Brain stimulation is a promising method for treating

Brain stimulation is a promising method for treating epilepsy. However, its clinical translation has been impeded by the obscure nature of its underlying mechanisms and the scarce effective targets (Udupa and Chen, 2015). Our results collectively showed that entorhinal LFES have an antiepileptic effect via non-selectively activating entorhinal CaMKIIα-positive neurons or its projection fibers and consequently inducing a global GABAergic inhibition in the hippocampus. This mechanism can partly explain that why delayed LFES has no antiepileptic effect and why the strength of antiepileptic effect of LFES with 2min duration were similar to that with 15min duration (Fig. 1b). Supportively, entorhinal LFES had a stronger antiepileptic effect in another pilocarpine induce spontaneous chronic epileptic mouse model in the LFES period than that in the LFES withheld period (Fig. 1i); and we did not find any long-term antiepileptic effect 1h after 3 times of 2-min LFES treatments, when LFES or optical stimulation was withheld. Consistent with recent studies (Suthana et al., 2012), but differing from them in terms of stimulation frequency, repeated entorhinal LFES and low-frequency photo-activation of entorhinal PNs relieved the memory impairments in several behavioral tests in kindled mice (Fig. S3). Thus, the underlying entorhinal-hippocampal “glutamatergic-GABAergic” circuit may also be relevant for understanding the functions of the entorhinal cortex and the hippocampus in situations other than TLE. For example, lesioning the neurons in entorhinal layer III may results in hippocampal hyperactivity and impairs memory (Brun et al., 2008; Schlesiger et al., 2015). In addition, entorhinal LFES seemed to produce a stronger anti-epileptic effect than that of photo-stimulation of focal hippocampal GABAergic neurons, which may be due to the spatial limitations of focal photo-stimulation. Taken together, our results also confirmed that entorhinal LFES may be an efficacious option for galanin stimulation treatment of temporal lobe epilepsy. In addition, the traditional view of epileptic network activity mainly attributes it to excessive excitation and impaired inhibition in the brain, such as the loss of INs or the hyper-activation of surviving PNs. However, our results, as discussed above, present a different and more complex view of the epileptic network wherein excessive inhibition or impaired excitation of PNs in specific brain areas (such as in the EC) may also contribute to epileptic networks. Our results suggest a novel and important idea that excitation in specific brain areas can be inhibiting in another part of the circuit, and vice versa. This idea is different from previous theories that propose a phenomenon of “rebound excitation” to explain how “excessive inhibition” contributes to epilepsy in difference disease models, such as febrile status epilepticus (Chen et al., 2001) and other epilepsy models (Klaassen et al., 2006). In addition, our view of epileptic networks has several implications. First, it supports the view that the preferential loss of layer III PNs (Bartolomei et al., 2005; Du et al., 1995) and hypometabolism in the EC (Goffin et al., 2009; Guo et al., 2009; Wang et al., 2014) promote hippocampal epileptogenesis and establish seizures in vivo. Second, our proposal regarding epileptic networks may provide some clues for exploring why some excitatory drugs have antiepileptic effects in TLE models (Armstrong et al., 2009) while some inhibitory drugs induce rather than suppress seizures (Perucca et al., 1998; Schuele et al., 2005). Overall, our results show that entorhinal CaMKIIα-positive neurons and the hippocampal INs form a “glutamatergic-GABAergic” antiepileptic circuit, which is involved in brain stimulation control of hippocampal seizures. Our work may facilitate the clinical translation of brain stimulation treatments for patients with epilepsy, and it may also add a new component to the current epileptic network theory.