Dentate EEG Spikes and Associated ... - Buzsaki Lab

JOURNALOF NEUROPHYSIOLOGY

Vol. 73, No. 4, April 1995. Printed in U.S.A.

Dentate EEG Spikes and Associated Interneuronal Population Bursts in

the Hippocampal Hilar Region of the Rat

ANATOL BRAGIN, GABOR JAND& ZOLTAN NADASDY, MARTIN VAN LANDEGHEM, AND GYiiRGY BUZSAKI Center for Molecular and Behavioral Neuroscience, Rutgers, The State Universitv of New Jersev. Newark, New Jersey 07102

SUMMAR’Y AND CONCLUSIONS

1. This paper describes two novel population patterns in the dentate gyrus of the awake rat, termed type 1 and type 2 dentate spikes (DS 1, DS2). Their cellular generation and spatial distribution were examined by simultaneous recording of field potentials and unit activity using multiple-site silicon probes and wire electrode arrays.

2. Dentate spikes were large amplitude (2-4 mV), short duration ( <30 ms) field potentials that occurred sparsely during behavioral immobility and slow-wave sleep. Current-source density analysis revealed large sinks in the outer (DSl ) and middle (DS2) thirds of the dentate molecular layer, respectively. DSl and DS2 had similar longitudinal, lateral, and interhemispheric synchrony.

3. Dentate spikes invariably were coupled to synchronous population bursts of putative hilar interneurons. CA3 pyramidal cells, on the other hand were suppressed during dentate spikes.

4. After bilateral removal of the entorhinal cortex, dentate spikes disappeared, whereas sharp wave-associated bursts, reflecting synchronous discharge of the CA3-CA1 network, increased several fold.

5. These physiological characteristics of the dentate spikes suggest that they are triggered by a population burst of layer II stellate cells of the lateral (OS1 ) and medial (DS2) entorhinal cortex.

6. We suggest that dentate spike-associated synchronized bursts of hilar-region interneurons provide a suppressive effect on the excitability of the CA3-CA1 network in the intact brain.

INTRODUCTION

Various population patterns, as reflected by spontaneous field potentials and rhythms, are present in the hippocampal formation, including theta activity and associated gamma pattern (40- 100 Hz), hippocampal sharp waves ( SPW) , and the SPW-associated high-frequency (200 Hz) oscillation ( “ripple” ) , sleep spindles, and delta waves of sleep (Bland 1990; Bragin et al. 1993; Buzsaki 1986; Buzsaki et al. 1983, 1992, 1994; Lopes da Silva et al. 1990; O’Keefe and Nadel 1978; Traub and Miles 1991; Ylinen et al. 1995a,b). Hippo- campal rhythmic slow activity (theta) is the most studied hippocampal pattern and has been implicated in several func- tions, ranging from sensory processing to the voluntary control of movement (Grastyan et al. 1959; Vanderwolf 1969). In the rat, it is associated with exploratory patterns, such as walking, turning rearing and sniffing and theta is the hall- mark of the paradoxical phase of sleep. During consumma- tory behaviors, behavioral immobility, and slow-wave sleep theta is replaced by a mixture of intermittent waves, usually referred to as large-amplitude irregular activity (Vanderwolf

1969). One of the physiologically characterized patterns of the large amplitude irregular activity is a short-duration (40- 120 ms) sharp wave (SPW) present in the CA3-CAl-subiculum-entorhinal cortex circuitry (Buzsaki 1986; Buzsaki et al. 1983; Chrobak and Buzsaki 1995; Suzuki and Smith 1987). The immediate cause of SPW in the CA1 region is the synchronous discharge of a large number of CA3 pyramidal neurons and the consequent near-simultaneous depolarization of CA1 pyramidal cells. In conduction with the stratum radiatum SPWs, there are high-frequency field oscillations ( “ripples” ) ( 0’ Keefe and Nadel 1978) present in the CA1 pyramidal layer and deep layers of the entorhinal cortex (Buzsaki et al. 1992; Chrobak and Buzsaki 1994; Ylinen et al. 1995).

In the experiments presented here, we describe two new population patterns in the dentate gyrus that emerge in con- junction with a synchronous discharge of inhibitory interneurons of the hilar region. We hypothesize that these population patterns, termed dentate spikes, serve to decrease the network excitability of the CA3 recurrent collateral system during nontheta behaviors in the intact animal.

METHODS

Animals and surgery *

Forty-eight male and female rats (250-450 g) of the Sprague- Dawley strain were used in this study. The rats were anesthetized with a mixture (4 ml/kg) of ketamine (25 mg/ml), xylazine ( 1.3 mg/ml), and acepromazine (0.25 mg/ml). Pairs of stainless steel wires (100 ,u,m in diameter) with 0.5-mm vertical tip separation were placed in the angular bundle on the right or both sides to stimulate the medial perforant path afferents to the hippocampus (AP, -7.0 mm from bregma; L, 3.5 mm from midline; V, 3.0 mm). Another electrode pair was placed into the ventral hippocampal commissure (AP, -0.8; L, 0.5; V, -4.2) to stimulate the commissural afferents to the CA l-3 regions and the dentate gyrus.

Three different recording electrodes were used: stationary wire electrodes, microelectrode arrays, and multisite recording silicon probes. Stationary electrodes (2-4 60+m tungsten wires) were implanted in the strata pyramidale and radiatum of CA1 and the molecular layer and hilus of the dentate gyrus unilaterally or bilaterally (AP, -3.0; L, 2.6; V, -2.4, -3.0). Microelectrode arrays consisted of four to eight tungsten wires (20 or 60 ,um in diameter). Two or three 20qm wires or a single 60-pm wire was inserted into a parallel array of fused silica tubes with 0.3-mm horizontal separations. The wires protruded 3-4 mm from the guiding tubes. The 20-pm wires within a single silica tube were glued together with varnish. A 3.5 X 1 mm slot was drilled into the skull above the

0022-3077/95 $3.00 Copyright 0 1995 The American Physiological Society 1691

1692 A. BRAGIN, G. JAND6, 2. NADASDY, M. VAN LANDEGHEM, AND G. BUZSk

dorsal hippocampus along the longitudinal or the traverse (dentate- CA3) axis of the structure. For simultaneous recording of field potentials and unit activity in different layers, silicon probes micro- machined with thin-film technology (Wise and Najafi 1991) .were used in 10 rats. The recording sites (5 x 15 pm2, sputtered iridium) were spaced 100 pm. The thickness of the silicon shank was 15 pm throughout. Five or 16 recording sites were available (80 ,wm wide at the base, narrowing to 15 ,um at the tip). In three rats, an epidural screw electrode, driven into the skull above the frontal cortex (AP, 2.5; L, 2.5)) was used to record neocortical electroen- cephalogram (EEG) .

The parallel wire arrays and silicon probes, attached to a movable headstage, were inserted into the neocortex or corpus callosum during surgery. After recovery, the tips were lowered gradually into the hippocampus. During the experiment, the evoked field potentials helped guide the microelectrodes. Two stainless steel watch screws driven into the bone above the cerebellum served as indifferent and ground electrodes.

Recording and data processing

Four 4-channel MOSFET input operational amplifiers, mounted in the female connector, served to eliminate cable movement arti- facts (Buzsaki et al. 1989a). The movement of the rat was recorded by a sensitive magnet-coil velocity detector attached to the trans- parent homecage. Physiological data were recorded wide band and sampled with 12-bit precision. The data were stored on optical disks. Field potentials were recorded either continuously (1 kHz per channel sampling) or collected as dentate spike-triggered ep- ochs (400 ms) together with unit discharges ( 10 kHz per channel sampling) in the absence of overt movements. Still alert behavior was defined as immobility with eyes open. Drowsiness was defined by an immobile sleeping posture with eyes open and the head resting on one or both forepaws. Finally, slow-wave sleep was characterized by sleeping posture with eyes closed and the domi- nance of large amplitude delta waves in the hippocampal EEG.

All analysis was carried out off-line on 486/33 and an IBM RS 6000 computers. The recorded data were digitally filtered at 120 dB/octave to select the frequency of interest: unit activity (500 Hz-5 kHz), dentate spikes ( 1- 1,000 Hz) and SPW-associated high-frequency field ripples ( 100-300 Hz). The dentate spikes were detected by a window discrimination program and their peaks were determined by a peak detection algorithm. These derived pulses were used as the zero point for the construction of field averages and cross-correlograms. The power of EEG was calculated from 25-s segment EEG during different behaviors.

