• Limbic seizures;
  • Kindling;
  • Thalamus;
  • GABA;
  • Glutamate;
  • Epilepsy


  1. Top of page
  2. Materials and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

Purpose: Studies have suggested that the medial dorsal nucleus of the thalamus plays a role in the behavioral expression of limbic seizures, but it is unclear whether this region is a key component for the primary seizure circuitry or a path for seizure spread from one region to another. This study was undertaken to determine the potential role of this region in limbic seizure activity.

Methods: Adult male rats received kindling stimulation either under urethane anesthesia or while awake. Glutamate or its agonists or the GABA antagonist bicuculline or agonist muscimol were infused into the medial dorsal nucleus. In another series, kindling acquisition was compared among three thalamic sites as well as with the amygdala and hippocampus

Results: Drugs that enhanced excitatory drive or blocked GABA resulted in significant prolongation of electrographic seizure activity compared to saline infused controls. Enhanced GABA activity resulted in a significant reduction of seizure duration. Infusion of the compounds lateral to the medial dorsal nucleus did not affect seizure duration. In the kindling studies the medial dorsal region is the only thalamic nucleus from which hippocampal seizures can be induced, but with an elevated afterdischarge threshold compared to the two limbic sites. However, the seizures generalized more rapidly from the medial dorsal region.

Conclusions: This study demonstrates that the medial dorsal nucleus and other dorsal midline nuclei have a significant role in the primary seizure circuits of limbic seizures as well as in spread of seizure activity to other regions.

Seizures occur in neuronal networks, and there is good evidence that the various components in that network play different roles (Avoli & Gloor, 1982; Lothman et al., 1991; Meeren et al., 2002). Understanding the make up of a seizure network and the role each component plays is important for selecting targets for therapeutic intervention.

Spike and wave seizures of absence epilepsy have provided a strong conceptual basis for the idea of a seizure network with different roles for the separate components. The primary network has been defined as a reciprocal circuit involving the neocortex and thalamic relay nuclei with the thalamic reticular nucleus serving as a key modulator (Avoli & Gloor, 1982; Meeren et al., 2002). In this functional model, the cortex provides the excitatory drive, and the thalamus serves to organize this drive into the spike wave pattern.

The role of the thalamus in other types of epilepsy is understood less well. Recordings from human epilepsies from this region are not performed very often, but there is clear evidence of thalamic involvement in the seizures, although the role of the thalamus in the seizures is not well defined (Guye et al., 2006). Recent human clinical trials have targeted various thalamic nuclei for therapeutic electrical stimulation in attempts to control the epilepsy (Kerrigan et al., 2004; Andrade et al., 2006), with mixed success. However the success or failure of this approach with a limited array of stimulation protocols cannot define the role of this region with its complex connections. Animal studies have also not had overwhelming success (Lado, 2006).

Mesial temporal lobe or limbic epilepsy is a specific form of epilepsy that is associated with anatomical changes in a number of regions including the hippocampus, entorhinal cortex and amygdala (Du et al., 1993; Hudson et al., 1993; Margerison & Corsellis, 1966). Recent evidence suggests that there are anatomic changes in the medial dorsal (MD) region of the thalamus associated with this form of epilepsy (Juhász et al., 1999; Bertram et al., 2001; Natsume et al., 2003), and there are initial suggestions from animal models that it may be involved with seizure activity from the earliest stages (Bertram et al., 2001). There is, however, a recent study that suggests that thalamic involvement is apparent only in the later stages of kindling if the rats are kindled in the amygdala (Blumenfeld et al., 2007).

There have been a number of previous studies (kindling and status epilepticus) that have suggested that the MD region could play a role in seizure spread (Lothman & Collins, 1981; Patel et al, 1988; Hirayasu & Wada, 1992; Cassidy & Gale, 1998; Bertram et al., 1998, 2001). In the present study, we examined whether modulation of synaptic activity in the MD region could influence limbic seizure activity to delineate further the potential role of this area in limbic seizures. We also examined the ability to induce seizures from the thalamus using the kindling model to determine whether this region could be a source for seizure initiation. The overall goal of these experiments was to define the role of the MD region in limbic seizures and to determine whether the seizures could be modified by influencing synaptic activity.

Materials and Methods

  1. Top of page
  2. Materials and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

All studies were performed under protocols approved by the Animal Care and Use Committee of the University of Virginia. The experimental procedures will be divided among studies performed under anesthesia and studies performed on awake rats.

Studies under anesthesia

Modulation of seizures

These experiments were performed to examine the ability of different receptor specific agonists and antagonists to alter electrically induced seizure activity when the compounds are infused into the midline dorsal thalamus. This model for rapid kindling under anesthesia (Stringer et al.,1989; Bertram et al., 2001) was chosen because the seizure activity was stable, and the seizures could be induced in a pattern so that infusion of small volumes into a highly focal area would likely have an immediate effect if the infusion site were an appropriate target.

Adult male Sprague-Dawley rats (Hilltop Laboratories, Scottsdale, PA, U.S.A.; 250-350 g) were used in this experiment. Under urethane anesthesia (1.2 g/kg, i. p.), the rats were placed on a multi-arm stereo-tactic frame, and body temperature was maintained at 37°C by a water blanket controlled by a rectal thermistor.

The stimulating and recording electrodes and the cannula for drug infusion were inserted stereotactically using target coordinates from a standard atlas (Paxinos & Watson, 1986). The recording electrodes were glass micropipettes filled with 0.9% NaCl and 1% Fast Green. One was in the midline thalamus (1.8–2.2 mm posterior to bregma, 0.6–0.8 mm left of midline, 5.7–6.3 mm below the dura, with a 5° arm angle from vertical axis to avoid the sagittal sinus). Another one was placed in CA1 (5.5 ± 0.1 mm posterior to bregma, 4.8–5.1 mm right of midline and 2.5–3.0 mm below the surface). A bipolar twisted pair stainless steel stimulating electrode was placed in the mid to ventral CA3 (5.5 mm posterior to bregma, 4.3–4.5 mm left of the midline, 4.5 mm below the dura). The depths of the stimulating and recording electrodes were adjusted to achieve the maximal response. The drug infusion cannula was positioned into midline thalamus only 0.3 mm posterior to the midline thalamus-recording electrode at the same depth. In a few experiments, the infusion cannula was placed more laterally (2.2–3.5 mm posterior to bregma, 3.5 mm left to the midline) to examine the specificity of the response to drug infusion in the midline nuclei. The drug infusion cannula was at the 5.7–6.3 mm below the dura at the same depth as the recording electrode, with a 5° arm angle from vertical axis to avoid the sagittal sinus.

