographed using a light microscope at x100 and examined to permit visible tracks to be correlated with the arrangement of cannulae reconstructed from the postimplant CT. The aim of this procedure was to identify which tetrode was responsible for each visible track and thereby to know which brain structures were sampled by each tetrode.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
As described in METHODS, recordings were obtained from three rhesus macaques, an aged female (monkey 1), a middle-aged male (monkey 2), and a middle-aged female (monkey 3). Each was implanted chronically with a 12-tetrode hyperdrive (Fig. 1A), which remained in place for 3 mo, during which time recordings were made daily. The electrodes were usually advanced each day by small increments (100–400 µm in most cases), but sometimes an electrode was left in place for as long as a week in cases where multiple high-amplitude units could be recorded. After recordings were completed, the monkeys were necropsied, and histological procedures were carried out.
In Nissl-stained sections, the temporal lobe regions penetrated by the tetrodes appeared generally healthy. In several cases, hippocampal layers were pushed downward 1–2 mm by the electrodes. In one case where an electrode remained at a fixed location within the CA3 layer for >80 days (tetrode 6 from monkey 1), most pyramidal cells had disappeared within a radius of 1 mm around the electrode tip. Such cell loss was not observed near the tips of tetrodes that were moved regularly. The dura and the tissue around the cannulae near the tops of the tracks appeared healthy for the subjects included in this study. Two of the monkeys showed transient neurological signs (pupil dilation and a tendency to favor the contralateral hand) for 1–2 wk after hyperdrive implantation, probably indicating that the cannulae caused some damage or swelling to the tissue they passed through. There were no signs of infection of the dura or brain tissue caused by these implants.
For monkey 1, the necropsy was carried out while the electrodes were still implanted, and all tetrode tracks could be identified, with their ends clearly distinguishable, as confirmed by Flouro-Jade staining of lesion-damaged neurons (Fig. 1, C and D). The histology showed that 5 of the 12 tetrodes encountered the hippocampus: tetrodes 4 (Fig. 1E) and 8 (Fig. 1F) passed tangentially through the CA1 layer; tetrode 3 (Fig. 8A) passed through the CA3 layer; and tetrode 5 (Fig. 8B) passed sequentially through the CA3 region, CA4/hilar region, dentate gyrus, a large cell-free space, and then into the CA1 layer near the point where it merges with the prosubiculum. For each of these tetrodes, depths at which unit activity was encountered correspond to depths at which cell layers are observed in the Nissl-stained sections. Thus units recorded from these tetrodes could be assigned with confidence to specific subregions of the hippocampus, some to CA1 and others to CA3. Tetrode 5 probably encountered cells from the fascia dentata (FD) as well, but because the FD is continuous with the hilar region and CA3, these areas are grouped together in this report. Two of the tetrodes proved to be especially valuable: tetrode 8, because it was inside the CA1 cell body layer for 45 consecutive days, moving tangentially, and picked up sharp waves during the entire time period; and tetrode 5, because it followed a trajectory moving perpendicularly through the hippocampal fissure and then the CA1 layer.
|
|
|
|
|
|
For monkey 3, as for monkey 1, the necropsy was carried out while the electrodes were still implanted; however, in this case, the cannulae could not easily be followed down into the brain using a CT scan because they were splayed; also, the marker lesions were not helpful because the electrodes were mostly extended beyond the deepest cortical layers at the time the lesions were made. The histological material showed that seven tetrodes penetrated hippocampal layers.
Figure 3 shows, in schematic form, the hippocampal levels traversed by tetrode tracks in each monkey. The placement from monkey 1 was several millimeters more posterior than for the other two monkeys, but all tracks from these monkeys went through the anterior half of the curved structure of the hippocampus.
The account that follows focuses in several respects on the data from monkey 1, because for that monkey, the histology was sufficient to specify which hippocampal areas were recorded from by each tetrode in each session. For the other two monkeys, identification of particular recordings as hippocampal depends in part on the histology but also in part on unit and EEG properties. Although the data from these two monkeys cannot be used directly to establish regional unit firing characteristics, they provide corroboration for the results obtained from monkey 1.
Unit activity
Table 1 summarizes the available unit data. The maximum number of units isolated from a single hippocampal tetrode in any recording session was 20 (Fig. 4A) from monkey 2; the maximum yields for monkeys 1 and 3 were 16 and 14 units, respectively. There were 9 data sets yielding 16–20 cells from a single tetrode and 33 data sets yielding 10–15 cells. The majority of data sets, however, yielded between one and five cells. It should be noted that, in extracellular recordings, units can only be identified if they show some level of activity during the recording session. Because hippocampal units are often very quiet, it is possible that additional, unrecognized cells may have been present. This applies particularly to the data from monkey 3, who rarely entered the drowsy/inactive behavioral states associated with higher firing rates spread across a larger fraction of the population. When a cell was recorded on multiple days, it was counted separately for each day, in terms of the numbers given in Table 1 and the analyses shown in Figs. 4–6. (This was necessary because it is often impossible to know with certainty whether a specific unit is identical to one recorded on the previous day.)
|
)—these units fired at overall rates <1 Hz, but with many ISIs <20 ms. Figure 4B shows basic properties of two such units (cells 1 and 2): the mean spike waveform on each of the four tetrode channels; an autocorrelation plot; and an ISI histogram. The central peaks in the autocorrelations indicate a tendency to fire in bursts, and the ISI histograms (using a logarithmic axis) show a large fraction of intervals in the range 4–20 ms. These features are shared by similar plots made using data from rat hippocampal pyramidal cells.
