Hippocampal field potentials evoked by stimulation of multiple limbic structures in freely moving rats

Hippocampal field potentials evoked by stimulation of multiple limbic structures in freely moving rats

Nmvmcience Vol. 4. pp. 1467 to 1478 perganm press Ltd 1P79, Printed in Great Britain HIPPO~AMPAL FIELD POTENTIALS EVOKED STIMULATION OF MULTIPLE LIMB...

3MB Sizes 0 Downloads 6 Views

Nmvmcience Vol. 4. pp. 1467 to 1478 perganm press Ltd 1P79, Printed in Great Britain



S. E. Fox and J. B. RANCK,JR Department of Physiology, Downstate Medical Center, State University of New York, Brooklyn, New York, U.S.A. Abstract-A preparation is described in which both single unit activity and field potentials evoked by stimulation of three sources of agerents can be recorded simult~eously from three moveable microelectrodes within the hippocampal formation in freely moving rats. Field potentials recorded in the hippocampal cellular layers during slow wave sleep in response to stimulation of the ventral hippocampal commissure, the entorhinal cortex and the medial septaf nucleus are described. Most of the field potentials in these freely moving animals are identical in form to those described in acute preparations. Some responses which differ from their descriptions in the literature are also discussed. Once electrodes are lowered into place, field potentials are stable for many days. This prep~ation allows most of the classical hip~campal ex~a~llul~ eIec~ophysiology to be carried out in a freely moving animal. Since field potentials and single unit activity can now be so readily recorded chronically in the mammalian central nervous system, many studies can be carried out which were previously impossible. Behavioral states are observable and can be subjected to experimental manipulation. When the ‘state’ of the nervous system is changing rapidly (e.g. during behavioral tasks, or in drug or seizure studies) this preparation with well-known anatomical and electrophysiological inter-relationships between its multiple recording and stimulating electrode sites increases the chance of being able to interpret changes which may occur in many synaptic systems simul~eously.

THIS paper describes a preparation which leads to some new approaches to ~p~~rn~ el~~ophysiology in freely moving rats. Electrodes are chronically implanted to stimulate three discrete populations of afferents and record both Geld potentials and unit activity simultaneously from three major subdivisions of the hippocampal formation (CA& CA3-4 and the fascia dentata). This technique expands the number of possible ex~rimen~ m~ipulations which can practicably be applied in an attempt to elucidate the functional organization of hippocampal formation in different behaviors. The basic preparation and some field potential studies will be described here. Comparisons will be made with work by others in anesthetized animals, and to work in freely moving animals. Unit recordings were made simultaneously during these experiments and those data wilf be reported in detail in another paper (S. E. Fox & J. B. RANCK, unpublished observations). The highly organized anatomy of the hippocampal for~t~on has led to many productive el~trophysiological studies of its functional organization, using laminar analysis of field potentials. Nippocampal field

Abbreviations: EC, entorhinal cortex or angular bundle; EPSP, excitatory ~s~~aptic potential; XPSP, inhibitory postsynaptic potential; MSN, medial septal nucIeus or nucleus of the diagonal band of Broca; VHC, ventral hip~amp~ commissure.

evoked by stimulation of several different aRerent ~pulations have been described in detail by ANDBRSEN & CO-WORKWS (ANDER~EN,BLACKSTADT & LIMO, 1966~; ANDIXRSEN, HOLMQVIST & VCNRHOEVE,1966b; ANDER~EN& Low, 1966). Most expotentials

trinsic afferents are excitatory and with high intensities or frequencies of stimulation produce negative ‘~pulation-spikes’ in the cellular layers. Populationspikes represent summed extracellular syn~~onous

unitary discharges of hippocampal projection ceils (ANDERSEN, BLISS & SKREDE, 1971~). Whether produced ort~odro~~ly or anti~o~~ly, they are Followed by a long-duration positive wave in the cell layer. At least part of this positive wave presumably reflects local recurrent ~~ibition rne~at~ by interneurons with synapses on the somas of projection cells (ANDEIQEN,ECCLES& LOYNING,1964u,b). Until recently studies in un~esthet~ed animals have been confined to ‘spontaneous’ slow wave activity or units (RANCK,19736; VANDFRWOLF, 1969; WIN~N, 1974). Little reference will be made to single unit activity in this paper. There have recently been a few studies of evoked potentials in the hippocampus of unan~thet~ed animals. In general, they have shown results similar to those in anesthetized animals. Chronic studies have shown that facilitation in the entor~n~-bite system occurs at stimulus frequencies as low as 0.2/s and can last for days or weeks (BLISS% GARDNER-MEDWIN, 1973; DOUGLAS t GOD DARD, 1975). WINSON& ABZUG(1978) have show


S. E. Fox and J. H. RANCK


that amplitudes of certain components of field potentials from the stimulation of afferents from the entorhinal cortex depend upon the behavioral state. The results reported here demonstrate the simi-

larity of hippocampal responses in this preparation evoked by entorhinal and commissural stimuli to those recorded in anesthetized animals. The hippocampal field potential in response to stimulation in


FIG. 1. Synaptic organization and electrode placement. (A) Cross-section of the hippocampus and the fascia dentata in a lamellar plane. The layers of cell bodies are shaded, schematic representations of pyramidal cells of CA1 and CA3, and a granule cell of the fascia dentata are shown with their dendritic arborizations. Associated interneurons are represented as filled circles, synaptic endings as open triangles. Anatomical pathways are heavy lines with arrows indicating the orthodronic direction as they leave the picture. See text for details. Point (a) corresponds approximately to point (a) in (B). (B) Dorsal view of hippocampal formation and septum with electrode placement and fiber systems. The shaded area represents a cross-section of the hippocampus (HIPPO) and the fascia dentata in a lamellar plane, as in (A). The three recording microelectrodes are represented as E,, E2 and E3 in CA3, CA1 and the fascia dentata respectively, E4 is a fixed macroelectrode with its tip in the contralateral fascia dentata (FD). The potential difference for El through E4 is recorded with respect to a single fixed reference electrode (R) at a distance. Fixed stimulating electrodes (S, through S,) are indicated in the medial septal nucleus (MSN), the medial entorhinal cortex (MEC) and the ventral hippocampal commissure (VHC), respectively. Cell bodies are represented by filled circles, other representations, as in (A). Other abbreviations: CONTRA, contralateral; EC, entorhinal cortex; LSN, lateral septal nucleus; NDBB, nucleus of the diagonal band of Broca.

