• BOLD technique;
  • Functional MRI;
  • Anterior thalamus;
  • Retrosplenial cortex;
  • Dentate gyrus;
  • Pentylenetetrazol;
  • Ethosuximide


  1. Top of page
  2. Abstract
  6. Acknowledgments

Summary: Purpose: Functional imaging of animal models makes it possible to map the functional neuroanatomy contributing to the genesis of seizures. Pentylenetetrazol (PTZ)-induced seizure in rats, a relevant model of human absence and of generalized tonic–clonic epilepsy, was used to stimulate seizure activity within 30 s of administration while collecting continuous, high-resolution, multislice images at subsecond intervals.

Methods: Pilot studies were conducted to establish a quick and effective PTZ model for the imaging experiments. PTZ was then used to stimulate seizure activity in rats while collecting multislice functional MRI (fMRI) images from the entire forebrain at 4.7 Tesla. Ethosuximide (ESM) also was used to block seizure activity.

Results: Within 2–4 s of PTZ administration, a rapid increase in blood oxygen level–dependent (BOLD) signal intensity was noted in the thalamus, especially the anterior thalamic nuclei. Activity in the anterior thalamus peaked ∼15 s before seizure onset and was more than twofold greater than that in all other thalamic areas. The retrosplenial cortex showed a twofold greater increase in activity as compared with other cortical areas, also peaking at ∼15 s. The dentate gyrus was twice as active as other hippocampal areas but peaked just before seizure onset. Treatment with ESM blocked seizures, decreasing PTZ-induced activation in most forebrain areas. The anterior thalamus and retrosplenial cortex were essentially blocked by pretreatment with ESM.

Conclusions: The anterior thalamus, retrosplenial cortex, and dentate gyrus show the greatest increases in BOLD signal activity before seizure onset. Neurons in these areas may contribute to the neural network controlling the initiation of generalized tonic–clonic seizure.

Although data from chemical-stimulation, lesion, and electrophysiologic studies have identified multiple brain areas involved in the initiation and propagation of generalized tonic–clonic seizures (1), the precise neural pathway integrating these areas is unknown. Unlike focal seizures that originate from a single locus, generalized seizures arise from a complex integration of cortical and subcortical signals contributing to the characteristic synchronized paroxysmal brain activity (2). Coherence studies on EEG signals have focused on the reciprocal interactions between the cortex and thalamus in the genesis of generalized seizures. For example, Sherman et al. (3,4) identified the anterior thalamic nucleus as having a strong association with cortical EEG activity, acting as an essential gate in the propagation of pentylenetetrazol (PTZ)-induced generalized tonic–clonic seizure activity from subcortical areas to the cortex. Electrophysiologic studies like this are essential for understanding the genesis of generalized seizures because they provide a window of observation measured in tens to hundreds of milliseconds; however, they are limited to a small number of brain areas that can be sampled at a time. Functional magnetic resonance imaging (fMRI), although having a lower temporal resolution than electrophysiology, offers a method for observing global changes in neural activity that could be used to draw a map of putative neural substrates involved in the genesis of generalized seizures.

Recently the technical problems associated with imaging of fully conscious animals in ultra-high field MR spectrometers were resolved (5–7). Studying brain function in animals with the high temporal and spatial resolution of noninvasive MRI makes it possible to follow activation of neural pathways in a variety of behavioral and neurologic models ranging from sexual arousal in monkeys (8) to generalized absence seizures in rats (9,10). In the latter studies, activation of the corticothalamic circuitry responsible for generating and maintaining the spike–wave discharge in absence seizures was confirmed by using fMRI. The present study addresses the feasibility of using fMRI in fully conscious rats to follow the genesis of PTZ-induced generalized tonic–clonic seizures.

Van Camp et al. (11) performed the first simultaneous EEG and fMRI study on curarized rats given subcutaneous PTZ. Acquisition time for a multislice gradient-echo sequence with averaging was ∼1 min. Within 10 min of PTZ injection, animals showed behavioral signs typical of absence seizures, with facial twitching and head movement, followed 20–30 min later by episodes of tonic–clonic seizures. Bilateral spreading of both positive and negative blood oxygen level–dependent (BOLD) signal changes occurred over most of the cortex, hippocampus, amygdala, and brainstem. The areas of activation observed in these imaging studies were not unlike those identified in deoxyglucose metabolism studies and immunostaining for c-fos protein after PTZ seizures (12–17). Because of the slow onset and progressive development of seizure activity and the long image-acquisition times in this study, it was not possible to identify neural substrates involved in the genesis of PTZ-induced seizures. The present study was designed to image the functional neuroanatomy contributing to the genesis of generalized seizures by collecting multislice images at subsecond intervals. Because sensory feedback from the whole body contributes to the seizure threshold, these studies were done in nonparalyzed, fully conscious animals (18).


