INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Brain rhythms occurring at various frequencies appear to represent a basic mode of operation of large neural networks, allowing for the selective processing of information related to behavioral and perceptual states. Theta-band oscillation and synchrony in the hippocampal formation (HPC) and related limbic structures is recorded as an extracellular field potential consisting of a sinusoidal-like waveform with an amplitude 2 mV and a narrowband frequency range of 3-12 Hz in mammals. The asynchronous activity termed large-amplitude irregular activity (LIA) is an irregular waveform with a broadband frequency range of 0.5-25 Hz (Leung et al. 1982). Many populations of cells in the HPC and related structures exhibit discharge properties that are precisely related to hippocampal theta field activity. Such theta-related cells comprised two distinct populations termed theta-ON and theta-OFF, first described in acute preparations using extracellular recordings by Colom et al. (1987), followed by a detailed cell classification paper by Colom and Bland (1987) and subsequently used to classify theta-related cells in the HPC in a number of studies (Bland and Colom 1988, 1989; Bland et al. 1996; Colom et al. 1991; Konopacki et al. 1992; Smythe et al. 1991). Theta-ON and -OFF cells have also been recorded in the medial septal nucleus and nucleus of the diagonal band of Broca (MS/vDBB) (Bland et al. 1990, 1994; Colom and Bland 1991; Ford et al. 1989), the entorhinal cortex (Dickson et al. 1994, 1995), cingulate cortex (Colom et al. 1988), caudal diencephalon (Bland et al. 1995; Kirk et al. 1996), rostral pontine region (Hanada et al. 1999), the superior colliculus (Natsume et al. 1999), the basal ganglia (Hallworth and Bland 1999), the red nucleus (A. Dypvik and B. H. Bland, unpublished data), and the neocortex (Lukatch and MacIver 1996; but see review by Buzsaki 2002 that does not acknowledge the literature on theta-ON and theta-OFF cells). The preceding data suggest, because many regions of the brain from the lower brain stem to the cerebral cortex display "theta-related" neuronal activity, that theta-ON and -OFF cells may represent a general organization of the cellular mechanisms underlying "theta band" oscillation and synchrony (Bland 2000; Bland and Oddie 1998). The morphological identity of theta-ON and -OFF cells is therefore crucial to the understanding of the cellular interactions involved in the generation of theta field activity. Earlier studies by Fox and Ranck (1975, 1981) provided indirect evidence that theta cells (thetaon cells in our scheme) in the HPC were interneurons, a view supported by some recent work on identified cells (see review by Buzsaki 2002). On the other hand, Bland and Colom (reviewed in Bland and Colom 1993) have proposed, also based on indirect evidence, that a subpopulation of HPC projection cells (pyramidal and granule cells) were theta-ON cells and a subpopulation of HPC interneurons were theta-OFF cells.

Theta-band oscillations may also be recorded intracellularly in some populations of cells in the HPC, during the simultaneous occurrence of the extracellular theta field oscillations. In agreement with previous work (Leung and Yim 1991), we have adopted the term membrane potential oscillations (MPOs) to designate the slow intracellular oscillations that occur at theta frequencies in subsets of hippocampal cells. The occurrence of MPOs in HPC pyramidal cells, dentate granule cells, and interneurons has been well documented (Bland et al. 1988; Chapman and Lacaille 1999; Fox 1989; Fox et al. 1983; Fujita and Sato 1964; Konopacki et al. 1992 Leung and Yim 1986, 1988, 1991; Leung and Yu 1998; MacVicar and Tse 1989; Munoz et al. 1990; Nunez et al. 1987, 1990a-c; Ylinen et al. 1995). At present, there is little agreement on how intracellular theta (MPOs) is generated, some arguing that inhibitory postsynaptic potentials (IPSPs) make the major contribution, whereas others argue the case for excitatory postsynaptic potentials (EPSPs) or intrinsic mechanisms (Bland and Colom 1993). As well, there is a lack of understanding of the relationship between MPOs and the rhythmic discharge properties of HPC cells. Do the mechanisms underlying the rhythmic discharge properties of the cell produce the MPOs or do the mechanisms underlying MPOs control the rhythmical cell discharges? The objectives of the present study were to determine the contributions made by MPOs in the control of the rhythmic discharge properties of HPC cells meeting the criteria for being classified as theta-ON and -OFF cells and to establish their morphological identity by intracellular labeling. To do this, we investigated the effect of four levels of depolarizing and hyperpolarizing constant current injections in HPC cells in vivo, applied during the spontaneous occurrence of both theta and LIA HPC field activities, on cellular MPOs and the rate and pattern of cell discharges. Whenever possible, after completion of the experimental protocol, an attempt was made to label cells through the intracellular injection of Neurobiotin.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects and surgical procedures

The data were obtained from 38 male black-hooded rats (125-130g) supplied by the Life and Environmental Sciences Animal Care Facility at the University of Calgary. The rats were initially anesthetized with Halothane while tracheal and jugular cannulae were inserted. Halothane was then discontinued, and urethan was administered via the jugular cannula to maintain an appropriate level of anesthesia during the remaining surgical and experimental procedures. The rats were placed in the stereotaxic instrument with the plane between Bregma and lambda leveled to horizontal. Body temperature was maintained at 37°C, and heart rate was monitored constantly throughout the experiment. An uninsulated tungsten wire placed in the cortex, anterior to Bregma, served as the indifferent electrode and the stereotaxic frame was connected to ground. A tungsten microelectrode (0.2-0.5 M) for recording hippocampal (HPC) field activity was placed in the right dorsal hippocampal formation in the dentate molecular layer (3.0-3.3 mm posterior to Bregma, 2.0-2.5 mm lateral to the midline and 2.4-2.8 mm ventral to the dural surface). Intracellular recordings in the left HPC were made at the same posterior and lateral coordinates, starting at the alvear surface and continuing ventral to the lower (endal) blade of the dentate granule cells. Procedures for preparing the site for intracellular recordings described previously (Bland and Konopacki 2000) are briefly the following: a dental acrylic circle was made on the skull encompassing the whole recording area on the left side of the skull. This served to form the pool for holding either glycerine or, later when recording, distilled water. At this point a bone window was drilled out using a 1/4 carbide drill bit, the dimensions ~3.0 mm long by 2.0 mm wide. The well was filled with glycerine to keep the brain moist. Using the same carbide drill bit and holding the bone fragment down with a pair of forceps, a hole was drilled in the center of the bone and rough edges removed from the bone fragment by polishing with the drill bit. Dura was removed using a No. 26 syringe with a small hook in the end and a pair of fine forceps, and the position of the large blood vessels observed because this determined the exact location of the bone fragment. With the layer of glycerine in the pool and using fine forceps, the bone fragment was placed upside down in the hole in the skull with the long side oriented to the midline and gently pushed under the skull at the midline enough to allow it to be pulled back under the lateral edge. The location of the bone fragment was finessed with respect to blood vessels, as discussed in the preceding text.