Hippocampal field potentials in freely moving rats the medial septal nucleus-diagonal also described. EXPERIMENTAL Anatomical

band of Broca is



To appreciate the rationale for the choice of stimulating and recording sites one must have an understanding of the synaptic organization of this system. The following description is not meant to be exhaustive, but does represent a large part of the known connectivity. The dorsal hippocampal formation of the rat contains two separate densely packed cellular layers: (1) the pyramidal cell layer {hip~campus proper or Ammon’s horn) composed of areas CAL to CA4 according to the system of LORENTEDE N6 (1934), and (2) the granule cell layer of the fascia dentata. These cellular layers are rolled into two interlocked U-shaped structures (see Fig. 1A and B). The pyramids of Ammon’s horn have basal dendrites in the stratum oriens and long apical dendrites which approach the embryonic pial surface (the hip~campal fissure). The granufe cells of the fascia dentata have no basal dendrites, but have broad apical dendritic arborizations which end at the fissure in the dorsolateral blade and the pial surface in the ventromedial blade (RAM~N Y CAIAL, 1893).

The medial septal nucleus and the nucleus of the diagonal band of Broca project to various specific laminae of the hippocampal formation via the fornix/~mbria (SWANSON& COWAN, 1976). The entorhinal cortex and other retrohippocampal areas send their axons via the angular bundle/perforant path (!&WARD, 1976). The alTerents from the entorhinal cortex densely innervate the outer two-thirds of the apical dendrites of the fascia dentata granule cells, and also have less dense terminations on the distal apical dendrites of pyramidal cells. The granule cells synapse on the proximal apical dendrites of the CA3 and 4 pyramidal cells with complex ‘mossy fiber’ contacts (LORENTEDE N6, 1934). The CA3 and 4 pyramids (1) project out through the fornix/fimbria to the lateral septal nucleus (SWANSON& COWAN, 1977), (2) send collaterals via the fimbria and ventral hippocampal commissure to specific Iaminae in the contralateral hippocampus (CAI-4) and the fascia dentata (GFTTLIEB & COWAN, 1973), (3) have ‘Schaffer’ collaterals to specific laminae of CA1 (SCHAFFER, 1892; HJORTH-SIMONSEN, 1973), and (4) may send ipsilateral association connections to the fascia dentata (ZIMMER,1971). CA1 pyramidal cells (1) project to the subiculum and the entorhinal cortex via the alveus (SWANSON,WYSS& COWAN, 1978), and/or (2) project to the lateral septal nucleus through the fornix/fimbria (HJORTH-SIMONSEN, 1973; SWANSON& COWAN, 1977). There are a relatively small number of interneurons (short axon cells) of various morphological types scattered throughout the hippo~mpal formation (L.ORENTE DE N6, 1934) most of whose afferents are unknown, but which definitely include recurrent collaterals of the hippocampal projection cells. These interneurons project to most laminae, however, one type (the basket cell) has dense plexuses of terminations on the somas of projection cells and has received particular attention. ANDERSEN, BLINK & SKREDE(1971b) defined hippocampal ‘lameilae’ electrophysiologically in the rabbit. They demonstrated that stimulation of the four-component excitatory chain: perforant path fibers from the entorhinal cortex, mossy fibers of the fascia dentata granule cells,