Current source density (CSD) analysis

Complete accounts of the theoretical basis of CSD analysis have been presented earlier (Freeman and Nicholson 1975; Mitzdorf 1985). Dentate spikes were first separated into groups (see RE-

SULTS) and averaged using their identified positive peaks (~1 = 50- 200). Smoothing of the averaged potential profiles was accomplished by convoluting the potential as a function of depth with a three-point rolling average of the voltage in depth. The second spatial derivative was calculated from the smoothed data points. Although some resistivity differences are present in the different hippocampal layers (Holsheimer 1987)) in practice these are not large enough to significantly modify the calculated distribution of current generators. Therefore isotropy of the extracellular space is assumed in the CSD analysis. The results thus are presented as the second derivative of potential as a function of depth, and will be referred to as CSD. The second differences of the voltage profile were divided by the square of the step size in centimeters and by the estimated tissue impedance (300 0 * cm) to convert them into CSD estimates (Brankack et al. 1993). CSD measurements were

plotted as a function of depth at selected time points or at all time points. In the latter case, color plots were constructed with color intensities reflecting sinks and sources in a time (abscissa) versus depth (ordinate) coordinate system. The exact anatomic layers, corresponding to the vertical scale of the CSD maps, were reconstructed with the aid of the histologically identified recording tracks and evoked potentials. The depth profiles of the perforant path and commissurally evoked responses have been well studied in the rat (Brankack et al. 1993; Buzsaki and Czeh 198 1; Deadwyler et al. 1975; Leung 1979). Currents sinks and sources associated with the activation of these known anatomic afferents provided precise landmarks for the identification of the recording sites. In addition, unitary activity in the CA1 pyramidal layer provided further help for the depth calibration of the electrodes.

Spike separation and analysis

A Haar transformation was performed on the digitally filtered (500 Hz-5 kHz) traces to locate the occurrence of spike events (Yang 1988). The spike events were identified by the factorial description of their shape ( “feature” ) characteristics. Event sort- ing was carried out by an IBM RS6000 using a perceptron version of the incremental conceptual clustering procedure (SpikePercep- tron, MUA Technology BT, P&s, Hungary). For visual display, the spike occurrences within the same target cluster were superimposed and projected relative to the reference event (Ylinen et al. 1995a). The cross-correlograms were calculated from the cumula- tive number of spikes. Isolation of single units within and across clusters was verified by clear refractory periods (2-3 ms) in their interspike interval histograms.

Several isolated units were identified by physiological criteria. Units that discharged at a shorter latency than the population spike and responded with two or more action potentials in response to perforant path or commissural stimulation were classified as interneurons (Buzsaki and Eidelberg 1982; Fox and Ranck 198 1) . These cells typically fired at high rates ( > 15 Hz) and discharged in rhythmic groups at the field theta frequency. Units that displayed spontaneous complex burst patterns were classified as pyramidal cells or mossy cells (Ranck 1973; Soltesz et al. 1993). Complex- spike cells typically fired at < 1 Hz. Physiological criteria for the identification of granule cells are not universally accepted (Buzsaki and Czeh 1992; Mizumori et al. 1989) and criteria for the separation of the various subgroups of interneurons in the hilar region (Amaral 1978; Han et al. 1993) are not yet available.

Unit activity was cross-correlated with the peaks of the dentate spikes or with high-frequency field oscillation of the CA1 region (ripples), with the field events serving as reference. Firing rates in a given time epoch (e.g., - lo- 10 ms of the peak of the dentate spikes) were compared with shuffled spikes, obtained outside of the dentate spike events, and statistical significance was assessed by nonpaired t-tests.

Entorhinal cortex lesion

In a group of seven rats, the entorhinal cortex was removed bilaterally. The electrodes were placed first, and the lesion was made after the completion of the physiological tests. The lesion was made under halothane (2.5%) gas anesthesia. The bone above the entorhinal and perirhinal cortex was removed and the dura was cut. The gray matter and the underlying white matter was aspirated under microscopic vision. The resulting cavity was filled by gel- foam and the wound was closed.

Colchicine lesion

In seven rats, colchicine toxin (2 pg in 0.5 ~1) was injected in the dentate gyrus of the dorsal hippocampus at the following

HIPPOCAMPAL DENTATE SPIKES 1693

coordinates: AP, -2.0 and 3.0; L, 1.5 and 2.2; and V, 3.0 and 3.0. These rats were equipped with fixed electrodes (~1 = 4) or movable electrode arrays (n = 3) 1 mo after the toxin injection. The goal of these experiments was to examine the survival of dentate spikes following toxin damage of the granule cell population.

Histological procedures

After completion of the experiments, the rats were anesthetized deeply and perfused through the heart first with cacodylate-buffered saline (pH 7.5) followed by a cacodylate-buffered fixative containing 4% paraformaldehyde and 5.9% calcium chloride (pH 7.5). Brains were left in situ for 24 h, removed, and then postfixed in the same solution for 1 wk. The brains were sectioned with the probes left in the brain on a vibrotome at 100 pm in the coronal plane. The sections were stained with the Gallyas silver method (Gallyas et al. 1993). Briefly, the sections were dehydrated with propanol and placed in an esterifying solution (98% propanol, 1.2% sulfuric acid) at 56OC for 16 h. After rehydration and sec- tioning, they were processed according to the following procedure: pretreatment in 8% acetic acid for 10 min, wash in water for 1 min, physical development with tungstosilicic acid for 10 min, and wash in 1% acetic acid. Finally, the sections were dehydrated, mounted on slides, and coverslipped. Selected sections were stained with cresyl violet.

16

FIG. 1. Simultaneous recording of field activity ( 16 recording sites; lOO- pm tip intervals) in the CA1 -dentate axis of the dorsal hippocampus during awake immobility. For clarity only every second traces are displayed. Sharp wave (SPW) in the CA1 region and type 1 and type 2 dentate spikes (DS 1 and DS2, respectively) are indicated by arrow. Double arrow in 4 indicates SPW-associated high-frequency (200 Hz) field oscillation ( “ripple” ) in the CA1 pyramidal layer; 14 is the granule cell layer. Temporal proximity of the 3 transient events (SPW, DS 1 and DS2) as shown here was very rare.

t .- E

2; n

T

B 800 1 T

Irre@u-, sharp transients occurred Juring immobility. grooming, drinking and slow-wave sleep in the hippocampal formation ( Fig. 1 ). Sharp w;tvcs (SPW ) in the CA I r-egion h;tve been described in dct;til car-licr (Buzsaki 1986; Buzsaki ct al. 1983, 1992; dc Curtis ct al. I99 I ; Suzuki and Smith 1987). SPW represent synchronous discharges of the neurons in the CA3 and hilar regions with resulting current sinks at the termination zones of the associational paths, i.e., the stratum radiatum of CA I and the inner molecular layer of the dentate gyrus (Buzsaki 1986; Ylinen et al. 1995a). In addition to the wider SPW, large-amplitude ( l-4 mV), short-duration ( 1 O-40 ms) transients of positive polarity were observed in the hilus of the dentate gyrus, and we termed them dentate spikes. The behavioral correlates of dcntatc spikes wcrc identical to SPW. They ~~~LII-I-c~ irrcgu- larlv and at about the s;mc incidcncc as SPW at 0.0 1 -0.Y s frequencies during immobility, drowsiness, and slow-wave sleep. Although the incidence of dentate spikes varied as a function of behavior and associated EEG power in the delta band (Fig. 2), the morphological features of dentate spikes, described below, did not depend on behavior. Most dentate spikes were collected in the still alert rat and during slow- wave sleep.

Two types c! f dentate spikes

On the basis of their distinct depth profile and wave duration, dentate spikes could be clearly distinguished from the wider SPWs, which are associated with population bursts of the CA3-CA1 network. SPWs and dentate spikes rarely occurred together in time and SPW virtually never followed dentate spikes within 200 ms.