  1. Top of page
  2. Abstract
  6. Acknowledgments

Male Sprague–Dawley rats weighing 300–350 g were obtained from Charles River Laboratories (Charles River, MA, U.S.A.). Animals were housed in pairs and maintained on a 12:12-h light/dark cycle (lights on at 900 h) and provided food and water ad libitum. All animals were acquired and cared for in accordance with the guidelines published in the Guide for the Care and Use of Laboratory Animals (National Institutes of Health Publications no. 85–23, revised 1985).

Developing a seizure paradigm

Critical to these functional imaging studies was the development of a seizure paradigm that was rapid in onset, predictable, and robust. PTZ was chosen to elicit seizures because it is the prototypical chemical agent for stimulating generalized tonic–clonic seizures in a variety of animals, including humans (19). Because PTZ–induced seizures are blocked with ethosuximide (ESM), a drug used to control generalized absence epilepsy, the PTZ animal model has been used to screen clinical drug candidates for treatment of human epilepsy (20). However, a major problem with using fMRI to monitor PTZ-induced changes in brain activity is the route of administration of this epileptogenic drug. A bolus intravenous injection of PTZ, although effective in triggering clonic seizures in <1 min, causes significant decreases in blood pressure and increases in heart rate, complicating the interpretation of BOLD signal changes (21,22). Consequently, we chose to administer PTZ through the lateral cerebroventricle (ICV) to circumvent many of these problems.

Dose–response data on ICV-injected PTZ was collected from four animals. Before testing, animals were anesthetized with Domitor (1 mg/kg, medetomidine, Pfizer) and ketamine (10 mg/kg, ketamine). An ICV catheter of PE 10 (0.61 mm OD) tubing was implanted into the lateral ventricle through a burr hole in the skull [bregma coordinates: AP, –1 mm; ML, 2 mm; DV, 4 mm (23)] and secured with cyanoacrylate. Subcutaneous EEG electrodes were placed over the anterior and posterior regions of each cerebral hemisphere along with a reference electrode in the neck (Ives EEG Solutions, London, ON, Canada). Anesthesia was then reversed with Antisedan (5 mg/kg, atipamezole), and the animal was placed in an observation cage and video monitored. Each animal received four doses of PTZ (1, 2, 2.5, and 4 mg) dissolved in artificial CSF (145 mM NaCl, 2.7 mM KCl, 1 mM MgCl2 6H2O, 1.2 mM CaCl2 2H2O, 2 mM Na2HPO4, pH adjusted to 7.4), at 30-min intervals in a counterbalanced study. The PTZ was dissolved into artificial CSF and injected in a volume 20 μl. All behavioral responses were videotaped.

For the actual imaging studies, five animals were prepared as noted in the simulated environment minus the EEG leads and femoral catheter. During the fMRI study, it was not possible to collect simultaneous EEG recordings and brain images because of the rapid duty cycle of the pulse sequence causing EEG artifacts. Imaging was done in a 4.7-T, 40-cm spectrometer (Bruker Biospin, Billerica, MA, U.S.A.) with a 12-cm, 25 g/cm gradient insert. Anatomic images were taken with a fast spin echo sequence with the following parameters: TR, 2 s; effective TE, 48 ms; eight NEX; eight echos; 256 × 256; 2.8-cm FOV; six slices; and 1.5-mm slice thickness. Functional images were acquired with a single-shot spin-echo, echo-planar sequence with identical geometry as the anatomic sequence and imaging parameters as follows; TR, 600 ms; TE, 60 ms; 64 × 64 matrix; 260 repetitions; 2.8-cm FOV; six slices; and 1.5-mm slice thickness.

An imaging session began with the collection of the anatomic dataset. Functional imaging started ∼30 min after anesthesia reversal, beginning with a vehicle injection of 20 μl of artificial CSF. Images were collected continuously at 0.6-s intervals over 2.6 min. A series of 100 baseline images were collected over the first minute, followed by ICV CSF injection and collection of 160 images. After 20 min, this procedure was repeated again but with 1.5-mg PTZ. In a second study, five animals were prepared for imaging as noted earlier but pretreated with an intraperitoneal injection of 250 mg/kg of ESM 1 h before scanning. Animals were given an ICV injection of CSF followed by PTZ. All animals were killed after the study by an overdose of sodium pentobarbital (100 mg/kg).