Note that the term complex spike cell has led to some misunderstanding. In some parts of the brain, such as the cerebellum, cells referred to by this term have a tendency to fire stereotyped bursts, consisting of numerous spikes in sequence, with ISIs falling within a narrow range (Ito 1984). The CS cells found in the rat hippocampus do not behave in this way—they fire many bursts, but the number of spikes in a burst is quite variable, and the intervals between spikes are irregular (Harris et al. 2001). The CS cells recorded from the primate hippocampus in this study were similar to those in the rat in this respect.
In addition to CS cells, a much smaller number of higher-firing-rate cells were recorded from inside the hippocampus. However, very few of these cells had large enough action potential amplitudes for high quality unit isolation. (Consequently, they do not appear in the histograms in Fig. 4C, which count only units that appeared to be well isolated.) Their prevalence was difficult to estimate but was at least an order of magnitude smaller than the fraction of complex spike cells. (Note that it is usually easy to recognize the presence of a high-firing-rate cell even when the action potentials it generates are quite small. In constrast, cells that fire only a few action potentials usually cannot be distinguished from background activity unless the action potentials are well above background amplitudes.) The plots for cell 3 in Fig. 4B show the average waveform, autocorrelation, and ISI histogram for one of the better-isolated non-CS cells from the sample. Some of these non-CS cells had comparatively narrow spike widths, in the range 0.15–0.25 ms, but many had spike waveforms similar to those of complex spike cells. In the rat hippocampus, the majority of putative interneurons have narrower spike-widths than pyramidal cells (Fox and Ranck 1975).
Most of the hippocampal CS cells fired at very low overall rates. More than 40% of the cells showed firing rates <0.1 Hz when averaged over sessions lasting an hour or more (Fig. 4C). The population average firing rate was only 0.29 Hz. For the monkey in which CA1 and CA3/DG recordings could be distinguished (monkey 1), the overall rates differed little between the two areas. For comparison, recordings from a group of nearby neocortical locations (mainly areas TE and IP) produced overall population average firing rates nearly 10 times higher, with only 6% of the sample firing at time-averaged rates <0.1 Hz.
Most hippocampal CS cells showed no significant task-related modulation, staying continuously silent except for an occasional isolated single spike or burst. However, a fraction did show reliable activation, in a few cases consisting of suppression of background activity, but more commonly consisting of brief periods of enhanced firing, during which the firing rates (averaged across trials) could rise to 10 Hz or more. Quantifying the types or levels of unit responses is beyond the scope of this study, but a general estimate is that substantially <10% of hippocampal CS cells showed clear task-related activation in our experimental conditions.
In Fig. 5, hippocampal units are compared with units from nearby cortical areas (TL, TE, TF, IP, and entorhinal regions EC and ELC) in terms of three properties: average spike width, measured from peak to valley on the channel with largest spike amplitude; firing rate averaged across the session; and fraction of burst spikes, defined as spikes occurring <20 ms after the previous spike. In Fig. 5, A–C, scatterplots are shown for data from monkey 1, with units divided into CA1, CA3/DG, and neocortical categories. In Fig. 5, D–F, scatterplots are shown that combine data from all three monkeys, with units divided into hippocampal, neocortical, and entorhinal categories. Entorhinal units, all of which came from monkey 2, are plotted separately because their spike train properties were quite distinctive.
In all of these plots, the majority of hippocampal units fall into a cluster. The edges of the cluster are not sharp, but can be described approximately by a combination of three criteria: spike width >0.25 ms; average firing rate <1 Hz; and >10% of spikes belonging to bursts. If this is taken as a working definition of a CS cell, across the three monkeys, 80% of well-isolated hippocampal units are classified as CS cells (496 of 618): 21% of neocortical units (84 of 402) and 0% of entorhinal units (1 of 173). Although 20% of hippocampal units did not meet the CS criteria, many of these had values close to the cut-offs. On the other hand, many of the neocortical units that met the criteria would have been difficult to distinguish from hippocampal units in terms of any basic spike train property. Thus the neocortex does contain cells with spike train properties similar to those of hippocampal pyramidal cells; however, they are a minority, mixed with numerous regular-spiking cells and other types of bursting cells.