Electrophysiological recordings

Intracellular recordings in HPC cells were made with glass microelectrodes (80-120 M) filled with 2 M potassium acetate and 2% Neurobiotin (Vector Laboratories). Details of these procedures have been published previously (Bland and Konopacki 2000; Konopacki et al. 1992). Cells meeting the criteria for classification as phasic theta-ON or phasic theta-OFF were selected for subsequent analysis. Theta-ON cells were defined as cells that increased their activity during theta field activity as reflected by an overall mean increase in discharge rate or as a linear positive increase in discharge rate in relation to increasing frequencies of simultaneously-recorded theta field activity. Theta-OFF cells were defined as cells that decreased their activity during theta field activity as reflected by an overall mean decrease in discharge rate (to 0 in many cases) or as a linear negative increase in discharge rate as theta frequency declines. A further criteria relates to the pattern of cell discharges. A given theta-related cell discharges in one of two characteristic patterns during theta field activity. A rhythmic discharge pattern was defined as two or more cell discharges occurring per extracellular theta wave and was termed phasic because the discharges occurred with a consistent phase relation to each cycle of theta field activity. The second pattern is either regular or irregular discharges termed tonic because they consisted of a nonrhythmical discharge pattern with no observable phase relation to theta field activity. Further criteria were developed on the basis of intracellular recordings of theta-related cells in the hippocampal formation. At theta field onset, phasic theta-ON cells underwent a depolarizing shift of the membrane potential along with prominent theta frequency MPOs that were highly coherent with the field oscillations. Rhythmic cell discharges occurred on the positive peaks of the MPOs. At theta field onset, phasic theta-OFF cells underwent a hyperpolarizing shift of the membrane potential, prominent MPOs appeared that were highly coherent with the field oscillations, and cell discharges ceased. As theta frequency slowed, the cell would discharge single spikes, gradually making a brief transition to rhythmic cell discharges (Bland et al. 1988; Konopacki et al. 1992). Signals were lead through an active bridge circuit (Axon Instruments, Axoclamp 2A) allowing simultaneous injection of current and measurement of membrane potential (V_m). The bridge balance was monitored and adjusted as necessary throughout the recording procedures. These and all other signals were displayed on a digital oscilloscope (Tektronix TDS 420) and stored on an FM tape recorder (Teac XR 30) for subsequent off-line data analysis. Once a recording was considered stabilized the experimental protocol was initiated. First, recordings were taken during the spontaneously occurring HPC field activities of synchrony (theta) and asynchrony or large-amplitude irregular activity (LIA), ensuring that 1 min in duration of each type was accumulated. Next a series of hyperpolarizing (100, 200, 300, and 400 pA) and depolarizing (100, 200, 300, and 400 pA) constant current pulses were administered during 30 s of each of the spontaneously occurring HPC field activities of theta and LIA. In our first such series of experiments, the current pulses were administered in random order. In later experiments, the protocol consisted of administering the hyperpolarizing pulses in ascending magnitude, followed by the depolarizing pulses in ascending magnitude. Analysis revealed no differences between the two stimulation protocols in terms of their effects on MPOs and cell discharge rate and pattern. After completion of this part of the protocol, a series of short-duration hyperpolarizing current pulses (100-ms duration, 100, 200, 300, and 400 pA) and depolarizing current pulses (100, 200, 300, and 400 pA) were administered in ascending magnitude, respectively, for purposes of carrying out standard electrophysiological measurements. Procedures for intracellular staining were modified slightly from the procedures previously described (Kita and Armstrong 1991). Positive pulses were applied (5 nA, 2 Hz, 100-ms duration) for 1-10 min and the electrode was left inside the cell for 2-3 min before it was withdrawn.

Data analysis

Data were analyzed using the Axon Instruments Axoscope 7, Microcal Origin 4.1, and Data Waves 6.1 software. Spike height and duration was determined from the first action potential evoked at threshold levels of depolarization. Resting membrane potentials were defined as an average of a series of a minimum of 30 membrane potentials measured between the spike discharges occurring during the LIA field conditions. The membrane potential values for the spontaneous theta control (no current) condition and the spontaneous theta conditions plus intracellular current injections were defined as average membrane potentials measured at the midpoint of the MPOs when present. Input resistance was provided as the slope of the linear regression line fitted through the linear portion of the current-voltage plots derived from the family of hyperpolarizing and depolarizing current pulse injections. MPOs, defined as the slow intracellular membrane potential oscillations occurring at extracellular theta frequencies, were measured at the positive and negative peaks (see Fig. 2A) using the cursor facility in the Axoscope 7 program, and calculated as the statistical average of a minimum of 50-60 individual MPOs measured in each experimental condition. Additional data analyses included: fast Fourier transforms of the field activities of theta and LIA, autocorrelation histograms and first-order interval spike histograms of the cell discharges during theta and LIA, cross-correlation's between MPOs and theta field activity, and quantification and statistical analysis using t-tests of the spike discharges occurring during each experimental condition.

