JN AJP: Heart and Circulatory Physiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J Neurophysiol 78: 2602-2615, 1997;
0022-3077/97 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (14)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Chapman, C. A.
Right arrow Articles by Racine, R. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chapman, C. A.
Right arrow Articles by Racine, R. J.

The Journal of Neurophysiology Vol. 78 No. 5 November 1997, pp. 2602-2615
Copyright ©1997 by the American Physiological Society

Converging Inputs to the Entorhinal Cortex From the Piriform Cortex and Medial Septum: Facilitation and Current Source Density Analysis

C. Andrew Chapman and Ronald J. Racine

Department of Psychology, McMaster University, Hamilton, Ontario L8S 4K1 Canada

    ABSTRACT
Abstract
Introduction
Methods
Results
Discussion
References

Chapman, C. Andrew and Ronald J. Racine. Converging inputs to the entorhinal cortex from the piriform cortex and medial septum: facilitation and current source density analysis. J. Neurophysiol. 78: 2602-2615, 1997. The entorhinal cortex receives sensory inputs from the piriform cortex and modulatory inputs from the medial septum. To examine short-term synaptic facilitation effects in these pathways, current source density (CSD) analysis was used first to localize the entorhinal cortex membrane currents, which generate field potentials evoked by stimulation of these afferents. Field potentials were recorded at 50-µm intervals through the medial entorhinal cortex in urethan-anesthetized rats and the one-dimensional CSD was calculated. Piriform cortex stimulation evoked a surface-negative, deep-positive field potential component in the entorhinal cortex with mean onset and peak latencies of 10.4 and 18.4 ms. The component followed brief 100-Hz stimulation, consistent with a monosynaptic response. CSD analysis linked the component to a current sink, which often began in layer I before peaking in layer II. A later, surface-positive field potential component peaked at latencies near 45 ms and was associated with a current source in layer II. Medial septal stimulation evoked positive and negative field potential components which peaked at latencies near 7 and 16 ms, respectively. A weaker and more prolonged surface-negative, deep-positive component peaked at latencies near 25 ms. The early components were generated by currents in the hippocampal formation, and the late surface-negative component was generated by currents in layers II to IV of the entorhinal cortex. Short-term facilitation effects in conscious animals were examined using electrodes chronically implanted near layer II of the entorhinal cortex. Paired-pulse stimulation of the piriform cortex at interpulse intervals of 30 and 40 ms caused the largest facilitation (248%) of responses evoked by the second pulse. Responses evoked by medial septal stimulation also were facilitated maximally (59%) by a piriform cortex conditioning pulse delivered 30-40 ms earlier. Paired pulse stimulation of the medial septum caused the largest facilitation (149%) at intervals of 70 ms, but piriform cortex evoked responses were facilitated maximally (46%) by a septal conditioning pulse 100-200 ms earlier. Frequency potentiation effects were maximal during 12- to 18-Hz stimulation of either the piriform cortex or medial septum. Occlusion tests suggested that piriform cortex and medial septal efferents activate the same neurons. The CSD analysis results show that evoked field potential methods can be used effectively in chronically prepared animals to examine synaptic responses in the converging inputs from the piriform cortex and medial septum to the entorhinal cortex. The short-term potentiation phenomena observed here suggest that low-frequency activity in these pathways during endogenous oscillatory states may enhance entorhinal cortex responsivity to olfactory inputs.

    INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References

The entorhinal cortex receives inputs from many cortical regions, including a prominent projection from layer II and III cells of the piriform, or primary olfactory, cortex, which terminates in the superficial layers (Boeijinga and Van Groen 1984; Kretteck and Price 1977; Luskin and Price 1983; Room et al. 1984; Witter et al. 1989). Layer II and III cells of the entorhinal cortex provide the hippocampal formation with much of its cortical sensory input (Germroth et al. 1989; Steward and Scoville 1976), but intracellular recordings reveal high levels of gamma -aminobutyric acid (GABA)-mediated inhibition, which limits activity in these entorhinal cortical neurons (Colino and Fernandez de Molina 1986; Finch and Babb 1980; Finch et al. 1988; Jones 1993; Jones and Heinemann 1991).

Postsynaptic responses in the entorhinal cortex are enhanced during repetitive low-frequency stimulation of efferents from the amygdala, subiculum, and hippocampus (Deadwyler et al. 1975; Finch et al. 1986; Jones 1987; Jones and Heinemann 1991; Racine and Milgram 1983), suggesting that the frequency of afferent input may play a critical role in activating neurons in this site (Jones 1993, 1995). The effective frequencies range from 2.5 to ~20 Hz, and these enhancements, sometimes referred to as frequency potentiation or facilitation (Andersen and Lomo 1967), may reflect mechanisms that are also active during endogenous oscillatory states. Electroencephalographic (EEG) activity in the olfactory system during periods of odor sampling is characterized by rhythmic oscillations in the beta (15-35 Hz) and gamma (35-80) frequency bands (Boeijinga and Lopes da Silva 1988; Bressler 1984; Freeman 1978). The cholinergic and noncholinergic projections from the medial septum to layers II-IV of the entorhinal cortex (Alonso and Kohler 1984; Gaykema et al. 1990; Insausti et al. 1987) also may govern temporal patterns of activation in the entorhinal cortex. These inputs contribute to the local generation of entorhinal cortex theta frequency (4-12 Hz) activity, which is coherent with the hippocampal theta rhythm (Alonso and Garcia-Austt 1987a,b; Bland and Colom 1993; Dickson et al. 1994; Mitchell and Ranck 1980; Mitchell et al. 1982).

The inhibitory and facilitatory mechanisms that govern the responsiveness of the entorhinal cortex to piriform cortex inputs have not been well described in behaving animals under normal levels of inhibition and neuromodulatory input. Further, although the medial septum is known to powerfully enhance hippocampal responses to entorhinal cortex inputs (Fantie and Goddard 1982; Robinson and Racine 1986), the effect of the medial septum on entorhinal cortex responsivity to sensory afferents has not been investigated.

Piriform cortex stimulation results in the depolarization of entorhinal cortex layer II neurons and an associated surface-negative field potential that reverses near layer II in the cat (Boeijinga and Van Groen 1984; Van Groen et al. 1987), the rat, and the isolated guinea pig brain in vitro (Alonso et al. 1990; de Curtis and Llinas 1993; de Curtis et al. 1991). In the cat, stimulation of posterior piriform cortex sites results in a surface-positive field potential, which also reverses near layer II and is associated with layer II unit responses but which is attributable to synaptic activation of dendritic processes in layer III (Boeijinga and Van Groen 1984).

Current source density (CSD) analysis, a method for spatially localizing membrane currents resulting from synchronous synaptic activation in laminar cortical tissue (Freeman and Nicholson 1975; Haberly and Shepherd 1973; Mitzdorf 1985; Nicholson and Freeman 1975), has been used to identify the entorhinal cortex layer I and II current sinks resulting from olfactory bulb and piriform cortex stimulation in the cat and guinea pig (de Curtis et al. 1991; Van Groen et al. 1987). A definitive localization of evoked membrane currents also is needed for piriform cortex and septal inputs in the rat because of the usefulness of this preparation for the study of olfactory processing and synaptic plasticity.