The kindling stimuli (10 s, 30 V, 50 Hz monophasicsquare wave with pulse duration 0.5 ms) were delivered every 5 min through a constant voltage stimulus isolation unit (Winston Electronics). The stimulating electrode (insulated twisted pair stainless steel bipolar) had a tip separation of 1 mm, with the more ventral tip of the pair being the negative pole. After 10 baseline stimulus trains, single and paired-pulse (20 ms intervals at maximal amplitude response) (five responses averaged) evoked responses in CA1 following CA3 stimulation were obtained to determine if any potential change in the afterdischarge duration (ADD) could be the result of the drug's interfering with the CA3-CA1 synaptic response. Then, drug or normal saline (2μl over 3 min) was infused into midline or lateral thalamus by an infusion pump. After infusion the stimulus trains were delivered at the same rate 11 more times. Immediately following the infusion and before each of the subsequent four stimulations, the CA1 evoked potential induced by single or paired CA3 stimulation was obtained. The duration of afterdischarge induced by each CA3 repetitive stimulus train before, during and after infusion was measured from the cessation of the stimulation to the last burst discharge.

The drugs used in this experiment included the GABAA antagonist bicuculline, the glutamatergic agonists glutamate, alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA), and N-methyl-D-aspartic acid (NMDA). All were prepared in normal saline. Normal saline was used for vehicle control.

Studies in awake animals

Modulation of seizures

These experiments were performed to determine whether limbic seizures could be altered by the focal application of a GABA agonist or antagonist in animals that had behavioral accompaniment to the electrographic seizures.

Animals had electrodes implanted under ketamine/xyla- zine anesthesia. A hippocampal-stimulating electrode was placed in the mid ventral hippocampus HC (from bregma AP −3.6 mm, ML −4.9, DV −5.0 from dural touch point). In addition, a guide cannula (Plastics One) was placed into the dorsal midline thalamic region (AP −2.3 mm, ML −1.0 mm, 10° from vertical, DV −5.7 mm). This approach was chosen as it gave the best trajectory for consistently placing the infusion cannula between the two MD nuclei. The animals were allowed to recover for 1 week before starting a standard kindling protocol of 5–6 stimulations daily given hourly from 10 a.m. and 4 p.m. (2 s, 50 Hz, 1 ms square wave biphasic 400 μ A peak to peak) using a constant current stimulator (A-M Systems). Once the animal had achieved a stable kindled state (consistent stage 4 or 5 seizures for at least two stimulation sessions) the drug was administered. Two baseline stimulations were given on the test day, and the infusion cannula was inserted through the guide cannula. A slow infusion of the GABA agonist muscimol or antagonist bicuculline was given over a maximum of 3 min. Five minutes after completion of the infusion, the next stimulation was delivered. Afterdischarge duration as well as behavioral seizure score were compared to the baseline responses.

Following the last stimulation of the test day, the animals were killed and the brains placed in 4% paraformaldehyde for later histological confirmation of the placement of the infusion cannula. The placements were evaluated by a blinded observer, and they were correlated to the efficacy of the infusion in reducing afterdischarge duration and behavioral seizure score.

Kindling acquisition study

These experiments were performed to determine whether it is possible to induce seizures from the MD nucleus, and how these kindled seizures compare to seizures kindled from the amygdala and hippocampus, two traditional sites for limbic kindling. In this series of experiments, we examined the ease of inducing a seizure at each site (afterdischarge threshold), the development of seizure activity (afterdischarge duration) and the development of increasing behavioral severity (behavioral seizure score). The intent of this experiment is to provide greater insight into the role of the MD and neighboring midline thalamic nuclei in the initiation and propagation of limbic seizures, to determine whether it could serve as an initiator of seizures, or whether it is more likely a central point on the path to generalization.

In the initial experiments each animal received two twisted pair electrodes, one in the dorsal midline thalamic region and the other in either the hippocampus or the amygdala. In each animal, kindling stimuli were given at only one site through an electronic switch that allowed stimulation and subsequent recording and the other electrode was used to record EEG activity. In a small number of animals (n = 3) bipolar electrodes were placed in the hippocampus, amygdala, and MD in order to observe the relationship among the three sites during stimulated seizures. Two rats were kindled from the amygdala and one from the HC.

Afterdischarge threshold (ADT) was determined using a 10 s 50 Hz, 1 ms biphasic square wave stimulus train that was delivered every 2 min until an afterdischarge of at least 3 s occurred. Thereafter, the animals were stimulated every hour, 5–6 times a day every other day, 3 days a week at afterdischarge threshold. Afterdischarge duration and behavioral seizure score were recorded for each stimulation, and the changes were recorded in the sequence in which they occurred.

Regional thalamic specificity of limbic seizures

To determine whether the MD region was the primary thalamic area involved in limbic seizure activity, we performed two experiments. First, to determine which thalamic nuclei were involved in limbic seizures during the course of kindling, we recorded from different thalamic nuclei while kindling from the hippocampus. Second, to determine which thalamic regions were capable of driving limbic seizures we attempted kindling from three different thalamic sites. The thalamic sites studied were the midline dorsal region (predominantly MD and paraventricular nuclei), the ventromedial nucleus (VM, primarily motor) and the ventrolateral (VL, primarily sensory).

Bipolar twisted pair stainless steel stimulating and recording electrodes were implanted in the midventral hippocampus and one of the thalamic nuclei/regions (VM: AP −2.5 mm, ML −1.7 mm, DV 6.4 mm; VL: AP −2.8 mm; ML −3.0 mm, DV −5.5 mm). Kindling stimulations were given as described above. Initial stimulations were given to determine afterdischarge thresholds for each of the sites. For hippocampal stimulation and thalamic recording, the responses were analyzed for duration of afterdischarge at each site and the relative onset of electrographic seizure activity in relation to the hippocampus. For thalamic stimulation, the responses were analyzed for the development of behavioral changes that outlasted the stimulation (stimulus forced versus stimulus independent behavior) and the ability to recruit the hippocampus into electrographic seizure activity. Stimulus forced behavior refers to a clear change in behavior (e.g., behavioral arrest, circling, repetitive unilateral limb activity) that begins immediately following the initiation of stimulation and stops with stimulation cessation. This term was developed to indicate that stimulation could cause behavioral changes independent of seizure induction.