EEG
Sharp waves resembling those seen in rats could be observed in all three monkeys. Monkey 3, however, was rarely in the drowsy/sleeping behavioral state associated with sharp waves and yielded too few events for detailed analysis to be possible. Sharp waves from the two monkeys with large numbers of events (monkeys 1 and 2) were analyzed. These events, consisting of a high-frequency ripple oscillation (100–120 Hz) accompanied by a large slow wave lasting 300 ms, could be observed simultaneously on all intra-hippocampal electrodes, as well as electrodes located adjacent to the hippocampus, but only minimally or not at all on electrodes located laterally in neocortical areas (such as TE, TF, and IPa). Figure 6 shows several examples; an enlarged view of the ripple component of one sharp wave can be seen in Fig. 6F. Note that the slow wave component of the sharp waves comes mainly after the high-frequency ripple. In typical recordings from the dorsal hippocampus of rats, in contrast, the ripples tend to appear centered on the slow deflections. (Positive polarity is plotted upward in this and subsequent figures.)
To analyze the incidence of sharp waves across time and their relations to unit activity, behavioral states, and behavioral events, we made use of our experience (gained from rodent studies) that high-frequency ripples are the most reliable marker of sharp waves (Csicsvari et al. 2000; Kudrimoti et al. 1999; Skaggs and McNaughton 1998); see METHODS for a description of the algorithm. As explained below, in monkey 1, high-amplitude ripples were observed only when electrodes were located in or near the CA1 layer. Fortuitously, one of the tetrodes from monkey 1 (Fig. 1F, number 8), passed tangentially through the CA1 layer, remained within it for >40 days, and showed large-amplitude ripple events. Sharp waves are global events; therefore the ripples observed on tetrode 8 could be used as triggering events to detect sharp waves for purposes of analyzing their temporal pattern, as well as their relationship to electrode location and to EEG patterns on other electrodes. For monkey 2, high-amplitude ripples were observable on one of the electrodes over a period of 2 wk and could be used in a similar way.
Figure 6, C–E, indicates that sharp waves were associated with bursts of increased unit activity, lasting 100–200 ms, in both CA3 and CA1. Figure 6, C and D, shows separately the responses of CA3 and CA1 complex spike units from monkey 1. Because CA3 and CA1 could not be distinguished for monkey 2, all hippocampal CS cell responses from this monkey are combined in Fig. 6E. Only a small number of putative interneurons were encountered in this study, but in general they showed either very weak responses to sharp waves or responses similar to those of complex spike cells, as shown in Fig. 6, G and H, which show the responses of two simultaneously recorded putative interneurons (from the same tetrode) from monkey 2.
Monkey 2 showed a higher incidence of sharp waves than did monkey 1. In monkey 1, the highest number observed in one session was 620 sharp waves across 135 min, a rate of 4.6/min. Sessions with overall rates <2/min were, however, more common. Monkey 2 showed as many as 896 events in 98 min, for a rate of 9.2/min, and the overall rate was rarely <4/min in sessions where clear ripples were available. Both monkeys showed broadly similar behavior patterns, alternating between periods of task-performance and drowsiness. These rates of sharp wave occurrence can be compared with data from sleeping rats, where sharp waves occur at a rate of 1/s, but in a very irregular temporal pattern (Gerrard et al. 2001; Skaggs and McNaughton 1998). As described above, monkey 1 was an aged female, whereas monkey 2 was a middle-aged male, but because differences in electrode location and behavior can strongly affect the counted number of sharp waves and because of the limited number of subjects, conclusions should not be drawn at this point about age- or sex-related differences in numbers of sharp waves on the basis of these data. Also, it may be worth noting that 1) the rate of sharp waves in drowsy rats is somewhat lower than in rats that are fully asleep, but still substantially higher than the rate observed in these monkeys, and 2) in comparisons of young and old rats, no differences could be observed in the temporal distribution of sharp waves (Gerrard et al. 2001). The clear impression from these data, however, is that sharp waves in these monkeys were an order of magnitude less frequent than would be expected from rats in comparable behavioral states (except that the rats were not restrained).
Although the overall rate of sharp waves in these monkeys was never >0.2 Hz averaged across a session, they occasionally occurred in rapid sequences, separated by 1 s or less, as shown in Fig. 6A. In monkey 2, sharp waves sometimes occurred in rhythmic sequences, with a frequency of 4 Hz. These almost never lasted for >2 s. Such rhythmicity was not observed in monkey 1. Note that the amplitudes of the ripples and sharp waves in the examples shown in Fig. 6, 0.2 mV for ripples and 1 mV for the slow component, are comparable with values commonly seen in rats (Buzsáki 1986), despite the larger brains of the monkeys. Amplitudes, however, vary greatly as a function of recording location.
In rats, sharp waves mainly occur during the LIA EEG state, which is associated with immobility, and particularly with drowsy/slow-wave sleep states (Buzsáki et al. 1986; Vanderwolf 1969). In monkey 1, a comparable relationship could be observed between sharp waves, hippocampal EEG patterns, and behavior, as shown in Fig. 7. At times when the monkey was drowsy and/or quiescent, the hippocampal EEG amplitude was relatively high (increased as much as twofold at some locations), the level of CA1 complex spike cell activity was increased as much as threefold (Fig. 6, D and E), and occasional sharp waves were observed. For monkey 2, the same relationships appeared to be present, although the transitions between states were not as sharp (data not shown). A possible explanation is that the VPC task performed by this monkey did not require a high level of alertness. (Notes taken by observers indicate that most of the time the monkey was drowsy and only marginally attentive, rarely in a highly alert state.) For monkey 3, on the other hand, the behavior was almost always alert and attentive, with only occasional periods of drowsiness. Even so, the few sharp waves observed from this monkey came from periods when the monkey appeared to observers to be drowsy. Note that in Fig. 7, many of the sharp waves appear near the ends of periods of increased unit activity and closed eyes, close to the time of transition to a different pattern. This effect could also be observed in some other data sets from monkey 1 but was not consistent, and could not be observed at all in monkey 2.