Imaging techniques and cellular identification

At the termination of a recording session, all animals were administered an overdose of pentobarbital sodium and perfused transcardially with 50 ml phosphate-buffered saline (PBS, 0.1 M, pH 7.4) followed by 300 ml of 4% paraformaldehyde in PBS. The brain was removed and immersed in the fixative for an additional 4 h and then cryoprotected, overnight, in a solution of 30% sucrose in PBS. The brains were frozen, and coronal sections were cut at 50 µm on a sliding microtome. Every section through the hippocampal formation was collected in PBS and then incubated for 2 h in a 1:500 solution of avidin-HRP in PBS. After three washes in PBS, the Neurobiotin-avidin-HRP complex was visualized by incubating the sections in a chromogen solution consisting of 10 ml 0.1 M Tris-buffered saline containing 5 mg diaminobenzidine, 40 µg nickel ammonium sulfate, and 10 µl 30% hydrogen peroxide. When the reaction was complete, the sections were washed three times in PBS and then mounted onto gelatin-coated glass slides and allowed to dry. The sections were dehydrated in an ascending series of ethanol, cleared in xylene and cover slipped using Permount. The labeled cells were identified using a Zeiss Axioplan 2 microscope at ×20, and digital images were captured and serially reconstructed using Open lab (v. 3.0, Improvision) and Adobe Photoshop (v. 6.0, Adobe) running on an Apple G4.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cells classified as phasic theta-ON cells

CA1 PYRAMIDAL CELLS. We describe here the results of intracellular recordings of cells (n = 18) in the HPC of urethan-anesthetized rats, classified as phasic theta-ON cells (comparison of group mean discharge rates for theta vs. LIA significant at P < 0.05, paired t-test). These cells discharged in a rhythmic pattern during theta but not during LIA, and MPOs were recorded only during the theta field condition, their onset signaled by a 5- to 10-mV depolarizing shift in the membrane potential. Each rhythmic discharge occurred on the depolarizing phase of the MPO and was phase locked to the simultaneously occurring extracellular theta waves. The average resting membrane potential was 62.6 ± 0.91 (SE) mV, average spike height 62.5 ± 1.8 mV, and average input resistance 31.3 ± 1.6 M. Of the 18 cells classified as phasic theta-ON cells, 7 were successfully labeled and morphologically identified as CA1 pyramidal cells. Unlabeled cells were inferred to be pyramidal cells as well based on their anatomical location in the CA1 pyramidal layer and their electrophysiological characteristics, which did not differ from those of the labeled cells.

View larger version (39K): [in this window] [in a new window]   Fig. 1. A: relationships between spontaneously occurring hippocampal (HPC) field activity and CA1 pyramidal cell discharges in the no current control condition. Top: the HPC field activity recorded from the molecular layer of the dentate region, bottom: the discharge pattern of a CA1 pyramidal cell (102) classified as a phasic theta-ON cell (positivity up in all traces). The 1st half of the panel shows the rhythmic cell discharge pattern and membrane potential oscillations (MPOs) occurring during HPC theta and the 2nd half of the panel shows the irregular cell discharge pattern occurring during HPC large-amplitude irregular activity (LIA). Note the absence of MPOs. Gamma oscillations (mean frequency = 70 Hz) were visible in all cell traces during the occurrence of both theta and LIA. B: fast Fourier transform of the theta field in A, top left. C: autocorrelation histogram of the rhythmic cell discharges associated with the theta field in A, top left. The intervals of regular peaks in the histogram (340 ms) coincided with the theta field frequency. D: interspike interval histogram of cell discharges during theta field activity. The bimodal distribution shows the intraburst and interburst intervals respectively, indicating a rhythmic discharge pattern. E: fast Fourier transform of the LIA field activity in A, top right, showing the absence of a peak in power at any narrowband frequency. F: autocorrelation histogram of the irregular cell discharges associated with the LIA field in A, top right, showing the absence of peaks. G: interspike interval histogram of cell discharges during LIA field activity showing the absence of a bimodal distribution, indicating the irregular cell discharge pattern.

View larger version (36K): [in this window] [in a new window]   Fig. 2. A: segment of spontaneous theta from left half of Fig. 1A with cell trace amplified to show the slow MPO and method of measurement using cursors (MPO = 8.6 mV; positivity up in all traces). B: segment of spontaneous LIA from Fig. 1A, right half, showing absence of MPOs. C: intracellular depolarizing current pulse (200-pA, 100-ms duration) applied to cell shown in Fig. 1A during spontaneously occurring theta. Note the simple spike pattern. D: current- voltage plot of the cell shown in Fig. 1A, input resistance = 22.5 M.

View larger version (51K): [in this window] [in a new window]   Fig. 3. The control (no current) condition (top middle) and hyperpolarizing (left) and depolarizing (right) intracellular constant current injections during the occurrence of spontaneous theta field activity, highlighting the effects on the average value of the membrane potential, the average MPO amplitudes and cell discharge patterns of a CA1 pyramidal phasic theta-ON cell (154). Spikes truncated (positivity up in all traces).

View larger version (24K): [in this window] [in a new window]   Fig. 4. A: graph summarizing the effects of hyperpolarizing and depolarizing constant current injections and the control (no applied current) condition on the cell discharge pattern and rate during the theta and LIA conditions for cell 154 shown in Fig. 3. B: graph summarizing the effects of hyperpolarizing and depolarizing constant current injections and the control (no applied current) condition, on the MPO amplitudes during the theta condition for cell 154. Data in both A and B are plotted against the average membrane potential values produced by the current injections and the average membrane potential value occurring during the control (no applied current) condition. Standard error bars showing positive half only in A.

View larger version (23K): [in this window] [in a new window]   Fig. 5. Graph summarizing the group data for all CA1 pyramids classified as phasic theta-ON cells (n = 18), plotted against the average membrane potential value for the control (no applied current) spontaneous theta condition and the average membrane potential values produced by the constant current injections. The left axis shows the mean cell discharge rate and the right axis shows the average MPO amplitude. The plot of MPO amplitudes revealed an inverted U-shaped curve asymmetrically distributed around the value of the membrane potential during the spontaneous theta condition. Rhythmic cell discharges only occurred during theta and only when MPO amplitudes were in the range of ~4.0-7.0 mV. All data plotted as means ± SE.