We have applied CSD analysis techniques to urethan-anesthetized rats to localize the evoked membrane currents activated by each of these inputs and to identify the associated field potential components. These experiments allowed the interpretation of field potentials subsequently obtained in studies of short-term facilitation and frequency potentiation in chronically prepared animals. Paired-pulse tests, in which inputs were stimulated with two pulses at a variable interval, allowed the net effect of facilitatory and inhibitory mechanisms evoked by the first pulse to be measured by monitoring the changes in the size of the response to the second pulse. Delivering the pulses to separate sites allowed heterosynaptic interactions between septal and piriform cortex inputs to be assessed. Some of these results have been reported in abstract form (Chapman and Racine 1995).

    METHODS
Abstract
Introduction
Methods
Results
Discussion
References

Surgery

ACUTE PREPARATIONS. Male Long Evans hooded rats (340-460 g) were anesthetized with urethan (1.5 g/kg ip) and placed in a stereotaxic frame with bregma and lambda on the horizontal plane. The temperature of the animal was monitored with a rectal probe and kept within 36.5-37.5°C with a heating lamp. Bipolar twisted wire Teflon-coated stainless steel stimulating electrodes (125 µm exposed tips) were lowered relative to bregma into the right piriform cortex (P 3.6 mm, L 6.5 mm, and V 8.5-9.0) and medial septum (A 0.2 mm, L 0-0.2 mm, and V 6.0 mm). One tip of each bipolar stimulating electrode extended 0.5 mm beyond the other. The stereotaxic instrument was grounded, and a stainless steel jeweller's screw placed above the left frontal cortex served as a reference electrode.

Monopolar recording electrodes were stainless steel 00 insect pins insulated with Epoxylite up to a flat exposed tip 40-60 µm in diameter. In three animals, Teflon-coated tungsten electrodes cut blunt to a diameter of 50-80 µm were used. To record field potentials at multiple intervals in the plane perpendicular to the surface of the medial entorhinal cortex, the electrodes were placed in a mechanical micromanipulator (Narishige) and advanced on the saggital plane 5.2 mm lateral to the midline at an angle 40° above horizontal. This angle was used by Mitchell and Ranck (1980) and by Alonso and Garcia-Austt (1987a) to examine theta-frequency activity across entorhinal cortex lamina. Coordinates for the surface of the entorhinal cortex were 0.2 mm anterior and 2.6 mm dorsal to the interaural line. Warmed mineral oil was applied at the site of electrode penetration to prevent drying of the exposed cortex. After depth profile recordings were completed, anodal current (50 µA) was passed through the recording electrode for 15 s, and the location of the resulting iron deposits were later identified using the Prussian blue reaction to confirm the recording depths.

CHRONIC PREPARATIONS. Male Long Evans hooded rats (320-460 g) were anesthetized with 0.9 mg/kg ketamine and 0.05 mg/kg xylazine and placed in a stereotaxic frame with the skull surface on the horizontal plane. The level of anesthesia was monitored closely, and 10-20% supplemental doses were administered as required. Bipolar electrodes identical to those used in acute experiments were implanted in the right piriform cortex and medial septum using the coordinates cited above and in the medial entorhinal cortex (P 8.8 mm, L 5.0 mm to bregma, and 0.1 to 0.2 mm above the ventral skull). The vertical placements of the stimulating electrodes were adjusted to minimize current thresholds for evoked field potentials, and the vertical placement of the entorhinal cortex electrode then was adjusted to maximize monopolarly recorded field potential amplitudes. The entorhinal cortex field potentials were recorded both bipolarly and monopolarly through each pole of the bipolar electrode. Electrode leads were connected to gold-plated Amphenol pins and mounted in a 9-pin connector. The assembly was embedded in dental cement and anchored to the skull with stainless steel jeweller's screws. One screw in the contralateral frontal bone served as a ground and reference electrode for monopolar recordings. Animals were housed individually in 18 × 20 × 35-cm stainless-steel hanging cages, and a 2-wk recovery period, during which animals were monitored for wound infection, preceded experimental testing.

Stimulation and recording

Electrical stimuli were generated with a Grass S88 stimulator, and photoelectric stimulus isolation units (Grass SIU6B) were used to deliver 0.1-ms biphasic constant current pulses to the piriform cortex and medial septum. Evoked field potentials in the entorhinal cortex were analogue filtered (0.3 Hz to 3 kHz, half-amplitude), and amplified using a Grass (Model 12) amplifier. Field potentials were digitized at 10 kHz with a 12-bit A/D board for storage on computer hard disk.

ACUTE DEPTH PROFILE RECORDING. The recording electrode was advanced slowly to the superficial layers of the entorhinal cortex, and the vertical positions of the stimulating electrodes were adjusted to optimize the current thresholds for evoked responses. Test-pulse intensities then were adjusted to evoke entorhinal cortex responses ~75% of asymptotic levels (piriform cortex, 475-725 µA; medial septum, 500-1,000 µA). The recording electrode again was advanced slowly while monitoring responses evoked by piriform cortex stimulation until the electrode contacted the skull at the surface of the entorhinal cortex. A 20-min period preceded subsequent testing. Evoked field potentials were recorded at each of 41 depths spaced at 50-µm intervals as the recording electrode was retracted 2.0 mm from the cortical surface. It was found in the first experiments that paired-pulse stimulation of either the piriform cortex or medial septum with an interpulse interval of 70 ms resulted in response facilitation. Field potentials evoked by paired-pulse stimulation of these sites were recorded at each depth to characterize the membrane currents associated with both the conditioning and test responses. Ten responses were recorded at each depth, one every 10 s, and were later averaged.

Polysynaptic responses are evoked less reliably than monosynaptic responses due to variability in evoked cell firing and summation of disynaptic potentials (Berry and Pentreath 1976). Polysynaptic responses tend to fail at stimulation frequencies >50 Hz, whereas monosynaptic responses usually follow frequencies of >= 100 Hz. After depth recordings, frequency of following tests were conducted in three animals by delivering 1.0-s trains of stimulation pulses at frequencies between 60 and 125 Hz. In one animal, responses to five trains at each stimulation frequency were recorded at a depth of 250 µm and averaged. Tests were inconclusive for medial septal evoked responses due to large early field components attributable to hippocampal activation (see below).

INPUT/OUTPUT TESTS. Animals were habituated to a 30 × 40 × 30-cm wooden chamber with a Plexiglas front and a wire-grid floor, and all chronic recordings were collected while animals were in a quiet, resting state. Input/output (I/O) tests were conducted 1 and 3 days before paired-pulse and frequency potentiation tests to ensure that evoked potentials were stable, and again 1 day after the completion of testing to monitor possible long-term changes in the responses. In I/O tests, 10 responses to piriform cortex and medial septal stimulation at each of 10 stimulus intensities (16-1,259 µA) were recorded. There was a 10-s interpulse interval, and the 10 responses at each intensity were averaged.