In rats, sharp waves originate with increases in CA3 unit activity, but the EEG components—the slow deflection and ripple—are generated mainly downstream, in CA1 (Chrobak and Buzsáki 1996; Ylinen et al. 1995). Although data from this study are not extensive enough to support strong conclusions about whether the same thing happens in monkeys, there are indications that it might. The evidence is shown in Fig. 8. Tetrode 5 from monkey 1 was particularly valuable, because of the trajectory with which it passed through the hippocampus, which replicated in reverse the trajectory most commonly used for studies of rat sharp waves. As described above, ripples recorded from tetrode 8 every day could be used as triggering events to construct a depth profile for tetrode 5, as it was lowered through the hippocampus. The profile shows a strong peak near the level of the fissure and a phase reversal in the CA1 layer; the same pattern as is observed in rats. Neither this nor any other tetrode showed phase reversals when passing through CA3 or the dentate gyrus. Average responses evoked at several other locations by ripples on tetrode 8, shown in Fig. 8A, are also consistent with CA1 being the primary generator. Furthermore, as stated above, for monkey 1, ripples were observed at all CA1 sites but not at CA3 sites. (A depth profile of ripple amplitude for tetrode 5 is shown in the inset of Fig. 8B).
In contrast to sharp waves, the other dominant EEG pattern observed in the rat hippocampus, theta rhythmicity (Green and Arduini 1954; Vanderwolf 1969), was not observed in these monkeys. The hippocampal EEG is extraordinarily complex, and there were many short episodes of apparent rhythmicity over a wide range of frequencies, but nothing resembling the long trains of rhythmic oscillations associated with locomotor behavior and REM sleep in rats.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Pyramidal cells from the macaque hippocampus are complex spike cells
The conclusion that the great majority of cells from the macaque hippocampus are complex spike cells was supported unequivocally for the six hippocampal-penetrating electrodes from monkey 1, for which each session could be matched precisely with an anatomical location on the basis of histology. It is highly probable that these complex spike cells were pyramidal cells, which make up >90% of the neurons in the CA3 and CA1 layers. For monkeys 2 and 3, the histological reconstruction was less precise, but it was found that most tetrodes that were observed to penetrate the hippocampus in the histological material encountered zones of low-firing-rate CS cells as observed in the unit recording data.
The finding that primate hippocampal pyramidal cells are CS cells is significant for at least two reasons. First, it gives criteria for distinguishing the hippocampus from neighboring cortical areas in future studies. Second, and more importantly, it provides essential support for computational models of the hippocampal role in memory that depend on sparse coding, the property that the fraction of neurons responding to any given event is small. Sparsity is required by the majority of existing models (Baum et al. 1988; Marr 1971; McNaughton and Nadel 1990; Rolls and Treves 1990; Treves and Rolls 1991; Tsodyks and Feigel'man 1988), because it is necessary for storage of large numbers of memory patterns without interference. These models store memories by enhancing the strength of synapses whose presynaptic and postsynaptic elements are active simultaneously. The sparser the population activity, the fewer synapses are strengthened by any single event, and the more events can be stored before the pool of available synapses is "used up."
These findings may clarify the interpretation of a number of other studies of primate hippocampal unit responses. In older studies (Miyashita et al. 1989; Watanabe and Niki 1985), and some more recent studies as well (Wirth et al. 2003), the majority of units recorded from the vicinity of the hippocampus were reported to fire at relatively high rates—5 Hz or more—and few to possess complex spike cell characteristics. In several recent studies, however, substantial numbers of hippocampal units, and in particular units with certain types of responses, were reported to fire at overall rates well <1 Hz, with long periods of silence (Hori et al. 2005; Rolls 1999). The results of this study indicate that such units may have been hippocampal pyramidal cells, but that units with much higher rates are unlikely to have been. It should be noted that units with low levels of activity are particularly difficult to recognize using acutely implanted electrodes, which must be moved large distances at the beginning of each recording session, leaving little time to observe the accumulation of activity at any given depth.
It does not seem likely that the differences in unit characteristics reported by each group can be accounted for in terms of task variations. In this study, two tasks were used: a PPAT that was essentially a go/no-go task with predictors for two of the monkeys and a type of passive viewing task for the third monkey. Most of the studies referenced above used object recognition tasks, or object-place association tasks, requiring monkeys to make motor responses when presented with stimuli meeting certain conditions. Although the tasks have varied in detail in a number of ways, most were structured along the same general lines as the go/no-go task in this study, and it would be surprising if they produced major differences in the population statistics of hippocampal activity.