Schaffer collaterals of CA3 pyramids, and alvear fibers of CA1 pyramids (see Fig. 1A) excited cells in a plane approximately normal to the septotemporal axis of the hippocampal formation. This has recently been largely confirmed in the rat (RAWLINS& GREEN,1977), although the lamellar nature of the Schaffer collaterals is questionable in this species (see Discussion). The approximate orientation of these lamellae is indicated in Fig. 1B by the plane of the section and the series of dotted lines. Anatomical confirmation of narrow bands of projections exists only for the mossy fibers (BLACKSTAD,BRINK, HEM & JEUNE,1970); the Schaffer collaterals, for example, spread several millimeters along the scptotemporal axis (HJQRTHSIMONSEN, 1973; SWANSONet af., 1978). The hippocampus is not organized into di~ontinuous units, but rather a ‘iamelia’ is a functional concept which utilizes the fact that most of the fibers tend to run in somewhat the same direction. Transmission through the lamellar chain of hippocampal pathways following entorhinal stimulation can be monitored simultaneousiy using three recording electrodes: in the fascia dentata, areas CA3 and CAL With stimulating electrodes in the entorhinal cortex, the medial septal nucleus and the ventral hippocampal commissure arranged ‘on beam’ (so that they maximally excite fibers projecting to the level of the recording electrodes) one can investigate much of the classical hippocampai electrophysiology. This approach takes advantage of both the laminar and lamellar organization of the hip~campus in a manner similar to in vitro studies on hippocampal slices. A somewhat similar approach using multiple recording electrodes and multiple stimulating electrodes has been used by CASEY& KEENE (1973) to investigate responses of medial thalamic neurons to stimulation of the medial forebrain bundle and the nucleus gigantocelIularis in freely moving rats. WINSON& ABZUG (1978) and BLISS & G~DN~-MEDWIN (1973) stimulated the entorhinal cortex or the angular bundle and recorded responses at a single hippocampal electrode in CAl, CA3 or the fascia dentata. The major advantage of the use of multiple stimulating and recording electrodes is that data can be collected stimultaneously on several systems. Surgery and placement of the electrodes Fifteen Long-Evans (hooded) rats weighing 350-450 g were anesthetized with pentobarbital (40 mg/kg, intraperitoneally) with supplementation as necessary. Electrodes for stimulating and recording were stereotaxically implanted using aseptic techniques. In all cases, small burr holes were made in the skull, and dura and bone edges were covered with Vaseline to protect them from later application of dental acrylic. A small triple plug of Delrin with Silastic membranes on the bottom held three moveable microelectrode carriers modified after RANCK (1973a) (see Fig. 2). The microelectrodes were etched tungsten or stainless steel, insulated with Epoxylite to within IOlm of the tips. The carriers were aligned in a plane and the microelectrodes (angled at 22” intervals) were aimed to converge at a point. With the center microelectrode extended to within 2.0-2.5 mm of the point of convergence, the triple array was implanted so that the three microelectrodes lay approximately in a tamellar plane. The device was lowered while watching an oscilloscopic display for unit activity, so that the tip of the extended microelectrode was placed in the CA1 pyramidal layer. This was initially determined by the presence


S, E. Fox and J. B.



sluin&ss I - 72 Stointrss

Silastio Mwnbfone





mm lot., +0.6



A-P 1

mm. Rortret to IntfPi - Aural ti

FIG. 2. Triple microelectrode array and orientation with respect to the lamellar plane. The entire Figure is a section in the plane of implantation: approximately in a lamellar plane. The plane of the microelectrodes is rotated from the parasagittal plane as follows: (1) The rostra1 end is rotated laterally by about 20” from the parasagittal. (2) The ventral end is tilted medially about 10” from the parasaggital. Microelectrodes are soldered into stainless steel screws and electrical contact is made (via a mercury drop and a stainless steel nut) with an Amphenol connector. The three microelectrodes lie in a plane and if fully extended converge at a reference point in the diencephalon. The coordinates given are anterior to the interaural line and dorsal to a mouth bar which is 5 mm above the interaural line.

of spontaneous unit activity and later confirmed by the form of the evoked field potentials after placement of the stimulating electrodes. The other microelectrodes (when lowered) would then intersect the cellular layers of CA3 and the fascia dentata. Stimulating electrodes were pairs of paraltef 150~ nichrome wires cut off square and cemented together with the tips separated by 0.7-1.0 mm in the longitudinal dimen-

’ The reason for choosing stimulating sites in MSN-DB and EC-AB are obvious. They are the sites of origin of major groups of afferents to the hippocampus. The ventral hippocampal commissure contralateral to the recording sites was chosen in order to stimulate purely commissural connections in a region of maximal efficacy (a dense bundle of axons), while minimizing the possibility of stimulating ipsilateral fibers passing through the fornixifimbria system to and from the area of recording. From the anatomical literature (e.g. GOITL~EXX & COWAN, 1973) one would expect antidromic activation of cells in CA34 and orthodromic responses in CAl, CA34 and the fascia dentata.

sion. These bipolar stimulating electrodes were under el~trophysiolo~c~ control (for maximal response in CA1 at low intensity) in: (1) the ventral hippocampal commissure (WC) about 1 mm from the midline, contrafateral to the recording array, (2) ipilateraffy in the medial entorhinal cortex or angular bundle (EC-AB), and (3) in the medial septal nucleus or the nucleus of the diagonal band of Broca (MSN-DB) (see Fig. lB).’ Hip~amp~ macroelectrodes (150 fl nichrome cut off square and inserted on one or both sides), neck muscle electromyographic (EMG) electrodes (150 pm nichrome), neocortical ei~tr~n~ph~o~aphic (EEG) electrodes (stainless steel screws inserted *through the skull) and a silver ground wire (inserted into the muscle at the side of the head) were brought together in a 7-pin Cannon plug. Ail electrode holders, connectors and several stainless steel screws set in frontal, temporal and interparietal bones were then emkdded in a matrix of dental cement. The scalp was sutured tightly around the headgear and the animai was allowed to recover for 5-7 days before electrophysiological studies were begun.

FIG. 3. Rat H47 ready for stimulation and recording. Three moveable recording microelectrodes have been lowered into the brain. Stimulating connections have been made through Amphenol contacts embedded in the acrylic matrix. A counterweighed 14-channeI headstage is rigidly attached to the head by a Cannon connector which also connects other fixed electrodes. Jumper connections from the microelectrodes to the headstage are rigid wires. See the text for further details.