PAIRED-PULSE TESTS. Both single- and double-site paired-pulse tests were conducted by delivering two stimulation pulses at interpulse intervals ranging <= 1,000 ms. The first of the two pulses is the conditioning pulse (C pulse) and the second is the test pulse (T pulse). First, 10 responses to single-pulse stimulation of the site receiving the T pulse were recorded, and then 10 responses to double-pulse stimulation at each interpulse interval were recorded. Pulses were set to intensities that resulted in responses ~75% of asymptotic levels in I/O tests except for the double-site-test C pulses, which were set to evoke maximal responses. The peak amplitudes of averaged responses to T pulses were expressed as a percent increase from the single-pulse conditions. The results were averaged across animals and plotted as a function of interpulse interval (Racine and Milgram 1983).

OCCLUSION TESTS. Separate occlusion tests were conducted using pulse intensities that resulted in either asymptotic or ~50% of asymptotic response amplitudes during I/O tests. Field potentials were recorded and averaged for 10 single-pulse stimulations and then for 10 combined stimulations. Because peak latencies of evoked responses vary across animals and were longer for medial septal evoked responses, the medial septum was stimulated 5-20 ms before the piriform cortex to synchronize entorhinal cortex activation. Double-pulse responses expected on the basis of linear summation were calculated by summing single-pulse responses corrected by the interpulse interval, and the result was compared with the actual mean response to combined stimulation. The differences between expected and actual peak amplitudes for weak and strong pulses were analyzed using a matched samples Student's t-test.

FREQUENCY POTENTIATION TESTS. In frequency potentiation testing, trains of 10 pulses were delivered at each of 11 frequencies between 4 and 30 Hz using pulse intensities that resulted in response amplitudes that were ~75% of asymptotic levels in I/O tests. Longer trains resulted in negligible further growth in preliminary tests. Responses to each frequency were recorded in ascending order with a 30-s intertrain interval until eight samples at each frequency were obtained. For each stimulation frequency, the field responses to each pulse in the trains were averaged, and the mean amplitude was expressed as a percent increase from the mean of the first-pulse responses. Data were averaged across animals, and response amplitudes for the 1st, 2nd, 6th, and 10th responses were plotted as a function of stimulation frequency.

The tests described above were conducted during a 4-day period in the following order: single-site paired-pulse tests, double-site paired-pulse tests, occlusion tests, and frequency potentiation tests.

Histology

Animals with chronically implanted electrodes were deeply anesthetized with chloral hydrate. All animals were perfused through the heart with 0.9% saline followed by 10% formalin, and brains were stored in 10% formalin and 20% glucose solution. Frozen, 40-µm-thick sections were placed on gelatin-coated slides for staining. Coronal sections were used to identify electrode locations with the exception of recording electrode trajectories in depth profile experiments for which saggital sections were used. After the Prussian blue reaction (5% KFeCN in 1% HCl), sections were stained with Neutral red. Light photomicrographs of recording electrode tracks were obtained with a Zeiss Axioskop microscope, and electrode locations were plotted on sections taken from the atlas of Paxinos and Watson (1986).

CSD analysis

Extracellular field potentials result from the volume average of many small membrane currents resulting from neuronal activity. The extracellularly recorded field potential (phi ) is related to membrane currents (Im) by the equation
<IT>I</IT><SUB>m</SUB> = −∇σ∇φ
where sigma  is a tensor describing the conductivity of the extracellular space, and nabla  is a gradient operator that quantifies the net rate of divergence or compression at each point in extracellular space (Mitzdorf 1985; Nicholson and Freeman 1975). Im is negative for current sinks and positive for current sources. The above equation holds when the extracellular electrical field obeys Ohm's law and is affected negligibly by capacitive, magnetic, and inductive effects (Mitzdorf 1985) and the error induced on the basis of these assumptions has been estimated to be <10% (Holsheimer 1987; Mitzdorf 1985; Nicholson and Freeman 1975).

In laminar cortical structures that are isotropic (conductivity is not different in different directions) and homogeneous (conductivity does not vary with position), simultaneous excitation within a cortical lamina results in extracellular currents that flow primarily in the direction perpendicular to the cortical plane, and the one-dimensional CSD can be computed by
−<IT>I</IT><SUB>m</SUB> = σ<SUB>z</SUB>⋅∂<SUP>2</SUP>φ/∂<IT>z</IT><SUP>2</SUP>
where z is the direction perpendicular to the cortical plane from which the voltage gradient is sampled (Mitzdorf 1985). Because the potential gradient is proportional to current flow (V = IR), alterations in the potential gradient reflect changes in the amount of current flow that must be due to Im. The second derivative of the voltage gradient (partial 2phi /partial z2) therefore reflects the location of membrane currents and is used alone to estimate the CSD in arbitrary units when gradients in conductivity (sigma z) are likely negligible (Ketchum and Haberly 1993; Leung et al. 1995; Rodriguez and Haberly 1989; Van Goen et al. 1987).

CSD analysis was applied to the field potentials recorded at multiple depths in the entorhinal cortex in the acute preparations described above. Because the second derivative is affected strongly by high-frequency noise, a preliminary spatial smoothing of the evoked potentials was performed, and contributions to the depth profiles with periods of <250 µm were smoothed digitally by convolving the data with weights derived from a Blackman window. Smoothing at this frequency was determined empirically not to mask nor shift the location of consistent sink-source distributions. The second derivative of the smoothed evoked potentials was computed with respect to cortical depth by differentiating a seventh-degree polynomial interpolated through the data points. Results were expressed in arbitrary units in both surface and topographical plots.

    RESULTS
Abstract
Introduction
Methods
Results
Discussion
References

Depth profiles and CSD analysis

Evoked potentials suitable for CSD analysis were obtained for piriform cortex stimulation in 15 of 18 animals tested, and in 9 of these animals, CSD analysis also showed clear current sinks in the entorhinal cortex after medial septal stimulation. Reliable CSD results were obtained for medial septal stimulation only in later experiments when the field potential component corresponding to an induced entorhinal cortex sink was identified and maximized during stimulation electrode placement.

STIMULATION AND RECORDING SITES. Effective medial septal stimulation electrode placements were clustered in or near the medial septum, but some of the stimulation current may have spread to activate low-threshold substrates in the nucleus of the diagonal band, the anterior commissure and/or the fornix-fimbria. The positions of the piriform cortex stimulating electrodes were distributed rostrocaudally and mediolaterally near the dense layer of pyramidal cells in layer II (Fig. 1A). Recording electrode trajectories passed through the caudal hippocampal formation and the ventral aspect of the medial entorhinal cortex, which is ~950 µm thick at these coordinates (Fig. 1B). Electrode penetrations were nearly perpendicular to entorhinal cortex lamina, and recording sites typically included portions of the subiculum and dentate molecular layer.