Data analysis

Each subject was individually analyzed for motion artifacts and improper cannula placement. No criteria were used for excluding any of the animals from the final data analysis. Each animal was registered or aligned to a fully segmented rat brain atlas that delineates >1,200 distinct anatomic subvolumes within the brain, based on 2D atlas textbooks (23,24). These detailed regions are collected into 12 major regions of the brain (e.g., amygdaloid complex, cerebrum, cerebellum, hypothalamus). The alignment process began by outlining the brain perimeters for each slice of the anatomy image sets. A marching cube algorithm with automated linearization creates accurate 3D surface shells for each subject (25,26). This enhanced surface-generation strategy eliminates the characteristic stair-stepped behavior of the marching cube algorithms while simultaneously increasing the accuracy of the geometry representation. These anatomy shells were aligned to the atlas shell. The affine registration involved translation, rotation, and scaling in all three dimensions, independently.

The matrices that transformed the subject's anatomy shells to the atlas space were used to embed each slice within the atlas. All transformed pixel locations of the anatomy images were tagged with the segmented atlas major and minor regions, creating a fully segmented representation of each subject. The inverse transformation matrix [Ti]−1 for each subject (i) was also calculated. An interactive graphical user interface (GUI) facilitated these shell alignments (27). Approximately 20 min per subject was required to create the slice perimeters, run the marching cube, align the geometries, and create the final segmented anatomy.

Statistical t tests were performed on each subject within the original coordinate system. The control window was the first 100 data acquisitions (1 min). The stimulation window was the first 50 data acquisitions (30 s) after PTZ injection before seizure onset. The t test statistics used a 95% confidence level, two-tailed distributions, and heteroscedastic variance assumptions. Because of the multiple t test analyses performed, a false-positive detection–controlling mechanism was introduced (28). This subsequent filter guaranteed that, on average, the false-positive detection rate was below our cutoff of 0.05. The formulation of the filter satisfied the following expression:

  • image

where P(i) is the p value based on the t test analysis. Each pixel (i) within the ROI containing (V) pixels was ranked based on its probability value. The false-positive filter value q was set to be 0.05 for our analyses, and the predetermined constant c(V) was set to unity, which is appropriate for data containing gaussian noise, such as fMRI data (28). These analysis settings provided conservative estimates for significance. Those pixels deemed statistically significant retained their percentage change values, (stimulation mean minus control mean) relative to control mean. All other pixel values were set to zero.

The composite BOLD activation pattern from all subjects is presented mapped to the segmented atlas, as shown in Figs. 3 and 5. The BOLD response maps of the composite are somewhat broader in their spatial coverage than an individual subject. Consequently, the magnitudes of the statistical values were commensurately reduced. This outcome was a natural consequence of the inability to align each subject within a pixel resolution. However, the subjects were aligned very well at the resolution of the ROIs. The time-history graphs for each composite region were based on the weighted average of each subject for that region.

  • image
  • image

where N is the number of subjects.


Figure 3. Blood oxygen level–dependent (BOLD) activation map after intracerebroventricular pentylenetetrazol (PTZ). Shown is the pattern of PTZ-induced changes in BOLD signal before seizure onset (first 30 s after PTZ injection). The composite map of all subjects (n = 5) is shown registered to six contiguous coronal sections of the segmented atlas. The scale bar shows the percentage change in BOLD signal intensity.

Download figure to PowerPoint


Figure 5. Blood oxygen level–dependent (BOLD) activation map in the presence of ethosuximide (ESM). Shown is the pattern of pentylenetetrazol (PTZ)-induced changes in BOLD with ESM blockade of seizure (first 30 s after PTZ injection). The composite map of all subjects (n = 5) is shown registered to six contiguous coronal sections of the segmented atlas. The scale bar shows the percentage change in BOLD signal intensity.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  6. Acknowledgments

The process of developing this model involved the validation and dose determination of the ICV injection, determining the stress and physiologic effects, and investigating the vehicle effects of the injection. From these dose–response studies, it was determined that the 1.5-mg dose of PTZ would predictably trigger clonic seizure activity ∼30 s after administration without significantly altering respiratory or cardiovascular function (Figs. 1 and 2, Table 1). A typical response consisted of a myoclonic jerk of the head and neck clonus starting in the face and spreading to forelimbs, culminating in whole body clonus and loss of posture with twisting–writhing. Tonus or tonic flexion with forelimb extension and hindlimb flexion was not observed in any of the animals tested at the 1.5-mg dose. EEG recordings showed repetitive hypersynchronous discharges which increased in frequency (Fig. 2). Behavioral and EEG measures of seizure activity lasted for ∼3–5 min, as confirmed by video monitoring (see supplemental material). Animals pretreated with 250 mg/kg of ESM 1 hour before PTZ injection showed complete blockage of behavioral and EEG measures of seizure activity (Fig. 2).