1694 A. BRAGIN, G. JAND6, Z. NADASDY, M. VAN LANDEGHEM, AND G. BUZSAKI

A DSI B DS2 C

r t slope of EPSP

-.x3-- ix1

--9 DS2

amplitude of DSl , DS2 (mV) slope of EPSP (x 0.5 mV/msec)

FIG. 3. Depth profiles of two types of dentate spikes (DSl, DS2). A and B: averaged DSI and DS2 (400 ms) recorded simultaneously from 16 sites in the CAl-dentate hilus axis (n = 30 events each, in a single rat). Note the different levels of polarity reversals for DSl and DS2. The wider “base” of DSl results from averaging the spindle-like, short oscillatory patterns of individual DSl. Trace 2 corresponds to the CA1 pyramidal layer, 16 is in the molecular layer of the ventral leaf. C: averaged amplitude plots of DSI and DS2 at their peaks and the slope of the perforant path evoked responses as a function of depth from 4 rats with similar medio-lateral and longitudinal electrode positions m the CAI-dentate axis. Note the similar polarity reversals of DS2 and the slope of the evoked extracellular postsynaptic potentials [(slope of excitatory ,postsynaptic potential (EPSP) 1. D: corresponding hippocampal layers. r, stratum radiatum of CA 1; hf, hippocampal fissure, m, molecular layer; g, granule cell layer; h, hilar region.

Quantitative evaluation of the laminar distribution of dentate spikes was carried out on simultaneously recorded potentials from 16&e silicon probes or from successive recordings with wire electrodes at different depths. In the latter case, stationary electrodes were placed in the hilus and molecular layer of the dentate gyrus and averages of dentate spikes were simultaneously obtained from the stationary wires and the movable electrodes at 100 pm steps. The averages obtained from the stationary electrodes were used to normalize the amplitude of the dentate spikes recorded by the movable electrodes. Based on wave morphology and voltage-versus-depth profiles, two types of dentate spikes could be distinguished. The first type (DS 1) showed a grad- ual decrease in amplitude in the molecular layer-hilar axis with a reversal in the outer third of the molecular layer (Figs. 1 and 3). The voltage gradient across the granule cell layer was 2-6 mV/mm. In addition, DSl occasionally consisted of three to five repetitive spikes of increasing-decreasing amplitude at 70-100 Hz with a large spike in the middle (Fig. 1). In contrast, the second type of sharp transient (DS2) had a fast rise time (0.3-0.5 mV/ms) and the short- duration spike ( 15-25 ms at the base and 8- 15 ms at half- amplitude) was often followed.by a longer, small amplitude negative wave ( 100-300 ms; Fig. 3). DS2 were positive in

the hilus and negative in the molecular layer with a sharp reversal in the inner molecular layer of the dentate gyrus. The voltage gradient across the granule cell layer was 12= 18 mV/mm. The depth profile of DS2 closely matched the laminar distribution of the evoked responses to medial perforant path stimulation. Based on these distinctive features DSl and DS2 could be distinguished in records with multiple-site electrodes, or with at least two electrodes straddling the granule cell layer.

The differences in depth profiles of DSl and DS2 also were reflected by the CSD profiles and maps. Current-source density analysis of DSl revealed an inward current (sink) in the outer third of the molecular layer (Fig. 4). Both the broader base and the superimposed spike (see Fig. 3) showed similar CSD distribution with depth in color-coded CSD maps (not shown). DS2 had a fast sink-source pair, with the sink located in the middle third of the molecular layer. The spatial position of the sink associated with DS2 and perforant path-induced EPSCs of the granule cells were virtually identical. The sinks of DSl and especially of DS2 were coupled with sources in the granule cell layer (Fig. 4). These distinct depth profiles were observed in eight out of eight rats equipped with multisite recording probes.

Both types of dentate spikes were present in every rat


DSI DS2 PP

r

hf

m

g

h

sink source sink

30 0 -30 I I I source sink

FIG. 4. Current source density (CSD) depth profiles of 2 types of dentate spikes (DSl and DS2) and evoked responses to stimulation of the perforant path (PP) inputs. Voltage measurements for CSD analyses were taken at the peak of the dentate spikes and 1 ms after the onset of the evoked responses, respectively. Superimposed curves reflect the stability of current distributions during 5 (DSl and DS2) or 3 (PP) successive averages of 10 events each. Note different sink and source pairs for DSl and DS2 and the similarity between DS2 and PP-evoked responses. Corresponding hippocampal layers are shown on the left. Intracellularly labeled and reconstructed pyramidal cell and granule cell provide further landmarks for the recording positions. o, stratum oriens of CAl; p, pyramidal layer; r, stratum radiatum; hf, hippocampal fissure; m, molecular layer; g, granule cell layer; h, hilar region.

tested. However, DSl occurred from 1.5 to 3 times more frequently than DS2. Although not investigated systemati- cally, it appeared that this ratio increased with the age of the rat. Quantification of this relationship is subject of a separate paper.

Spatial synchrony of dentate spikes

Synchrony between simultaneously recorded events was evaluated in three different ways. First, dentate spikes recorded simultaneously from different sites were averaged using the peak of the dentate spike from a selected site as the trigger. These derived pulses also were used to construct cross-correlograms between dentate spikes and unit activity recorded from different sites. Second, dentate spikes recorded from different sites were averaged by isolated single units ( spike-triggered averages). Third, pulses derived from dis- criminated dentate spikes of different locations were cross- correlated. DSl and DS2 showed similar spatial coherence characteristics and therefore they are discussed together.

Longitudinal coherence of dentate spikes was assessed by placing four to eight electrodes in the dorsal hilus along the longitudinal axis of the hippocampus (n = 6 rats) and by simultaneous recordings from the hili of the dorsal and ventral hippocampus (n = 3 rats). Dentate spikes recorded as far as 1.5mm intervals in the dorsal hilus were essentially synchronous (Fig. 5 ) . Cross-correlograms between field and unit activity were similar when dentate spikes were recorded from the same or different electrode as the units. Comparison of histological location of the electrode tips with electrical activity suggested that lateral or vertical mismatch of the electrodes relative to the center of the hilus contributed sig-

nificantly more to the variability than distance. Peaks of field versus unit cross-correlograms were essentially flat, however, when the events were shuffled.

Dentate spikes in the ventral hippocampus were similar

1

-200 0 msec 200

FIG. 5. Longitudinal synchrony of dentate spikes. A : averages of simultaneously recorded dentate spikes from 5 recording sites (400-ms traces). Distance between electrodes 1 and 5 (shown in 0) was 1.5 mm. B: averaged dentate spikes from the ventral hippocampus. Recording sites a, b, and c are shown in C. D: cross-correlogram of dentate spikes recorded simultaneously from the septal and temporal portions of the hilus in a different rat. The recording positions were similar to sites 2 and b in C. cc, corpus callosum; rf, rhinal fissure.

f-Al 203.2 ms -. 1.

--- ’

Y. M. YRN LANDEGHEM. AND G. BUZS~KI

228.4 ms

time (msec)

-0.5 0 2 mv


in shape and amplitude. Because of the curvature of the hippocampal axis in the temporal region, it was more diffi- cult to position the electrode tips at similar positions. Never- theless, dentate spikes in the ventral hilus were also synchronous (Fig. 5) and were coupled tightly to the discharges of the hilar neurons (not shown). When dentate spikes with twice the amplitude of the background activity were discrim- inated from the dorsal and ventral portions of the hippocampus, their cross-correlograms showed significant peaks at time 0. These findings indicate that neurons located as far as 8 mm in the septal and temporal parts of the hippocampus can discharge within 10 ms during the dentate spike.

The synchrony of dentate spikes in the mediolateral direction was assessed by moving an array of eight electrodes in the coronal plane (~2 = 7 rats). A pair of stationary electrodes was placed in the hilus and the molecular layer. Locally recorded field potentials from each electrode were averaged by the peak of dentate spikes recorded from the hilus. The averages obtained from the stationary electrode in the hilus were used to normalize the amplitude variations at different depths. The electrode array was moved at 100~p/m steps and data were collected from a total of 160-200 locations. Evoked responses and neuronal discharges were used to de- fine the position of the recording electrode array. The resulting averaged dentate spikes were then used to construct two-dimensional voltage maps. As Fig. 6 illustrates, there was little activity outside the dentate region during DSI. The largest amplitude positivity (hot colors) occurred in the granule cell layer surrounded by smaller amplitude negativity (cool colors). The spatial extent of DS2 was similar to DSI, although voltage maps were not constructed.

When electrodes were placed in a symmetrical fashion in both hilar regions, dentate spikes in the two hemispheres occurred virtually synchronously (Fig. 7). Although occasionally lower amplitude (<2 mV) dentate spikes were observed in isolation in one hemisphere, in most cases dentate spikes in the two hemispheres occurred together. Closer in- spection of the original records revealed that small delays (0- 10 ms) were often present in the individual records, but there was a 0 time lag between the peaks of the averaged dentate spike recorded from the opposite hemispheres. Unit activity simultaneously recorded from both hilar regions (n = 11 rats) had similar peaks independent of whether the reference event for the cross-correlation was the ipsilateral or the contralateral dentate spike (Fig. 7).