View larger version (49K):
[in this window]
[in a new window]
  		FIG. 1. A: locations of medial septal (MS) and piriform cortex (PIR) stimulating electrodes used in depth profile experiments are shown on representative coronal sections taken from the atlas of Paxinos and Watson (1986). B: trajectories (2.0 mm) of recording electrodes through the entorhinal cortex are indicated on tracings of representative saggital sections. , tracks from which field potentials evoked by both piriform cortex and medial septal stimulation were recorded. - - -, tracks from which only piriform cortex-evoked field potentials were recorded. ···, track from which data shown in Figs. 3 and 5 were obtained. A, amygdala; ml, corpus callosum; CPu, caudate/putamen; fmj, forcepts major of the corpus callosum; GrDG, dentate gyrus granule cell layer; HiF, hippocampal fissure; Mol, dentate gyrus molecular layer; Pyr, piriform cortex; S, subiculum. C: photomicrograph of a representative saggital section. Cortical divisions and layers are indicated schematically at right. Calibration bar, 500 µm.

PIRIFORM CORTEX STIMULATION. Field potentials. A prominent negative deflection was observed in field potentials evoked in the superficial layers of the entorhinal cortex by piriform cortex stimulation. The superficial negativity clearly followed stimulation frequencies of 80 Hz. Although consecutively evoked responses tended to merge and were obscured by stimulation artifacts during 90- and 100-Hz stimulation, individual deflections occurred at appropriate latencies after the final pulses in these trains (Fig. 2). Similar deflections were not observed after 125 Hz trains, suggesting a failure of the synaptic response between frequencies of 100 and 125 Hz.


View larger version (22K):
[in this window]
[in a new window]
  		FIG. 2. Frequency of following test for entorhinal cortex field potentials evoked by 1.0-s piriform cortex stimulation trains at the indicated frequencies. Recordings were made at a depth of 250 µm. Individual responses to each pulse were less distinct at stimulation frequencies >80 Hz and were not observed during 125-Hz stimulation. Stimulus artifacts have been reduced by low-pass filtering to improve clarity. Downward deflections are negative.

The surface-negative component was largest in layers I and II, at depths between 150 and 350 µm below the surface, and reversed to a positive deflection in the deep layers of the entorhinal cortex (Fig. 3A). Onset latencies averaged 10.4 ms (range, 6-15 ms), the mean latency to peak was 18.4 (range, 16.5-23 ms), and the mean peak amplitude was 0.55 mV (range, 0.16-1.1 mV). At similar latencies, a second negative-going potential usually was present in the molecular layer of the dentate gyrus (11 of 15 cases) and declined in amplitude toward the deep layers of the entorhinal cortex.


View larger version (54K):
[in this window]
[in a new window]
 

View larger version (80K):
[in this window]
[in a new window]
  		FIG. 3. Field potentials and current source density (CSD) analysis for piriform cortex stimulation. A: averaged, unsmoothed field potentials recorded at 50-µm intervals between 0.0 and 1.85 mm from the surface of the entorhinal cortex. Interpulse interval was 70 ms, and the pulse intensity was 650 µA. Negativity is downward. B: CSD analysis of the field potentials in A is represented in surface and topographical plots. Dotted line in topographical plot indicates sink-source transitions, and cortical layers and divisions are indicated at right.

The surface-negative component was followed by a weaker and longer-lasting positive component at similar depths. It began at latencies >25 ms and reversed into a negative deflection in the deep layers of the entorhinal cortex at depths between 450 and 550 µm (Fig. 3A).

Paired-pulse stimulation in all animals resulted in a facilitation of responses to the second pulse relative to the responses evoked by the first pulse. The peak amplitude of the superficial negativity was enhanced by 146% on average (range, 112-174%), and the negativity recorded in the dentate gyrus usually doubled in amplitude (Fig. 3A).

CSD analysis. The typical CSD profile included three sink-source pairs that corresponded to the field components described above (Fig. 3B). The depths and latencies of the sink-source distributions were very similar across animals so that it was possible to average the CSD results after normalizing them to the amplitude of the superficial sink (Fig. 4A).


View larger version (48K):
[in this window]
[in a new window]
  		FIG. 4. Mean CSD analysis results indicating the location of membrane current sources and sinks in the entorhinal cortex after paired-pulse stimulation of the piriform cortex (A, n = 15) and medial septum (B, n = 9). In each case, results were standardized to the peak amplitude of the superficial sink.

The surface-negative field potential component was associated with a current sink that peaked at a mean latency of 18 ms in the lower portion of layer II, 250-350 µm below the cortical surface. The associated current source peaked in layers V and VI at depths near 750 µm. In 7 of 15 cases, a smaller, more superficial sink occurred just before the layer II sink deep in layer I and was associated with a layer II source (Fig. 3B). This result is expressed in the averaged CSD results as a descending initial current sink that merges with the larger layer II sink (Fig. 4A).

The negative potential recorded in 11 cases at sites near the dentate gyrus was associated with a current sink in the outer dentate molecular layer at depths near 1.8 mm. Although the mean peak latency of this component (18 ms) was similar to that of the surface-negative component, the mean onset latency of this component was ~3 ms later (Figs. 3B and 4A).

The long-latency surface-positive component was associated with a source in layer II and a sink in layers IV and V in 12 of 15 cases. The source had an onset latency near 30 ms and a peak latency between 45 and 50 ms, and it peaked at depths 50 µm more superficial than the early sink. Although the early sinks in the entorhinal cortex and dentate molecular layer were enhanced by paired-pulse stimulation, this late sink-source pair was not markedly affected by paired-pulse stimulation.

MEDIAL SEPTAL STIMULATION. Field potentials. Five field potential components were observed in response to medial septal stimulation in most animals. The earliest components tended to decay in amplitude with distance from the hippocampal formation, whereas the later components reversed near the cortical surface (Fig. 5A). The earliest component at latencies of 2-4 ms was a positive spike-like component (10 of 15 cases). It was largest in the dentate molecular layer and reversed into a slight negative deflection near the hippocampal fissure. A positive component, peaking at latencies of 5-8 ms in the dentate molecular layer was observed in all animals. The early spike-like component sometimes was superimposed on the rising portion of this field component, which diminished with distance from the dentate gyrus. A negative component, peaking with a mean latency of 15.6 ms, also was observed in the dentate molecular layer. The amplitude of this component was reduced in three animals in which the recording track passed through the granule cell layer. This positive then negative dentate field potential pattern is similar to that observed by others after medial septal stimulation (Andersen et al. 1961; McNaughton and Miller 1984; Robinson 1986).