View larger version (40K): [in this window] [in a new window]   Fig. 6. Relationships between spontaneously occurring HPC field activity and CA1 pyramidal cell (226) discharges in the no current control condition. Left, top: is the HPC field activity recorded from the molecular layer of the dentate region; bottom: the discharge pattern of a CA1 pyramidal cell classified as a phasic theta-ON cell. (positivity up in all traces). The 1st half of the panel shows the rhythmic cell discharge pattern and MPOs occurring during HPC theta, and the 2nd half of the panel shows the irregular cell discharge pattern occurring during HPC LIA. Note the absence of MPOs. Gamma oscillations (mean frequency = 70 Hz) were visible in all cell traces during the occurrence of both theta and LIA. Right: the intracellular injection of Neurobiotin into this CA1 pyramidal cell resulted in the labeling of 2 CA1 pyramids. These results were observed when injecting 4 of 7 (57%) successfully labeled cells. Top: a low-power magnification showing the location of the cells in the CA1 cell pyramidal layer. Bottom: a higher power magnification showing the details of the cell morphology Calibration bar = 50 µ.

BISTRATIFIED INTERNEURONS. The left side panel in Fig. 7 shows an example of one of two bistratified cells (216) recorded during the simultaneous occurrence of either HPC theta (left half of the panel) or LIA (right half of the panel). We were unable to measure MPOs of these cells during the occurrence of either theta or LIA field activity. Both cells had a pronounced afterhyperpolarization (see Fig. 7) and showed the highest number of rhythmic cell discharges per theta wave (mean = 6.5 ± 1.2 Hz) compared with all other cells in the study. The mean discharge rate during theta field activity was 24 ± 3.1 Hz and during LIA was 20 ± 2.7 Hz. The right side panel shows that cell 216 was located in the CA1 pyramidal layer (as was the other cell) and was identified as a bistratified interneuron.

View larger version (46K): [in this window] [in a new window]   Fig. 7. Relationships between spontaneously occurring HPC field activity and cell discharges of a bistratified cell (216) classified as a phasic theta-ON cell, in the no current control condition. Top panel, top trace: the HPC field activity recorded from the molecular layer of the dentate region, bottom trace: the discharge pattern of the cell (positivity up in all traces). The 1st half of the panel shows the rhythmic cell discharge pattern occurring during HPC theta and the 2nd half of the panel shows the irregular cell discharge pattern occurring during HPC LIA. Note the absence of MPOs. Bottom left: a single spike at a fast sweep to illustrate the prominent afterhyperpolarizing potential. Bottom right: the intracellular injection of Neurobiotin resulted in the labeling of a bistratified interneuron; top: low-power magnification showing the location of the cell in the CA1 cell pyramidal layer; bottom: is a higher-power magnification showing the details of the cell morphology. Calibration bar = 50 µ.

View larger version (46K): [in this window] [in a new window]   Fig. 8. The control (no current) condition (top middle) and hyperpolarizing (left) and depolarizing (right) intracellular constant current injections during the occurrence of spontaneous theta field activity, highlighting the effects on the average value of the membrane potential, the average MPO amplitudes and cell discharge patterns of bistratified phasic theta-ON cell 216. Spikes truncated (positivity up in all traces).

View larger version (24K): [in this window] [in a new window]   Fig. 9. Graph summarizing the effects of hyperpolarizing and depolarizing constant current injections and the control (no applied current) condition, on the cell discharge pattern and rate during the theta and LIA conditions for cell 216 shown in Figs. 7 and 8. Data are plotted against the average membrane potential values produced by the current injections and the average membrane potential value occurring during the control (no applied current) condition. SE bars showing positive half only.

Cells classified as phasic theta-OFF cells

We describe here the results of intracellular recordings of cells (n = 22) in the HPC of urethan-anesthetized rats classified as phasic theta-OFF cells. These cells were mostly silent during theta and discharged during LIA while MPOs were recorded only during the theta field condition, their onset signaled by a 5- to 10-mV hyperpolarizing shift in membrane potential. The cell discharges that did occur during low-frequency theta discharged on the depolarizing phase of the MPOs and were phase locked to the simultaneously occurring extracellular theta waves. Nine of the phasic theta-OFF cells were recorded from the stratum pyramidale, 11 were recorded from the dentate granule cell layer, and 2 were recorded in the hilus. The average resting membrane potential of the CA1 layer basket cells was 66.6 ± 0.91 mV, average spike height 60.5 ± 1.8 mV, and average input resistance 29.4 ± 6.1 M. Of the nine cells recorded in the CA1 pyramidal cell layer and classified as phasic theta-OFF cells, one was successfully labeled and morphologically identified as a CA1 pyramidal layer basket cell. Unlabeled cells were inferred to be CA1 pyramidal layer basket cells as well based on their anatomical location in the CA1 pyramidal layer, electrophysiological characteristics, and responses to current manipulations, which did not differ from that of the labeled cell. The average resting membrane potential of the dentate layer cells was 65 ± 4.91 mV, average spike height 58 ± 1.8 mV, and average input resistance 32 ± 3.1 M. Of the 11 cells recorded in the dentate layer, 3 were labeled and identified as dentate granule cells. Unlabeled cells were inferred to be granule cells as well, based on their anatomical location in the denate granule layer and their electrophysiological characteristics, which did not differ from that of the labeled cells. Of the two cells recorded in the hilus, one was labeled and identified as a mossy cell interneuron.