Figure 1. Pentylenetetrazol dose–response after intracerebroventricular injection. A: Onset of EEG spike discharges and behavioral clonic movements. B: Time to recovery.

Download figure to PowerPoint


Figure 2. EEG recordings of pentylenetetrazol (PTZ)-induced seizures and blockade with ethosuximide (ESM). Shown are bilateral EEG traces at various intervals after PTZ injection in the absence and presence of ethosuximide. Samples taken at 30 s after an injection of PTZ show spike discharges that increase in frequency over time. Spike discharges were absent after ESM treatment.

Download figure to PowerPoint

Table 1. Physiologic changes caused by vehicle, pentylenetetrazol, and pentylenetetrazol + ethosuximide
 Respiration RateHeart RateBlood Pressure
  1. Pentylenetetrazol caused a significant change in respiration rate after 30 s (p< 0.01) but had no other effects. Standard deviation is in parentheses. *p value >0.01.

Vehicle (n = 4)
 Pre Injection141 (9.2)  414 (28.2)121 (7.2) 
 30s Post Inj.148 (11.7)`398 (35.3)117 (3.3) 
PTZ (n = 4)
 Pre Injection135 (21.8)409 (14.5)119 (18.7)
 30s Post Inj. 177 (31.1)*418 (33.2)106 (23.9)
ESM+PTZ (n = 4)
 Pre Injection124 (17.3)409 (17.0)113 (32.7)
 305 Post Inj.138 (30.3)413 (22.9)108 (28.5)

To control for the effect of restraint stress and auditory stimulation on seizure threshold, four animals were surgically prepared as described and secured in a commercially available MR-compatible restraint device with built-in imaging coils (Insight Neuroimaging Systems, Worcester, MA, U.S.A.). In addition, a catheter was placed in the femoral artery, a pneumatic thoracic force transducer positioned over the chest, and a pulse oximeter (Nonin Medical, Plymouth, MN, U.S.A.) connected to the tail. While restrained and exposed to a taped recording of the noise from an imaging experiment, animals were injected ICV with 1.5 mg of PTZ in 20 μl of artificial CSF. Heart rate, blood pressure, respiratory rate, and EEG were continuously recorded. Again, seizure activity was routinely elicited within ∼30 s from the time of PTZ injection. No significant change in blood pressure or heart rate within 30 s of the ICV injection was seen in vehicle controls, PTZ-treated animals and PTZ-plus-ESM animals. The respiration rate increased in the first 30 s from 135 ± 13 to 177 ± 23 (p < 0.01) for the PTZ-treated group (Table 1) but remained unchanged for control and ESM-treated animals.

Imaging of ICV vehicle injections of 20 μl of artificial CSF produced changes in signal intensity along the cerebroventricles immediately after injection. Analysis of this signal change showed that the injection volume spread down the ventricular system in <3 s from its site of origin in the lateral ventricle. This vehicle injection had a modest but significant 2–3% increase in BOLD signal in all of the regions of interest (ROIs). The increase in BOLD signal was transient, lasting for only 5–7 s. Ipsilateral and contralateral analysis of bilateral ROIs showed no lateralization effect from the unilateral injection, including the anterior cingulate nearest the cannula. Although the cannula was placed in the rostral end of the lateral ventricle, the first ROIs to activate with PTZ alone were more caudal: the anterior thalamus and retrosplenial cortex (Fig. 4). Data from the vehicle injection indicate that vehicle has a minimal contribution to the increases seen after PTZ administration.


Figure 4. Pentylenetetrazol (PTZ)-induced changes in blood oxygen level–dependent (BOLD) signal intensity over time. Shown are composite time course data for PTZ-induced changes in BOLD signal in the absence (n = 5) and presence (n = 5) of ethosuximide. Changes are presented as an aggregate of whole-brain activity (top panels) and with subsets of major brain areas (lower panels).

Download figure to PowerPoint

An activational map showing PTZ-induced positive BOLD signal changes is shown in Fig. 3. All major brain areas (e.g., thalamus, hippocampus, amygdala, and cerebrum) showed a significant increase in BOLD signal (Fig. 4). The most salient feature in the forebrain activation pattern was the rapid and robust increase in BOLD signal in the anterior thalamus, retrosplenial cortex, and dentate gyrus (Fig. 4). Within 3–4 s of PTZ injection, the anterior thalamus showed a rapid rise in BOLD signal, peaking at ∼15 s before seizure onset. Within 4–6 s of PTZ injection, the retrosplenial cortex was activated, also peaking at ∼15 s before seizure (Fig. 4). The anterior thalamus and retrosplenial cortex stood out because their change in BOLD signal intensity was more than twofold greater than that in other thalamic and cortical areas. Activity in the dentate gyrus was also twice as high as in other hippocampal areas but peaked ∼25 sec just before seizure onset (Fig. 4). The amygdala in general was activated by PTZ treatment, but no particular nuclear area distinguished itself.