View larger version (59K):
[in this window]
[in a new window]
 

View larger version (81K):
[in this window]
[in a new window]
  		FIG. 5. Field potentials and CSD analysis for medial septal stimulation. A: averaged, unsmoothed field potentials recorded at 50-µm intervals between 0.0 and 1.85 mm from the surface of the entorhinal cortex. Interpulse interval was 70 ms, and the pulse intensity was 650 µA. Negativity is downward. Arrow, late surface-negative component, which is enhanced after the second pulse. Note the small spike-like field component superimposed on the deep-negativity in response to the second pulse. B: CSD analysis of the field potentials in A is represented in surface and topographical plots. Dotted line in the topographical plot indicates sink-source transitions, and cortical layers and divisions are indicated at right.

At longer latencies near 20 ms, a weak surface-positive component reversed into a shallow negative deflection in layer II at depths of 300-350 µm. This weak component, however, was clearly discriminable in only six animals. A stronger surface-negative component peaked with a mean latency of 29.3 ms in layer II at an average depth of 230 µm. It reversed into a positive component in layer III at a depth of 350 µm. This latter component returned to baseline levels at latencies near 50 ms. Paired-pulse stimulation increased the amplitude of all field potential components and made the weaker later components more distinct. The early spike-like component was enhanced only slightly, however (Fig. 5A).

CSD analysis. Sink-source pairs in the entorhinal cortex after medial septal stimulation were sufficiently similar to allow averaging of the CSD profiles (Figs. 4B and 5B). The early spike-like and negative field potential components corresponded to current sources and sinks in the hippocampal formation. The current sink associated with the spike-like component peaked in the subiculum at a depth of 1,400-1,700 µm, and the source was located in the dentate gyrus. The negative field component was associated with a large sink in the dentate molecular layer near the hippocampal fissure which had a mean peak latency of 15.6 ms. The positive field potential component, however, which decayed roughly linearly with distance from the dentate gyrus, was not clearly associated with a current source. A current source was observed at a depth of 1,300-1,500 µm, but this source was more prolonged and was associated with a current sink in the subicular-angular bundle region, 900-1,200 µm below the cortical surface (7 of 9 cases).

The late, surface-negative field potential component was associated with a current sink with onset and peak latencies of 14.1 and 30.3 ms. The sink began throughout most of layers II to IV and peaked in layer II near the border with layer III at a depth of 350 µm. The associated current source was located in layers V and VI. The superficial sink was preceded by a weak source at similar depths; this correlated with the weak surface-positive component. Paired-pulse stimulation enhanced both the superficial sink and the weak source preceding it. Sink-source reversals for these currents remained at the same depths, but the peaks in these currents shifted to deeper sites near the reversals. The peak of the superficial sink moved deeper into layer III, and its onset became associated with a current source in layer I. The earlier superficial current source also was localized more clearly to deep layer III after paired-pulse stimulation (Figs. 4B and 5B).

Paired-pulse and frequency potentiation

HISTOLOGY. The positions of chronically implanted stimulating and recording electrode tips were all located in or near the targeted structures (Fig. 6).


View larger version (68K):
[in this window]
[in a new window]
  		FIG. 6. Locations of chronically implanted medial septum (MS)- and piriform cortex (PIR)-stimulating electrodes and of recording electrodes in the entorhinal cortex (EC) are shown on representative coronal sections taken from the atlas of Paxinos and Watson (1986).

INPUT/OUTPUT TESTS. Of the 15 animals in which electrodes were chronically implanted, 14 rats showed entorhinal cortex field responses to piriform cortex stimulation and 12 rats showed responses to medial septal stimulation. Field potentials recorded during I/O tests in the superficial layers of the entorhinal cortex were similar to those obtained in the acute preparations and showed no long-term changes after paired-pulse and frequency potentiation testing (Fig. 7). The mean onset latency of the field potential evoked by piriform cortex stimulation was 10.9 ms (range, 6.7-15.4 ms), and the component had a mean peak amplitude of 0.51 mV (range, 0.22-1.0) at a latency of 16.2 ms (range, 14.3-19.0 ms) with strong stimulation pulses. Field potentials evoked by medial septal stimulation showed positive and negative deflections at latencies <15 ms and a longer-latency component due to membrane currents in the entorhinal cortex (Fig. 7A2). The later potential had a mean onset latency of 22.6 ms (range, 16.6-29.5) and a broad, low-amplitude peak (mean, 0.29 mV; range, 0.1-0.87), which returned to baseline at a mean latency of 58.4 ms (range, 50-63 ms).


View larger version (27K):
[in this window]
[in a new window]
  		FIG. 7. Input/output (I/O) tests of entorhinal cortex responses to piriform cortex (left) and medial septal (right) stimulation. Negativity is upward in this and all subsequent figures. A: field potentials recorded during I/O tests in response to the stimulation intensities (indicated at left) are shown for a representative animal. right-arrow, late surface-negative field potential component evoked by medial septal stimulation. B: mean peak amplitudes of field potential components are shown as a function of test-pulse intensity both before (Pre-PP/FP) and after (Post-PP/FP) paired-pulse and frequency potentiation testing. Data have been standardized to the response to the highest test-pulse intensity during the last pretesting I/O test. Bars indicate twice the standard error of the mean in this and all subsequent figures.

FACILITATION EVOKED BY PIRIFORM CORTEX STIMULATION. Field responses to the second pulse in pairs of piriform cortex pulses were facilitated maximally (248%) when the C-T interval was 30 and 40 ms. Facilitation was reduced at C-T intervals <30 ms and also fell off sharply as the C-T interval was increased to 50 ms. There was a more gradual decline in facilitation as the C-T interval was increased to 500 ms (Fig. 8A1).


View larger version (25K):
[in this window]
[in a new window]
  		FIG. 8. Mean peak amplitudes of responses to test-pulses during paired-pulse tests are shown as a function of interpulse-interval and are expressed as a percent increase from the single-pulse condition. Panel titles indicate the order in which the C and T pulses were delivered in each of the 4 tests (e.g., in A2, the C pulse was delivered to the piriform cortex and the T pulse was delivered to the medial septum). Inset: sweeps show representative averaged field potentials recorded during the single-pulse condition (------) superimposed on mean responses to T pulses at the interpulse intervals resulting in maximal facilitation (- - -). Sweep duration, 50 ms; vertical calibration, 0.5 mV. right-arrow, latencies at which the responses evoked by medial septal stimulation were measured.

Responses to T pulses delivered to the medial septum were facilitated also by prior stimulation of the piriform cortex at interpulse intervals <500 ms (Fig. 8A2). Although the peak facilitation effect was smaller (59%), the largest facilitation effect was observed for C-T intervals of 30 and 40 ms, similar to the results for double-pulse stimulation of the piriform cortex.

FACILITATION EVOKED BY MEDIAL SEPTAL STIMULATION. Paired-pulse stimulation of the medial septum caused the greatest facilitation of the long-latency, surface-negative entorhinal cortex response (149%) at interpulse intervals near 70 ms. Lesser facilitation was observed as the C-T interval was decreased to 30 ms and as it was increased to 700 ms (Fig. 8B1). There was a facilitation of the early positive component and a depression in the early negative component at interpulse intervals near 50 ms, consistent with the effect of paired-pulse septal stimulation on dentate gyrus field potentials (data not shown) (Robinson and Racine 1982, 1986).