Correlations between dentate spikes und unit activity e

To reveal the neuronal populations contributing to the dentate spikes, action potentials of physiologically identified single neurons and multiple-unit discharges were cross-correlated with the peak of the dentate spike. Units that were recorded 2200 p,rn below the reversal of the perforant path evoked response were regarded as ‘ ‘hilar cells’ ’ . A subgroup of these units with complex-spike bursts were categorized

as mossy cells (Soltesz et al. 1993 ) or CA3c pyramidal cells (Fox and Ranck 198 1; Ranck 1973). Four of these complex spike cells were putative mossy cells because they were localized to the hilus proper. Nine neurons were identified positively as interneurons, based on their multiple action potential response to perforant path and/or commissural stimulation (e.g., Fig. 7C, inset). Identification of granule cells was not attempted in this study, because of the lack of reliable criteria (Buzsaki and Czeh 1992; Mizumori et al. 1989). Putative granule cells recorded within 100 pm of the reversal of the perforant path response fired very rarely during immobility, thus preventing the construction of interpret- able cross-correlograms.

The relationship of complex spike cells and the remaining hilar cell group to the dentate spike events was different. All hilar cells (n = 75)‘) including the nine positively identified interneurons and hilar multiple-unit groups (n = 50) increased their activity during both DS 1 and DS2. In the experiment shown in Fig. 8, four fast-firing hilar cells were recorded simultaneously along with pyramidal neurons of the CA3b and CA1 regions. In association with DS2, hilar cells showed peaks in their cross-correlograms. In contrast, multiple-unit activity in the CA3b site decreased, whereas unit activity in the CA1 region was not altered in this animal. In two out of nine further cases, however, decreased firing of CA1 pyramidal cells was also evident. DSl or DS2-associated decrease of unit activity was found in 3 out of 14 single or multiple CA3 neurons and one out of the four putative mossy cells. It must be emphasized, that lack of a clear depression in the other cases might be because of the low firing rates of the complex spike neurons. Importantly, complex spike cells never increased their discharges in either CA3 or CA1 regions during the dentate spikes.

To date, we have not identified any cell that discharged only during the peak of DS 1 but not DS2 or vice versa. However, the shape of the dentate spike-versus-unit cross- correlograms were characteristically different for the two events. In association with DS2 the discharge peak was preceded by a depression of firing for 30- 100 ms (Figs. 7-9). In contrast, histograms associated with DS 1 did not show such a depression (Fig. 9).

Spike-triggered averaging of the local unfiltered potentials by the physiologically identified interneurons also revealed a reliable relationship between unit discharges and local field potentials in the hilus. Putative interneurons discharged significantly more frequently during the dentate spikes as evidenced by the spike-triggered field averages. The shape of these averages was often similar to the algebraic sum of averaged DS 1 and DS2, suggesting that the interneurons fired during both events. Shuffled spikes resulted in flat averages.

Relationship between dentute spikes und sharp waves

The two types of dentate spikes appeared mutually exclusive because they virtually never occurred together within a

FIG. 6. Two-dimensional potential maps displayed at different time points of the type I dentate spike (DS 1; bottom r-i~@) extracted from a movie for studying the time-space spread of the field events. The potential maps are superimposed on coronal sections of the dorsal hippocampus at the level of the recording electrodes. Pyramidal and granule cell layers are outlined. Potentials were collected from 160 recording positions (20 lOO+m steps of 8 horizontally spaced electrodes at 300+m intervals). Figures on the maps ctrrespond to the time scale (bottom right). Note dominating positive voltage in and around the granule cell layer.

1698 A. BRAGIN, G. JANDG, Z. NADASDY, M. VAN LANDEGHEM, AND G. BUZSAKI

A right

C . right

ms k 200

B- left vs right

I II II II I I IlLIll I I

-200 ms 200

DI

-200 ms 200

400-ms time window. In addition, SPWs and associated high-frequency (200 Hz) field ripples in the CA1 pyramidal layer were never observed after the dentate spikes ( <200 ms) nor were they followed by DS2 (<200 ms). These observations are based on 12 rats in which SPWs and dentate spike were simultaneously recorded. In contrast, 5 - 15% of DSl were preceded by SPW events. This relationship was revealed by cross-correlograms of DS 1 and CA1 ripples as well as DS 1 and neuronal discharges of CA1 pyramidal neurons, respectively (Fig. 9, B and C). Increased activity of CA1 neurons and associated 200-Hz field ripples occurred 60- 120 ms before the peak of DS 1. Conversely, when CA1 ripples were used as reference events, unit discharges of hilar neurons both preceded and followed them (Fig. 927). Increased hilar neuronal activity before SPW reflected recruitment of these neurons by the CA3 population burst events (Buzsaki 1986; Buzsaki et al. 1983; Scharfman 1995; Strowbridge et al. 1993). The slightly increased neuronal firing after the SPWs likely corresponded to the triggered DSl events. Despite the occasional coupling between SPW and DSl, it must be emphasized that the occurrence of SPW was not a prerequisite for the occurrence of DSl. The above findings also indicate that the hilar network can be synchronized by two distinct mechanisms: from the entorhinal cortex during dentate spikes (feed-forward drive) and from the CA3 region during SPW (feed-back drive).

Effect of colchicine lesion on dentate spikes

Colchicine injection into the dentate gyrus of the dorsal hippocampus resulted i n nearly complete elimination of the

FIG. 7. Bilateral synchrony of the dentate spikes. A: field activity recorded from symmetrical positions of the right and left hili. B : cross-correlogram between dentate spikes (DS2) recorded from the opposite hemispheres. Note most events at around time 0. C: cross-correlogram between DS2 from the right hilus and intemeuronal activity (right inset) recorded from the same electrode. The neuron responded repetitively ( l ) and below the threshold of the population spike to stimulation of the perforant path (t , left inset). D : cross-correlogram between DS2 from the right hilus and multiple unit activity (inset) recorded from the contralateral hilus. Note similar peaks and suppression of unit activity before the peak of the dentate spike (time 0).

granule cells in one rat and partial damage in the remaining three animals. However, both the cresyl violet staining and the Gallyas silver impregnation methods (Gallyas et al. 1993) revealed that not only granule cells but a large portion of the hilar neurons also were damaged by the toxin. The amplitude of dentate spikes, in general, was considerably lower in these rats than in the intact animals. The toxin effect especially was convincing in two cases where the longitudinally placed electrode arrays surveyed areas with and without granule cells. In areas with complete elimination of granule cells, the amplitude of dentate spikes was two- to threefold lower than the simultaneously recorded spikes from areas with intact granule cells. These observations suggested that the currents underlying the dentate spikes are generated mostly by the granule cell population. However, the immediate course of neuronal synchrony could not be determined because of simultaneous damage of the hilar cell population.

Effect waves

of entorhinal lesions on dentate spikes and sharp

Histological analysis of the brains revealed that in five of the seven rats the entorhinal cortex was removed completely (Fig. 10). In all cases, the extent of the lesion was larger than the boundaries of the entorhinal cortex and involved damage to the perirhinal and/or occipital cortical areas, as well. In two cases, a small portion of the medial entorhinal- cortex remained in connection with the angular bundle. Evoked responses from both perforant path and the commissural inputs were decreased on the day of the lesion but both recovered to normal values 24 h later. The threshold of the


DS2 vs CA1 ripple a n=90

H2

f

H2 I I

H3

7r H4

---r 2 ms

CA3

CA1

H4 I I 1111

CA3

CA1

FIG. 8. Simultaneous recording of field and unit activity from the hilus (H), CA3 and CA 1 regions during DS2 ( l , inset). From the hilar recording electrode 4 units could be separated (left column HI -H4: wide band averages). Independence of the units were verified by the refractory periods in the autocorrelograms (not shown). Single traces of multiple units are shown for the CA3 and CA1 recordings sites. Right colu~zn : averaged field potentials (rz = 50) and cross-correlograms (400-ms traces). Time 0: peak of DS2. Note increased activity of the hilar neurons (HI -H4) and decreased activity of CA3 pyramidal cells (arrow) during DS2.

perforant path.response increased progressively day by day, but small amplitude evoked field patterns in response to high intensity stimulation were observed for at least 7 days. In