At the onset of generalized seizure activity, a decrease in BOLD signal intensity was found for most ROIs (Fig. 4). In the case of the parietal cortex and retrosplenial cortex, this decrease was precipitate and robust. Indeed, all cortical areas showed robust decreases in BOLD signal during seizures that exceeded most other forebrain areas. The thalamus, particularly the anterior and dorsolateral areas, appeared to sustain their BOLD activation during seizures. The same was true of the dentate gyrus.

The ESM blockade of PTZ-induced seizures occurred in the face of modest increases in BOLD activity in all of the major brain areas (Fig. 5). The onset of activity in all ROIs was delayed by 10–12 s as compared with PTZ treatment alone. The medial dorsal thalamus was still responsive to PTZ in the presence of ESM, whereas the anterior and lateral dorsal thalami were effectively blocked (Fig. 4). All cortical areas showed sustained activity over the first minute of the data-acquisition period after PTZ treatment. Both the dentate and CA1 showed sustained activity, but CA3 collapsed into negative BOLD signal ∼40 s after PTZ treatment. Interestingly, only the cortical nucleus of the amygdala showed an immediate response to PTZ in the face of ESM blockade that was similar to PTZ alone.


  1. Top of page
  2. Abstract
  6. Acknowledgments

This study describes a method for predictably triggering generalized clonic seizure activity in awake rodents that is compatible with BOLD imaging. The acquisition of continuous, high-resolution, multislice images at subsecond intervals makes it possible to image neural substrates activated by PTZ before the onset of seizure activity. These PTZ-sensitive brain areas are all potential candidates contributing to the genesis of generalized seizures. Based on their rapid and robust change in BOLD signal intensity, the hippocampus, particularly the dentate gyrus, anterior thalamus, and the retrosplenial cortex, appears to play a prominent role in triggering seizure activity. Pretreatment with ESM delays the onset of activity in these areas and reduces their BOLD signal change. The anterior thalamic nuclei are essentially blocked as their activity is decreased from a 20% change in BOLD signal intensity to <2%.

The sensitivity of hippocampal and cortical neurons to PTZ-induced seizures was shown in the landmark article of Morgan et al. (13) measuring expression of c-fos messenger RNA and c-Fos protein. Several subsequent articles using immediate-early gene expression as a measure of neuronal activity underscored the sensitivity of these brain areas to PTZ-induced seizure (14,29). Adult rats limited to clonic seizures, as reported here, show strong c-Fos labeling in the different hippocampal subfields, piriform, cingulate, entorhinal, olfactory and parietal cortices, septum, amygdala, dorsal striatum, globus pallidus, and hypothalamus (15,17). This pattern of neuronal activation is much greater than that reported here by using BOLD imaging and most likely reflects the sampling time. The BOLD activation map show in Fig. 4 reflects neuronal activity before the onset of seizure, whereas measures of c-fos activity are routinely acquired 1–4 h after seizure. The low-to-moderate BOLD-activation pattern in areas like the piriform, entorhinal, and olfactory cortices would suggest they are not involved in the genesis of seizures but are more involved in the expression and maintenance of seizures.

The robust activation of the anterior thalamic area is not unexpected because it is know to be a critical neuroanatomic substrate, together with the mammillary bodies, in the regulation of PTZ-induced seizures. Multiunit electrical activity increases dramatically in the mesencephalon and thalamus ∼8–10 s before the onset of paroxysmal EEG discharges in cats treated with intravenous PTZ (30). Conversely, mesencephalic and prethalamic lesions reduce PTZ-induced electrical activity in cats (31), suggesting an ascending mesencephalic/thalamic pathway contributing to the genesis of generalized seizure. Studies by Mirski et al. (32–34) extended this early work in cats to focus on an epileptogenic pathway in guinea pig consisting of the mammillary bodies and anterior thalamus, as critical for PTZ-induced seizure activity. Manipulating the mammillary bodies and their connections to the anterior thalamus via the mammillothalamic tracts, with lesions, chemical injections, and frequency-specific electrical stimulations, blocks or reduces PTZ seizures. Similar experimental manipulations of the anterior thalamus also block or reduce PTZ seizures (35). The anterior thalamus appears to be the gateway to the cortex regulating PTZ-induced paroxysmal electrical activity, as evidenced by EEG coherence studies (3,4).