CA1 PYRAMIDAL LAYER BASKET CELLS. Figure 10, A and B, shows examples of a CA1 layer basket cell (229) recorded during the simultaneous occurrence of either LIA (left half of the panel) or HPC theta (right half of the panel). MPOs were recorded during the theta field condition only. In Fig. 10A, theta frequency was slightly higher (4.3 Hz) and the cell failed to discharge. In Fig. 10B, theta frequency was lower (3.4 Hz) and the cell began to discharge, first single discharges and then a doublet. Figure 10C shows the response of this cell to a 200-pA depolarizing intracellular current pulse, applied during the occurrence of spontaneous theta field activity. Figure 10D reveals that the intracellular injection of Neurobiotin into cell (229) resulted in the labeling of a cell identified as a basket cell interneuron. The top panel is a low-power magnification showing the location of the cell in the CA1 cell pyramidal layer. The bottom panel is a higher-power magnification showing the details of cell morphology (axons not drawn). The axons of the basket cell were visually confirmed to form a plexus restricted to the CA1 pyramidal layer. Figure 10E presents the current-voltage plot along with the linear regression curve (input resistance = 29.4 M).

View larger version (38K): [in this window] [in a new window]   Fig. 10. A and B: relationships between spontaneously occurring HPC field activity and the cell discharges of a CA1 layer basket cell (229) classified as a phasic theta-OFF cell, in the no current control condition. Top trace in each panel is the HPC field activity recorded from the molecular layer of the dentate region; bottom trace is the discharge pattern of the cell (positivity up in all traces). The 1st half of the panels in A and B show the irregular cell discharge pattern occurring during HPC LIA. Note the absence of MPOs. The 2nd half of the panel in A shows the complete cessation of cell discharges during theta field activity with at a higher frequency (4.3 Hz) and the occurrence of MPOs. The 2nd half of the panel in B shows that as theta field frequency slowed to 3.4-Hz, phase-locked cell discharges began to occur. Again, MPOs were recorded during theta field activity. C: intracellular depolarizing current pulse (200-pA, 100-ms duration) applied to the cell during spontaneously occurring theta. D: the intracellular injection of Neurobiotin into cell 229 resulted in the labeling of a cell identified as a CA1 pyramidal layer basket cell. Top: a low-power magnification showing the location of the cell in the CA1 cell pyramidal layer. Bottom: a higher-power magnification showing the details of cell morphology. Calibration bar = 50 µ. E: current-voltage plot of the cell shown in Fig. 1A, input resistance = 29.4 M.

View larger version (38K): [in this window] [in a new window]   Fig. 11. The control (no current) condition (top middle) and hyperpolarizing (left) and depolarizing (right) intracellular constant current injections during the occurrence of spontaneous theta field activity, highlighting the effects on the average value of the membrane potential, the average MPO amplitudes and cell discharge patterns of CA1 pyramidal layer basket phasic theta-OFF cell (229). Spikes truncated (positivity up in all traces).

View larger version (21K): [in this window] [in a new window]   Fig. 12. A: graph summarizing the effects of hyperpolarizing and depolarizing constant current injections and the control (no applied current) condition on the cell discharge pattern and rate during the theta and LIA conditions for cell 229 shown in Figs. 11 and 12. B: graph summarizing the effects of hyperpolarizing and depolarizing constant current injections and the control (no applied current) condition on the MPO amplitudes during the theta condition for cell 154. Data in both A and B are plotted against the average membrane potential values produced by the current injections and the average membrane potential value occurring during the control (no applied current) condition. SE bars showing positive half only in A.

MOSSY CELL INTERNEURONS. Figure 13 (left side panel) shows a transition from LIA to theta field activity (top trace) and the accompanying cell discharges of a labeled mossy cell (203) located in the hilus (right side panel). The cell discharged at a mean rate of 7.0 ± 1.1 Hz during LIA and no MPOs were recorded. The beginning of theta was marked by a hyperpolarizing shift in the membrane potential along with the presence of clear MPOs and the cessation of cell discharges. As the membrane potential became more depolarized, a few cell discharges occurred. In the interest of identifying the cell, the stimulation protocol was not carried out. Figure 13, right, shows that injecting Neurobiotin into cell 203 resulted in the labeling of a mossy cell interneuron in the hilus of the dentate gyrus. Similar relations between cell discharges and spontaneous theta and LIA were observed for one other unlabeled cell recorded from the hilar region.

View larger version (32K): [in this window] [in a new window]   Fig. 13. Relationships between spontaneously occurring HPC field activity and the cell discharges of a mossy hilar cell (203) classified as a phasic theta-OFF cell in the no current control condition. Top trace: the HPC field activity recorded from the molecular layer of the dentate region; bottom trace: the discharge pattern of the cell (positivity up in all traces). The 1st half of the panel shows the irregular cell discharge pattern occurring during HPC LIA. Note the absence of MPOs. The 2nd half of the panel shows the initial cessation of cell discharges during theta field activity followed by a few phase-locked discharges as theta frequency reduced. MPOs occurred along with theta field activity. Right: the intracellular injection of Neurobiotin into the cell. Right, top: a low-power magnification showing the location of the mossy cell in the hilar reion of the dentate gyrus. Bottom: a higher-power magnification showing the details of the cell morphology. Calibration bar = 50 µ.

GRANULE CELLS. Figure 14 (left side panel) shows a transition from LIA to theta field activity (top trace) and the accompanying cell discharges of a labeled granule cell located in the lower blade of the denate granule layer (right side panel). The cell discharged at a mean rate of 3.0 ± 4.1 Hz during LIA and no MPOs were recorded. The beginning of theta was marked by a hyperpolarizing shift of 5.2 mV in the membrane potential along with the presence of clear MPOs and the cessation of cell discharges. These data are representative of all 11 granule cells recorded. We found these cells very difficult to hold for long periods so in the interest of identifying the cell, the stimulation protocol was not carried out. Figure 14, right, shows that injecting Neurobiotin into cell 190 resulted in the labeling of three granule cells in the lower blade of the granule cell layer. The injection of Neurobiotin into granule cells 218 and 220 both resulted in the labeling of two granule cells (data not shown).