DSl vs CA1 ripple b n=160

DSl vs CA1 units

I I I

C

1 Spike/bin

DSl vs hilus units d

I n=160

DS2 vs hilus units e I

n=90

ripple vs hilus units

-200 0 msec 200

FIG. 9. Relationship between dentate spikes, hilar neuronal activity and SPW-associated bursts in the CA1 region. A and B : cross-correlograms between the 2 types of dentate spikes (DSl and DS2) and SPW-wave associated high-frequency oscillation (ripple) in the CA1 pyramidal layer. Time 0: peak of dentate spikes. C: DS I vs. multiple unit discharge of CA1 pyramidal neurons. Note that SPW-associated activity in CA1 preceded DSl by - 100 ms. D and E: cross-correlograms between the 2 types of dentate spikes (DSl and DS2) and multiple unit activity in the hilus. Note that unit activity is suppressed before the occurrence of DS2 and increased during both DS 1 and DS2. F: cross-correlogram between SPW-associated ripples recorded from the CA1 pyramidal layer and multiple unit activity in the hilus. The early increase (J ) is due to the SPW-associated population burst of CA3 and hilar region neurons. The late increase of neuronal activity ( N ) is likely due to DS 1 -concurrent unit activity that often followed the SPW bursts. Number of sweep are indicated above the histograms.

the two rats with partial lesions, the perforant path evoked responses survived to the end of the observation period. The commissurally evoked responses also predicted the extent of the lesion. In the intact rat, commissural stimulation evoked a late ( 20-25 ms latency) response in the dentate gyrus (not shown). This late reverberatory potential reflects sequential

1700 A BRAGIN, G. JAND6, Z. NADASDY, M. VAN LANDEGHEM, AND G. BUZSAKI

FIG. IO. Entorhinal cortex lesion. A : ventral view of the brain showing the extent of the lesion, including the whole extent of the entorhinal cortex, part of the perirhinal cortex and part of the occipital cortex. Arrowheads indicate the corpus callosu~n. b, brainstem. H: coronal section of the brain shown in A. C: terminal degeneration of the perforant path fibers in the molecular layer of the dentate gyrus and in the stratuln lacunosum-molecu- lare of CA1 and CA3 regions (arrowheads).

activation of the CA 1 -entorhinal cortex-dentate gyrus circuitry (Deadwyler et al. 1975). This late potential perma- nently disappeared in rats with complete lesions but survived in the two partially damaged rats.

Changes in theta activity and associated gamma waves

after entorhinal cortex damage have been described earlier (Bragin et al. 1995). In contrast to the intact rat, large amplitude fast EEG activity was no longer a characteristic feature of the hilar region. Large interictal spikes (>4 mV) occurred for l-4 days after the lesion in every rat. These interictal spikes were largest on the second postoperative day and their depth profiles and form were virtually identical with the perforant path-evoked responses. Tentatively, we assume that these large transients were brought about by spontaneous synchronized discharges of the perforant path fibers due to the demyelination process. Apart from these large-amplitude sharp spikes during the early postoperative days, dentate spikes were observed only rarely during the postlesion observation period (21 days; Fig. 11). The surviving dentate spikes were small amplitude sharp events (<l mV) and emerged after the first postoperative week (Fig. 11, A and B). However, based on the depth of polarity reversal and shape these were similar to DS 1 (Fig. 11 B) . In the two rats with partial lesion, dentate spikes survived and, despite the large extent of the entorhinal cortex damage, their incidence and amplitude returned to the preoperative levels after 2 wk. These findings suggest that the entorhinal cortex plays a primary role in triggering dentate spike events.

In contrast to the drastic reduction of the dentate spikes, SPWs occurred more frequently after entorhinal cortex damage (Fig. 11, A and C). Not only did the total number of SPWs increase but the incidence of double SPWs (2 events within a 400-ms time window) also increased significantly. The amplitude of the SPWs, however, was not significantly different from the preoperative level. Occasionally, SPWs occurred rhythmically at 3-6 Hz for several seconds. In addition, postlesion SPWs often were followed by a “tail” of 50- 100 Hz oscillation (Fig. I ID).

DISCUSSION

This study revealed the occurrence of short-duration and large-amplitude field transients in the hilar region of the intact rat brain. These field transients, termed dentate spikes, invariably were associated with the synchronous bursting of the hilar-cell population. The dentate spikes and associatecl population burst\ of the putative hilar intcrncurons occurred virtually simultaneously along the long axis of the hippocampus and in the two hemispheres and emerged independently from the SPW-associated bursts of the CA3 region.

Cellular-synaptic generution of dentate spikes

Two types of dentate spikes (DSI and DS2) could be distinguished. They had different voltage-versus-depth profiles and different spatial distribution of their current sinks. After bilateral removal of the entorhinal cortex, dentate spikes were virtually eliminated. Both types of dentate spikes had similar spatial synchrony and were associated with a synchronous population burst of putative dentate interneurons. However, the hilar-region population burst was consis- tently preceded by a decreased discharge probability of hilar interneurons before the occurrence of DS2 but not before DS 1. Frequently, DS 1 appeared as a brief oscillatory event, whereas DS2 was an isolated event followed by a longer negative field event. DS 1 and DS2 appeared to be mutually


B DS before DS 7 days after

40 a 1 dentate spikes

a, 8o Q

60

bef 2 3 4 7 10 1421

bef 7d bef 7d

SPW 7 days after

exclusive, because they very seldom occurred together in a 400-ms time window.

Most of the observed characteristics of DS2 support the hypothesis that they are triggered from a synchronous burst of layer II stellate cells of the medial entorhinal cortex. Based on the voltage-versus-depth profile and a large sink of DS2 in the middle molecular layer, we assume that extracellular negativity at this level reflects synchronous depolarization of granule cells, basket and chandelier cells and hilar interneurons by the medial entorhinal input (Gemroth et al. 1989; Han et al. 1993; McNaughton and Barnes 1977; Steward 1976). Overall, these findings suggest that a dominant source of the extracellular currents underlying DS2 derives from synchronous excitatory postsynaptic currents imping- ing upon the dendrites of granule cells in the middle third of the dentate molecular layer.

Other observations, on the other hand, suggest that the current source observed in the granule cell layer during DS2 is not simply a passive return current. In agreement with previous observations (Buzsaki and Eidelberg 1982)) several putative interneurons responded earlier and at a lower current threshold than the granule cell population

FIG. 1 1. Effects of entorhinal cortex lesion on dentate spikes and SPW bursts. A : frequency of dentate spikes (DS 1 + DS2) and SPWs per minute before and various days after bilateral removal of the entorhinal cortex in a single rat. B: single DS 1 events before and 7 days after the lesion. Only 7 traces are shown. Arrows, polarity reversal of DSl between 3 and 4. Amplitude of the rarely occurring dentate spikes was substantially decreased after the lesion. C: group data for 5 rats with complete bilateral lesion of the entorhinal cortex before (bef) and 7 days (7d) post-lesion. Note the reciprocal relationship between the incidence of dentate spikes and SPWs. D SPWs in the CA1 -dentate gyrus axis recorded 7 days after entorhinal cortex lesion. Such “double” SPWs were very rare before the lesion. Note the high-frequency (50- 100 Hz) oscillatory tail on the descending part of the SPW ( arrows ) .

spike evoked by stimulation of the perforant path. The anatomic basis of this observation is that basket cells and several hilar cell types extend their dendrites into the molecular layer of the dentate gyrus (Amaral 1978; Han et al. 1993; Scharfman 199 1) . These neurons therefore could be driven monosynaptically by layer II stellate cells of the medial entorhinal cortex during DS2. Independent of whether basket cells and hilar interneurons were activated by the perforant path, granule cells, or by other means, their terminals on the somatic/perisomatic region of granule cells likely produced a concerted inhibition of the latter neurons. We may hypothesize, therefore, that at least part of the extracellular current flow, associated with the source in the granule cell layer, arises from synchronously active inhibitory postsynaptic currents (IPSCs) on the granule cells. This conclu- sion is also supported by the survival of lower amplitude dentate spikes after bilateral removal of the dentate gyrus. Further discussion about the role of somatic/perisomatic inhibition in the generation of field events is hampered by the lack of consensus regarding the direction of inhibitory current flow in granule cells. Patch-clamp experiments suggest that y-aminobutyric acid-A (GABAJ -mediated

1702 A. BKAGIN, G. JAND6, Z. NADASDY, M. VAN LANDEGHEM. AND G. BUZSiiKI

IPSCs in granule cells are depolarizing because the resting membrane potential in these cells is considerably more negative than the chloride equilibrium potential (Soltksz and Mody 1994; Staley et al. 1992). In experiments with sharp electrodes, on the other hand, presynaptic activation of identified single interneurons resulted in short-latency hyperpolarization of the granule cells from the resting membrane potential (Buhl et al. 1994). Because similar measurements are not yet available in vivo, these in vitro experiments do not directly support or refute the interpretation of the currents underlying DS2. We conclude, therefore, that the major electromotive forces underlying DS2 corre- spond to active inward and outward synaptic currents at the dendrites and somata of granule cells, respectively. Because these active currents are segregated spatially, their summa- tion in the extracellular space should give rise to a large extracellular field.

