Temporal lobe epilepsy with amygdala enlargement: a subtype of temporal lobe epilepsy

The amygdala are located in the medial temporal lobe of human brain. They receive extensive nerve input from many different brain areas, such as many sensory areas, hippocampus, hypothalamus, thalamus, frontal lobe and so on [1]. Therefore, they can influence neuroendocrine, emotional, and cognitive aspects of biologic information processing. The amygdala play a crucial role in mediating affective behavior in humans and primates [2]. Some functional imaging studies showed that emotional information processing, for instance angry or happy facial expressions, activated the amygdaloid complex [3],[4]. In patients with major depression, there is increased metabolism in limbic areas, indicating an overactivation of the amygdala might be a trait marker of depression [5]. These findings demonstrate that the amygdala are critical brain structures for the emotional evaluation of specific sensory input on the background of individual experience.

Emotional disturbances and psychiatric problems are frequently encountered in patients with temporal lobe epilepsy (TLE) [6]. Previous studies demonstrated significant amygdala enlargement (AE) was associated with dysphoric disorder and psychosis in TLE patients without hippocampal sclerosis (HS) [7],[8]. Furthermore, one study proposed that psychosis in patients may develop on the background of the dysphoric disorder of TLE [8]. A neurophysiological study using intracranial recordings showed that 5% of mesial TLE (MTLE) patients had seizure onset in the amygdala [9]. Previous imaging studies reported that isolated unilateral amygdala damage was observed in about 8% of MTLE patients [10]. Therefore, the amygdala may not only be involved and affected in MTLE but may also be regarded as an epileptogenic focus of MTLE.

Recently, some studies reported AE was observed in "magnetic resonance imaging (MRI) negative" TLE [7],[8],[11],[12]. "MRI negative" TLE was defined by the absence of a neocortical lesion, normal hippocampal volumetry, and absence of any increased signal in the mesial temporal lobes on routine visual assessment, despite we cannot preclude completely that "MRI negative" TLE may have HS identified after epilepsy surgery [13]. However, some earlier studies found inconsistent results. For instance, the study performed by Kalviainen R et al. showed that the amygdala were decreased in volume by at least 20% in 19% of TLE patients [14]. Moreover, isolated amygdala sclerosis has been described in some more earlier studies [15],[16]. Up to date, a role for the amygdala as a focus of epilepsy has not been explicitly established, unlike TLE with HS. Considering that the above findings are inconsistent and the prior studies are confined to small number of TLE patients, this study aims to investigate the clinical significance of AE in "MRI negative" TLE patients detected by MRI and compare the amygdala volume of the same patient before and after treatment by a larger sample size.

Up to date, there have been few amygdala volumetric studies, partly because of the amygdala's poorly demarcated anatomical boundaries [17],[18]. However, the method of amygdala volume measurement used in this study is now well established [7],[8],[11],[12]. The amygdala volumetry was determined in both patients and controls by the same epilepsy specialist (R-J Lv), therefore there was no inter-rater difference or bias. This study is a larger sample size report of TLE patients with increased amygdala volume. We obtained a similar finding in TLE patients with later onset age to the study performed by Mitsueda-Ono T et al. [12]. Our normal ranges and amygdala volume ratios are consistent with recently reported data [12].

In the present study, we found all TLE patients with enlarged amygdala ipsilateral to the IEDs except two patients with bilateral AE showed bilateral IEDs. Although only two patients had ictal EEG, most patients in our study presented unilateral IEDs ipsilateral to AE. Some studies indicated that IEDs in TLE can correlate very well with ictal onset zone, and about ninety percent focal spikes were ipsilateral to epilpeptogenic zone [19],[20]. Therefore, we have reason to speculate that most patients of our study presented localized anterior-inferior IEDs always consistent with the side of AE is not accidental phenomena. It may have potential hint that the seizures of 31 patients in our study were lateralized to the hemisphere containing the enlarged amygdala. There was some evidence demonstrating the effect of seizures to MR images which resolved and turned to normal after some time [21],[22]. Thus, the decreased volume of the enlarged amygdale at follow-up visit may suggest that AE may be epilpeptogenic zone conversely. The results of this study, combined with previous electrophysiological and radiological studies [11],[12],[23],[24], indicate that AE may function as an epileptogenic focus in this subgroup of TLE patients. However, we did not get direct recording from the amygdala by means of depth electrodes and also we did not have ictal single proton emission computed tomography (SPECT), thus we could not definitively demonstrate that the enlarged amygdala was epilpeptogenic zone. This needs further study according to more direct testimony such as invasive intracranial electrodes recording.