View larger version (33K): [in this window] [in a new window]   Fig. 14. Relationships between spontaneously occurring HPC field activity and cell discharges of a denate granule cell (190) classified as a phasic theta-OFF cell in the no current control condition. Top trace: the HPC field activity recorded from the molecular layer of the dentate region; bottom trace: the discharge pattern of the cell (positivity up in all traces). The 1st half of the panel shows the irregular cell discharge pattern occurring during HPC LIA. Note the absence of MPOs. The 2nd half of the panel shows the complete cessation of cell discharges during theta field activity. MPOs occurred along with theta field activity. Right: the intracellular injection of Neurobiotin into the cell resulted in the labeling of 3 dentate granule cells. Right, top: a low-power magnification showing the location of the cells in the lower blade of the dentate gyrus. Bottom: a higher-power magnification showing the details of the cells morphology. Calibration bar = 50 µ.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cells classified as phasic theta-ON cells

In the present study, all 18 CA1 pyramidal cells responded with a simple cell discharge pattern to depolarizing current pulses. Jensen et al. (1996) demonstrated that CA1 pyramidal cells fell into two categories according to their response to depolarizing current pulses: nonbursters (simple spikes) and bursters (complex spikes). Nonbursters generated stimulus-graded trains of independent action potentials with nondecrementing amplitudes, whereas bursters generated clusters of three or more closely spaced spikes of descending amplitude riding on a distinct depolarizing envelope.

The present study has provided evidence supporting the following conclusions concerning cells classified as phasic theta-ON cells: 1) morphologically identified hippocampal CA1 pyramidal cells formed a subset of cells meeting the criteria for classification as phasic theta-ON cells. Previous studies (Fujita and Sato 1964; Leung and Yim 1988, 1991; Nunez et al. 1987, 1990a) have clearly demonstrated the rhythmical discharge properties of CA1 pyramidal cells; 2) MPOs occurred only during theta field activity, their onset signaled by a 5- to 10-mV depolarizing shift in membrane potential; 3) the amplitude of membrane potential oscillations in CA1 pyramidal phasic theta-ON cells was voltage dependent and frequency was voltage independent; 4) there were no phase changes observed during current injections; however, amplitude analysis of MPOs revealed an inverted U-shaped curve asymmetrically distributed around the average value of the membrane potential occurring during the spontaneous theta (no current) control condition; 5) the rate of rhythmic cell discharges in the CA1 pyramidal phasic theta-ON cells during the theta condition was precisely controlled within a critical range of membrane potential values from approximately 57 to 68 mV, corresponding to a range of MPO amplitudes of ~4-7 mV. Outside the critical range, rhythmical discharges were abolished. And 6) there appear to be some interneurons that meet our criteria for being phasic theta-ON cells. These are the bistratified cells first described by Buhl et al.(1994). However, MPOs were not recorded in these cells during theta field activity and they responded in a different manner to depolarizing constant current injection.

Mechanisms underlying the generation of MPOs in identified CA1 pyramidal cells

The present study demonstrated that MPOs occurred only during spontaneous theta field activity and not during the spontaneous occurrence of LIA field activity. Previous data on the mechanisms underlying the generation of MPOs may be summarized as supporting three main conclusions: they are generated by inhibitory postsynaptic potentials (IPSPs); they are generated by excitatory postsynaptic potentials (EPSPs); and they are generated by intrinsic properties of the cell membrane. Evidence supporting the IPSP hypothesis came from observations that the laminar profile of antidromically evoked IPSPs in the CA1 region was similar to that of the theta rhythm of anesthetized animals (Artemenko 1973; Fox et al. 1983). Leung and Yim (1986) provided stronger support by demonstrating that MPOs recorded in CA1 pyramidal cells reversed in phase with respect to extracellularly recorded theta at a reversal potential corresponding to GABA-mediated IPSPs. Further support for the importance of IPSPs has come from studies by Fox (1989), Soltesz and Deschenes (1993), and Ylinen et al. (1995). The main support for the EPSP hypothesis has derived from two main observations: MPOs do not reverse with membrane voltage or the intracellular injection of Cl and the MPO was larger with hyperpolarization. (Fujita and Sato 1964; Nunez et al. 1987, 1990a). The main support for the intrinsic generation hypothesis has come from studies using in vitro hippocampal preparations. Leung and Yim (1988, 1991) demonstrated theta frequency MPOs could be induced in hippocampal cells solely by depolarization and even when synaptic transmission was blocked by low Ca²⁺, low Na⁺, and tetrodotoxin. In the present study, the demonstration of an inverted U-shaped function following the injection of depolarizing and hyperpolarizing constant current pulses also argued against the dependence of MPOs on synaptic activity. There were no phase reversals observed between the MPOs and the extracellularly recorded theta field activity and hyperpolarization of the membrane potential did not increase the size of the MPOs. Nunez et al. (1990a) also reported that the injection of hyperpolarizing current pulses into CA1 pyramidal cells recorded in vivo failed to produce any phase reversals, although they did report that the amplitude of the intracellular theta (MPOs) increased. In the present study, we did not observe an increase in amplitude during hyperpolarizing current injections; in fact, increasing hyperpolarization eventually totally abolished MPOs (as did increasing levels of depolarization). On the other hand, Soltesz and Deschenes (1993) did observe an ~180° phase shift in CA1 cells that were hyperpolarized from 65 to 85 mV, and similar phase shifts were reported by Ylinen et al. (1995). In the present study, we failed to see a phase reversal in three cells hyperpolarized in the range from 85 to 90 mV. A possible explanation for the difference between these two studies and the present study may lie in the way the CA1 pyramidal cells were selected for data analysis. Indeed, in both the Solstez and Deschenes and Ylinen et al. studies only low-firing (<1 and 2 Hz, respectively) CA1 pyramidal cell were selected for analysis compared with discharge rates of 10-12 Hz during theta in the present study. Soltesz and Deschenes (1993) and Ylinen et al. (1995) also reported that the frequency of intracellular theta rhythm in CA1 pyramidal cells was independent of membrane potential, a result supported by the present study. The absence of any measurable MPOs during the 300- and 400-pA applied current conditions could be due to the fact that the membrane potential would be largely determined by the spike conductances, which may short-circuit those responsible for the generation of MPOs. If this was the case, then rhythmic spike discharges would be abolished due to the large current-induced depolarization and not the abolishment of the MPOs. In the current study, we had no direct experimental evidence to determine which explanation was correct. It is also of interest to note that in the present study the four levels of depolarizing constant current failed to produce MPOs during spontaneously occurring LIA field activity, in agreement with the results of Ylinen et al. (1995). On the other hand, in the in vitro studies of Leung and Yim (1998), they were able to manipulate MPO frequency with intracellular current injections. We have no explanation for the differences except to note that our experiments were carried out in vivo. In addition, it's possible that our current steps were too large and thus missed producing the appropriate membrane potential for activating the membrane currents underlying the generation of MPOs.