Hippocampal field potentials in freely moving rats Behavior Long bouts of relatively stable behaviors were required since averaging of many responses was necessary for most of the eI~~ophysiolo~~1 studies. Since the amplitudes of field potential components have been shown to be dependent upon behavior (WINSON & ABZUG, 1978)all of the present data were collected in slow-wave sleep. Electrographic criteria for slow-wave sleep included: hippocampai large-amplitude irregular activity, largeamplitude slow waves in neocortical EEG and small amplitude EMG. Recording


After rats recovered from surgery a counterweighted headstage containing 14 field effect transistors (each in source-follower coloration) was rigidly attached to the Cannon plug on their headgear, The output of each recording electrode was passed through one of these headstages to minimixe movement artifacts. Both recording and stimulating connections were led through a 21-channel mercury commutator so that movement of the rat was unimpeded. Differential recording with respect to a stainless steel screw inserted through the frontal or temporal bone minim~ed stimulus artifacts. Rofloff of amplifiers (Grass 7PSll) was -6 dB/octave below 0.1 Hz and above 10 kHz. For narrow-band (unit) recording the outputs of these amplifiers were passed through active filters whose rolloffs were - I2 dB/octave below 5W Hz. The outputs of 3 ind~ndent const~t~rrent stimulus isolation units (Grass PSIU6) were capacitatively coupled (0.47 pF) to the stimulating electrodes. Isolation units were driven by modified Tektronix PGSOS pulse generators. The three sites (MSN-DB, EC-AB, VHC) were stimulated in rotation at 0.1 Hz. This low rate insured minimal interaction of stimuli (3.3 s inter-stimulus interval) and eliminated the development of long term potentiation (10s between 2 stimuli to the same pathway). Stimulus intensity was usually zz 1 mA; durations were usually 0.1-1.0 ms. In all cases intensities were carefully kept below threshold for an overt behavioral response or interruption of ongoing behavior. The lowest threshold behavioraf response was often scrotal contraction. This had to be watched for carefully, since it could occur at intensities of stimulation which did not even awaken the rat from slow wave sleep. During stimulation the recording microelectrodes were lowered to their final places in the cellular layers. In a few cases microelectrodes were placed in apical dendritic layers. Figure 3 shows rat H47 ready for stimulating and recording. Data collection

violet for confirmation locations.

of recording

and stimulating


All three recording microelectrodes intersected the hippocampus in 9 of the 15 rats. In the rest only two microelectrodes were usable. The field potential threshold was used as an indication of how ‘on beam’ stimulating electrodes were. In 12 rats the threshold for field potentials from VHC stimulation was 5 100 ,uA. In 12 rats the threshold for field potentials from EC-AB stimulation was I 200 4 (6 I 100 fi). In only 5 animals was the threshold for MSN-DB st~ulation s700 fi. There are several reasons for this high threshold for response from MSN-DB. In some early preparations the blood supply to the medial septal nucleus was sometimes disrupted by placements on the midline (see C~YL.E,1975), resulting in infarcts localized around the stimulating electrode. A more lateral placement was adopted in the nucleus of the diagonal band of Broca, but thresholds were still high. Strength/duration analysis indicated that the elements responsible for the most consistent type of field potential from MSN-DB had a chronaxie on the order of 1 ms. Initially, 100 or 200~s pulses had been used, n~sitating high stimulus currents for activation. Pulse durations were increased to 1 ms for MSN-DB stimulation, and threshold intensities decreased. The results of a strength/duration analysis of the three stimulation sites is shown in Fig. 4. The

1400 h !





and analysis

Data was either analyzed on-line with a storage oscilloscope and Polaroid photographs, or recorded on a 7-channel FM tape recorder for later analysis. Only one or two tracks were made through the cell layers with each microelectrode (collecting both unit and field potential data). At the conclusion of experiments, tips of electrode tracks were marked by small lesions or by depositing iron from stainless steel electrodes and utilixing the Prussian, blue reaction. Animals were anesthetized with pentobarbital and perfused with lOy_ formahn in saline; and brains were removed. Brains were allowed to stand overnight in 2% potassium ferrocyanide in formal-saline for the Prussian blue reaction. Coronal frozen sections (50 pm) were cut, mounted on slides and stained with cresyl





FIG. 4. Strength-duration curves for the three stimufating sites. Gpen circles represent the threshold of the field potential in the CA1 pyramidal cell layer to stimulation of the ventral hippocampal commissure (VHC). Filled circles represent the threshold of the field potential in the fascia dentata granule cell layer (FD) to stimulation of the entorhinal cortex (EC). Fifled squares represent the threshold of the field potential in the FD to stimulation of the medial septal nucleus (MSN). Chronaxies are about 200, 300 and 1000 ms for VHC, EC and MSN stimulation, respectively.


S. E. Fox and J. W. RA~L‘(x

chronaxie of about 300~s for EC-AB stimulation is Chronaxies for stimulation of grey matter have often been reported in the range from 200 to 700 ps {RANCK,1975). The chronaxie of about 200~~ for VHC stimulation is unusual. Myelinated fibers in the CNS generally have chronaxies in the range from 50 to 100 ps (RAN~K, 1975). The chronaxie of 1 ms for MSN-DB stimulation suggests that unmyelinated fibers, dendrites, somas or some combination were being stimulated. not surprising.


of jield

again by (4) a long-duration positivity, at least part of which reflects summed extracellular currents due to IPSPs (ANDERSEN et al., 1964n). It is common to speak of an ‘EPSP, population-spike, IPSP sequence’ for the field potentials in response to orthodromic stimulation of monosynaptic excitatory afferents to

the hippocampal formation. This is a useful .but obviously over-simpli~ed explanation of the response. The field potentials have many less consistent components






2== 3.