Given the findings in cats and guinea pigs reporting a mesencephalic/limbic pathway contributing to the genesis of PTZ-induced seizures, it was surprising to observe that the mammillary bodies (not shown; Figs. 3 and 5 activation maps) in this study were ostensibly devoid of BOLD signal activity. Several possible explanations account for this discrepancy. In the present study, PTZ was given in the lateral cerebroventricle in fully conscious rats, triggering seizures in seconds. In other studies, PTZ was always given intravenously in cats or intraperitoneally in guinea pigs under different experimental conditions with variable delays in seizure onset. The differences in drug administration, animal species, and experimental preparations may all contribute to this discrepancy.

The possibility exists that the mammillary bodies are not as critical to seizure initiation as once thought. In this imaging study, it can be observed that areas immediately adjacent to the mammillothalamic tracts (e.g., posterior hypothalamus, midline thalamic nuclei, ventromedial and anteromedial thalamus) are activated by PTZ before seizure onset. If these areas are important in the initiation of seizures, any electrolytic lesion aimed at the mammillothalamic tracts might affect their function and reduce seizure activity. This possibility was noted by Mirski and Ferrendelli (36). In an earlier publication, these authors reported no apparent increase in neuronal metabolic activity in the mammillary bodies as measured with [14C]2-deoxyglucose uptake in guinea pigs despite PTZ-induced EEG epileptiform discharges (37). However, when pretreated with ESM followed by PTZ, a significant accumulation of label occurred in the mammillary bodies, suggesting that this brain area may be involved in the anticonvulsant action of ESM. The findings from this metabolic activation study were corroborated here, as PTZ alone caused no change in BOLD signal intensity in the mammillary bodies, but in the presence of ESM, a significant increase in signal contrast was noted.

No evidence from our studies indicates that cortical activation precedes subcortical activation during the genesis of PTZ-induced generalized seizures. Data from sleep–wake studies and seizure studies point to a corticothalamic connection regulating the generation of cortical EEG patterns (38,39). Partial coherence estimations show that the cortex and anterior thalamus are synaptically correlated during cortical EEG and behavioral seizure expression (4). Time-delay calculation using EEG activity suggests the anterior thalamic nucleus leads in propagation of epileptiform activity (3).

The decrease in BOLD activity during the start of the clonic episode (∼30 s; Figs. 4 and 5), particularly in the cortical areas, is noteworthy and most likely does not reflect a decrease in neuronal activity. Caspers and Speckman (40) collected simultaneous recordings of paroxysmal electrical activity and tissue po2 at the surface of the cerebral cortex during PTZ-induced seizure. Seizure initiation was always preceded by a decrease in tissue po2. This transient tissue hypoxia caused by PTZ seizure can also be measured by a change in reduction/oxidation state of cytochrome oxidase in rat brain before systemic changes in blood pressure or heart rate (22). The presence of the resulting paramagnetic deoxygenated hemoglobin would cause proton dephasing and a decrease in BOLD signal intensity (41). The most dramatic negative BOLD signal at the putative onset of seizures occurred in the parietal and retrosplenial cortices and the basolateral amygdala. This would suggest that these areas have an exceedingly high PTZ-induced metabolic demand without appropriate compensatory blood flow.

The general effect of ESM blockade of PTZ-induced seizure is an overall reduction in BOLD signal intensity. The cortices and hippocampal subfields all show significant and sustained activation above baseline but without initiation of seizures. As noted earlier, a key neural substrate in PTZ-induced generalized clonic seizures is the anterior thalamus. The suppression of PTZ-induced changes in BOLD activity by ESM in this brain area may be responsible, in part, for the blockade of seizures. Limbic and association cortices connect to the dentate gyrus of the hippocampus through the angular bundle of the entorhinal cortex. In turn, the hippocampus has direct efferent connections to the anterior thalamus. Reducing the level of activity in the neural populations connecting to the anterior thalamus may be sufficient to suppress seizure activity. It also is possible that ESM blockade of the anterior thalamus prevents feed-forward recruitment and synchronization of limbic and association cortices critical for seizure initiation.