All data are presented as means ±SEM. Comparisons were made across groups using an ANOVA. For ANOVAs that showed a significant difference, pairwise comparisons were made using a post hoc Student-Newman-Keuls (SNK) test. A paired t-test was used for comparison between data obtained immediately before infusion of drug and immediately after. For comparing relative proportions a Fisher exact test was used. Significance was set at the p < 0.05 level.


  1. Top of page
  2. Materials and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

Studies under anesthesia

Afterdischarge distribution

The afterdischarges in the midline thalamus and CA1 following CA3 stimulation were quite similar (Fig. 1), with a series of repetitive burst discharges found at both recording sites. In general, the duration of afterdischarge bursts were equal in duration in the thalamus and hippocampus, although on very rare occasions one site or the other had a more prolonged run of terminal bursts. In no instance did we find an afterdischarge at one site without an afterdischarge at the other.


Figure 1. Effect of bicuculline infusion into the MD results in prolonged seizure activity. (A) Tracings from the MD and HC at baseline, immediately following the infusion and after recovery. There is significant transient lengthening of the afterdischarge after bicuculline infusion. The high-amplitude activity at the onset of the tracing in all channels represents the 10 s stimulation artifact from CA3 stimulation. (B) Evoked potentials in CA1 (same recording site as in A) from contralateral CA3 stimulation at baseline and immediately following infusion, demonstrating that the drug did not have a direct effect on the area of CA1 from which the seizures were recorded. In each sweep there is an initial 5mV deflection that is the calibration pulse followed by the brief high-amplitude stimulation artifact, the evoked response with a negative population spike, the second stimulation artifact and the second, now suppressed, evoked potential without the population spike. (C) Mean electrographic seizure durations (measured in CA1 from end of stimulation to end of last burst discharge) over course of experiments following midline and lateral bicuculline infusions (0.1mM) as well as midline saline infusions, demonstrating that the changes are specific to the drug infused in midline (*p < 0.05 from seizures immediately preceding infusion). (D) Dose response of midline infusion of two concentrations of bicuculline. Values are taken the first stimulation following infusion (response to second stimulation is usually longer). (*p < 0.05, **p < 0.01 t-test, compared to saline infused controls).

Download figure to PowerPoint

Evolution of afterdischarges and effects of drug infusion

The first train of CA3 stimuli induced 1–4 bursts of CA1 and medial thalamic afterdischarges that usually lasted several seconds. The duration of the afterdischarge bursts gradually and steadily increased with each subsequent stimulation. Following the 10th CA3 stimulus train, the drug was injected into the midline thalamus at the nominal rate of 2μl/min over 3 min. One minute after the infusion was completed the stimulations resumed.

Figure 1A demonstrates a clear transient prolongation of afterdischarge duration after bicuculline infused into midline thalamus. Figure 1B shows that there was no change in the CA1 response to paired CA3 stimulation after bicuculline infused into midline thalamus, suggesting the prolongation of afterdischarges duration following bicuculline is not due to the effect of bicuculline on the CA1 primary responses. Because the route from CA3 to the thalamus is polysynaptic no consistent evoked response to a single stimulation could be recorded from the thalamus following CA3 stimulation to demonstrate the effect of bicuculline on evoked responses in the thalamus. It is likely that the seizure activity seen in the MD region following CA3 stimulation was likely driven from another site.

The time course of mean afterdischarge duration throughout the experiment is shown in 1C, showing that there is a clear peak of duration in the two afterdischarge bursts following bicuculline infusion into midline thalamus. Infusion of saline into the same site had no effect, and the specificity of the midline thalamus for this effect was demonstrated by the lack of an effect of bicuculine infusions 3 mm lateral to the midline infusion site. Figure 1D shows the dose–response effect of bicuculline infusion into midline thalamus on CA1 afterdischarge duration induced by CA3 repetitive stimulation. The duration of CA1 afterdischarges increased from preinfusion 5.44 ± 1.09 s to 7.92 ± 1.62 s (first stimulation after start of the infusion) during infusion of 0.1 mM bicuculline (n = 8, p < 0.05, paired t-test). The effect of 0.5 mM bicuculline infusion on the duration of CA1 afterdischarges is greater (from preinfusion 6.46 ± 1.54 s to 14.17 ± 3.49 s during infusion, n = 8, p < 0.01 paired t-test compared to baseline and, to 0.1 mM, p < 0.05, t-test).

Infusion of glutamatergic drugs had similar effects. AMPA infusion into midline thalamus (0.1 mM) induced transient increases in the duration of CA1 afterdischarges (Fig. 2A) (from preinfusion 7.10 ± 0.95 s to 10.04 ± 1.50 s during infusion, n = 12, p < 0.001). L-glutamate or NMDA infusion into the midline thalamus has similar effects. The duration of CA1 afterdischarges increased from 5.09 ± 0.78 s of preinfusion to 11.00 ± 4.48 s following L-Glutamate (1 mM, n = 8, p < 0.05). NMDA infusion (50?M) increases the duration of CA1 afterdischarges from preinfusion 9.11 ± 2.00 s to 13.13 ± 2.45 s during infusion (n = 8, p < 0.05) (Fig. 2B). Infusion of AMPA lateral to the MD nucleus had no effect.


Figure 2. Effect of glutamate agonists on CA1 afterdischarge duration. (A) Infusion of 0.1mM AMPA into medial dorsal nucleus results in prolongation of the afterdischarge, with the duration immediately returning to baseline. Data are from one animal from beginning to end of experiment. Duration is measured from the CA1 recording from the end of the stimulation to the end of the last burst discharge. (B) Group averages of changes in the afterdischarge duration from infusions of glutamate, AMPA, and NMDA. All infusions into the midline resulted in significant prolongations of the afterdischarge as seen in the change in the absolute duration (top) and in the percent change (bottom). Infusion of AMPA lateral to the medial dorsal nucleus had no significant effect suggesting that the effect of these agonists is specific to the midline region(*p < 0.05, **p < 0.001, paired t-tests compared to preinfusion baseline).

Download figure to PowerPoint

Studies on awake animals

Infusion effect on kindled seizures

These studies were performed to determine whether infusion of drugs specific for the GABA-A receptor into the MD region of the thalamus could alter the duration and behavioral severity of kindled limbic seizures in awake animals in order to understand better the role this region plays in the seizures.

For muscimol infusion, all animals were first kindled to stable stage 5 seizures that occurred consistently on at least two consecutive stimulation days. On the test day all animals had two stage 5 seizures with the first two stimulations. After the second stimulation, the infusion cannula was placed in the guide and the infusion was completed 5 min before the next stimulation was given. Muscimol was given with a total volume of 1μl over 3 min with a total dose of 50 nanomoles. At the end of the day the brain was removed and sectioned, and the placement of the cannula in a particular thalamic nucleus was determined using the Paxinos and Watson atlas.