Relationship between MPOs and rhythmic cell discharges at theta frequencies in identified CA1 pyramidal cells

The present study demonstrated that the occurrence of MPOs in a critical range of amplitudes and the rhythmical cell discharges at theta frequencies was highly correlated. These data provide evidence that once extrinsic synaptic inputs enable the theta state, voltage-dependent intrinsic MPOs functioned to control the rhythmic discharge properties of CA1 hippocampal pyramidal cells classified as phasic theta-ON cells. Thus intrinsically generated MPOs allow CA1 phasic theta-ON cells to be tuned selectively to theta-band frequencies. The question remains as to whether MPOs control the rhythmical discharge properties or vice versa. We believe our data provide evidence that once synaptic inputs enable the theta state, voltage-dependent MPOs control the rhythmic discharge properties of phasic theta-ON cells morphologically identified as CA1 pyramidal cells. We are aware of only one other published study in which the data suggested that MPOs actually modulate rhythmical CA1 pyramidal cell discharges rather than vice versa (Leung and Yim 1998). Although the present data are not unequivocal, there were several observations supporting our contention: 1) we observed on occasions during the control (no current) spontaneous theta condition that MPOs may occur without cell discharges but never observed rhythmic cell discharges in the absence of an MPO; 2) in some phasic theta-ON cells, we observed MPOs persisting at lower levels of current produced hyperpolarizations along with a complete inhibition of cell discharges; 3) nonrhythmic cell discharges still occurred when MPOs were very small or totally abolished; 4) nonrhythmic cell discharges occurred during the LIA field condition in the complete absence of MPOs; and 5) the application of depolarizing current during the LIA field state was not sufficient to produce MPOs or rhythmic cell discharges. We therefore concluded that MPOs in a critical range of amplitudes between 4 and 7 mV serve to entrain cell discharges into a rhythmic pattern.

The present research had the advantage of studying spontaneously occurring field activities in the whole animal. Thus all experimental manipulations occurred during the spontaneous cycling between "physiologically occurring" theta and LIA. Furthermore, theta field frequency in each animal was essentially clamped as a variable. That is, in a given experiment, theta frequency was stationary, varying <0.5 Hz for the entire duration of the experiment. Across all experiments, theta field frequencies ranged from 3 to 4.5 Hz. MPOs occurred only during theta field activity and at the same frequency as the extracellular theta frequency. Previous work using extracellular recording techniques demonstrated that phasic theta-ON cells in both the CA1 and dentate regions of the hippocampal formation increased the number of rhythmic discharges per theta wave as theta field frequency increased in both acute (Colom and Bland 1987; Colom et al. 1987) and freely moving animals (Bland et al. 1983; Sinclair et al. 1982). In the present experiments, despite the fact that theta frequency remained constant, CA1 phasic theta-ON cells increased the number of rhythmic cell discharges in response to increases in membrane depolarization in a range from approximately 67 to 58 mV, corresponding to MPO amplitudes from 4 to 7 mV. Thus while "physiologically occurring" synaptic inputs were driving the cells at a given theta frequency, injecting depolarizing or hyperpolarizing current was capable of modulating the rate of rhythmic spike discharges within a critical range of MPO amplitudes. In contrast, while "physiologically occurring" synaptic inputs were producing LIA field activity with no accompanying MPOs, injecting depolarizing or hyperpolarizing current modulated the rate of irregular spike discharges. The present data therefore indicated that the relationship between the intrinsic membrane properties and extrinsic theta-related synaptic inputs is a complex one. It would be of interest to repeat the present experiments with the addition of stimulating the ascending brain stem hippocampal synchronizing pathways to produce a range of theta frequencies.

Our finding that the injection of Neurobiotin in 57% of our recorded cells resulted in the labeling of two pyramidal cells was very similar to the value of 63% of Lucifer yellow-coupled CA1 pyramidal cells recorded during theta field activity (Munoz et al. 1990).

Cells classified as phasic theta-OFF cells

In the present study, 22 cells met the criteria for classification as phasic theta-OFF cells and provided evidence to support the following conclusions: 1) morphologically identified CA1 pyramidal layer basket cells, mossy hilar cells, and granule cells formed a subset of cells meeting the criteria for classification as phasic theta-OFF cells; 2) MPOs occurred only during theta field activity, their onset signaled by a hyperpolarizing shift of 5-10 mV in membrane potential; 3) the amplitude of membrane potential oscillations in CA1 pyramidal layer basket cells was voltage dependent and frequency was voltage independent; 4) the phase of membrane potential oscillations in CA1 pyramidal layer basket cells underwent an ~180° phase reversal when the membrane potential was depolarized to around 65 mV; and 5) the occurrence and rate of rhythmic cell discharges in the CA1 pyramidal layer basket phasic theta-OFF cells during the theta condition was precisely controlled within a critical range of membrane potential values from approximately 62 to 60 mV, corresponding to a range of MPO amplitudes of ~7-7.5 mV. Outside the critical range, cell discharges were absent or occurred singly.