Responses to T pulses delivered to the piriform cortex were facilitated also by prior stimulation of the medial septum, but the weaker heterosynaptic facilitation effect peaked at much longer interpulse intervals of 150 and 200 ms. The facilitation was 46% above single-pulse levels at its peak and was more variable and not reliably observed at C-T intervals <100 ms (Fig. 8B2).

OCCLUSION TESTING. Occlusion tests assessed whether piriform cortex and medial septal inputs may activate the same cells in the entorhinal cortex. When both sites were stimulated with weak pulses that evoked responses ~50% of the asymptotic response amplitude in the single-pulse conditions, responses to combined stimulation were almost equivalent to the sum of the single-pulse responses. In contrast, combined stimulation with pulses that evoked maximal single-pulse responses resulted in field potentials smaller than those expected on the basis of summing the single pulse responses (Fig. 9). The difference between the observed and expected results was roughly equivalent to the amplitude of single-pulse medial septal evoked responses, suggesting that strong piriform cortex stimulation served to occlude the effect of medial septal stimulation by maximally depolarizing the same entorhinal cortex neurons. The difference between the observed and expected peak amplitudes was significantly greater when strong rather than weak pulses were used (t = 4.80, P < 0.001).


View larger version (23K):
[in this window]
[in a new window]
  		FIG. 9. Results of occlusion testing with low- (A) and high (B)-intensity stimulation pulses are shown for 4 animals representative of the range of response morphologies observed. Negativity is upward. Left: single-pulse responses to piriform cortex (PIR) and medial septal (MS) stimulation with responses to septal stimulation shifted left to match the peak latencies for piriform cortex evoked responses. Right: responses to double-site stimulation (observed) at an interpulse interval equal to the latency shift applied in the left panels. Double-pulse responses expected on the basis of summing the single-pulse evoked responses (expected) are shown for comparison.Observed peak amplitudes were smaller than expected amplitudes for high-,but not for low-, intensity stimulation. Horizontal calibration, 10 ms; vertical calibration, 0.5 mV.

FREQUENCY POTENTIATION. The field potentials evoked by both piriform cortex and medial septal stimulation were enhanced during low-frequency, 10-pulse stimulation trains, and the effects were strongest at stimulation frequencies between 12 and 18 Hz (Fig. 10). Responses evoked by the second pulse in each piriform cortex train were similar to those observed at comparable C-T intervals during paired-pulse testing, but further enhancements in slope and amplitude occurred during continued 6- to 20-Hz stimulation. The tuning curves were rather broad, and peaked on average at 14 Hz. Frequency potentiation effects evoked by medial septal stimulation were weaker and more variable, and both the early hippocampal field components and the later negative entorhinal cortex components were altered. Frequency potentiation of the entorhinal cortex field response peaked for 12-Hz trains and growth during the train was less pronounced than during piriform cortex stimulation. Tuning curves fell off sharply at higher frequencies, and responses after 30-Hz stimulation usually were depressed (Fig. 10B2). In two animals, there was a development of two distinct late field components with different latencies. The negative early component was depressed in all animals at stimulation frequencies >8-12 Hz, and the earlier positive component was enhanced at these frequencies in some animals.


View larger version (41K):
[in this window]
[in a new window]
  		FIG. 10. Frequency potentiation tests of entorhinal cortex responses to piriform cortex (left) and medial septal (right) stimulation. A: average response to the 1st(------) and last pulses (- - -) in each 10-pulse train are superimposed for a representative animal at each of the indicated stimulation frequencies. Stimulus artifacts due to the 2nd pulses are visible in traces for 25- and 30-Hz stimulation. Negativity is upward. B: group averages of field potential amplitudes are shown as a function of pulse frequency for responses to piriform cortex (B1; n = 14) and medial septal stimulation (B2; n = 12). Mean response amplitudes are shown for the 1st (pulse 1), 2nd (pulse 2), 6th (pulse 6), and 10th (pulse 10) pulses in each train.

    DISCUSSION
Abstract
Introduction
Methods
Results
Discussion
References

The current sinks and sources in the medial entorhinal cortex that generate evoked field potentials in the rat medial entorhinal cortex after piriform cortex and medial septal stimulation have been localized here to the superficial layers of the entorhinal cortex using one-dimensional CSD analysis techniques. Piriform cortex stimulation resulted in a current sink peaking in layer II, and medial septal stimulation resulted in a more prolonged and spatially distributed current sink, which peaked close to the layer II-III border. These findings provide the basis for interpreting field potentials recorded at single cortical depths in behaving animals and therefore allow the examination of synaptic responses in these pathways in behaving animals under normal levels of inhibition. We have used this preparation to measure the relative sizes and time courses of synaptic facilitation effects induced in the entorhinal cortex by stimulation of these pathways and also have monitored the responsiveness of the entorhinal cortex to a range of repetitive low-frequency stimulation trains (Zucker 1989).

CSD analysis results can be affected by variations in the conductivity of extracellular space, which can occur near areas where soma or myelinated fibers are packed tightly (Mitzdorf 1985). Although the calculated locations of sinks and sources are affected minimally by conductivity gradients in both the piriform cortex (Haberly and Shepherd 1973) and area CA1 of the hippocampus (Holsheimer 1987; Leung et al. 1995), the apparent location of layer I membrane currents in the entorhinal cortex after piriform cortex stimulation may have been shifted slightly toward deeper sites in layer II by reduced conductivity in layer I. Similarly, reduced conductivity deep in the subiculum may have shifted the current sink in this area evoked by medial septal stimulation toward more superficial sites (Fig. 4B). Further, recording electrode locations could not be visually verified during surgery, so that CSD results might have been shifted to deeper sites if recordings from the subdural space contributed to the most superficial recording. Adjustments in traces based on histological results were required in only two cases, however, and we estimate the error in surface determination to be <25 µm.

Piriform cortex efferents

Stimulation of the piriform cortex resulted in a surface-negative field potential in the entorhinal cortex similar to those observed in previous acute and in vitro preparations (Fig. 4A) (Alonso et al. 1990; Boeijinga and Van Groen 1984; de Curtis and Llinas 1993; de Curtis et al. 1991; Van Groen et al. 1987). The ability of the field potential component to follow frequencies of 100 Hz (Fig. 2) is consistent with it being a monosynaptic response. The conduction velocities of the fastest fibers mediating these responses were estimated to be 0.50 and 0.48 m/s for the acute and chronic preparations based on onset latencies near 11 ms and an interelectrode distance of 5.2 mm. Similarly slow estimates have been obtained in the cat for the anterior (0.7 m/s) and posterior (0.4 m/s) piriform cortex projections on the basis of unit firing latencies (Boeijinga and Van Groen 1984).