A total of 20 rats were used in this muscimol infusion experiment. Thirteen of the animals had cannulas that were in the target zone (MD, paraventricular nuclei within the borders of the centromedian nucleus). There was a strong relation between placement of the cannula and reduction of seizure duration and, to a lesser extent, the suppression of behavioral accompaniment (Fig. 3A–B). Infusions in the target zone resulted in a significant reduction in seizure duration and behavioral seizure score (Fig. 3C) whereas infusion outside the zone showed no change in the duration and a slight but nonsignificant reduction in behavioral severity. Another way to view the data is to divide the animals into two groups: those that had a reduction in the afterdischarge duration of at least 20 s (successful infusions) and those that did not. All of the rats with successful muscimol infusions (n = 10) had the cannula within the MD and paraventricular region, and all were at or very close to midline (Fig. 3B). The failed infusions with no change in the afterdischarge duration (n = 10) were mostly above or below these regions. Three were technically in the region, but they were either very posterior or at the inferior margin, suggesting that much of the infusate did not reach the bulk of the midline dorsal thalamic region (Fig. 3B). The proportion of rats with successful infusions that were in the region (10/10) versus those without successful infusions with cannulas that were in the region (3/10) was significantly different (p < 0 001, Fisher exact). As noted above the difference in behavioral changes was less dramatic, although only animals with cannulas in the target zone achieved a significant reduction (Fig. 3C). These findings suggest that the midline region is specific for the modulation of seizure activity, but that seizure spread may involve more regions of the thalamus.


Figure 3. Effect of muscimol infusion on fully kindled seizures. (A) Hippocampal recording of afterdischarge in a fully kindled rat before, immediately after and 2 h after the infusion of muscimol. There is a clear but transient reduction of the afterdischarge duration following the infusion. Blank segment before onset of seizure activity is when EEG was switched to stimulation before returning to recording mode. (B) Location of the infusion cannulas for successful and unsuccessful infusions. Numbers indicate Paxinos and Watson positions behind bregma. The successful infusion cannulas were centered in the midline dorsal thalamic region Dots indicate placement of successful infusions (reduction of ADD by more than 20 s). X's indicate the placement of unsuccessful infusion cannulas. Unsuccessful infusions were either outside this region or at its periphery. The three unsuccessful placements in the target zone were at −2.3 or −3.6. Three others were in or at the edge of the ventricle. (C) Means of afterdischarge duration and behavioral seizure scores of the two groups. There is a difference in the effect on afterdischarge duration. Although both groups showed a reduction in behavioral seizure score, it only achieved significance for the group with the cannula inside the target zone.(*p < 0.01, paired t-test).

Download figure to PowerPoint

For the bicuculline infusion, which we hypothesized would worsen the seizures, the rats were stimulated on two days, one day apart, to assure that afterdischarges could be reliably obtained at the original stimulation threshold. The reason for using animals that had not achieved a fully kindled (consistent Stage 5 seizures) was to allow for a demonstrable worsening of the seizures as a drug effect. On the second day, when the animals were still having nonmotor seizures, two initial stimuli were given. Starting 50 min following the second stimulation, the bicuculline was infused at a rate of 1μl/min for 2 min at a concentration of 1 mM. Five minutes following the infusion of bicuculline, a third stimulation was given. Four of the eight animals entered into status epilepticus with continuous EEG seizure activity and intermittent convulsions that lasted at least several hours before termination of the seizure activity with a systemic benzodiazepine. We recorded only in the hippocampus as there was not sufficient room for an infusion cannula and bipolar recording electrode in the MD region. The electrographic seizure activity was continuous and there were intermittent convulsions. The other four rats receiving bicuculline had no overt behavioral or electrographic changes. In comparing the four rats that entered into status epilepticus with the four that had unsuccessful bicuculline infusions, the primary difference was in the placement of the cannulas, with all of the successful ones placed within the confines of the target zone as defined above, whereas the placement was lateral in the unaffected group (p < 0.05, Fisher's exact test).

Kindling acquisition studies

The purpose of these experiments was to determine whether the midline thalamus was capable of initiating limbic seizures and if the involvement of the midline region was specific for limbic seizures. By examining the afterdischarge thresholds, the progression of seizure durations and the rate of acquisition of behaviorally more severe seizures, we can determine the relative susceptibility of each site to seizure induction and the potential role that each location plays in the seizure circuits.

There were significant differences in ADT across the groups (Table 1) (ANOVA, p < 0.001). ADTs of the amygdala and hippocampus were the same. The ADTs at the three thalamic sites were significantly greater than at the two limbic sites (p < 0.001, SNK—post hoc), but the MD ADT was significantly less than those of two lateral relay nuclei (p < 0.005, SNK). Of note, the initial afterdischarges following MD stimulation always included the hippocampus or amygdala. Stimulation of the two lateral thalamic nuclei resulted only in a local afterdischarge that was very brief, and never involved the limbic sites. Hippocampal involvement following stimulation of these two sites was rarely seen, only after a minimum of 16 stimulations, only intermittently and then late in the evolution of the afterdischarge (see below).

Table 1.  Thalamic indling comparisons
Kindling stimulation site (n)Afterdischage threshold (μA ± S.E.M)Stims to 1st stage 5 seizure (±S.E.M)
  1. aSignificantly different from amygdala and hippocampus p < 0.001 (ANOVA post hoc SNK).

  2. bSignificantly different from mediodorsal p < 0.005 (ANOVA post hoc SNK).

Amygdala (5) 64 ± 108.2 ± 0.4b
Hippocampus (6)70 ± 912.7 ± 2.2a, b
Mediodorsal (10) 411 ± 46a3.6 ± 0.3a
Ventrolateral (5)  680 ± 12a, b>20a, b
Ventromedial (8)  589 ± 67a, b>20a, b

All animals achieved a stable kindled state from stimulation of the MD region, the hippocampus and the amygdala, but there were clear differences in the evolution of the kindled seizures between the MD nucleus and the two limbic sites. The hippocampus and the amygdala had significantly lower afterdischarge thresholds, and the increase in afterdischarge duration was faster than the afterdischarges elicited by stimulation of the thalamic site (Fig. 4A). It was interesting to note that at the beginning of a stimulation day (2–3 days after the previous stimulation day) the ADD of the MD induced seizures was reduced from the previous stimulation session but recovered quickly and continued to show a steady increase. As shown in Figure 4A, this transient reduction of ADD lessened over time as the ADDs lengthened. Compared to traditional limbic site stimulation, the increase in behavioral severity was much faster with MD stimulation, with all animals having a brief motor seizure (stage 3) with the first stimulation, and all animals having at least one kindled motor (stage 4 or 5) seizure by the third stimulation (Fig. 4B). Whether stimulating from the amygdala or hippocampus or from the dorsal midline, the midline thalamus and the limbic site were involved in the afterdischarge from the beginning (Fig. 5). Data for the amygdala acquisition are not shown, but paralleled hippocampal acquisition, with the exception of the earlier development of motor seizures (Table 1).