Mechanisms underlying the generation of MPOs in identified CA1 pyramidal layer basket cells

The present study demonstrated that MPOs occurred in phasic theta-OFF cells only during spontaneous theta field activity and not during the spontaneous occurrence of LIA field activity, the same findings reported above for phasic theta-ON cells. However, in contrast to phasic theta-ON cells, the onset of MPOs in all cells identified as phasic theta-OFF cells was signaled by a hyperpolarizing shift of 5-10 mV in the membrane potential. These data support our earlier paper (Bland et al. 1988) that was the first to our knowledge to report this finding in cells recorded in hippocampal slices and subsequently in acute in vivo preparations (Konopacki et al. 1992). Furthermore, the amplitude of the MPO was voltage dependent, increasing in amplitude and reversing in phase when the membrane was depolarized below the average membrane potential occurring during spontaneous theta field activity. Contrary to the findings for phasic theta-ON cells, these observations support the view that MPOs in basket cells are synaptically mediated and that chloride-dependent IPSPs play a role in their generation. This, along with the finding that the frequency of MPOs was voltage independent in basket cells, supports the previous findings of Ylinen et al. (1995). Ylinen et al. also reported that basket cells discharged in a manner related to theta that would meet our criteria of being theta-ON cells, exactly the opposite of our findings but consistent with the earlier inferences of Buzsaki et al. (1983) and Fox and Ranck (1975) that interneurons discharge at higher rates during theta compared with the nontheta condition. We have no explanation for this difference except to suggest that subsets of CA1 pyramidal layer basket cells may differ in their relationship with theta field activity.

Relationship between MPOs and rhythmic cell discharges at theta frequencies in identified CA1 pyramidal layer basket cells

As discussed in the preceding text, MPOs in basket cells occurred during theta field activity only and begin with a hyperpolarizing shift in the membrane potential. Thus for much of the time, cell discharges are prevented from reaching firing threshold. However, as the membrane becomes depolarized, correlated with the brief transition between slow frequency theta to LIA, theta-OFF cells begin to show phase locked single cell discharges, followed by an increasing number of rhythmic discharges per theta wave. These results are compatible with earlier findings that phasic theta-OFF cells code the declining theta field frequency (Colom and Bland 1987). The cell discharges under these conditions were correlated with the MPOs in that they occurred on the depolarizing phase of the MPOs and increased the number of rhythmic discharges per MPO in relation to the increasing amplitude of the MPOs.

Mossy cells and granule cells classified as theta-OFF cells

Because of the great difficulties we experienced in being able to maintain lengthy stable recordings in these cells, we made the decision to collect enough data during the spontaneous field conditions to classify the cells and then attempt to stain them for identification. The 2 mossy cells and 11 granule cells classified as theta-OFF cells all behaved in the same manner at the transition from LIA to theta field activity as the basket cells described in the preceding text. That is, MPOs in mossy cells and granule cells occurred during theta field activity only and began with a hyperpolarizing shift in the membrane potential. Cell discharges only occurred as the membrane potential began to depolarize. These findings are in essential agreement with those reported by Soltesz et al. (1993) for mossy cell interneurons and Ylinen et al. (1995) for granule cells although these authors did not refer to them as theta-OFF cells.

General discussion

IDENTIFICATION OF THETA-RELATED CELLS. Colom et al. (1987) first introduced the nomenclature of theta-ON and -OFF to distinguish two types of HPC cells that were essentially reciprocally related to theta field activity. As the name implies, theta-ON (theta) cells discharged at higher rates during theta compared with LIA and theta-OFF cells were often silent during theta and discharged during LIA. Further study extended these findings and resulted in a classification scheme for theta-related cells (Colom and Bland 1987). Fox and Ranck (1975, 1981) presented evidence that theta (theta-ON) cells were interneurons while Bland et al. (1980) argued that projection cells were theta (theta-ON) cells. Based on several lines of evidence, Bland and Colom (1989) proposed that theta-OFF cells were inhibitory interneurnons that received inhibitory GABAergic projections from the medial septum (see also Smythe et al. 1992). Ylinen et al. (1995) have provided evidence that CA1 layer basket cells are related to theta in the manner that we would classify as theta-ON cells and that granule cells are related to theta in a manner we would classify as theta-OFF cells. The findings of the present paper serve to illustrate that the relationship of cells in the hippocampal formation to the field states of theta and LIA are more complex than was previously envisioned by theta researchers. Indeed, as we previously hypothesized, some interneuron subtypes behave as theta-OFF cells (basket cells and mossy interneurons) but some behave as theta-ON cells (bistratified cells). On the other hand, contrary to our previous proposal, some projection cells (granule cells) behave as theta-OFF cells while at least a subset of others (CA1 pyramidal cells) behave as theta-ON cells.

Relationships between theta-ON and theta-OFF cells

Colom et al. (1987) emphasized the reciprocal relationship between the discharge rates of phasic theta-ON and phasic theta-OFF cells that occurred over a range of theta field frequencies. That is, phasic theta-ON cells increased their discharge rates as theta frequency increased and phasic theta-OFF cells decreased their discharge rates as theta frequency increased (to 0 in many cases). These observations were based on extracellular unit recordings. The intracellular data in the present paper provides further evidence in support of this reciprocal relationship (see also Cobb et al. 1995). Increasing levels of membrane depolarization in both CA1 pyramidal phasic theta-ON cells and CA1 layer basket phasic theta-OFF cells resulted in increasing discharge rates during theta and LIA field activity. However, in phasic theta-ON cells, discharge rates were higher during theta compared with LIA and in phasic theta-OFF cells discharge rates were higher during LIA compared with theta. Furthermore, at the highest levels of membrane depolarization, phasic theta-ON cell discharges became arrhythmic while phasic theta-OFF cell discharges became rhythmic. Finally, the data supported the view that the main currents underlying the voltage-dependent MPOs in CA1 pyramidal phasic theta-ON cells are intrinsic, while those underlying the MPOs, at least in CA1 layer basket phasic theta-OFF cells, were mainly synaptic, mediated by chloride currents. In conclusion, the findings of the present paper support the model of theta generation proposed by Bland and Colom (1993).