Figure 5 shows examples of the field potentials recorded in the cellular layers of hip~campus evoked by stimuli to VHC, EC-AB and MSN-DB Many of these are the same as the field potentials described in anesthetized preparations. The largest, most consistent, and thereby most interpretable of these are: (1) the CA1 and CA3 responses to VHC stimulation (Fig. 5A-C), and (2) the fascia dentata response to EC-AB stimulation (Fig. 5E. 2). These have been described in detail in the literature (ANDERSEN & LOMO, 1966; LIIMO, 1971). The field potential in the CA3 pyramidal cell layer in response to VHC stimulation consists of: (1) a small biphasic incoming fiber potential which is not always discernible, (2) a largeamplitude, constant latency negative population-spike presumably representing antidromic invasion of pyramidal cells (ANDERSENet al., 197la), (3) a second population-spike (whose latency decreases with increasing stimulus intensity) presumably due to the orthodromi~ monosynapti~ activation of pyramidal cells (ANDERSENet al., 1966rr), followed by (4) a long-duration positive wave which at least partially reflects summed extracellular currents due to inhibitory postsynaptic potentials (IPSPs) in pyramidal cells (ANDERSEN et al., 1964~). The field potentials in the CA1 p~amidal cell layer in response to VHC stimulation and in the fascia dentata granule cell layer in response to EC-AB stimulation are somewhat similar to each other in general form. They consist of: (1) a small biphasic incoming fiber potential, (2) a positive wave presumably at least partially reflecting summed extra~l1ular currents produced by distal excitatory postsynaptic potentials (EPSPs) (ANDERSEN et u/., 1966~; LOMO, 1971), (3) one or more large-amplitude negative monosynaptic population-spikes (ANDERSEN et al., 1971a), followed

which have yet to be explained

basic neural

cA)I Jj+_


in terms of

However, this does not

(Cl, 2 3

FIG. 5. Forms of field potentials. (A) CA1 response to ventral hippocampal commissure stimulation (350 PA). Trace 1: CA 1 pyramidal cell layer. Trace 2: CA1 apical dendritic layers. Three sweeps. Calibrations: 15 ms, trace 1: 6 mV; trace 2: 12mV. (B) Response in the CA3 pyramidal cell layer to ventral hippocampal commissure stimulation (1.5 mA, different animal from (A)). Three sweeps: Calibrations: 15ms, 4.4 mV. (C) Simultaneously recorded responses to increasing intensities of stimulation of the ventral hippocampal commissure. Intensities: 40, 60, 80, 100, 150, 200, 250, 300, 400 and 500 PA. Trace 1: CA3 pyramidal cell layer. Trace 2: CA1 pyramidal cell layer. Trace 3: _ 300 pm dorsal to CA1 pyramidal cell layer. Calibrations: 6 ms, trace 1: 8.8 mV: traces 2 and 3: 4.4 mV. Note the stability of the latency of the first negative spike in CA3 independent of stimulus intensity. Also note that positivity following population-spikes is inverted to negativity dorsal to CAl, (D) Most common response in the CAL pyramidal cell layer to stimulation of the entorhinal cortex. Stimulus intensity: tOO& Trace 1: narrow band record. Trace 2: wide band record. Calibrations: 6ms, trace 1: 650 pV; trace 2: 5 mV. Note unit action potential on short latency small negative wave. (E) Uncommon responses to stimulation of the entorhinal cortex (1.5 mA) recorded simultaneously. Trace 1: CA1 pyramidal cell layer. Trace 2: fascia dentata granule cell layer. Trace 3: CA3 pyramids cell layer. Three sweeps. Calibrations: 9 ms, 4.4 mV (ail traces). Note that the fascia dentata population-spike seems to be volume conducted to both CA3 and CAl. Two-pronged negativity in CA3 (trace 3) may represent mono- and disynaptic activation. Note also the highly variable latency of the population-spike in CA1 suggesting the response is multisynaptic. The negative wave following the volume conducted fascia dentata populationspike may reflect monosynaptic activation (see also D). (F) Simultaneously recorded responses to stimulation of the medial septal nucleus (2.1 mA). Trace 1: CA1 pyramidal cell layer. Trace 2: fascia dentata granule cell layer. Trace 3: CA3 pyramidal ceil layer. Three sweeps. Calibrations: 15 ms, 4.4 mV fall traces). Note the population-spike in CAL but also small, possibly antidromic (no preceding field potential) population-spike in CA3. See text for further details. Stimulus duration 0.2 ms in all cases, A-F.

Hippocampal field potentials in freely moving rats preclude drawing some conclusions as to the most prominent effects. The interpretation of field potentials other than those discussed above is not as clear cut, but there is some indication of what they may be. The response in the fascia dentata granule cell layer to stimulation of the VHC shows no large population-spike (Fig. 6A, FD) even though short-latency unit discharges are seen in narrow band records. DEADWYLER,WEST, COTMAN& LYNCH (1975) obtained the same result in anesthetized rats. They concluded that although the input is initially excitatory, the density of commissural terminals on dentate granule cells is too low to discharge synchronously large numbers of them. CA1 field potentials in response to EC-AB stimulation are interpretable, but do not entirely agree with the literature. There is an initial negative or biphasic spike which is presumably volume conducted from the fascia dentata. It is consistently followed at only slightly longer latency by a negative wave in the cell layer upon which unit action potentials are routinely recorded. On the basis of latency this may represent direct monosynaptic activation from the EC-AB (see Fig. 5D). In some animals a highly variable relatively large amplitude ‘EPSP, population-spike, IPSP’ sequence follows the initial field potential in the cell layer. This is apparently the tri-synaptic activation described by ANDERSENet al., (1966b) (Fig. 5E, 1). CA3 field potentials in response to stimulation of the EC-AB were often too complex to interpret (Fig. 5E, 3),


presumably due to volume conduction from the nearby fascia dentata. The evoked potentials from the MSN-DB are difficult to interpret. At high intensities stimulus spread is a problem and the antidromic activation of hippocampal units from the lateral septal nucleus stimulation, both antidromic and orthodromic activation of fibers in the pre-commissural fornix and stimulation of fibers of passage from the midbrain are three obvious possible sources of contamination of the data. The most consistent pattern seen in MSN-DB stimulation in four of the five relatively low threshold placements is a CA3/fascia dentata field potential with long latency (lo-30 ms). It consists of a positive wave in the cell layers sometimes followed by a smaller long lasting ( - 100 ms) negative wave. The field potentials recorded in CA1 and CA3 are smaller than in the fascia dentata, and may be volume conducted (see Fig. 6C). Occasionally ‘EPSP, population-spike, IPSP’ sequences were recorded in CA1 in responses to stimulation of the medial septal nucleus (see Fig. 5F), but these could have been due to current spread to the lateral septal nucleus and the antidromic activation of CA3 and CA4 cells with Schaffer collaterals to CAl. Stability of recordings

The effect of large movements (> l/2 mm or so) of one of the recording microelectrodes on the position of the other microelectrodes is difficult to -predict.