Figure 4. Kindling acquisition HC versus MD stimulation. (A) Mean ADD over the course of kindling. Vertical axis seconds, horizontal sequential stimulation number. Evolution of the afterdischarge following stimulation of either site. The HC begins with a longer ADD but the two sites then develop similar ADDs for the first 10–12 stimulations, after which the HC induced ADD is much longer until close to 30 stimulations when they are again equal. Arrows indicate fall in ADD following 1 or 2 days without stimulation that is sometimes seen in this stimulation protocol. (B) BSS progression starts at a higher level and plateaus much more rapidly with MD stimulation. (n = 10 for MD and six for HC).

Download figure to PowerPoint


Figure 5. Thalamic involvement in stimulated limbic seizures. Nonmotor seizures at the beginning of kindling induced by stimulation of the amygdala (A) or hippocampus (B). In both cases, the MD is involved from the outset with repetitive spiking present as soon as the stimulation was stopped. Recordings begin with the onset of stimulation as seen with thehigh frequency artifact of 2 s duration. For the two stimulation sites, there is significant artifact resulting from disconnecting the relevant electrodes during stimulation. The end of stimulation in these channels is shown by the heavy vertical lines indicated by the heavy arrows. Actual recording in these sites began several seconds later when the electrodes were reconnected. Following amygdala stimulation the first MD discharges are seen on the downslope immediately after the stimulation artifact indicated by the light vertical arrow. Initial MD spiking after HC stimulation (B) is shown on the upslope following the stimulation artifact (vertical arrow). Of note, all sites are involved with seizure activity from the beginning of kindling, independent of the site of stimulation.

Download figure to PowerPoint

Specificity of midline thalamic region

To determine whether the ability to initiate a limbic seizure from the thalamus was specific to the midline, kindling was performed with bipolar electrodes placed either in the VM (n = 8) or VL (n = 5) thalamic nuclei and these electrodes were paired with a hippocampal electrode. In each animal, the thalamic electrode was used either as a recording or stimulating electrode. With hippocampal stimulation the thalamic electrodes were used to observe for an afterdischarge, the timing of an afterdischarge in relation to the hippocampal afterdischarge and the progression of the afterdischarge duration with subsequent stimulations. In the case of stimulation with the thalamic electrodes, data were collected with regard to afterdischarge threshold, the duration of the induced afterdischarge and the involvement of the hippocampal electrode.

The progression of ADDs and behavior was very limited in the two lateral nuclei (Fig. 6A). There was often a stimulus forced behavior associated with the lateral stimulation, but it did not progress as is typical for kindling nor did it ever evolve to continue after the stimulation was completed (Fig. 6B). Even when motor seizures did arise with an afterdischarge following lateral thalamic stimulation, they appeared later and did not achieve a stable and reliable pattern. Motor seizures could be elicited from VM but were unlike the limbic motor seizures, consisting primarily of focal limb clonus. Unlike stimulation of the midline region which resulted in an immediate HC afterdischarge from the outset, stimulation of the lateral two nuclei resulted in no HC afterdischarge in the initial sessions, and even in the later sessions, the hippocampus was only intermittently involved and then only late (Fig. 6A). In many cases, it was not possible to induce an afterdischarge consistently with lateral thalamic stimulation, especially from VL. With regard to recruitment specificity of the thalamic nuclei, hippocampal stimulation resulted in an immediate and synchronized afterdischarge in the MD region (Bertram et al., 2001) (Fig. 7A) In contrast the lateral nuclei showed no such early involvement, rather only after a number of stimulations when the animal had motor seizures (Fig. 7B and C).


Figure 6. Kindling with thalamic stimulation. (A) HC afterdischarge duration following stimulation of MD (n = 10) or VM (n = 8). The HC is involved from the very first stimulation of MD, and is generally synchronized with MD. VM stimulation induces no HC afterdischarge for the first 17 stimulus trains, and then the HC afterdischarges always appeared late in the course of the seizure following VM stimulation. In addition, HC recruitment from VM was erratic even after it started, present following one stimulation, absent for the next several. VL stimulation did not cause HC afterdischarges following up to 30 stimulations (data not shown). (B) Although stimulation of the three thalamic sites were associated with a motor response, for many of the animals the behavioral activity was seen only during stimulation (stimulus forced behaviors). For MD and VM the percentage of stimulation forced behaviors gradually decreased, but for VL, the behaviors were only stimulation forced through 20 stimulations. Compared to VL, the percentage of rats with stimulus forced behavior MD was significantly less at all points. VM had significantly fewer only after 20 stimulations (p < 0.05, Fisher's exact test).

Download figure to PowerPoint


Figure 7. Thalamic seizure activity following HC stimulation. In all figures the flat line in the HC channel is the 10 s stimulation (between the two switch artifacts) followed by an approximate 2 s offline period while the electrodes are repolarizing. (A) Early HC stimulation with a nonmotor seizure shows MD involvement very early in the course of kindling and stimulation. In this and following parts, the vertical arrow indicates the approximate onset of seizure activity and the oblique arrowheads movement artifact from wet dog shakes. (B) Early in the course of kindling (nonmotor seizures) there is no involvement of VM following HC stimulation. (C) Later with the onset of motor seizures there is a gradual recruitment of VM as the seizure progresses (vertical arrow).

Download figure to PowerPoint


  1. Top of page
  2. Materials and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

This study demonstrates several points. First, pharmacological manipulation of the midline thalamic region, including the MD and paraventricular nuclei, can modulate seizure activity initiated in the hippocampus. Second, although it is more difficult to stimulate a seizure from the MD region than from the hippocampus or amygdala, seizures stimulated from this area generalize very quickly, suggesting that this thalamic region may play an important role in the spread of the seizures. Third, the limbic seizures can be altered by a variety of receptor agonists and antagonists infused into the dorsal midline region, indicating that limbic seizures are influenced by modulation of synaptic activity in this area. Finally, the late involvement of several lateral thalamic nuclei in seizures induced by limbic stimulation as well as the difficulty in initiating limbic or other seizures by stimulation of these relay nuclei suggest that the midline thalamic region is the one that is specifically involved in the initial stages of limbic seizures. These observations expand and more fully define preliminary observations reported before and clearly indicate that these midline nuclei play a significant role in the circuitry of limbic seizures, including the initiation and spread.