The surface-negative potential was associated with a current sink in layers I-II and a source deep in the entorhinal cortex. Piriform cortex efferents to the medial entorhinal cortex are most prominent in layers I and II, and only a small number of fibers project to layer III (Room et al. 1984) so that the deep current source is likely due to passive current flow induced by active membrane currents in layers I and II. The superficial current sink always peaked in layer II, and in 7 of 15 cases, the sink first emerged in layer I before peaking in layer II, similar to findings for the ventrolateral entorhinal cortex of the urethanized cat after either olfactory bulb or piriform cortex stimulation (Van Groen et al. 1987). Larger layer I currents were expected on the basis of previous research (Van Groen et al. 1987), but the pattern observed here is consistent with dendritic activation of layer II neurons, which arborize heavily in layers I and II (Lingenhohl and Finch 1991), and strong somatic depolarization in layer II. A surface-positive component like that observed by Boeijinga and Van Groen (1984) in the lateral entorhinal cortex of the cat, which they attribute to activation of currents deeper in layers II and III, was not observed in the present study. This is likely due to the sparse distribution of piriform cortex projections to layer III in the medial entorhinal cortex sites studied here (Boeijinga et al. 1982; Luskin and Price 1983; Room et al. 1984).

The later source in layer II that peaked at latencies near 45 ms corresponded to a surface-positive field potential component. In the cat, a similar long-latency layer I-II source is associated with both reduced unit activity and a paired-pulse depression at interpulse intervals of 40 ms, suggesting that the source is due to active outward membrane currents (de Curtis et al. 1991; Van Groen et al. 1987). Finch and Babb (1980) have recorded inhibitory postsynaptic potentials in principal cells of the entorhinal cortex at similarly long latencies after stimulation of the fornix and hippocampus. Inhibitory cells in the superficial layers of the entorhinal cortex are fast-spiking and have extensive basket-like axonal arborizations in layer II (Finch et al. 1986; Jones and Buhl 1993; Kohler et al. 1985), and there is both anatomic (Wouterlood et al. 1985) and electrophysiological (Finch et al. 1988) evidence suggesting that they are activated predominantly in a feedforward manner by olfactory efferents. A net depression of field potentials never was observed in behaving animals during paired-pulse piriform cortex stimulation in this study or in that by Racine and Milgram (1983), but there were sharp reductions in the amount of facilitation at interpulse intervals from 40 to 70 ms that correspond to the duration of the layer II source observed in the CSD experiments (Fig. 8A1).

The sink in the dentate molecular layer observed after piriform cortex stimulation is consistent with the findings of Canning et al. (1995), who describe a sink in the outer molecular layer with an onset latency near 13 ms after piriform cortex stimulation in the rat (Fig. 3B). Others also have observed negative dentate field potentials after stimulation of piriform cortex and olfactory bulb in the rat and cat (Habets et al. 1980; Schwerdtfeger et al. 1990; Wilson and Steward 1978; Woolley and Barron 1968). Monosynaptic activation of the dentate gyrus might occur if the piriform cortex test-pulses directly stimulated the most rostral aspects of the entorhinal cortex. Such direct activation could occur only if the entorhinal cortex extends further rostrally than normally recognized (see Burwell et al. 1995). The distribution of electrode positions used here, and the lack of direct projections from the piriform cortex to the hippocampal formation, are also consistent with a disynaptic mediation of the dentate gyrus sink. These current sinks showed a later onset, but similar peak latencies, to those in the entorhinal cortex, suggesting that a large portion of entorhinal cortex neurons are activated early during the entorhinal cortex sink and that the slowest piriform cortex efferents do not contribute significantly to further disynaptic activation of the dentate gyrus. This is also consistent with the activation of a population of low-threshold entorhinal cortex neurons by single piriform cortex test-pulses (Alonso and Klink 1993). Finally, the paired-pulse facilitation of the dentate gyrus sink at a 70-ms interpulse interval (Figs. 3B and 4A), is similar to that observed by Canning et al. (1995) for intervals of 50 ms and indicates that synaptic facilitation effects in the entorhinal cortex can have a significant impact on the activation of dentate gyrus neurons.

Medial septal efferents

The late surface-negative field potential evoked by septal stimulation was associated with a current sink in layers I-IV that peaked near the border between layers II and III ~50 µm deeper on average than the piriform cortex evoked sink (Fig. 4B). The fastest fibers mediating the response are estimated to have conduction velocities of 0.57 and 0.45 m/s based on a mean onset latencies of 17.9 and 22.6 ms in the acute and chronic preparations, respectively, and a 10.2 mm curved path between the electrodes. The longest-duration portions of the sink were focused near layer IV as the more superficial currents weakened. This general distribution suggests that medial septal stimulation initially activates substrates in layers II-IV (Alonso and Kohler 1984; Gaykema et al. 1990; Insausti et al. 1987) and that the resulting currents in layer IV have a more prolonged duration.

In addition to the superficial current sink, there was a shorter-latency current source in layers III and IV, and both were enhanced by paired-pulse stimulation (Fig. 4B). Paired-pulse facilitation of the superficial current sink in behaving animals was largest for the 70-ms interpulse interval, which also was used to evoke field potentials during depth profile recordings. The short-latency source may be due to active inhibitory currents if the partly GABAergic nature of the septo-hippocampal pathway is mirrored by septo-entorhinal projections. Alternatively, this source could result from passive current flow induced by the deeper subicular sink (observed in 7 of 9 animals). The subicular sink also was enhanced by paired-pulse stimulation and may be mediated by medial septal efferents that innervate the ventral subiculum (Gaykema et al. 1990).

The early positive then negative field potential components, which peaked at latencies near 6 and 16 ms, were observed in all recording sites, but the results of CSD analysis showed that these field components were volume conducted from the hippocampal formation to recording sites in the entorhinal cortex (Fig. 5). The positive component was not associated with a clear current source, consistent with its nearly linear decay with distance from the hippocampal formation, but the negative component was associated with a large current sink in the dentate molecular layer. This field potential pattern and the enhancement of the negative component by paired-pulse stimulation at a 70-ms interval (Fig. 5A) are consistent with the findings of others for dentate field potentials evoked by septal stimulation (Andersen et al. 1961; Chandler and Crutcher 1983; McNaughton and Miller 1984; Mosko et al. 1973; Robinson and Racine 1986). Dentate granule cells fire during the negative wave (McNaughton and Miller 1984) and population spike activity similar to the 2- to 4-ms latency spike seen here also has been observed at earlier latencies (Andersen et al. 1961; Robinson 1986; Robinson and Racine 1982). We show here that the earlier spike recorded at 2- to 4-ms latency is associated with a current source in the molecular layer and a current sink in subicular white matter (Fig. 4B) and is therefore unlikely to be due to a volley in the angular bundle.