(A) Day 0


FIG. 6. Stability of field potentials. Field potentials were recorded simultaneously from CAl, CA3 and the fascia dentata (FD) cellular layers on 3 different days: Day 0 (when the microelectrodes were first lowered into their respective positions), Day 1 (24 h later), and Day 3 (48 h after Day 1). For all sets of responses CA3 is on the top, CA1 is in the center, and FD is on the bottom. Each trace represents three consecutive sweeps. (A) Response to ventral hippocampal commissure stimulation (1.25 mA, 0.1 ms). Calibrations: 15 ms, 5 mV. (B) Response to entorhinal cortex stimulation (925 fi, 0.1 ms). Same calibrations as (A). (C) Response to stimulation of the medial septal nucleus (35Ofi 1.0 ms). Calibrations: 30 ms, 2.5 mV. Note that responses do not change significantly after the first 24 h. The amount of 6OHz noise on the FD electrode increased somewhat during the series, but responses remained constant.


S. E. Fox and J. B. RANCK

Sometimes the form of the field potentials remains stable, indicating that the tip locations of the other microelectrodes remain very nearly stationary with respect to the brain. Often, however, their location changes somewhat. Adjustments in their tip positions of 100 pm or less are usually sufficient to return field potentials to their original form. Once microelectrodes have been lowered into the hippocampus and are not moved more than a few hundred microns, the system is quite stable. Field potentials remain quite constant over long periods of time. Figure 6 shows field potentials evoked from VHC, EC-AB and MSN-DB stimulation immediately after the final lowering of the microelectrodes, and 1 and 3 days thereafter. Since BARMES,MCNAUGHTON,GODDARD, DOUGLAS & ADAMEL(1977) have reported that the amplitude of population-spikes has a circadian cycle, all records for this series were taken at the same time of day. However, BARNESet al. (1977) did not consider the behavior of the animals, so the results are difficult to interpret. No systematic study of circadian cycles was attempted here. It is clear from the data in Fig. 6 that after the first 24 h, neither stimulating nor recording electrodes move very much with respect to the brain, DISCUSSION Co~~iso~

with previous reports

This study demonstrates the similarity of many hippocampal field potentials in freely moving rats to their characteristics in anesthetized preparations. All responses to VHC stimulation and the fascia dentata response to EC-AB stimulation are consistent and correspond to their des~iptjon in the literature (ANDERSENer al., 1966~; ANDER~EN& LIMO, 1966; DEADWYLM~ et al., 1975; LQMO,1971). Changes were observed in the field potentials in response to the stimulation of entorhinal afferents to the hippocampus in different behaviors. These were not systematically explored, but the fragmentary data were in agreement with the results of WIN.X~N& ABZUG (1978). Specifically, population-spikes in the dentate granule cell layer in response to stimulation of .the EC-AB were largest during slow-wave sleep, while extracellular ‘EPSPs’ in the stratum moleculare of the fascia dentata were smallest during slow-wave sleep. In contrast to the pattern in the fascia dentata in response to EC-AB stimulation, both population spikes in the pyramidal cell layer and extracellular ‘EPSPs’ in the stratum radiatum of CA1 in response to activation of commissural afferents were largest in slow-wave sleep. WINSON & ABZUG (1978) demonstrated a similar pattern in CA1 in response to ipsilateral indirect activation of Schaffer collaterals. The surprisingly rare occurrence of trisynaptic activation of CA1 in response to EC-AB stimulation (ANDERSEN et al., 1966b) might be due to electrodes being off the lamellar ‘beam’. An alternative explanation is also possible. An el~trophysiologi~al study by

RAWLINS& GREEN (1977) suggests that the Schaffer collaterals in the rat may not project in a lamellar plane, but in a somewhat temporal direction. To an extent this agrees with the anatomical data (HJORTHSIMONSEN, 1973; SWANSONet al., 1978) which demonstrates considerable spread of Schaffer collaterals (45 mm) along the septotemporal axis, with the center of the terminal field in the stratum radiatum of CA1 being about 0.6-l .2 mm temporal to the origin. One would then expect the trisynaptic response in CA1 in response to EC-AB stimulation to be maximal at a more temporal site than the maximal CA1 monosynaptic response to the same stimulus. As a result, it is not unlikely that we unwittingly selected against a maximal trisynaptic response by implanting electrodes under pentobarb~tal anesthesia (blocking multisynaptic responses), and using as the criterion for correct placement of EC-AB stimulating electrodes a maximal short latency (monosynaptic) activation of CAl. In CA3 the latency differences between mono- and disynaptic responses are not great enough to definitely attribute activation to one or the other, without further experimental procedures (e.g. lesions of dentate granule cells). WIN~ON & ABZU~; (1978) report the occurrence of di- and trisynaptic responses in unanesthetized rats in CA3 and CA1 respectively, but do not comment on how common they were. In any case, the problem deserves further study. It is hard to account for the di~culty of obtaining low-threshold field potentials from MSN-DB. The above description of these potentials does not entirely correspond to the description found in the literature. Population-spikes were never observed in response to stimulation of the MSN-DB at moderate intensities (up to 1 mA). ANDERSEN,BRULAND& KAADA(1961) report population-spikes in the granule cell layer in response to medial septal nucleus stimulation in anesthetized rabbits. Again, urethane/chloralose additional studies are in order. Advantages qf rhe preparatior~