Behavioral state affects monkey hippocampal EEG and unit activity
In the rat hippocampus, three global states of neural population activity can be distinguished, commonly called the RSA, LIA, and SIA states, on the basis of the EEG patterns associated with each of them (Jarosiewicz et al. 2002; Skaggs and McNaughton 1998; Vanderwolf 1969). In the RSA state (which appears during active locomotor behavior—voluntary movement, as Vanderwolf termed it—and REM sleep), and in the SIA state (which appears intermittently during sleep), the population activity of hippocampal pyramidal and granule cells is very sparse. During any randomly chosen 1-s period, <5% of these cells fire action potentials. For the remaining small fraction of cells, firing rates are distributed more or less exponentially, with as many as 40–50 spikes/s emitted by the most active neurons (unpublished observations). For a typical neuron, the spike train consists of long periods of almost complete silence—sometimes lasting for minutes—punctuated by brief bursts of activity: sometimes weak and sometimes intense. In rats, these bursts of activity can occur when the rat passes through a particular place in an environment (Leutgeb et al. 2005; O'Keefe and Nadel 1978) and last from a minimum of 500 ms to a maximum that depends on how much time the rat spends near that place, although the intensity of firing tends to fluctuate if the period of activity lasts longer than a second. The overall population firing rate is not constant but increases almost linearly as a function of running speed (McNaughton et al. 1983).
In the rat LIA state, which occurs during slow wave sleep and assorted waking behaviors not involving locomotion, the structure of population activity is quite different. Most of the spike activity is concentrated within sharp waves, which occur at average rates of up to once per second but are spaced irregularly in time. Different sharp waves usually activate different subpopulations of neurons. Consequently, the overall population activity, accumulated over 1 s or longer periods, is more evenly distributed during LIA than during RSA or SIA.
In this study, the macaque hippocampus showed distinct population activity patterns during active/alert versus inactive/drowsy/sleeping states, corresponding to the LIA versus RSA distinction observed in rats. (No clear analog of the SIA state was observed.) The strongest evidence was obtained from monkey 1, in which eyes-closed periods were associated with increased hippocampal unit activity, increased hippocampal EEG amplitude, and the occasional appearance of sharp waves. Monkey 2 performed a passive visual task and was rarely in a highly alert state during recording. Nevertheless, when the monkey was not performing the task, hippocampal activity and EEG amplitude were increased, as was the incidence of sharp waves. Monkey 3, in contrast, was almost always highly alert and would not tolerate periods of inactivity. Consistent with this behavior pattern, few sharp wave events were observed from this monkey.
Monkey hippocampus generates sharp wave/ripple events
In monkey 1, high-amplitude EEG ripples were observed if and only if an electrode was located near the CA1 cell body layer. These CA1 ripples were associated with large slow EEG deflections recorded from electrodes in other parts of the hippocampus; the slow waves were much smaller at locations distant from the hippocampus. In monkey 2, strong ripples were recorded from one hippocampal-penetrating electrode as it passed through a layer of densely packed CS cells. Large slow waves associated with the ripples were recorded from several other hippocampal-penetrating electrodes. In monkey 3, very few sharp waves were observed on tetrodes that penetrated the hippocampus, but this is consistent with the observation that this monkey very rarely entered the drowsy/sleeping behavioral state conducive to sharp waves.
The finding that the primate hippocampus generates sharp waves was expected, because high-frequency ripples have been reported in the hippocampal region of human patients (Bragin et al. 1999; Staba et al. 2002, 2004). This finding is significant, however, for several reasons. These include the fact that theoretical accounts of memory consolidation assign an important role to sharp waves (McNaughton 1998; Sirota et al. 2003); in rats, sleep reactivation of behaviorally induced ensemble activity patterns has been associated specifically with sharp waves (Gerrard et al. 2001; Kudrimoti et al. 1999); and, more generally, because the prevalence of this type of activity across such different mammalian species suggest the functional importance of sharp waves.
Sharp waves in monkeys are probably generated mainly by the CA1 layer, as they are in rats (Buzsáki 1986; Chrobak and Buzsáki 1996; Ylinen et al. 1995). The evidence for this comes from monkey 1. First, high-frequency ripples were only recorded in the vicinity of the CA1 layer. Second, a phase reversal of the slow component was seen above the CA1 layer on an electrode that followed a dentate-gyrus/fissure/CA1 trajectory, resembling the phase reversal seen in rats (but in the opposite direction, because the inferiorly located hippocampus of the monkey is upside-down with respect to the dorsally located hippocampus of the rat). No phase reversals were observed in the vicinity of the CA3 layer.
Despite the overall similarity between rat and monkey sharp waves, there did, however, appear to be some quantitative differences: the ripple oscillations in monkeys had frequencies of 100–120 Hz compared with 130–200 Hz in rats (Csicsvari et al. 1999), and the overall incidence of sharp waves during drowsy/sleeping behavior appeared to be lower in monkeys. Also, the fact that the high-frequency ripple comes before the largest part of the slow wave differs from the typical situation in the dorsal hippocampus of the rat, where the two more or less coincide in time. However, because the current recordings were conducted near the anterior end of the hippocampus (which corresponds to the ventral hippocampus in the rat), whereas in rats most studies have involved the dorsal end, it cannot be ruled out that the differences observed here reflect location differences rather than species differences.