One apparent aspect of DS2 that the above reasoning fails to explain is the prominent decrease of the firing rate of hilar cells 30- 100 ms before the peak of the field DS2. This transient decrease of discharge rate in hilar interneurons may be because of a subcortically mediated inhibitory process that precedes the synchronous population burst in layer II neurons of the entorhinal cortex. A candidate mechanism for such subcortical inhibition of hilar and possibly entorhinal cortex interneurons is the septal GABAergic projection on hilar and possibly entorhinal cortex interneurons (Freund and Antal 1988; T. F. Freund, personal communication). An experimental verification of this hypothesis will require simultaneous recordings of dentate spikes and medial septal neurons. Alternatively, population bursts in the entorhinal cortex may emerge from a rebound of synchronous hyperpolarization of layer 11 stellate cells. In the latter case, the decreased firing of hilar cells precedin, (7 DS2 is a disfaciliation event due to a transient decrease of the entorhinal drive. Although the cellular basis of such hyperpolarization-rebound bursts in the entorhinal cortex has yet to be demonstrated, such intrinsic mechanisms are known to form the basis of network neuronal bursts in other parts of the brain (Llinis 1988; Steriade et al. 1993).

The cellular-synaptic generation of DS 1 appears different from that of DS2. Typically, DS 1 was a brief oscillatory event with one large and two to four small surrounding waves at - 100 Hz. The voltage-versus-depth profiles and CSD findings as well as the entorhinal cortex lesion experiment are compatible with the hypothesis that a major part of the currents underlying DS 1 are generated by the lateral entorhinal cortical synapses in the outer molecular layer (McNaughton and Barnes 1977; Steward 1976). Acceptance of the suggestion that the field events in the dentate gyrus underlying DSl reflect depolarization of the distal dendrites of granule cells, brought about by the population burst events of the entorhinal cortex, carries the explicit assumption that the intrinsic and/or circuit properties of layer II neurons are different in the medial and lateral entorhinal cortex, because the wave forms and unit correlates of DS 1 and DS2 were different.

The importance of extrahippocampal drive in the triggering of dentate spikes is supported further by observations that sharp transients in the dentate hilar region, likely analogous to the dentate spikes. were reduced by

damaging the amygdala in the cat (Par6 et al. 1994). The role of the entorhinal input in the genesis of dentate spikes also may be interpreted from a different view- point. It may be assumed that population bursts in the entorhinal cortex served merely as a trigger for initiating the synchronous discharge of the hilar network. In hippocampal slices, bath application of 4-aminopyridine in- duces burst discharges in hilar neurons even when N- methyl-u-aspartate and cr-amino-3-hydroxy-5methyl-4- isoxazolepro-pionic acid-mediated excitatory neuro- transmission is blocked pharmacologically (Michelson and Wong 199 1; Muller and Misgeld 1990; Sol&z and Mody 1994). Parallel to bursting of hilar cells, synchronously occurring giant IPSPs are seen in granule cells and CA3 pyramidal cells accompanied by positive field potentials. These in vitro observations are compatible with the dentate spike-concurrent large positive field potentials in the granule cell layer, synchronous discharges of hilar-region interneurons and inhibition of CA3c pyramidal cells (Michelson and Wong 199 1 ). Both sets of findings may be explained by the hypothesis that population synchrony of inhibitory cells in the hilar region is due to depolarizing responses mediated by GABA receptors on inhibitory interneurons (Michelson and Wong 199 1 ). An alternative mechanism for the fast recruitment of hilar cells during dentate spikes is a temporary increase of the efYicacy of gap junctions. Katsumaru et al. ( 1988) have demonstrated that the majority of hilar- region interneurons, but not other cell types, are con- nected through gap junctions. Because the efficacy of gap junctions is modulated by pH, temperature (Church and Baimbridge 199 1 ) and possibly neurotransmitters, such a mechanism could provide an efficient and rapid means for the synchronization of interneurons during dentate spikes.

Individual dentate spikes could emerge from virtually any segment of the dentate region as evidenced by the latency and amplitude differences of simultaneously recorded single events. However, averages of field events and units histograms recorded from different locations had 0 time lags. When recordings were carried out from the septal and temporal ends of the hilar region many dentate spikes occurred synchronously. Spatial synchrony of the dentate spikes may be explained by the widespread projection of the perforant path to the dentate area (Amaral and Witter 1989) and/or by our recent observation that certain types of hilar interneurons have extremely large axonal arbors covering longer than one-third of the longitudinal extent of the dentate gyrus (Sik et al. 1994).

In the absence of theta activity, two kinds of irregular burst events occur in the intact hippocampus: SPWs and dentate spikes. SPWs are initiated in the CA3 region and invade the CA 1 region and retrogradely the dentate networks (Buzs;iki 198(i), whereas dentate spikes are initiated in the hilar regions and tend to suppress the emergence of SPW bursts.

A small percentage ( <20% ) of DS 1 were preceded by SPW-associated population bursts but such relationship was


not observed with DS2. Because layer V neurons of the medial entorhinal cells discharge synchronously with hippocampal SPW but layer II cells do not alter their firing patterns (Chrobak and Buzsaki 1994)) we have to assume either that the physiological connectivity between layers V and II connections are different in the medial and lateral parts of the entorhinal cortex or that the entorhinal cortex in not involved in SPW-triggered DS 1 events. An alternative expla- nation for the SPW-triggered DSl events is based on the population dynamics of the CA3-hilar region network. After the SPW-burst, CA3 pyramidal cells become silent due to the burst-induced long-lasting hyperpolarization in the pyramidal cells and to recurrent inhibition. Such transiently reduced excitation of hilar interneurons by the CA3 pyramidal cells and hilar mossy cells (Li et al. 1994; Scharfman 1994) may create conditions favorable for hilar network synchrony (disfacilitation induced rebound).

Synchronous discharge of a large number of hilar neurons within a narrow time window is likely to be an important event for the operation of the hippocampus. Full synchrony of these neurons is dependent on the driv- ing force of the entorhinal cortex. On the other hand, dentate spikes were never followed by SPWs or by an increased discharge of CA3 and CA1 pyramidal cells. Instead, dentate spikes tended to suppress pyramidal cell activity in the CA3 region. Such a scenario does not support the generally held view of the hippocampal circuitry as a series of unidirectionally excited groups of cells from the entorhinal cortex to dentate granule cells -+ CA3 + CA1 pyramidal cells and subicular neurons (Amaral and Witter 1989; Andersen et al. 197 1 ). Instead, it suggests that during the dentate spikes the net output of the dentate gyrus to the CA3 region in the intact brain is mainly inhibitory. This view is further supported by the increased excitability of the CA3-CA1 network, as reflected by the higher incidence of SPWs, after entorhinal cortex lesion. Suppression of the recurrent CA3 network can be accomplished by feed-forward excitation of interneurons in the CA3 region (Frotscher 1989) or by a direct inhibitory action of hilar interneurons on the CA3 pyramidal cells. The anatomic substrate for the latter possibility has yet to be demonstrated. However, in line with such a suggestion, we have recently recorded intracellularly from a hilar chandelier cell, which fired bursts of action potentials during dentate spikes. Reconstruc- tion of the in vivo labeled cell revealed extensive axonal arborization in the fascia dentata, hilus and the CA3c region (Sik et al. 1994).

Based on the relationship between dentate spikes and SPWs, it may be suggested that a possible physiological function of dentate spikes and associated population bursts of hilar interneurons is to delay the occurrence of SPW- concurrent network bursts in the CA3-CA 1 -subiculum-layer V entorhinal cortex circuitry. Thus dentate spikes may be conceived as a “disable” signal which can prevent the occurrence of the powerful feedback from the hippocampus to the neocortex by the entorhinal cortex. The latter mechanism has been postulated to play a critical role in transferring information from the hippocampus to neocortical areas (Chrobak and Buzsaki 1994).