These data demonstrate that one can record field potentials in the hippocampal formation of freely moving animals of equal quality and very similar form to those recorded from anesthetized animals. It is somewhat surprising that perfectly normal electrophysiological responses can be recorded from a preparation with such a large number of stimulating and recording electrodes chronically implanted. One might have expected more damage. This preparation has many advantages. The animal maintains its own body temperature, blood gases, intracranial environment and other variables which must be externally controlled in acute preparations. The state of the animal is unambiguous both in terms of ‘health’ and functional brain state. Anesthesia or paralysis may simply mask changes in the normal internal functional state of the brain, which would otherwise be observed as a behavioral change. Contrary to expectations, stimulus artifact is not a major problem.

Hippocampal field potentials in freely moving rats Cross talk was minimai, presumably because connections had to be relatively short and neatly organized for the chronic preparation. One can record stable field potentials over many days and presumably weeks once the electrodes have been lowered into place. Since responses are stable one is able to study the effect of a variety of ex~rimen~ manipulations which may require several days to accomplish. Changes in the electrophysiological characteristics of this system can suggest testable hypotheses regarding the cellular mechanisms of those changes. Changes in synaptic mechanisms would imply changes in the information processing occurring between different behavioral ‘states’ controlled by the experimenter. Where time is limited and parameters are constant changing (e.g. in drug or seizure studies) simultaneous data from all of these systems is important, and cannot be obtained in chronic preparations with single electrodes. Single-unit activity can be recorded from two or three units simult~eously from the same or different areas of the hippocampus to. study interactions between them, or with respect to some other variables. Studies of the interactions of the effects of stimulating different afferent populations can be done. Microstimulation through recording electrodes can be used to demonstrate orthodromic activation by intrinsic afferents, and characterize projection cells antidromically. This preparation also allows the construction of laminar profiles of field potential components. There are, however, some disadvantages of chronic prep~ations. There is ~mewhat less ~exibility as compared with acute preparations. Usually only a one-dimensional movement of the electrodes is possible, so recording and stimulating sites and angles of approach should be chosen with care. Visual control over the placement of electrodes is lost. One must depend upon stereotaxic coordinate systems which are somewhat variable between animals. Presumably only extracellular recording is possible in a freely moving animal. Even if glass micropipettes ‘floating’ in the brain were used, tips would have to be large in order to keep tip impedance low enough to avoid movement artifacts. It is valuable to be able to use el~~ophysiologicai tools to test for changes in the patterns of information processing in a system in different behaviors, and when one does this, one must control those behaviors as carefully as possible. To the extent that ‘brain state’ is a function of behavior, behavior is a con-


founding variable in studies which do not directly control it. This variable may only be masked by procedures such as paralysis, brainstem transection or even general anesthesia. If one is not interested in the effects of behavior on the system under study, then it should be maintained in as steady a state as is possible, by using unanesthetized animals whose behavior is controlled. Stereotyped global behaviors which are homogeneous and easily monitored should be chosen, (A motionless and awake state is not welldefined and therefore, is not a particularly good category, since it includes freezing, drowsiness and several levels of arousal in between.) These behaviors should be able to be brought under experimental control so that they occur at the discretion of the experimenter, not the subject. Controls for ‘level of arousal’ and circadian or hormonal cycles should be included where possible. Many others have recorded evoked field potentials and single unit activity in freely moving animals, but none prior to this study have combined all of the following techniques: (1) stimulation of multiple inputs with known action; (2) recording from multiple sites with known interconnections; (3) simultaneous recording of field potentials and single units; with (4) control of the behavior of the animals. All of these techniques are used not just to be encyclopedic but to increase the interpretability of any effects which are observed. In a subsequent paper (S. E. Fox & J. B. RANCK,unpublished observations) we will describe a first step toward applying these techniques to elucidate the functional relationships between two broad categories of units in the hippocampus: complex-spike cells and theta cells. One can imagine how these techniques might be further applied to sub-categories of these cefl types to further define functional relationships between single units and changes in those relationships associated with changes in behavior.

Acknowledgemems---This work was supported by Grants NS 10970, NS 12664 and NS 14497 from the National Institutes of Health, and BNS 77-09375 from the National Science Foundation to J. B. RANCJC, JR and N.I.H. Grant NS 05773 to V. E. AMASSIAN.

REFERENCES ANDER~EN P., BLACKSTAD T. W. & LIMO T. (1966a) Location and identification of excitatory synapses on hippocampal pyramidal cells. Expl Bruin Res. 1, 236248. ANDER~N P., BLISST. V. P. & SKREDEK. K. (1971~~)Unit analysis of hippocampal population spikes. Expl Bruin Res. 13, 208-221. ANDER~ENP., BLISS T. V. P. % SKREDEK. K. (1971b) Lamellar organization of hippocampal excitatory pathways. Expl Brain Res. 13, 222-238. ANLIER~EN P., BRULANDH. & KAADA B. R. (1961) Activation of the dentate area by septal stimulation. Actu physiol. sand. 51, 17-28.