A question may arise with regard to the use of the anesthetized animals as well as the awake animals in this study. Both models have been regularly used in the laboratory for many years to answer specific questions. The advantage of using kindling under urethane is that the seizures don't rapidly involve multiple areas (implying restricted circuit recruitment), are relatively short in duration, can be given very close together, and show only a gradual change in duration over time so that they offer a very stable pattern for intervention. Under anesthesia, the placement of electrodes and maintaining an exact placement is much easier than in awake, chronic animals. Further the multiple electrodes and the acute infusions are much easier to place with a multiple arm stereotactic frame. There are potential disadvantages to using urethane as well. The seizures are more difficult to elicit and are shorter in duration. This observation suggests that the urethane has a seizure suppressive effect that may limit the number of neuronal circuits that are involved. It is not known which circuits could be more susceptible to these suppressive effects, and if key circuits are the primary target of the drug. On the other hand the advantage may be that, if there are fewer circuits involved because of the anesthesia, that it is easier to isolate several. It is because of this issue of anesthesia-induced seizure suppression that the use of awake animals is necessary to corroborate any findings obtained under anesthesia. The awake animals offer the advantage of having full behavioral seizures that involve multiple circuits so that one can get a better overview of the likely interactions and roles of many different sites. The primary disadvantage is that because of the involvement of more brain areas with longer seizures, the seizures may recruit additional areas that create a larger thalamic target zone. The results of the muscimol infusion study in the awake kindled animals suggests that the thalamic targets for the initial seizure initiation of limbic seizures remain relatively restricted.

The circuitry of epilepsy has been of interest for many decades. That seizures spread in the brain in some pattern that may be predicted by the connectivity of one region to another or by the recruitment of adjacent areas has been surmised at least since the first descriptions of Jacksonian march. In this study, we examined regions that are involved in limbic seizures (hippocampus and amygdala) and midline thalamic sites with known connectivity to the amygdala and hippocampus (Herkenham, 1978; Krettek & Price, 1978; van Groen & Wyss, 1990; Turner & Herkenham, 1991; Ray et al., 1992; Su & Bentigvolio, 1990; Dolleman-van der Weel & Witter, 1996; van der Werf et al., 2002; McKenna & Vertes, 2004) to study the potential role that these sites play in seizure initiation and spread. The thought that the MD region may play a central role in limbic seizure activity is based on the key observations that it is always from the beginning involved in electrographic seizure activity involving the hippocampus (Bertram et al., 2001), and, as demonstrated in this study, that pharmacologic manipulation of this area has a clear effect on electrographic limbic seizure activity. The hypothesis is further supported by the observation that direct seizure induction from this site results in almost immediate spread of seizure activity as evidenced by the rapid motor involvement. Because MD has such a high afterdischarge threshold, however, it is unlikely that this region of the thalamus acts as a seizure initiator. However, once recruited into a seizure, it could, through its connections recruit other regions. This possibility requires further study.

As noted earlier, other investigators have examined the potential role of the MD nucleus in seizures. In those experiments they evaluated the effect of drug infusion into the MD region on the behavioral expression of induced seizures that was similar to ours (Patel et al., 1988; Hirayasu & Wada, 1992; Cassidy & Gale, 1998). These studies demonstrated a consistent effect on seizure behavior. Unfortunately, EEG was not recorded to determine the effect on seizure physiology, so one was not able to determine if the altered seizure behavior was the result of a treatment effect on electrographic seizure activity or of changes in seizure spread. The present study suggests the MD region can influence both. However, other studies examining the role of MD and its thalamic neighbors in other types of acute seizures (systemic pentylenetetrazole) found that infusion of GABAergic agents into the region significantly worsened the seizures (Miller et al., 1989; Miller & Ferrendelli, 1990). These observations emphasize that the role of any region in a seizure will depend on the seizure and its associated circuitry, as pentylenetetrazole-induced seizures are considered more of a model for generalized seizures.

There have also been studies in the past examining the effect of lesioning the thalamus on seizures induced by electrical stimulation of the amygdala. These reports, all from the same laboratory, showed a mixed effect on the kindling process as well as on established kindled seizures (McCaughran et al. 1978; Hiyoshi & Wada, 1988a, 1988b). Review of the studies, specifically the placement of the electrolytic lesions, suggested that the lesions were never complete and occasionally only marginally involved the MD nuclei. The data from the present study did show that the thalamic effects can be quite focal, which may explain the relatively mild effects of incomplete lesions. One study of kindling in the massa intermedia of the thalamus (a much broader midline region in the thalamus than examined in the present study) had similar results to ours. They did not examine the specificity of the midline region to show that the effect was strictly localized to a small region in the thalamus (Mori & Wada, 1992).

A similar functional anatomy for absence seizures, with the cortex providing the drive and the thalalmus supplying a necessary organization was proposed and strongly supported by studies using a feline model of absence seizures that was carried out several decades ago (Avoli & Gloor, 1982). They induced seizures by the focal or systemic application of penicillin, and, as in the present study, they found linked activity between the cortex and thalamus (in this case the lateral relay nuclei). They also found that both regions were necessary for the development of well-organized spike and wave activity. Of note, they could only drive the seizures from the cortex by penicillin application. In other words, the cortex provided the excitatory drive and the thalamus organized the activity. More recently, Meeren et al. (2002) demonstrated in a rat model of spontaneous spike and wave seizures a similar interaction between the cortex and thalamus, with the cortex likely providing much of the drive to initiate the seizures. The present results for limbic seizures are very similar. One is tempted to take these observations from two distinct seizure types and generalize them into a prototypical seizure model that is built around thalamocortical circuits. Each component in such a hypothesized circuit plays a distinct role in the seizure, and the physiology of the overall seizure activity is dictated by the physiology of the individual components (e.g., spike and wave activity or tonic activity).

Such a division of function and physiology in a seizure circuit may also suggest that there are potentially separate targets for therapy with entirely different physiology and pharmacology. Thus it is important to know what the circuits are, and more importantly, where control points lie so that therapy can be directed at those points and not globally as it is done now with systemically administered drugs. In this study, and in others, the MD and perhaps some of its midline thalamic neighbors appear to function as a control point. Pharmacological manipulation of this area modulates the duration of the seizure (primary circuit) and the behavioral severity of the seizure (seizure spread).