Dentate spikes and interictal spikes

Increased synchrony of the dentate cell population may lead to epileptic interictal spikes (Michelson and Wong 1991; Muller and Misgeld 1990, 199 1; Scharfman and Schwartzkroin 1990). Indeed, the polarity and form of the dentate spikes are similar to type 2 interictal spikes observed in several epilepsy models (Buzsaki et al. 1989b, 1991; Fujita et al. 1983; Wadman et al. 1983), whereas type 1 interictal spikes represent excessive recruitment of CA3 and CA1 pyramidal neurons (Buzsaki et al. 1983, 1989b; Wong and Traub 1983). The involvement of the different neuronal types (hilar neurons and pyramidal cells, respectively) and their opposite target effects (inhibition and excitation, respectively) in these population events may explain why different categories of interictal spikes suppress or promote epileptic seizures (Engel 1989; Stevens et al. 1972). Experimental support for this hypothesis will require identification of interictal events as SPW-like (type 1) or dentate spike-like (type 2) and correlation of their incidence with the occurrence of after- discharges.

Conclusions

In the intact rat large amplitude, positive field potentials are present in the dentate gyrus in the absence of theta waves. The entorhinal cortex is involved critically in triggering dentate spikes and synchronous discharges of layer II stellate cells in the lateral and medial entorhinal cortex are hypothesized to be the immediate course of two types of dentate spikes. At least part of the extracellular current flow underlying these intermittent field events is due to synchronous inward currents to the granule cell dendrites. Dentate spikes are invariably associated with a burst discharge of hilar interneurons. In the immobile rat and in slow-wave sleep, dentate spikes provide a physiological means to suppress the excitability of the CA3-CA1 network.

We thank K. Wise and J. Hetke for supplying the silicon probes, J. J. Chrobak, J. Goebel, I. Mody, and I. Soltesz for comments on the manuscript; and T. F. Freund, H. E. Scharfman, and X.-J. Wang for discussions.

This work was supported by National Institute of Health, Human Frontier Science Program, and the Whitehall Foundation.

Permanent addresses: A. Bragin, Institute of Experimental and Theoreti- cal Biophysics, Puschino, Russia; G. Jando, Department of Physiology, Medical School, P&s, Hungary; M. van Landeghem, Institute of Physiology 11, Heinrich Heine University, Dusseldorf, Germany.

Address for reprint requests: G. Buzsaki, Center for Molecular and Be- havioral Neurosciences, Rutgers University, 197 University Ave., Newark, NJ 07102.

Received 26 May 1994; accepted in final form 6 December 1994.

REFERENCES

AMARAL, D. G. A Golgi study of cell types in the hilar region of the hippocampus in the rat. J. Cornp. Neural. 182: 85 l-9 14, 1978.

AMARAL, D. AND WITTER, M. The three-dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience 3 1: 571-591, 1989.

ANDERSEN, P., BLISS, T. V. P., AND SKREDE, K. K. Lamellar organization of hippocampal excitatory pathways. Exp. Brain Res. 13: 222- 238, 1971.

1704 A. BRAGIN, G. JANDi), 2. NADASDY, M. VAN LANDEGHEM, AND G. BUZSAKI

BLAND, B. H. Physiology and pharmacology of hippocampal formation theta rhythms. Prog. New-obiol 0 26: l-54, 1990.

BRAGIN, A., JAND~, G., NADASDY, Z., HETKE, J., WISE, K., AND BUZSAKI, G. Gamma (40- 100 Hz) oscillation in the hippocampus of the behaving rat. J. New-ox?. 15: 47-60, 1995.

BRANKACK, J., STEWART, M., AND Fox, S. E. Current source density analysis of the hippocampal theta rhythm: associated sustained potentials and candidate synaptic generators. Brain Res . 6 15: 3 1 O-327, 1993.

BUHL, E., HALASY, K., AND SOMOGYI, P. Hippocampal unitary inhibitory postsynaptic potentials: diverse sources and number of synaptic release sites. Nature Lond. 368: 823-828, 1994.

BUZS,~KI, G. Hippocampal sharp waves: their origin and significance. Brain Rex 398: 242-252,1986.

BUZS~,KI, G. AND CZ~H, G. Commissural and perforant path interactions in the rat hippocampus: field potentials and unitary activity. Exp. Brain Res. 225: 234-247, 1981.

BUZSAKI, G. AND CZ~H, G. Physiological function of granule cells: a hypothesis. In: The Dentate Cyrus and its Role in Seizures, edited by C. E. Ribak, C. M. Gall, and I. Mody. Amsterdam: Elsevier, 1992, p. 28 l-290.

BUZSIL~KI, G. AND EIDELBERG, E. Direct afferent excitation and long-term potentiation of hippocampal interneurons. J. Neurophysiol. 48: 597 -

607,1982. BUZS~~KI, G., BICKFORD, R. G., RYAN, L. J., YOUNG, S., PROHASKA, O.,

MANDEL, R. J., AND GAGE, F. H. Multisite recording of brain field potentials and unit activity in freely moving rats. J. Neurosci. Methods 28: 209-217, 1989a.

BUZS~~KI, G., HORVATH, Z., URIOSTE, R., HETKE, J., AND WISE, K. High- frequency network oscillation in the hippocampus. Science Wash. DC 256: 1025-1027, 1992.

BUZSAKI, G., LEUNG, L., AND VANDERWOLF, C. H. Cellular bases of

hippocampal EEG in the behaving rat. Brain Res. Rev. 6: 139- 171, 1983.

BUZSAKI, G., PONOMAREFF, L. G., BAYARDO, F., RUIZ, R., AND GAGE, F. H. Neuronal activity in the subcortically denervated hippocampus: a chronic model for epilepsy. Neuroscience 28: 527-538, 1989b.

BUZSAKI, G., Hsu, M., SLAMKA, C., GAGE, F. H., AND HORVATH, Z. Emer- gence and propagation of interictal spikes in the subcortically denervated hippocampus. Hippocampus 1: 163 - 180, 199 1.

BUZSAKI, G., BRAGIN, A., CHROBAK, J. J., NADASDY, Z., SIK, A., Hsu, M., AND YLINEN, A. Oscillatory and intermittent synchrony in the hippocampus: relevance to memory trace formation. In: Temporal Coding in the Brain, edited by G. Buzsski, R. Llinhs, W. Singer, A. Berthoz, and Y. Christen. Berlin: Springer-Verlag 1994, p. 146- 172.

CHROBAK, J. J. AND BUZS~~KI, G. Selective activation of deep layer retrohip- pocampal neurons during hippocampal sharp waves. J. Neurosci. 14: 6160-6170, 1994.

CHIJRCH, J. AND BAIMBRIDGE, G. K. Exposure to high-pH increases the incidence and extent of dye coupling between rat hippocampus CA 1 pyramidal neurons in vitro. Neuroscience I I : 3289-3295, 199 I.

DI~AI~WYI~, S. A., Wr:sr, J. R., C~TMAN, C. W., AND LYNC‘II, G. Physiolog- ical studies of the reciprocal connections between the hippocampus and

entorhinal cortex. Exp. Neural. 49: 35-57, 1975. DE CURTIS, M., PAR& D., AND LLINAS, R. The electrophysiology of the

olfactory-hippocampal circuit in the isolated and perfused adult mammalian brain in vitro. Hippocampus 1: 34 l-354, 199 1.

ENGEL, J., JR. Seizures and epilepsy. Philadelphia, PA: F. A. Davis, 1989. Fox, S. E. AND RANCK, J. B ., JR. Electrophysiological characteristics of

hippocampal complex-spike cells and theta cells. Exp. Brain Res. 41: 299-313, 1981.

FREEMAN, J. A. AND NICHOLSON, C. Experimental optimization of current source-density technique for anuran cerebellum. J. Neurophysiol. 38: 369-382, 1975.

FREUND, T. F. AND ANTAL, M. GABA-containing neurons in the septum control inhibitory interneurons in the hippocampus. Nature Lond. 336:

170-173, 1988. FROTCHER, M. Mossy fiber synapses on glutamate decarboxylase-immuno-

reactive neurons: evidence for feed-forward inhibition in the CA3 region of the hippocampus. Exp. Brain Res. 75: 441-445, 1989.

FUJITA, Y., HARADA, H., TAKEUCHI, T., SATO, H., AND MINAMI, S. En- hancement of EEG spikes and hyperpolarizations of pyramidal cells in the kindled hippocampus of the rabbit. Jpn. J. Physiol. 33: 227-

238, 1983. GALLYAS, F., Hsu, M., AND BUZSAKI, G. Four modified silver methods for

thick sections of formaldehyde-fixed mammalian central nervous tissue: dark neurons, perikarya of all cells, microglial cells and capillaries. J. Neurosci. Methods 50: 159- 164, 1993.