Now that clear sharp waves have been observed in primates, it would also be important to test directly whether primates show the sort of reactivation of hippocampal activity patterns related to behavior during sharp wave periods that has been implicated in memory consolidation processes. (Kudrimoti et al. 1999; Wilson and McNaughton 1994). Such a demonstration, however, has the technical requirement of a simultaneously recorded ensemble of a minimum of 50 hippocampal pyramidal cells, of which a substantial number must be activated by the behavioral task. In addition, extended periods of sleep or drowsiness before and after the task period are required. These conditions were not met in the data available for this report; there is no obvious reason why they should not be achievable as the techniques are refined.
In addition to sharp waves, a very different type of EEG phenomenon is easily observed in rats and many other species: regularly oscillating theta waves with frequencies usually in the 4- to 12-Hz range. Early studies suggested that theta is considerably less prevalent in behaving or sleeping primates (Green and Arduini 1954), although it has been observed in monkeys anesthetized with urethane (Stewart and Fox 1991). The monkeys in this study also did not show salient hippocampal theta rhythmicity. The hippocampal EEG was very complex, with occasional short-lasting bouts of rhythmicity in many frequency bands, but there was nothing resembling the dominant rhythmic pattern seen in rats and cats during active locomotion and REM sleep. The inability to observe theta here may not, however, be conclusive, because on one hand, our monkeys did not sleep deeply enough to enter the REM state, and on the other hand, it is possible that the head restraint needed for recording suppresses theta, as whole body restraint does in rats (Foster et al. 1989).
Given that high-frequency ripples have also been described in the epileptic human hippocampus and temporal lobe (Bragin et al. 1999; Staba et al. 2002, 2004), it is likely that these observations carry over to humans as well. The similarity of sharp waves across a range of mammalian species encourages speculation that they play an important role in the functions of the hippocampal system and that theories of memory stabilization motivated by the properties of sharp wave/ripple events in rats may be applicable to primates as well, including humans.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: C. A. Barnes, University of Arizona, Evelyn F. McKnight Brain Institute, Life Sciences North, Rm 384, Tucson, AZ 85724 (E-mail: carol{at}nsma.arizona.edu)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Baum EB, Moody J, Wilczek F. Internal representations for associative memory. Biol Cybern 59: 217–228, 1988.[CrossRef][Web of Science]
Bragin A, Engel J Jr, Wilson CL, Fried I, Buzsáki G. High-frequency oscillations in human brain. Hippocampus 9: 137–142, 1999.[CrossRef][Web of Science][Medline]
Brasted PJ, Bussey TJ, Murray EA, Wise SP. Role of the hippocampal system in associative learning beyond the spatial domain. Brain 126: 1202–1223, 2003.
Burgess N, Becker S, Kind JA, O'Keefe J. Memory for events and their spatial context: models and experiments. Philos Trans R Soc Lond B Biol Sci 356: 1493–1503, 2001.
Buzsáki G. Hippocampal sharp waves: their origin and significance. Brain Res 398: 242–252, 1986.[CrossRef][Web of Science][Medline]
Buzsáki G. A two-stage model of memory trace formation: a role for "noisy" brain states. Neuroscience 31: 551–570, 1989.[CrossRef][Web of Science][Medline]
Buzsáki G, Czopf J, Kondakor I, Kellenyi L. Laminar distribution of hippocampal rhythmic slow activity (RSA) in the behaving rat: current-source density analysis, effects of urethane and atropine. Brain Res 365: 125–137, 1986.[CrossRef][Web of Science][Medline]
Chrobak JJ, Buzsáki G. High-frequency oscillations in the output networks of the hippocampal-entorhinal axis of the freely behaving rat. J Neurosci 16: 3056–3066, 1996.
Csicsvari J, Hirase H, Czurko A, Mamiya A, Buzsáki G. Fast network oscillations in the hippocampal CA1 region of the behaving rat. J Neurosci 19: RC20, 1999.
Csicsvari J, Hirase H, Mamiya A, Buzsáki G. Ensemble patterns of hippocampal ca3-ca1 neurons during sharp wave-associated population events. Neuron 28: 585–594, 2000.[CrossRef][Web of Science][Medline]
Eichenbaum HA. Cortical-hippocampal system for declarative memory. Nat Rev Neurosci 1: 41–50, 2000.[CrossRef][Web of Science][Medline]
Ekstrom AD, Kaplan JB, Ho E, Shattuck K, Kahana MJ. Human hippocampal theta activity during virtual navigation. Hippocampus 15: 881–889, 2005.[CrossRef][Web of Science][Medline]
Erickson CA, Desimone R. Responses of macaque perirhinal neurons during and after visual stimulus association learning. J Neurosci 19: 10404–10416, 1999.
Foster TC, Castro CA, McNaughton BL. Spatial selectivity of rat hippocampal neurons: dependence on preparedness for movement. Science 244: 1580–1582, 1989.