Although one can speculate about the mechanisms underlying the effect of intervention on seizure duration, the mechanisms of seizure spread are still unclear. The early involvement of the MD nucleus in limbic seizures and the rapid generalization of the seizure following stimulation of this midline region suggest that it is significantly involved, but there are a number of other hypotheses. One hypothesis would focus on the MD/midline nuclei as a primary path of spread. In considering its known connections (projection and reciprocal) it is easy to construct a scenario in which seizure activity would spread through the MD nucleus to the neocortex (Krettek & Price, 1977; Groenewegen, 1988; Ray & Price, 1992; Kuroda et al., 2004). The MD-limbic connections get the seizure started, and the seizure then spreads to other sites to which this thalamic nucleus is connected. Against the MD as a primary route of spread is the high afterdischarge threshold, an observation that raises the possibility that the drive for the seizure will more likely come from elsewhere. Another current hypothesis routes the seizure spread through the parahippocampal gyrus to other regions of the brain (Kelly & McIntyre, 1996; Kelly et al., 2002; Zhang et al., 2001). This scenario is based on the connectivity of this region, as well as on experimental evidence suggesting the spread of seizures through this site following stimulation. The amygdala similarly has widespread connections and could also serve as a path for generalization (Krettek & Price, 1978; Amaral & Price, 1984; Kita & Kitai, 1990). This issue needs further investigation.

There are several recent studies that, at first glance, are less supportive of the present conclusions. In a study examining the link between temporal lobe seizures and the thalamus in humans Guye et al. (2006) showed that there was thalamic involvement, but that the temporal relationship was much more variable. Although there were a few seizures that appeared to have the strong linkage between limbic sites (hippocampus and entorhinal cortex) and the thalamus, in other seizures there was less apparent synchrony at the time of seizure onset. However, the recordings came from multiple sites in the thalamus, including the pulvinar and several of the lateral relay nuclei with only a small number from the MD (three total in the MD and ten in the pulvinar and lateral nuclei). There were a few seizures that had a very tight linkage, but it was never made clear in the displayed EEGs where in the thalamus any given recording originated so that no conclusions can be drawn regarding the degree of linkage between a limbic site and a particular thalamic nucleus from the data as presented. It was clear that there was very good synchronization in at least one seizure, but specific details about the recording sites were not provided. As we have shown in this study, the relationship between the thalamus and the limbic sites varies with the thalamic nuclei, and if the recording site is in the "wrong" thalamic nucleus, the activity may be very poorly synchronized. Blumenfeld et al. (2007) have examined the relationship between the amygdala and the MD during the kindling acquisition. They showed a later involvement of the thalamus than what we have seen, an observation that would call into question our observations or suggest that our results may be more specific to the hippocampus. Reviewing their techniques, they used animals that were much smaller than ours, and the stereotactic coordinates used would put the thalamic electrodes on the outside edge of the MD. Assuming, as we have seen ourselves that placement can vary by as much as a half millimeter and that the target in the younger animals was smaller, it is possible that many of the electrodes were lateral to the MD. As we have seen thalamic involvement is highly focal, so that slight shifts can make large differences in recordings. Although they reported confirmation, electrode location was not shown.

In conclusion, this study has demonstrated that the MD/dorsal midline thalamic region is an integral part of the limbic seizure circuit and may play a role in seizure spread. Synaptic modulation in this region can alter limbic seizure activity, an observation that may have implications for targeting therapy. The kindling studies suggest that it is likely that the primary drive to initiate limbic seizures comes from the limbic sites, with the MD region acting as a key control point. Finally the data strongly support the hypothesis that, in the initial stages of seizure activity, thalamic involvement is specific to the MD region.


  1. Top of page
  2. Materials and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

This study was supported by NIH/NINDS grant NS25605.

We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. The authors have no known conflicts of interest associated with the material contained in this manuscript.