Frequency potentiation

Repetitive low-frequency stimulation of either the piriform cortex or the medial septum in behaving animals resulted in an enhancement of entorhinal cortex responses when stimuli were delivered in the 8- to 20-Hz range (Fig. 10). The most effective frequencies of 12-18 Hz do not correlate well with gamma-frequency (35-80 Hz) oscillations in the olfactory system. They are, however, near the beta (15-35 Hz) and theta (4-12 Hz) EEG activities in the entorhinal cortex (Boeijinga and Lopes da Silva 1988; Bressler 1984; Dickson et al. 1994; Mitchell and Ranck 1980). The 12- to 18-Hz frequency range also corresponds with the frequency range of oscillatory burst responses triggered by presentations of biologically significant olfactory stimuli (Heale et al. 1994). Frequencies ranging from 8 to 20 Hz also have been found to be effective in potentiating lateral olfactory tract inputs to the piriform cortex (10 Hz) (Mokrushin and Emel'yanov 1993) and polysynaptic hippocampal responses to olfactory bulb stimulation in vivo (8-20 Hz) (Cragg 1960).

This latter finding suggests that frequencies of activity from 8 to 20 Hz may be particularly effective in mediating polysynaptic activation of the hippocampal formation and is consistent with other studies showing that frequency potentiation effects in the afferent and intrinsic pathways of the hippocampal formation are induced readily by 8- to 20-Hz stimulation (Andersen and Lomo 1967; Finch and Babb 1980; Landfield et al. 1986; MacVicar and Dudek 1979; Munoz et al. 1991; Pitler and Landfield 1987). Further, low-frequency stimulation of the contralateral CA3 results in frequency potentiation of entorhinal cortex responses and a strong parallel enhancement of a long-latency dentate gyrus field response, which is abolished by entorhinal cortex lesions (Deadwyler et al. 1975). Polysynaptic transmission to the CA3 and CA1 regions is also frequency dependent because, whereas perforant path stimulation at 0.2 Hz evokes monosynaptic responses in CA3 and CA1, multisynaptic suprathreshold activation of these sites occurs reliably only during stimulation at frequencies in the 5- to 15-Hz range (Yeckel and Berger 1990). This type of frequency-dependence in entorhinal cortex efferents suggests that activity in cortical efferents to the entorhinal cortex may have its greatest impact on hippocampal activity during periods of oscillatory activity.

Convergence and interactions of piriform and septal inputs

Both piriform cortex and medial septal stimulation resulted in superficial current sinks in the entorhinal cortex that overlapped in layer II, consistent with the known anatomy of these pathways (Fig. 4). The convergence of these inputs onto the same neurons in layer II has not been confirmed with intracellular recordings in vivo, but the occlusion and heterosynaptic facilitation effects presented here provide indirect evidence supporting this hypothesis. Occlusion tests showed that postsynaptic responses evoked by weak stimulation of each input summed during simultaneous activation but did not sum for combined stimulation with strong pulses. These results suggest that strong activation of piriform cortex afferents masked or occluded postsynaptic responses evoked by medial septal stimulation onto the same neuronal population (Fig. 9).

The heterosynaptic facilitation effects shown by double-site paired-pulse tests are also consistent with the activation of common pools of principle neurons and inhibitory interneurons. Stimulation of one afferent may enhance subsequent responses at other synapses onto the same neurons by inducing intrinsic membrane oscillations that may increase the subsequent opening of voltage-dependent channels after synaptic activation (Alonso and Klink 1993). Paired-pulse depression of GABAergic transmission, as has been described for both the neocortex and hippocampus (Deisz and Prince 1989; Mott et al. 1993), also may contribute to the facilitation effects if both sites activate the same pool of interneurons.

Medial septal responses were facilitated by piriform cortex stimulation at interpulse intervals of 30 and 40 ms (Fig. 8A2); these were also the most effective intervals when both pulses were delivered to the piriform cortex. In contrast, piriform cortex responses were facilitated most effectively by septal stimulation at much longer interpulse intervals of 150-200 ms (Fig. 8B2). The longer time course of the heterosynaptic facilitation induced by medial septal stimulation may be due in part to cholinergic inputs from the medial septum (Gaykema et al. 1990) because cholinergic agonism has prolonged effects on the excitability of neurons in cortical areas (Cole and Nicoll 1984; McCormick and Prince 1986) and enhances NMDA currents in the CA1 region (Markram and Segal 1990).

The heterosynaptic facilitation of entorhinal cortex responses to piriform cortex stimulation suggests that the medial septum may play a significant role in the short-term regulation of responsiveness of the entorhinal cortex to olfactory inputs. Further, animals in this study were tested in a resting state, but short-term facilitation effects might be enhanced during behavioral states associated with greater activity in septo-entorhinal efferents. This suggested to us that the septum also may contribute to the induction of longer-term synaptic plasticity during theta rhythm activity, which may be the primary endogenous cause of short-term changes in synaptic responsivity (Alonso et al. 1990; Chapman and Becker 1995; de Curtis and Llinas 1993). We recently have found that entorhinal cortex field potentials in chronically prepared rats are potentiated for weeks after tetanization of the piriform cortex and that the amount of LTP induced can be enhanced by theta-frequency stimulation of the medial septum (personal observations).

    FOOTNOTES

  Address reprint requests to R. J. Racine.

  Received 5 March 1997; accepted in final form 2 July 1997.

    REFERENCES
Abstract
Introduction
Methods
Results
Discussion
References
0022-3077/97 $5.00 Copyright ©1997 The American Physiological Society



This article has been cited by other articles:


Home page
 J. Neurophysiol.Home page
S. D. Glasgow and C. A. Chapman
Local Generation of Theta-Frequency EEG Activity in the Parasubiculum
J Neurophysiol, June 1, 2007; 97(6): 3868 - 3879.
[Abstract] [Full Text] [PDF]

Home page
 J. Neurophysiol.Home page
D. A. Caruana, R. E. Sorge, J. Stewart, and C. A. Chapman
Dopamine Has Bidirectional Effects on Synaptic Responses to Cortical Inputs in Layer II of the Lateral Entorhinal Cortex
J Neurophysiol, December 1, 2006; 96(6): 3006 - 3015.
[Abstract] [Full Text] [PDF]

Home page
 J. Neurophysiol.Home page
D. A. Caruana and C. A. Chapman
Stimulation of the Parasubiculum Modulates Entorhinal Cortex Responses to Piriform Cortex Inputs In Vivo
J Neurophysiol, August 1, 2004; 92(2): 1226 - 1235.
[Abstract] [Full Text] [PDF]

Home page
 J. Neurophysiol.Home page
S. Kourrich and C. A. Chapman
NMDA Receptor-Dependent Long-Term Synaptic Depression in the Entorhinal Cortex In Vitro
J Neurophysiol, April 1, 2003; 89(4): 2112 - 2119.
[Abstract] [Full Text] [PDF]

Home page
 J. Neurophysiol.Home page
G. Biella, L. Uva, U. G. Hofmann, and M. De Curtis
Associative Interactions Within the Superficial Layers of the Entorhinal Cortex of the Guinea Pig
J Neurophysiol, September 1, 2002; 88(3): 1159 - 1165.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (14)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Chapman, C. A.
Right arrow Articles by Racine, R. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chapman, C. A.
Right arrow Articles by Racine, R. J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Visit Other APS Journals Online