Search tips
Search criteria 


Logo of epicurrLink to Publisher's site
Epilepsy Curr. 2016 May-Jun; 16(3): 164–165.
PMCID: PMC4913848

Seizure Activity and Intervention Efficacy Are Shaped by REMnants of Preceding Brain States


Brain State Is a Major Factor in Preseizure Hippocampal Network Activity and Influences Success of Seizure Intervention.

Ewell LA, Liang L, Armstrong C, Soltesz I, Leutgeb S, Leutgeb JK. J Neurosci 2015;35(47):15635–15648 [PubMed].

Neural dynamics preceding seizures are of interest because they may shed light on mechanisms of seizure generation and could be predictive. In healthy animals, hippocampal network activity is shaped by behavioral brain state and, in epilepsy, seizures selectively emerge during specific brain states. To determine the degree to which changes in network dynamics before seizure are pathological or reflect ongoing fluctuations in brain state, dorsal hippocampal neurons were recorded during spontaneous seizures in a rat model of temporal lobe epilepsy. Seizures emerged from all brain states, but with a greater likelihood after REM sleep, potentially due to an observed increase in baseline excitability during periods of REM compared with other brains states also characterized by sustained theta oscillations. When comparing the firing patterns of the same neurons across brain states associated with and without seizures, activity dynamics before seizures followed patterns typical of the ongoing brain state, or brain state transitions, and did not differ until the onset of the electrographic seizure. Next, we tested whether disparate activity patterns during distinct brain states would influence the effectiveness of optogenetic curtailment of hippocampal seizures in a mouse model of temporal lobe epilepsy. Optogenetic curtailment was significantly more effective for seizures preceded by non-theta states compared with seizures that emerged from theta states. Our results indicate that consideration of behavioral brain state preceding a seizure is important for the appropriate interpretation of network dynamics leading up to a seizure and for designing effective seizure intervention.

One of the holy grails of epilepsy research and treatment has been the identification of characteristic markers of pathological activity that may signal an oncoming seizure. Elucidation of such hallmarks could profoundly improve the ability to predict and prevent a seizure before it develops (1). It has been hypothesized that epileptogenic and hyperexcitable brain circuits exhibit changes in activity in the short time window immediately preceding seizures, thus setting the stage for the generation and propagation of seizure activity. Indeed, several studies have indicated altered activity of both hippocampal inhibitory interneurons and excitatory principal cells leading up to seizures in temporal lobe epilepsy (TLE) (2–4). Some of these changes, however, are also typical of neuronal activity displayed during certain brain states, and in particular, the transitions from one state to another.

Many neural networks oscillate at frequencies that are distinctive of certain activity levels and arousal states, and signified by changes in the oscillatory frequency of local field potential recordings. For example, 2–4 Hz oscillations, known as delta (δ), are associated with slow-wave sleep (SWS). Higher-frequency oscillations in the theta (θ) band, at 6–10 Hz, are characteristic of rapid eye movement (REM) sleep, during which dreaming occurs. (Note that theta activity in humans is designated at the slightly slower 4–7 Hz frequency). Importantly, theta activity is also observed during exploratory movement in awake states. As theta frequency neuronal activity patterns have recently been observed in the hippocampus prior to seizures (2), it would be important to determine if the specific behavioral state associated with the theta rhythm (i.e., asleep versus awake) shapes features of neural activity leading up to and during seizures.

In this study, young adult rats were subjected to status epilepticus induced by sequential low systemic doses of kainate, then observed for several months for the development of spontaneous behavioral seizures, and thus the presence of epilepsy. Six epileptic rats were implanted with multiple movable tetrodes, which permitted the recording of several sites in the hippocampus across CA1, CA3, and dentate gyrus. Both local field potential and single-unit activity were recorded while the rats were allowed to do as they please and during periods when the rats were induced to forage for food treats to drive the development of theta rhythm associated with movement. Thus the rats could be observed and recorded during non-theta (slow-wave sleep) and theta states (REM sleep and foraging), and, importantly, the theta states could be directly associated with the arousal state of the animal. This permitted the researchers to directly compare seizures and neural activity during the period immediately preceding the seizures across these different brain states in the same animals.

Interestingly, seizures in this model were disproportionately associated with REM sleep. Despite the fact that on average the rats spent 5% of time in REM sleep, 16 of 33 seizures examined (~48%) occurred in this brain state. The presence of seizures, however, did not interfere with the overall amount of time spent in REM sleep, indicating that the seizures were occurring at the end of the REM period or at times of transition back to SWS. This suggests that at times of transition between brain states, certain neurons and circuits may be especially vulnerable to seizures. Intriguingly, however, the patterns of neuronal firing-rate changes appeared to more closely reflect the underlying brain state rather than the presence of seizures. For example, inhibitory interneurons typically increased their firing rate at the end of SWS leading up to REM, maintained this high level through the REM period, and then decreased firing whether the transition was back to SWS or into a seizure. Thus, overall, the pattern of activity was similar for REM periods that were and were not associated with seizures. A similar lack of seizure-associated changes in firing rates was also observed for seizures during SWS, although a subpopulation of interneurons (21%) decreased firing immediately before seizure onset. Note that this is the opposite pattern observed in REM sleep. Furthermore, comparison of firing rates across the same cells during REM-associated versus foraging-associated theta activity indicated higher levels of excitatory principal cell firing and decreased activity in ~35% of interneurons in REM, suggesting a potential substrate for increased excitability during REM sleep. Thus, somewhat surprisingly, patterns of activity both immediately preceding and during seizures appeared not to reflect pathological shifts signaling an oncoming seizure but rather maintenance of the normal pattern of activity characteristic of the particular brain state from which the seizures arose.