In summary, this study shows it is technically feasible to use fMRI and BOLD imaging to follow changes in brain activity associated with the genesis of generalized clonic seizures. The fact that these studies can be done in awake animals is a major advantage because it avoids the confound of anesthesia. From these studies, it was possible to generate activational maps of the forebrain identifying neural substrates that may be directly involved in the genesis of PTZ-induced generalized clonic seizures. The anterior thalamic nuclei and the dentate gyrus of the hippocampus were distinguished from other brain sites by their rapid and robust activation profile. The prominence of these two areas in the neuroanatomy of generalized seizures is well documented. One potentially important finding gleaned from these imaging studies is the observation that the retrosplenial cortex is more active than other cortical areas before seizure onset. Its rapid and robust activation profile and dramatic collapse in BOLD signal intensity coinciding with the start of seizure suggest a brain area of high metabolic activity. The anterior thalamus, particularly the anteroventral nucleus and its caudal extension, the lateral dorsal nucleus, have extensive reciprocal connections with the retrosplenial (posterior cingulate) cortex (42). Of all of the midline cortical areas, the retrosplenial has the greatest efferent association network extending rostrally to the anterior cingulate and prelimbic areas and caudally to the presubiculum. Hence the retrosplenial cortex may play a critical role in integrating thalamocortical activity during the initiation of generalized clonic seizures.


  1. Top of page
  2. Abstract
  6. Acknowledgments

Acknowledgment:  This work was funded by a grant from the National Institute of Mental Health, Division of Neuroscience and Behavioral Science, R01 MH52280, to C.F. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIMH.