S. E. Fox and J. H. RANCK

ANDERsENP., ECrLESJ. C. & LOYNINC~ Y. (1964~) Location of postsynaptic inhibitory synapses on



J. Neurophysiol. 27, 592.-607. ANDERSENP., Ec‘CLES J. C?. L LOYNIN~~ Y. (1964h) Pathway of postsynaptic p~~s~u~.27, 608 -6 IO.

inhibition in the hippocampus. ,I. &‘enro-

ANDERsEN P., HOWW’ISTB. 8~VOORHOEVE P. E. (1966hl Excitatory synapses on hippocampal apical dendrites activated by entorhinal stimulation. Ac,ta physiol. scund. 66, 461-472. ANDERSENP. & LflMO T. (1966) Mode of activation of hippocampal pyramidal cells by excitatory synapses on dendrites, Expl. Brain Res. 2, 247.-260.

BARNEYC. A., MCNA~JGHTONB. L., GODDARDG. V., DOIJGLASR. M. & ADAMEC R. (1977) Circadian rhythm of synaptic excitability in rat and monkey central nervous system. Science, N.k: 197, 91-92. BLACKSTAD T. W., BRINK K., HEM J. & JECNE B. (1970) Distribution of hip~campal mossy fibers in the rat. An experimental study with silver impregnation methods. J. camp. Neural.138,433-449. BLISST. V. P. & GARDNER-MEDWINA. R. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the unanesthetized rabbit following stimulation of the perforant path. J. Physiol., Lond. 232, 357-374. CASEY K. L. & KEENEJ. J. (1973) Unit analysis of the effects of motivating stimuli in the awake animal: Pain and self-stimulation. In Brain G’nir Acti~if): During Behacior (ed. PHILLIPSM. I.), pp. 115-129. Thomas, Springfield. COYLE P. (1975) Arterial patterns of the rat rhinencephalon and related structures. Expf Neural. 49, 671-690. DEADWYLERS. A., WEST J. R., COTMANC. W. & LYNCH G. S. (1975) A neurophysiological analysis of commissural projections to dentate gyrus of the rat. J. Neurophysiol. 38, 167-184. DOUGLASR. M. & GODDARDG. V. (1975) Long-term potentiation of the perforant path-granule cell synapse in the rat hippocampus. Brain Res. 86, 205-215. GOTTLIEBD. I. & COWANW. M. (1973) Autoradiographic studies of the commissural and ipsilateral association connections of the hippocampus and dentate gyrus of the rat. J. camp. Neuroi. 149, 393-421. HIORT~-S~~ON~~ A. (1973) Some intrinsic connections of the hippo~ampus in the rat: an experimental analysis. J. camp. Neural. 147, 1455161. L!JMOT. (1971) Patterns of activation in a monosynaptic cortical pathway: The perforant path input to the dentate area of the hippocampal formation, Expl Brain Res. 12, 184.5. LORENTEDE Nb R. (1934) Studies on the structure of the cerebral cortex--II. Continuation of the study of the Ammonic system. J. Ps~chol. Neural., Leipzig 46, 113-177. RAM& Y CAJAL S. (1893). Uber die feinere struktur des Ammonshornes. Trans. by A. von Kolliker, 2. wiss. Zool. 56, 613-663. RANCK J. B., JR (1973a) A moveable mi~oeiectrode for recording from single neurons in unrestrained rats. In Brain linit Actil;ity During Behavior (ed. PHILLIPSM. I.), pp. 76-79. Thomas, Springfield. RANCK J. B., JR (1973h) Studies on single neurons in dorsal hippocampal formation and septum in unrestrained rats-l. Behavioral correlates and firing repertoires. Expl Neural. 41, 461-531. RANCK J. B., JR (1975) Which elements are excited in electrical stimulation of mammalian central nervous system: A review. Brain Res. 98, 417-440. RAWLINSJ. N. P. & GREEK K. F. (1977) Lamellar organization of the rat hippo~ampus. Expt Brain Res. 28, 335-344 SCHAFER K. (1892) Beitrag zur histologie der Ammonshorn formation. Arch. mikrosk. Anat. E~tw~ech. 39, 611-632. STEWARD 0. (1976) Topographic organization of the projections from the entorhinal area to the hippocampal formation of the rat. J. camp. Nc~urol. 167, 285 314. SWANSONL. W. & COWAN W. M. (1976) Autoradiographic studies of the development and connections of the septal area in the rat. In Adduces in Behauiorul Biology. Vol. 20: The Septal Nuclei (ed. DEFRANCEJ. F.), pp. 37-64. Plenum, New York. SWANSCIF: L. W. & COWAN W. M. (1977) An autoradiographic study of the organization of the efferent connections of the hippocampal formation in the rat. J. camp. Neural. 172, 49-84. SWANSON I_.W., WYSSJ. M. & COM’ANW. M. (1978) An autoradiographic study of the organization of intrahippocampal association pathways in the rat. J. camp. Neural. 181, 681.-715. VANDERWOLFC. H. (1969) Hippocampal electrical activity and voluntary movement in the rat. EleCrroeucePh. c/in. Nrurophrsiol. 26, 407-418. WINSONJ. (1974) Patterns of hippocampdl theta rhythm in the freely moving rat. ~~ec~~oenceph. c&g. ~e~ro~h~s~a~. 36, 291.-301. WINSON J. & ABZUG f. (1978) Neuronal transmission through hippocampal pathways dependent upon behavior. J. NeGophysiol. 41, 716732. ZIMMERJ. (1971) Ipsilateral aRerents to the commissural zone of the fascia dentata, demonstrated in decommissurated rats by silver impregnation. J. camp. Neural. 142, 393416. (Accepred 15 Murch 1979)