GERMROTH, P., SCHWERDTFEGER, W. K., AND BUHL, E. H. Morphology of identified entorhinal neurons projecting to the hippocampus. A light microscopical study combining retrograde tracing and intracellular injec-

tion. Neuroscience 30: 683 -69 1, 1989. GRASTYAN, E., LISSAK, K., MADAR&Z, I., AND DONHOFFER, H. Hippocam-

pal electrical activity during the development of conditioned reflexes. Electroencephal. Clin. Neurophysiol . I I : 409-430, 1959.

HAN, Z.-S., BUHL, E. H., L~RINCZI, Z., AND SOMOGYI, P. A high degree of spatial selectivity in the axonal and dendritic domains of physiologically identified local-circuit neurons in the dentate gyrus of the rat hippocam-

pus. Eur. J. Neurosci. 5: 395-410, 1993. HOLSHEIMER, J. Electrical conductivity of the hippocampal CA1 layers and

application to current source density analysis. Exp. Brain Res. 67: 402- 410, 1987.

KATSUMARU, H., KOSAKA, T., HEIZMAN, C. W., AND HAMA, K. Gap-junctions on GABAergic neurons containing the calcium-bindng protein par-

valbumin in the rat hippocampus (CA1 regions). Exp. Brain Res. 72: 363-370, 1988.

LEUNG, L. S. Potentials evoked by alvear tract in hippocampal CA 1 region of rats. I. Topographical projection, component analysis, and correlation with unit activities. J. Neurophysiol. 42: 1557- 1570, 1979.

LI, X.-G., SOMOGYI, P., YLINEN, A., AND BUZSAKI, G. The hippocampal CA3 network: an in vivo intracellular labeling study. J. Comp. Neural. 339: 18 l-208, 1994.

LLINAS, R. R. The intrinsic electrophysiological properties of mammalian neurons: insight into central nervous system function. Science Wash. DC 242: 1654-1664, 1988.

LOPES DA SILVA, F. H., WITTER, M., BOEIJINGA, P. H., AND LOHMAN A.

Anatomic organization and physiology of the limbic cortex. Physiol. Rev. 70: 453-511, 1990.

MCNAUGHTON, B. L. AND BARNES, C. A. D. Physiological identification

and analysis of dentate granule cell responses to stimulation of the medial and lateral perforant pathways in the rat. J. Comp. Neural. 175: 439- 454, 1977.

MICHELSON, H. B. AND WONG, R. K. S. Excitatory synaptic responses mediated by GABA receptors in the hippocampus. Science Wash. DC 253: 1420-1423, 1991.

MITZDORF, U. Current source-density method and application in cat cerebral

cortex: investigation of evoked potentials and EEG phenomena. Physiol. Rev. 65: 37-100, 1985.

MIZUMORI, S. J. Y., MCNAUGHTON, B. L., BARNES, C. A., AND Fox, K. B. Preserved spatial coding in hippocampal CA1 pyramidal cells during

reversible suppression of CA3 output: evidence for pattern completion in hippocampus. J. Neurosci. 9: 39 15-3928, 1989.

MIJLLER, W. AND Mrsmm, U. Inhibitory role of dentate hilus neurons in

guinea pig hippocampal slice. J. Neurophysiol. 64: 46-56, 1990. MIJLLER, W. AND MISGEI.~), IJ. Picrotoxin- and 4-aminopyridine-induced

activity in hilar neurons in the guinea pig hippocampal slice. ./. Neuro-

physiol. 65: 141-147, 1991. O’KEEFE, J. AND NADEL, L. The Hippocampus as a Cognitive Map in

Oxford, UK: Clarendon, 1978. PAR& D., DONG, J., AND GAUDREAU, H. Amygdalo-entorhinal relations

and their reflection in the hippocampal formation: generation of sharp potentials. J. Neurosci. In press.

RANCK, J. B., JR. Studies on single neurons in dorsal hippocampal formation and septum in unrestrained rats. I. Behavioral correlates and firing reper-

toires. Exp. Neurol. 42: 461-531, 1973. SCHARFMAN, H. E. Dentate hilar cells with dendrites in the molecular layer

have lower thresholds for synaptic activation by perforant path than granule cells. J. Neurosci. 11: 1660- 1673, 1991.

SCHARFMAN, H. E. Synchronization of area CA3 hippocampal pyramidal cells and non-granule cells of the dentate gyrus in bicucullin-treated rat hippocampal slices. Neuroscience 59: 245 -257, 1994.

SCHARFMAN, H. E. AND SCHWARTZKROIN, P. A. Consequences of prolonged afferent stimulation of the rat fascia dentata: epileptiform activity in area CA3 of hippocampus. Neuroscience 35: 505-5 17, 1990.

SIK, A., YLINEN, A., PENTTONEN, M., AND BUZSAKI, G. Hippocampal

interneurons: an in vivo intracellular study. Sot. Neurosci. Abstr. 148.15.

SOLTI~SZ, I., BOURASSA, J., AND DESCHI~NES, M. The behavior of mossy cells


of the rat dentate gyrus during theta oscillations in vivo. Neuroscience 57: 555-564, 1993.

SOLT~SZ, I. AND MODY, I. Patch-clamp recordings reveal powerful GABA- ergic inhibition in dentate hilar neurons. J. Neurosci. 14: 2365-2376, 1994.

STALEY, K. J., OTIS, T. S., AND MODY, I. Membrane properties of dentate gyrus granule cells: comparison of sharp microelectrode and whole-cell patch recordings. J. Neurophysiol. 68: 1346- 1358, 1992.

STERIADE, M., MCCORMICK, D. A., AND SEJNOWSKI, T. J. Thalamocortical oscillations in the sleeping and aroused brain. Science Wash. DC 262: 679-685, 1993.

STEVENS, J. R., LONSBURY B. L., AND GOEL, S. L. Seizure occurrence and interspike interval. Arch. Neurol. 26: 409-4 19, 1972.

STEWARD, 0. Topographic organization of the projections from the entorhinal area to the hippocampal formation of the rat. J. Comp. Neurol. 167: 285-314, 1976.

STROWBRIDGE, B. W., ANDERSON, N. L., AND SCWARTZKROIN, P. A. Ana- tomical evidence for reciprocal connections between CA3 pyramidal cells and mossy cells. Sot. Neurosci. Abstr. 23: 147.7, 1993.

SUZUKI, S. S. AND SMITH, G. K. Spontaneous EEG spikes in the normal hippocampus. I. Behavioral correlates, laminar profiles and bilateral synchrony. Electroencephal. Clin. Neurophysiol. 67: 348-359, 1987.

TRAUB, R. D. AND MILES, R. Neuronal Networks of the Hippocampus. Cambridge: Cambridge University Press, 199 1.

VANDERWOLF, C. H. Hippocampal electrical activity and voluntary movement in the rat. Electroencephal. Clin. Neurophysiol. 26: 407- 418, 1969.

WADMAN, W. J., LOPES DA SILVA, F. H., AND LEUNG, L. S. Two types of interictal transients of reversed polarity in rat hippocampus during kin- dling. Electroencephal. Clin. Neurophysiol . 55: 3 14- 3 19, 1983.

WISE, K. D. AND NAJAFI, K. Microfabrication techniques for integrated sensors and microsystems. Science Wash. DC 254: 1335- 1342, 199 1.

WONG, R. K. S. AND TRAUB, R. D. Synchronized burst discharge in disinhib- ited hippocampal slice. I. Initiation in CA2-CA3 region. J. Neurophysiol. 49: 442-458, 1983.

YANG, X. A totally automated system for detection and classification of neural spikes. IEEE Trans. Biomed. Eng. 35: 806-816, 1988.

YLINEN, A. SIK, A., BRAGIN, A., JAND~, G., AND BUZSAKI, G. Intracellular correlates of hippocampal sharp wave bursts in vivo. J. Neurosci. 15: 30-45, 1995.

YLINEN, A., SOLT~SZ, I., BRAGIN, A., PENTTONEN, M., STK, A., AND

BUZSAKI, G. Intracellular correlates of hippocampal theta rhythm in identified pyramidal cells, granule cells, and basket cells. Hippocam- pus. In press.

Dentate EEG Spikes and Associated ... - Buzsaki Lab

Documents

Transcript of Dentate EEG Spikes and Associated ... - Buzsaki Lab