Fox SE, Ranck JB Jr. Localization and anatomical identification of theta cells and complex spike cells in dorsal hippocampal formation in rats. Exp Neurol 49: 299–313, 1975.[CrossRef][Web of Science][Medline]
Gerrard JL, Kudrimoti H, McNaughton BL, Barnes CA. Reactivation of hippocampal ensemble activity patterns in the aging rat. Behav Neurosci 115: 1180–1192, 2001.[CrossRef][Web of Science][Medline]
Gothard KM, Skaggs WE, Moore KM, McNaughton BL. Binding of hippocampal CA1 neural activity to multiple reference frames in a landmark-based navigation task. J Neurosci 16: 823–835, 1996.
Green JD, Arduini A. Hippocampal electrical activity in arousal. J Neurophysiol 17: 533–557, 1954.
Halgren E, Ball TL, Crandall PH. Human hippocampal formation EEG desynchronizes during attentiveness and movement. Electroencephalogr Clin Neurophysiol 44: 778–781, 1978.[CrossRef][Web of Science][Medline]
Harris KD, Hirase H, Leinenkugel X, Henze DA, Buzsáki G. Temporal interaction between single spikes and complex spike bursts in hippocampal pyramidal cells. Neuron 32: 141–149, 2001.[CrossRef][Web of Science][Medline]
Hoffman KL, McNaughton BL. Coordinated reactivation of distributed memory traces in primate neocortex. Science 297: 2070–2073, 2002.
Hori E, Nishio Y, Kazui K, Umeno K, Tabuchi E, Sasaki K, Endo S, Ono T, Nishijo H. Place-related neural responses in the monkey hippocampal formation in a virtual space. Hippocampus 15: 991–996, 2005.[CrossRef][Web of Science][Medline]
Ito M. The Cerebellum and Neural Control. New York: Raven Press, 1984.
Jarosiewicz B, McNaughton BL, Skaggs WE. Hippocampal population activity during the small-amplitude irregular activity state in the rat. J Neurosci 22: 1373–1384, 2002.
Kahana M, Seelig D, Madsen J. Theta returns. Curr Opin Neurobiol 11: 739–744, 2001.[CrossRef][Web of Science][Medline]
Kudrimoti HS, Barnes CA, McNaughton BL. Reactivation of hippocampal cell assemblies: effects of behavioral state, experience, and EEG dynamics. J Neurosci 19: 4090–4101, 1999.
Leutgeb S, Leutgeb JK, Moser MB, Moser EI. Place cells, spatial maps and the population code for memory. Curr Opin Neurobiol 15: 738–746, 2005.[CrossRef][Web of Science][Medline]
Marr D. Simple memory: a theory for archicortex. Philos Trans R Soc Lond B Biol Sci 262: 23–81, 1971.
McClelland JL, McNaughton BL, O'Reilly RC. Why there are complementary learning systems in the hippocampus and neocortex: insights from the successes and failures of connectionist models of learning and memory. Psychol Rev 102: 419–457, 1995.[CrossRef][Web of Science][Medline]
McNaughton B. The neurophysiology of reminiscence. Neurobiol Learn Mem 70: 252–267, 1998.[CrossRef][Web of Science][Medline]
McNaughton BL, Barnes CA, O'Keefe J. The contributions of position, direction, and velocity to single unit activity in the hippocampus of freely-moving rats. Exp Brain Res 52: 41–49, 1983.[Web of Science][Medline]
McNaughton BL, Morris RGM. Hippocampal synaptic enhancement and information storage within a distributed memory system. Trends Neurosci 10: 408–415, 1987.[CrossRef][Web of Science]
McNaughton BL, Nadel L. Hebb-Marr networks and the neurobiological representation of action in space. In: Neuroscience and Connectionist Theory, edited by Gluck MA. Hillsdale, NJ: Lawrence Earlbaum Associates, 1990, p. 1–63.
Miyashita ETRY, Cahusac PMB, Kesner RP, Niki H, Feigenbaum JD, Bach L. Hippocampal neurons in the monkey with activity related to the place in which a stimulus is shown. J Neurosci 9: 1835–1845, 1989.[Abstract]
Murre JM. TraceLink: a model of amnesia and consolidation of memory. Hippocampus 6: 675–684, 1996.[CrossRef][Web of Science][Medline]
O'Keefe J, Nadel L. The Hippocampus as a Cognitive Map. Oxford, UK: Clarendon Press, 1978.
O'Reilly RC, Norman KA. Hippocampal and neocortical contributions to memory: advances in the complementary learning systems framework. Trends Cogn Sci 6: 505–510, 2002.[CrossRef][Web of Science][Medline]
Paxinos G, Huang X-F, Toga AW. The Rhesus Monkey Brain in Stereotaxic Coordinates. San Diego, CA: Academic Press, 2000.
Qin YL, McNaughton BL, Skaggs WE, Barnes CA. Memory reprocessing in corticocortical and hippocampocortical neuronal ensembles. Philos Trans R Soc Lond B Biol Sci 352: 1525–1533, 1997.
Rolls ET. Spatial view cells and the representation of place in the primate hippocampus. Hippocampus 9: 467–480, 1999.[CrossRef][Web of Science][Medline]
Rolls ET, Treves A. The relative advantages of sparse versus distributed encoding for associative neuronal networks in the brain. Network 1: 407–421, 1990.