  1. Top of page
  2. Abstract
  6. Acknowledgments
  • 1
    Avoli M, Gloor P, Kostopoulos G, et al. Generalized Epilepsy: Neurobiological Approaches. Boston : Birkhauser, 1990.
  • 2
    Blumenfeld H. From molecules to networks: cortical/subcortical interactions in the pathophysiology of idiopathic generalized epilepsy. Epilepsia 2003;44: 715.
  • 3
    Sherman DL, Tsai YC, Rossell LA, et al. Spectral analysis of a thalamus-to-cortex seizure pathway. IEEE Trans Biomed Eng 1997;44: 657664.
  • 4
    Mirski MA, Tsai YC, Rossell LA, et al. Anterior thalamic mediation of experimental seizures: selective EEG spectral coherence. Epilepsia 2003;44: 355365.
  • 5
    Ludwig R, Bodgdanov G, King J, et al. A dual RF resonator for high-field functional magnetic resonance imaging of small animals. J Neurosci Methods 2004;132: 125135.
  • 6
    Brevard ME, Duong TQ, King JA, et al. Changes in MRI signal intensity during hypercapnic challenge under conscious and anesthetized conditions. Magn Reson Imaging 2003;21: 9951001.
  • 7
    Febo M, Segarra AC, Tenney JR, et al. Imaging cocaine-induced changes in the reward system in conscious rats. J Neurosci Methods 2004;139: 167176.
  • 8
    Ferris CF, Snowdon CT, King JA, et al. Activation of neural pathways associated with sexual arousal in non-human primates. J Magn Reson Imaging 2004;19: 168175.
  • 9
    Tenney JR, Duong TQ, King JA, et al. Corticothalamic modulation during absence seizures in rats: a functional MRI assessment. Epilepsia 2003;44: 11331140.
  • 10
    Tenney JR, Duong TQ, King JA, et al. Functional MRI of brain activity in a genetic rat model of absence seizures. Epilepsia 2004;45: 576582.
  • 11
    Van Camp N, D'Hooge R, Verhoye M, et al. Simultaneous electroencephalographic recording and functional magnetic resonance imaging during pentylenetetrazol-induced seizures in rat. Neuroimage 2003;3: 627636.
  • 12
    Ben-Ari Y, Trembly E, Riche D, et al. Electrographic, clinical land pathological alterations following systemic administration of kainic acid, bicuculline or pentetrazole: metabolic mapping using the deoxyglucose methods with special reference to the pathology of epilepsy. Neuroscience 1981;6: 13611391.
  • 13
    Morgan JI, Cohen DR, Hempstead JL, et al. Mapping patterns of c-fos expression in the central nervous system after seizure. Science 1987;237: 192197.
  • 14
    Dragunow M, Robertson HA. Generalized seizures induce c-fos protein(s) in mammalian neurons. Neurosci Lett 1987;82: 157161.
  • 15
    Shehab S, Coffey P, Dean P, et al. Seizures induced by pentylenetetrazole and maximal electroshock. Exp Neurol 1992;118: 261274.
  • 16
    Mirski MA, Ferrendelli JA. Selective metabolic activation of the mammillary bodies and their connections during ethosuximide-induced suppression of pentylenetetrazol seizures. Epilepsia 1986;27: 194203.
  • 17
    Andre V, Pineau N, Motte JE, et al. Mapping of neuronal networks underlying generalized seizures induced by increasing doses of pentylenetetrazol in the immature and adult rat: a c-Fos immunohistochemical study. Eur J Neurosci 1998;10: 20942106.
  • 18
    Rodin E, Onuma T, Wasson S, et al. Neurophysiological mechanisms involved in grand mal seizures induced by metrazol and megimide. Electroencephalogr Clin Neurophysiol 1971;30: 6272.
  • 19
    Fisher RA. Animal models of the epilepsies. Brain Res Rev 1989;14: 245278.
  • 20
    De Deyn PP, D'Hooge R, Marescau B, et al. Chemical models of epilepsy with some reference to their applicability in the development of anticonvulsants. Epilepsy. Res 1992;12: 87110.
  • 21
    Kontos HA, et al. Role of tissue hypoxia in local regulation of cerebral microcirculation. Am J Physiol 1978;234: H582H591.
  • 22
    Hoshi Y, Tamura M. Oxygen Transport to Tissue XIII. New York : Plenum Press, 1992.
  • 23
    Paxinos G, Watson C. The Rat Atlas in Stereotaxic Coordinates. 2nd ed. New York : Academic Press, 1991.
  • 24
    Swanson LW. Brain Maps: Structure of the Rat Brain. 2nd ed. New York : Elsevier, 1998.
  • 25
    Wu Z, Sullivan JM Jr. Multiple material marching cubes algorithm. Int J Numer Methods Eng 2003;58: 189207.
  • 26
    Huang W, et al. Stair-stepped removal via automatic linearization for marching cubes formulations. Proceedings of the 12th International Society of Magnetic Resonance Medicine 2004: 756.
  • 27
    Zhang JQ, et al. Intermodality registration via Interactive graphics coupled with A.I.R. intramodality registration. Proceedings of the 11th International Society of Magnetic Resonance Medicine 2003: 3430.
  • 28
    Genovese CR, Lazar NA, Nichols T. Thresholding of statistical maps in functional neuroimaging using the false discovery rate. Neuroimage 2002;15: 870878.
  • 29
    Saffen DW, Cole AJ, Worley PF, et al. Convulsant-induced increase in transcription factor messenger RNAs in rat brain. Proc Natl Acad Sci U S A 1988;85: 77957799.
  • 30
    Velasco F, Velasco M, Estrada-Villanueva F, et al. Specific and nonspecific multiple unit activities during the onset of pentylenetetrazol seizures, I: intact animals. Epilepsia 1975;16: 207214.
  • 31
    Velasco F, Velasco M, Maldonado H, et al. Specific and nonspecific multiple unit activities during the onset of pentylenetetrazol seizures, II: acute lesions interrupting nonspecific system connections. Epilepsia 1976;17: 461475.
  • 32
    Mirski MA, Ferrendelli JA. Interruption of the mammillothalamic tract prevents seizures in guinea pigs. Science 1984;226: 7274.
  • 33
    Miller JW, McKeon C, Ferrendelli JA. Functional anatomy of pentylenetetrazol and electroshock seizures in the rat brainstem. Ann Neurol 1987;22: 615621.
  • 34
    Mirski MA, Rossell LA, Terry JB, et al. Anticonvulsant effect of anterior thalamic high frequency electrical stimulation in the rat. Epilepsy Res 1997;28: 89100.
  • 35
    Kusske JA, Ojemann GA, Ward AA Jr. Effects of lesions in ventral anterior thalamus on experimental focal epilepsy. Exp Neurol 1972;34: 279290.
  • 36
    Mirski MA, Ferrendelli JA. Interruption of the connections of the mammillary bodies protects against generalized pentylenetetrazol seizures in guinea pigs. J Neurosci 1987;7: 662670.
  • 37
    Mirski MA, Ferrendelli JA. Selective metabolic activation of the mammillary bodies and their connections during ethosuximide-induced suppression of pentylenetetrazol seizures. Epilepsia 1986;27: 194203.
  • 38
    Velasco M, Lindsley DB. Role of orbital cortex in regulation of thalamocortical electrical activity. Science 1965;149: 13751377.
  • 39
    Velasco M, Skinner JE, Asaro KD, et al. Thalamo-cortical systems regulating spindle bursts and recruiting responses, I: effect of cortical ablations. Electroencephalogr Clin Neurophysiol 1968;5: 463470.
  • 40
    Caspers H, Speckman EJ. Cerebral pO2, pCO2 and pH: changes during convulsive activity and their significance for spontaneous arrest of seizures. Epilepsia 1972;13: 699725.
  • 41
    Ogawa S, Lee TM, Kay AR, et al. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad.Sci U S A 1990;24: 98689872.
  • 42
    Domesick VB. Projections from the cingulate cortex of the rat. Brain Res 1969;12: 296320.