The mechanisms leading to amygdala enlargement in TLE are a matter of debate. A recent study reported the chronic and long lasting inflammatory processes with or without a self-limited course occurred in these TLE patients could lead to amygdala enlargement possibly [12]. Neuroinflammatory processes have been reported to be one of the causes of adult-onset TLE [25]. Autoimmune-mediated encephalitis has been characterized in middle-aged patients with TLE, and some patients demonstrated a self-limited course [25]-[27]. In our study, one patient had autoimmune limbic involvement and exhibited great improvements in seizure frequency and memory after immunotherapy, which suggested neuroinflammatory processes participated in amygdala enlargement possibly. Whereas, there were not paraneoplastic or autoimmune limbic involvement found in most patients, perhaps because the tests used for paraneoplastic or autoimmune were not comprehensive enough or owing to ethnic difference. That was the cause that we did not give immunotherapy to most patients. The emotional problems appeared before or simultaneously with seizure in this group of patients, which was similar to the report of limbic encephalitis [28], supporting that the etiology was autoimmune inflammation in some patients possibly. In addition, most patients (22 of 33 patients) showed significantly decreased volume of the enlarged amygdala without immune modulating treatment, which was similar to Mitsueda-Ono's study [12]. Some studies reported that seizure can induce brain lesions [29],[30]. Therefore, MRI scan were performed seven days after seizure at least to preclude seizure induced brain lesions as far as possible. Although we cannot preclude seizure related functional change completely because of the short follow-up time, the decreased volume of the enlarged amygdale at follow-up visit may suggest a self-limited course of neuroinflammation. In this study, the seizure onset was late, and FDG-PET scanning was performed within a few years after onset. This may imply that the epileptogenicity in the amygdala was mild and developed slowly. This assumption was consistent with the fact that seizures have been well controlled with a low dose of AEDs instead of immune modulating treatment in these TLE patients with AE. The assumption of long-lasting mild epileptic activity in the amygdale may be a clue to speculate that the possible etiology of AE was the chronic and long lasting inflammatory processes. However, the therapeutic effect declined gradually with follow-up time prolonging in some patients (11 patients), whose amygdala volume did not change after treatment. Therefore, the AE patients in our study could be heterogenous in etiology. Focal cortical dysplasia (FCD) cannot be precluded in some patients completely, especially those patients showing a decline in therapeutic efficacy. Similarly, a recent study found that FCD and low-grade tumors were major causes of late-onset TLE with AE [31]. The speculated reasons are very similar to the the most recent study [32]. However, it must be emphasized that these interpretations are very speculative and until now, there are no generally accepted models that can explain the mechanism leading to the phenomena of enlarged amygdala. This still need be studied further.

TLE with AE may define a subgroup of "imaging-negative" TLE patients that are clinically different to those with HS. Our patients were all late onset and exhibited fewer secondary generalized seizures compared to those patients with HS, consistent with Bruton's series [33]. With regard to ictal semiology, similar to Mitsueda-Ono's study [12], complex partial seizures were the most common sizure type. This group of patients had obvious affective disturbances from the onset of the diseases, mainly manifesting as dysthymia, in accordance with the findings of Tebartz van Elst L [7]. The affective disturbance was different from the general meaning of psychosis because they can be well controlled by antidepressant combined with AEDs. After AEDs treatment, most patients became seizure free or showed a dramatic improvement in seizure occurrence. Seizures are currently well controlled on anticonvulsant medication, a dissimilar situation to the usual natural history of TLE with HS, that was the reason we can not get further pathological results to confirm the mechanism leading to AE. In this group of patients, secondarily generalized seizures often occurred at night. To our knowledge, there are no similar findings reported in the literature and we currently have no plausible explanation for this finding. Further research is needed to replicate this finding.

These findings do have very important implications for the way we think about the diagnosis of "imaging-negative" TLE. TLE with AE probably represents a distinct nosological and probably less homogeneous syndrome which is most likely a subtype of TLE without ipsilateral HS. If there is no apparent HS in TLE patients, AE should be considered first.

Study concept and design (LRJ); acquisition of data (Drs LRJ, SXQ, CT, SZR, GHZ, RHT); analysis and interpretation of data (LRJ and SXQ); drafting of the manuscript (LRJ); Critical revision of the manuscript for important intellectual content (LRJ and SXQ); statistical expertise (LRJ); administrative, technical, and material support (LRJ, RHT and SXQ); study supervision (LRJ and SXQ). All authors read and approved the final manuscript.