In some patients, the seizure prodrome (the minutes, hours, or even days leading up to a seizure) consists of predictable changes in psychological functioning and behavior, including the development of irritability, anxiety, and aggression (34,35).

Directed, purposeful, aggressive, or violent behavior during seizures is generally considered rare. Papers in the literature typically include either cases in which seizures were inferred from the history, or others in which seizures were directly observed by trained professionals or recorded with sophisticated monitoring techniques.

Hindler reported the case of a 19-year-old nanny with a history of febrile convulsions up to the age of 7 and a positive family history for focal epilepsy in a younger sister (36). The patient had a more recent history of 5- to 10-minute episodes of strange sensations resulting in a feeling of emptiness and a sensation of the mind becoming separated from the body, followed by intense feelings of hate and aggression directed at the baby in her care. The episodes were triggered by the baby laughing and were only experienced while the nanny and the baby were alone. During one such episode, she swung the baby by the legs and the baby’s head hit a cabinet and perhaps the floor, leading to a fatal brain injury.

After she was incarcerated, her doctor observed her to have a “generalized epileptiform convulsion.” An electroencephalogram showed epileptiform abnormalities and she was treated with anticonvulsant therapy with no further abnormal experiences. Hindler concluded that the patient committed murder during an epileptic attack and that she had an unusual form of reflex epilepsy.

One hundred and thirty-two prisoners with epilepsy (including 11 who were convicted of murder, attempted murder, rape, attempted rape, and armed violence) were asked if they had ever committed an offense in association with a seizure (37). None of the prisoners admitted to a history of ictal violence, although three reported being arrested for disorderly conduct immediately after a seizure and one had a seizure while a robbery was already in progress.

Directed, purposeful ictal aggression during observed seizures has only been rarely reported (1,38,39,40). At the Montreal Neurological Institute, Gloor witnessed few, if any, cases of spontaneous or electrically induced ictal aggression (41). Similarly, Rodin witnessed several hundred spontaneous psychomotor seizures; only one seizure was accompanied by directed aggression, which Rodin concluded was goal directed and an expression of vengeance, rather than related directly to the seizure (40).

Rodin further reported 57 patients whose bemegride-induced “psychomotor automatisms” were photographed while they were seated in a chair undergoing electroencephalographic (EEG) testing (40). None of the patients became aggressive during seizures, although two patients were observed to suddenly lunge forward and “show a rather ferocious facial expression.” In one case, an unsuccessful attempt by the attending physician to keep the patient seated made the patient clench his fist, adopt a boxer-type stance, and appear imminently violent. The threatening behavior immediately ended once the patient was physically released, suggesting to Rodin “a defensive reaction due to being restrained rather than active trouble-seeking.” He postulated that
“the patient, if he were to be restrained during his confusional state, might react in a defensive manner that could be misinterpreted as a goal-directed assault.” Because the EEG results were not provided, it is possible that this may have been a case of postictal resistive violence (see subsequent text).

Ramani and Gumnit recorded focal seizures in 15 patients with epilepsy and significant histories of episodic aggressive behaviors (39). No patients showed ictal violent or aggressive behavior.

Delgado-Escueta et al. reviewed the cases of 5,400 patients with epilepsy who underwent inpatient evaluation (1). They further studied 19 patients who were believed to have ictal aggression, including 3 previously reported patients (31,42). On the basis of a review of 33 seizures, the authors concluded that only one patient committed an aggressive act during a seizure that could have harmed another person (the patient attempted to scratch her psychologist’s face). None of the other observed aggressive behaviors threatened the well-being of others, although some behaviors appeared hostile, such as assuming a karate posture and making striking motions, shouting, or spitting, or destroying inanimate objects such as curtains, chairs, and beds.

Other anecdotal reports of aggressive ictal automatisms have documented behaviors such as flailing or chopping movements of the arms, spitting, biting, or physical assaults on nearby objects (32,42,43), which were generally stereotyped and nondirected.

Nonconvulsive status epilepticus may present with psychiatric symptomatology, including psychotic features (44). A report nearly 40 years ago suggested that absence status epilepticus may be associated with hostility, aggression, agitation, and impulsive behavior (45).

Kutlu et al. (46) analyzed ictal behaviors during 282 recorded seizures in 48 patients with temporal lobe epilepsy (TLE) who subsequently underwent anterior temporal lobectomy and became seizure-free for at least 2 years. Ictal aggression occurred nearly twice as often in the 13 patients whose resected tissue showed pathological findings other than hippocampal sclerosis, such as tumors, cavernoma, and hamartoma, as in the 35 patients with hippocampal sclerosis. However, this difference was not significant when a patient with neocortical gliosis was excluded from the analysis.

Several studies utilizing video EEG monitoring showed that violent behavior may result from attempts at physical restraint during the postictal period, the so-called “resistive violence” (1,43,47). For example, one patient “exhibited angry facial expressions during ictal episodes and confused, nondirected, aggressive behavior postictally while attempting to free himself of the soft restraints” (39). Another patient became greatly agitated during the postictal period, with “undirected aggression, loud screaming, kicking and fighting against the restraints” (47). Tassinari et al. (48) noted that late ictal and postictal biting behavior was usually precipitated by the intrusion of the examiner’s hand near the patient’s face.