If the pattern of activity leading up to seizures reflects the underlying brain state, and is thus heterogeneous across seizures superimposed on different brain states, this may influence the efficacy of specific therapeutic interventions. A recent study demonstrated that optogenetic activation of parvalbumin-positive interneurons could reduce seizure duration in the intrahippocampal kainate mouse model of TLE (5). In the present work, this data set was reexamined to determine if the preceding brain state influenced the ability of this optogenetic stimulation to reduce seizure duration. Remarkably, seizures that arose from non-theta states (which were associated with decreased interneuron activity in the rat recordings) could be aborted by the activation of interneuronal activity, but seizures arising from theta states (associated with increased interneuron activity) were unaffected. This finding underlines the obligation to not assume that the same treatment will effectively stop all seizures arising from the same circuits.

As the authors of the present study acknowledge, these recordings were made in areas that were very likely not the sites of seizure onset, where more dramatic changes in cellular and circuit activity presumably could occur. Thus, it remains unclear whether the relationship of brain state and seizure activity observed here is true for specific seizure foci. In addition, these experiments focused on the dorsal hippocampus, but at least in the pilocarpine rat model of TLE, the ventral hippocampus appears to be activated earlier (6). Furthermore, the association of REM and increased seizure propensity might be somewhat particular to rats, as the seizures were more frequent during non-theta states in the mouse model. REM does not appear to be linked with increased temporal lobe seizure activity in humans (7), although the transition from sleep to wake may be a time of increased seizures in patients with idiopathic generalized epilepsies (8). The potential relationship of interictal discharges to seizures (9) in delta versus theta states also warrants further investigation. Nevertheless, this study emphasizes the importance of identifying specific individual neurons that may exhibit telltale signs of altered activity leading up to seizures, and thus could theoretically be targets for intervention, and of taking into consideration the behavioral states from which individual seizures arise when choosing potential therapies.


Editor's Note: Authors have a Conflict of Interest disclosure which is posted under the Supplemental Materials link.


1. Ramgopal S, Thome-Souza S, Jackson M, Kadish NE, Sanchez Fernandez I, Klehm J, Bosl W, Reinsberger C, Schachter S, Loddenkemper T. Seizure detection, seizure prediction, and closed-loop warning systems in epilepsy. Epilepsy Behav. 2014;37:291–307. [PubMed]
2. Fujita S, Toyoda I, Thamattoor AK, Buckmaster PS. Preictal activity of subicular, CA1, and dentate gyrus principal neurons in the dorsal hippocampus before spontaneous seizures in a rat model of temporal lobe epilepsy. J Neurosci. 2014;34:16671–16687. [PMC free article] [PubMed]
3. Toyoda I, Fujita S, Thamattoor AK, Buckmaster PS. Unit activity of hippocampal interneurons before spontaneous seizures in an animal model of temporal lobe epilepsy. J Neurosci. 2015;35:6600–6618. [PMC free article] [PubMed]
4. Truccolo W, Donoghue JA, Hochberg LR, Eskandar EN, Madsen JR, Anderson WS, Brown EN, Halgren E, Cash SS. Single-neuron dynamics in human focal epilepsy. Nat Neurosci. 2011;14:635–641. [PMC free article] [PubMed]
5. Krook-Magnuson E, Armstrong C, Oijala M, Soltesz I. On-demand optogenetic control of spontaneous seizures in temporal lobe epilepsy. Nat Commun. 2013;4:1376. [PMC free article] [PubMed]
6. Toyoda I, Bower MR, Leyva F, Buckmaster PS. Early activation of ventral hippocampus and subiculum during spontaneous seizures in a rat model of temporal lobe epilepsy. J Neurosci. 2013;33:11100–11115. [PMC free article] [PubMed]
7. Dinner DS. Effect of sleep on epilepsy. J Clin Neurophysiol. 2002;19:504–513. [PubMed]
8. Zambrelli E, Canevini MP. Pre- and post-dormitum epilepsies: idiopathic generalized epilepsies. Sleep Med. 2011;12(suppl 2):S17–S21. [PubMed]
9. Keller CJ, Truccolo W, Gale JT, Eskandar E, Thesen T, Carlson C, Devinsky O, Kuzniecky R, Doyle WK, Madsen JR, Schomer DL, Mehta AD, Brown EN, Hochberg LR, Ulbert I, Halgren E, Cash SS. Heterogeneous neuronal firing patterns during interictal epileptiform discharges in the human cortex. Brain. 2010;133:1668–1681. [PMC free article] [PubMed]

Articles from Epilepsy Currents are provided here courtesy of American Epilepsy Society