Schmued LC, Albertson C, Slikker W Jr. Fluoro-Jade: a novel fluorochrome for the sensitive and reliable histochemical localization of neuronal degeneration. Brain Res 751: 37–46, 1997.[CrossRef][Web of Science][Medline]
Sirota A, Csicsvari J, Buhl D, Buzsaki G. Communication between neocortex and hippocampus during sleep in rodents. Proc Natl Acad Sci USA 100: 2065–2069, 2003.
Skaggs WE, McNaughton BL. Replay of neuronal firing sequences in rat hippocampus following spatial experience. Science 271: 1870–1873, 1996.[Abstract]
Skaggs WE, McNaughton BL. Neuronal ensemble dynamics in hippocampus and neocortex during sleep and waking. In: Neuronal Ensembles, edited by Eichenbaum H, Davis J. New York: Wiley, 1998, p. 235–246.
Squire L, Stark C, Clark R. The medial temporal lobe. Annu Rev Neurosci 27: 279–306, 2004.[CrossRef][Web of Science][Medline]
Squire LR. Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans. Psychol Rev 99: 195–231, 1992.[CrossRef][Web of Science][Medline]
Staba RJ, Wilson CL, Bragin A, Fried I, Engel J Jr. Quantitative analysis of high-frequency oscillations (80–500 hz) recorded in human epileptic hippocampus and entorhinal cortex. J Neurophysiol 88: 1743–1752, 2002.
Staba RJ, Wilson CL, Bragin A, Jhung D, Fried I, Engel J, Jr. High-frequency oscillations recorded in human medial temporal lobe during sleep. Ann Neurol 56: 108–115, 2004.[CrossRef][Web of Science][Medline]
Stewart M, Fox SE. Hippocampal theta activity in monkeys. Brain Res 538: 59–63, 1991.[CrossRef][Web of Science][Medline]
Thompson LT, Best PJ. Long-term stability of hippocampal place cell activity recorded from freely-behaving rats. Brain Res 509: 299–308, 1990.[CrossRef][Web of Science][Medline]
Treves A, Rolls ET. What determines the capacity of autoassociative memories in the brain? Network 2: 371–398, 1991.[Web of Science]
Treves A, Skaggs WE, Barnes CA. How much of the hippocampus can be explained by functional constraints? Hippocampus 6: 666–674, 1996.[CrossRef][Web of Science][Medline]
Tsodyks MV, Feigel'man MV. The enhanced storage capacity in neural networks with low activity level. Europhys Lett 6: 101–105, 1988.[CrossRef][Web of Science]
Ulanovsky N, Moss CF. Hippocampal cellular and network activity in freely moving echolocating bats. Nat Neurosci 10: 224–233, 2007.[CrossRef][Web of Science][Medline]
Vanderwolf CH. Hippocampal electrical activity and voluntary movement in the rat. Electroencephelogr Clin Neurophysiol 26: 407–418, 1969.[CrossRef]
Watanabe T, Niki H. Hippocampal unit activity and delayed response in the monkey. Brain Res 325: 241–254, 1985.[CrossRef][Web of Science][Medline]
Wilson MA, McNaughton BL. Dynamics of the hippocampal ensemble code for space. Science 261: 1055–1058, 1993.
Wilson MA, McNaughton BL. Reactivation of hippocampal ensemble memories during sleep. Science 265: 676–679, 1994.
Wirth S, Yanike M, Frank LM, Smith AC, Brown EN, Suzuki WA. Single neurons in the monkey hippocampus and learning of new associations. Science 300: 1578–1581, 2003.
Wise SP, Murray EA. Role of the hippocampal system in conditional motor learning: mapping antecedents to action. Hippocampus 9: 101–117, 1999.[CrossRef][Web of Science][Medline]
Ylinen A, Bragin A, Nádasdy Z, Jandó G, Szabó I, Sik A, Buzsáki G. Sharp wave-associated high-frequency oscillation (200 Hz) in the intact hippocampus: network and intracellular mechanisms. J Neurosci 15: 30–46, 1995.[Abstract]
Zola SM, Squire LR, Teng E, Stefanacci L, Buffalo EA, Clark RE. Impaired recognition memory in monkeys after damage limited to the hippocampal region. J Neurosci 20: 451–463, 2000.
This article has been cited by other articles:
|
|
 |
|
  Z. Petanjek, B. Berger, and M. Esclapez Origins of Cortical GABAergic Neurons in the Cynomolgus Monkey Cereb Cortex, February 1, 2009; 19(2): 249 - 262. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Axmacher, C. E. Elger, and J. Fell Ripples in the medial temporal lobe are relevant for human memory consolidation Brain, July 1, 2008; 131(7): 1806 - 1817. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Le Van Quyen, A. Bragin, R. Staba, B. Crepon, C. L. Wilson, and J. Engel Jr Cell Type-Specific Firing during Ripple Oscillations in the Hippocampal Formation of Humans J. Neurosci., June 11, 2008; 28(24): 6104 - 6110. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