  1. Top of page
  2. Materials and Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References
  • Amaral DG, Price JL. (1984) Amygdalo-cortical projections in the monkey (Macaca fascicularis). J Comp Neurol 230:465496.
  • Andrade DM, Zumsteg D, Hamani C, Hodaie M, Sarkissian S, Lozano AM, Wennberg RA. (2006) Long-term follow-up of patients with thalamic deep brain stimulation for epilepsy. Neurology 66:15713.
  • Avoli M, Gloor P. (1982) Interaction of cortex and thalamus in spike and wave discharges of feline generalized penicillin epilepsy. Exp Neurol 76:196217.
  • Bertram EH, Zhang DX, Mangan P, Fountain N, Rempe D. (1998) Functional anatomy of limbic epilepsy: a proposal for central synchronization of a diffusely hyperexcitable network. Epilepsy Res 32:194205.
  • Bertram EH, Williamson JM, Scott C, Mangan PS, Zhang DX. (2001) The midline thalamus: alterations and a potential role in limbic epilepsy. Epilepsia 42:967978.
  • Blumenfeld H, Rivera M, Vasquez JG, Shah A, Ismail D, Enev M, Zaveri HP. (2007) Neocortical and thalamic spread of amygdala kindled seizures. Epilepsia 48:254262.
  • Cassidy RM, Gale K. (1998) Mediodorsal thalamus plays a critical role in the development of limbic motor seizures. J Neurosci 18:90029009.
  • Dolleman-van der Weel MJ, Witter MP. (1996) Projections from the nucleus reuniens thalami to the entorhinal cortex, hippocampal field CA1, and the subiculum in the rat arise from different populations of neurons. J Comp Neurol 364:637650.
  • Du F, Whetsell WO, Abou-Khalil B, Blumenkopf B, Lothman EW, Schwarcz R. (1993) Preferential neuronal loss in layer III of the entorhinal cortex in patients with temporal lobe epilepsy. Epilepsy Res 16:223233.
  • Groenewegen HJ. (1988) Organization of the afferent connections of the mediodorsal thalamic nucleus n the rat, related to the mediodorsal-prefrontal topography. Neuroscience 24:379431.
  • Guye M, Regis J, Tamura M, Wendling F, McGonigal A, Chauvel P, Bartolomei F. (2006) The role of corticothalamic coupling in human temporal lobe epilepsy. Brain 129:19171928.
  • Herkenham M. (1978) The connections of the nucleus reuniens thalami: evidence for a direct thalamo-hippocampal pathway in the rat. J Comp Neurol 177:589610.
  • Hirayasu Y, Wada JA. (1992) N-methyl-D-aspartate injection into the massa intermedia facilitates development of limbic kindling in rats. Epilepsia 33:965970.
  • Hiyoshi T, Wada JA. (1988a). Midline thalamic lesion and feline amygdaloid kindling. I. Effect of lesion placement prior to kindlingl. EEG Clin. Neurophysiol 70:325338.
  • Hiyoshi T, Wada JA. (1988b). Midline thalamic lesion and feline amygdaloid kindling. II. Effect of lesion placement upon completion of primary site kindling. EEG Clin Neurophys 70:339348.
  • Hudson LP, Munoz DG, Miller L, McLachlan RS, Girvin JP, Blume WT. (1993) Amygdaloid sclerosis in temporal lobe epilepsy. Ann Neurol 33:622631.
  • Juhász C, Nagy F, Watson C, Da Silva EA, Muzik O, Chugani DC, Shah J, Chugani HT. (1999) Glucose and [C-11]flumazenil positron emission tomography abnormalities thalamic nuclei in temporal lobe epilepsy. Neurology 53:20372045.
  • Kelly ME, McIntyre DC. (1996) Perirhinal cortex involvement in limbic kindled seizures. Epilepsy Res 26:233243.
  • Kelly ME, Staines WA, McIntyre DC. (2002) Secondary generalization of hippocampal kindled seizures in rats: examining the role of the piriform cortex. Brain Res 957:152161.
  • Kerrigan JF, Litt B, Fisher RS, Cranstoun S, French JA, Blum DE, Dichter M, Shetter A, Baltuch G, Jaggi J, Krone S, Brodie M, Rise M, Graves N. (2004) Electrical stimulation of the anterior nucleus of the thalamus for the treatment of intractable epilepsy. Epilepsia 45:346354.
  • Kita B, Kitai ST. (1990) Amygdaloid projections to the frontal cortex and the striatum in the rat. J Comp Neurol 298:4049.
  • Krettek JE, Price JL. (1977) The cortical projections of the mediodorsal nucleus and adjacent thalamic nuclei in the rat. J Comp Neurol 171:157191.
  • Krettek JK, Price JL. (1978) Amygdaloid projections to subcortical structures within the basal forebrain and brainstem in the rat and cat. J Comp Neurol 178:225254.
  • Kuroda M, Yokofujita J, Oda S, Price JL. (2004) Synaptic relationships between axon terminals from the mediodorsal thalamic nucleus and gamma-aminobutyric acidergic cortical cells in the prelimbic cortex of the rat. J Comp Neurol 477: 220234.
  • Lado FA. (2006) Chronic bilateral stimulation of the anterior thalamus of kainate-treated rats increases seizure frequency. Epilepsia 47:2732.
  • Lothman EW, Collins RC. (1981) Kainic acid induced limbic seizures: metabolic, behavioral, electroencephalographic and neuropathological correlates. Brain Res 218:299318.
  • Lothman EW, Bertram EH, Stringer JL. (1991) Functional anatomy of hippocampal seizures. Prog Neurobiol 37:182.
  • Margerison JH, Corsellis JAN. (1966) Epilepsy and the temporal lobes. A clinical, electroencephalographic and neuropathological study of the brain in epilepsy, with particular reference to the temporal lobes. Brain 89:499530.
  • McCaughran JA, Corcoran ME, Wada JA. (1978) Role of the nonspecific thalamus in amygdaloid kindling. Experiment Neurol 58:471485.
  • McKenna JT, Vertes RP. (2004) Afferent projections to nucleus reuniens of the thalamus. J Comp Neurol 480:115142.
  • Meeren HK, Pijn JP, Van Luijtelaar EL, Coenen AM, Lopes da Silva FH. (2002) Cortical focus drives widespread corticothalamic networks during spontaneous absence seizures in rats. J Neurosci 22:14801495.
  • Miller JW, Hall CM, Holland KD, Ferrendelli JA. (1989) Identification of a median thalamic system regulating seizures and arousal. Epilepsia 30:493500.
  • Miller JW, Ferrendelli JA. (1990) Characterization of GABAergic seizure regulation in the midline thalamus. Neuropharmacol 29:649655.
  • Mori N, Wada JA. (1992) Kindling of the massa intermedia of the thalamus in rats. Brain Res 575:148150.
  • Natsume J, Bernasconi N, Andermann F, Bernasconi A. (2003) MRI volumetry of the thalamus in temporal, extratemporal, and idiopathic generalized epilepsy. Neurology 60:12961300.
  • Patel S, Millan MH, Meldrum BS. (1988) Decrease in excitatory transmission within the lateral habenula and the mediodorsal thalamus protects against limbic seizures in rats. Exp Neurol 101:6374.
  • Paxinos G, Watson G. (1986) The rat brain in stereotaxic coordinates, 2nd ed. Academic Press, San Diego , CA .
  • Ray JP, Price JL. (1992) The organization of the thalamocortical connections of the mediodorsal thalamic nucleus in the rat, related to the ventral forebrain-prefrontal cortex topography. J Comp Neurol 323:167197.
  • Ray JP, Russchen FT, Fuller TA, Price JL. (1992) Sources of presumptive glutamatergic/aspartatergic afferents to the mediodorsal nucleus of the thalamus in the rat. J Comp Neurol 320:435456.
  • Stringer JL, Williamson JM, Lothman EW. (1989) Induction of paroxysmal discharges in the dentate gyrus: frequency dependence and relationship to afterdischarge production. J Neurophysiol 62:126135.
  • Su H-S, Bentivoglio M. (1990) Thalamic midline cell populations projecting to the nucleus accumbens, amygdala, and hippocampus in the rat. J Comp Neurol 297:582593.
  • Turner BH, Herkenham M. (1991) Thalamoamygdaloid projections in the rat: a test of the amygdala's role in sensory processing. J Comp Neurol 313:295325.
  • Van Der Werf YD, Witter MP, Groenewegen HJ. (2002) The intralaminar and midline nuclei of the thalamus. Anatomical and functional evidence for participation in processes of arousal and awareness. Brain Res Rev 39:107140.
  • Van Groen T, Wyss JM. (1990) Extrinsic projections from area CA1 of the rat hippocampus: olfactory, cortical, subcortical, and bilateral hippocampal formation projections. J Comp Neurol 302:515528.
  • Zhang X, Hannesson DK, Saucier DM, Wallace AE, Howland J, Corcoran ME. (2001) Susceptibility to kindling and neuronal connections of the anterior claustrum. J Neurosci 21:36743687.