Violent behavior may occur during a postictal psychosis, which typically begins within several days of a cluster of complex partial or secondary generalized seizures and includes hallucinations, paranoia, and delusions (44,49,50,51,52). The fear that accompanies the delusions or paranoia may lead to violent behavior, which may be self-directed (53,54).

Kanemoto et al. evaluated aggression in 30 patients with postictal psychosis and 30 patients matched for gender, age, and age of seizure onset with postictal confusion, but no postictal psychosis (54). All seizures had partial onset, and were directly observed and documented by trained professionals. A postictal psychosis was diagnosed on the basis of hallucinations, delusions, thought disorder, and affective changes that were clinically demarcated from seizures and the immediate postictal state. Violence was defined as well-directed, well-organized attacks against human beings because of irrational behavior that resulted in a severe injury or a life-threatening situation,
such as strangulation. Resistive violence as described in the preceding text was analyzed separately.

In their study, violent attacks occurred in association with 13 out of 57 (23%) episodes of postictal psychosis, compared with 1 out of 134 episodes of postictal confusion (p <0.001). In eight postictal psychotic episodes, violent behavior occurred within 24 hours of the onset of postictal psychosis. Another four patients in the postictal group attempted suicide during their postictal psychosis, compared with none of the controls (p <0.005). On the other hand, four control patients displayed resistive violence during the postictal state compared with none of the patients with postictal psychosis.

Blumer has described an interictal dysphoric disorder, characterized by marked irritability, with troublesome loss of control and often associated with intermittent depressive or euphoric moods, anergia, insomnia, anxiety, and fears (55,56). Interictal aggression has generally been studied in select groups of patients and prisoners with epilepsy. The results in these skewed populations suggest that impulsive, assaultive interictal behavior is seen in up to 56% of patients with TLE (57,58,59,60,61,62,63).

Of 100 children with TLE followed for a mean of 10 years, 36 had outbursts of unspecified “catastrophic rage” (64). Rage outbursts were associated with a history of head injury or cerebral infection, seizure onset under the age of 1 year, and a “disordered” home (defined as gross poverty, deceased mother before the proband was 10 years old, aggressive father, or psychosis or chronic neurosis in one or both parents). There was no relationship between seizure frequency and rage outbursts.

Not all studies find a high incidence of interictal aggression. Rodin (40) reviewed the charts of 700 patients with epilepsy seen at a tertiary referral center over a 5-year period and found that 34 (4.8%) had exhibited “destructive-assaultive” behavior. These aggression-prone patients tended to be young men of lower-than-average intelligence, with a history of behavioral difficulties dating back to school age, who had unspecified signs of diffuse organic disease. There was no association between seizure type or the location of onset and the likelihood of aggressive behavior. Hermann et al. later confirmed these findings (65).

Ramani and Gumnit (39) studied 19 patients with epilepsy and significant histories of episodic aggressive behaviors for an average of 6.3 weeks as inpatients with video telemetry. Two patients exhibited repeated aggressive behavior. There was no apparent temporal relationship between the aggression and ictal EEG patterns in one patient, who ripped loose a kitchen sink. Upon admission, 13 patients were irritable and socialized poorly, but this behavior was said to improve in the “supportive ward milieu.”

Mendez et al. (66) reviewed the records of 44 patients with epilepsy attending a university-affiliated neurology clinic who were referred for psychiatric evaluation because of violent interictal behavior (defined as the forceful infliction of injury or harm on persons or property). Most of the violent acts consisted of verbal or minor physical aggression. Compared with nonviolent age- and sex-matched patients with epilepsy, the violent patients did not differ on seizure variables such as type and frequency of seizures, auras, EEG findings, age of seizure onset, or antiepileptic drug (AED) treatment. The authors concluded that interictal violence in their sample was not associated with epileptiform activity or other seizure variables.

Herzberg and Fenwick (33) retrospectively reviewed the charts of patients admitted to the epilepsy unit of the Maudsley Hospital over a 10-year period with a diagnosis of TLE on discharge. Aggressive behavior was associated with male sex, early age of seizure onset, and a history of chronic behavioral difficulties. No relationship was found with specific EEG or computed tomography scan findings or a history of psychosis.

Kanemoto et al. (54) studied the incidence of interictal violence in 33 patients with localization-related epilepsy and acute interictal psychosis. Violent attacks occurred
in association with 3 of 62 interictal psychotic episodes, which was significantly less than the incidence of violent attacks that occurred in other patients during postictal psychoses (p <0.01).

Among adolescent and adult prisoners, the prevalence of EEG abnormalities and epilepsy has been reported to be as high as 65% and 18%, respectively (16,37,67,68,69,70). Because the prevalence of epilepsy in prisoners may be similar to the prevalence of epilepsy in the underprivileged, lower socioeconomic communities from which prisoners typically originate (71), these studies do not necessarily prove that epilepsy is more frequent among prisoners.

Several studies on prisoners have attempted to correlate epilepsy with violence. Lewis et al. (67) retrospectively reviewed records of incarcerated adolescent boys and found that the number of “psychomotor symptoms” correlated with degree of violence (as measured by a scale derived by the authors). By contrast, two other studies found that prisoners with epilepsy were not convicted of more serious or more violent crimes than age- and race-matched nonepileptic prisoners (37,68,70).