Genital automatisms in complex partial seizures

Methods.

We retrospectively studied 90 consecutive patients with focal epilepsy who were referred to our epilepsy monitoring unit because of medically refractory seizures. All patients underwent an extensive presurgical evaluation that included the following diagnostic procedures:

• 1. Prolonged video-EEG monitoring with scalp and sphenoidal electrodes was performed with a commercially available monitoring system (Pegasus Monitoring System, EMS Company, Korneuburg, Austria) for an average of 5 days. Focal slowing as well as the frequency and location of interictal spikes was assessed by visual analysis over the entire recording time. Ictal EEG changes were classified by location, morphology, and time evolution according to criteria proposed in previous reports.9,10 One author (F.L.), who was blinded for the EEG and imaging data of the patients, retrospectively studied all videotapes for the appearance of 1) genital automatisms, defined as fondling, scratching, or grabbing the genitals in a stereotyped and repeated manner during the seizure or immediately after EEG seizure offset; and 2) other apparently sexual automatisms, especially pelvic or truncal movements.

• 2. MRI was done on a 1.5-Tesla machine (Philips Gyroscan ACS-NT; Best, the Netherlands) using a circularly polarized head coil.

• 3. All patients underwent interictal and if possible ictal SPECT studies, formal neuropsychological testing, and a Wada test in order to lateralize language and memory dominance.

Finally, the correct localization of the epileptogenic zone was confirmed by seizure-free outcome after selective amygdalahippocampectomy in four patients and lesionectomy in the hippocampus in one patient.

Patients and results.

Seventy-four patients had temporal lobe epilepsy (TLE) and 16 patients had frontal lobe epilepsy. Genital automatisms as described above occurred in five patients with TLE but in none of the patients with frontal lobe epilepsy. Other sexual automatisms were not observed in either group.

Discussion.

We report five patients with genital automatism, defined as repeated fondling or grabbing of the genitals, during or immediately after complex partial seizures. The results of clinical seizure semiology, interictal and ictal EEG, and MRI showed the seizure onset zone to be located in the mesial part of the temporal lobe. Correct localization of the seizure onset zone was confirmed by good to excellent postoperative seizure control (class I or II according to the classification of Engel et al.11) after selective amygdalahippocampectomy in four patients and after removal of a ganglioglioma involving the hippocampus in one patient.

The localizing significance of sexual automatisms is controversial. In particular, it is unclear whether temporal or frontal brain structures are involved in their generation. Currier et al.6 reported two patients with sexual automatisms and proposed a temporal lobe origin of their seizures. The first patient occasionally lifted her shirt, spread her knees, and elevated her pelvis rhythmically during her seizures. The EEG showed left anterior temporal focal spiking and slow wave activity. A definite localization of the seizure onset zone was not examined further because the patient remained almost seizure free after regular use of antiepileptic medication. The second patient manipulated her genitals rhythmically during a seizure. EEG within a few minutes after this seizure showed definite left temporal spikes. Other authors reported seizures in a woman with somatosensory sensations in the genitals and automatisms characterized by the patient beating her chest and manipulating the perineum.5 EEG revealed right temporoparietal slow waves.

Other authors12 reported 12 patients with sexual ictal manifestations, all with TLE, and mentioned that these phenomena predominate in women. We could not confirm these latter findings because three of our five patients were men, although it should be mentioned that because of the small number of patients, no definite conclusion concerning a gender predominance can be drawn.

In contrast to the above-mentioned studies, in which a temporal lobe seizure onset was proposed as the origin of sexual automatisms on the basis of surface EEG findings without definite confirmation concerning the seizure onset zone, Spencer et al.7 reported four patients with sexual automatisms who were studied with depth electrodes. The automatisms consisted of rhythmic pelvic and truncal movements in three patients, with additional grabbing and manipulation of the genitals in one. After frontal lobe resection, two patients remained seizure free, and one patient again began to have seizures with temporal lobe onset 1 year after surgery. The fourth patient, by contrast, exclusively showed genital automatisms similar to those of our patients, which suggest a temporal lobe seizure onset. Indeed, depth electrode recordings performed in this particular patient showed independent interictal epileptiform activity in the left frontal and temporal lobes. Thus, one could speculate that these genital automatisms could be due either to seizure spread from the frontal to the temporal lobe or to seizure onset in the temporal lobe. However, no surgery was performed, and therefore the definite localization of the seizure onset zone could not be determined in this patient. The authors concluded that sexual automatisms during complex partial seizures were related to a frontal lobe seizure origin. Furthermore, Spencer et al.7 suggested that the findings of temporal lobe seizures in the above-mentioned studies were possibly incorrect because of erroneous localization of the seizure onset zone by scalp EEG and because of the lack of confirmation by pathologic examination and seizure-free outcome after surgery.

Williamson et al.8 studied 10 patients with frontal lobe epilepsy documented by depth electrode recordings, and observed hypermotoric activities such as pelvic thrusting, kicking, and rocking, sometimes combined with manipulation of the genitals. In agreement with Spencer et al.,7 these authors concluded that sexual automatisms are common in seizures of frontal lobe origin and may be helpful to distinguish temporal from frontal lobe epilepsy.

In contrast to these findings, our results indicate that sexual automatisms during complex partial seizures cannot be attributed exclusively to frontal lobe seizures. We established correct localization of the seizure onset zone by clinical seizure semiology, interictal and ictal EEG recordings, and high-resolution MRI scans as well as good to excellent outcome after surgical resection of mesial temporal structures. In our group of 16 frontal lobe epilepsy patients, sexual hypermotoric automatisms never occurred, which documents their low incidence.

To exclude the possibility that genital automatisms are a purposeful reaction to an aura, consisting of sexual feelings or excitement, all patients were given an aura questionnaire. Epigastric auras were most common, followed by cephalic aura, déjà vu, and testicular pain in one patient, but a sexual aura never was reported.

Concerning the two patients with genital automatisms after EEG seizure offset, we wondered whether they could eventually be due to spread of the epileptic activity into frontal lobe structures of the brain. However, our patients’ seizures did not show any clinical signs of frontal lobe involvement such as hypermotoric truncal or pelvic movements. Also, the EEG seizure pattern was confined to the temporal lobe, as evidenced by an amplitude ratio >2:1 between the temporal and the parasagittal chain in a bipolar longitudinal montage.9 However, because we did not perform invasive EEG, we cannot exclude the possibility of seizure spread to the frontal lobe in these two patients.

We therefore believe that sexual automatisms should be subdivided into two distinct subgroups. The first consists of discrete genital automatisms, like grabbing or fondling the genitals, which are related to a temporal lobe seizure onset. A possible role of the temporal lobe in sexual behavior has been reported by Klüver and Bucy,13 who demonstrated that bilateral temporal lobectomy in monkeys produced hypersexuality, including exhibitionistic masturbatory activity. Similar findings were later obtained in humans.14,15 The second subgroup, by contrast, consists of hypermotoric sexual automatisms like pelvic or truncal thrusting, eventually combined with manipulations of the genitals; these automatisms are common in frontal lobe seizures.

We conclude that sexual manifestations represent rare phenomena during complex partial seizures and do not appear exclusively in frontal lobe seizures. Specifically, discrete genital automatisms exclusively occurred in temporal lobe seizures in our patients. Because all kinds of sexual automatisms have been related to frontal lobe seizures so far, a distinction between genital and hypermotoric sexual automatisms may be useful in clinically defining the seizure onset zone correctly. In our patient group, genital automatisms ocurred only in temporal lobe seizures and were absent in seizures evolving from the frontal lobe.