Afflicted by the Gods: The Shared History and Neurobiology of Psychosis and Epilepsy

Why is it that all those who have become eminent in philosophy or politics or poetry or the arts are clearly of an [ill-tempered] temperament, and some of them to such an extent as to be affected by diseases caused by black bile, as is said to have happened to [Hercules] among the heroes? For he appears to have been of this nature, and that is why epileptic afflictions were called by the ancients ‘the sacred disease…’

—The Complete Works of Aristotle (1984), Edited by Jonathan Barnes

Hercules, demi-god of Greek mythology, endured a difficult life. The product of an affair between Zeus and Alcmene, he was cursed by Hera (always the jealous wife) with frequent episodes of madness and seizures. His eyes would roll back, he would foam at the mouth, and he would lose consciousness. Shortly after, he would become “mad” with violent and irrational behavior. He agonized with these spells throughout his life but, from each one, he emerged triumphant, stronger, and ready to conquer the next adventure.

In this mix of mythology, adultery, and a Greek demi-god’s six-pack abs, there is an underlying question: is Hercules suffering from two conditions—epilepsy and psychosis—or from one? These phenomena have mystified and enchanted spiritual leaders, authors, philosophers, and physicians for millennia. Remarkably detailed accounts of postictal psychosis were described in a collection of Babylonian (2000 BC) patients who started to see “horrible, alarming, or immoral ‘visions’ and [were] in a constant state of fear” shortly after seizure-like episodes (1). The ancient Greeks considered the motoric convulsion and the psychotic symptoms to be part of one malady, epilepsy, whose causes included being cursed by deities such as Selene, goddess of the moon (hence the term “lunatics”). For centuries after, seizures and psychosis were attributed to demonic possession or personal moral failings. Seizures have also been aggrandized: as a connection to the spiritual world (e.g., in Hmong communities) or a contribution to artistic endeavors (e.g., van Gogh, Kierkegaard, and Dostoyevsky). Mostly, though, they were treated as “madness”— through the 1800s, patients with psychosis, epilepsy, or both were largely confined to asylums.

The pioneers of our field—Charcot, Jackson, Babinski, Kraepelin, Freud—wrestled intensely with issues of nosology—how do we make sense of this complex phenomenology (2)? Based on careful observations of individuals with epilepsy and psychosis, they described two distinct trajectories: some patients would return to a clear, nonpsychotic mental state between seizures while others remained psychotic indefinitely. They postulated these to be distinct diseases. However, lacking appropriate tools, they had no means to test the hypothesis.

It wasn’t until the 1920s that scientists developed an objective way to measure brain activity: the electroencephalogram (EEG). The EEG offered immediate insights into diagnosis: epilepsy could now be described and quantified based on its electrical signature—it was clearly a disease of the brain. Psychosis, on the other hand, showed no such signature. This innovation fueled an emerging split in medicine—mind from brain (2). Those conditions that could be localized within the brain became the purview of neurology; syndromes that did not have objective biomarkers became the purview of psychiatry. In other words, neurology had a biological basis and psychiatry did not (2,3). Over the ensuing decades, the same pattern continued: newly identified brain pathologies became the purview of neurology while psychiatry, abandoning a medical model, became fixated on psychoanalytic theory (2).

A century later, the idea of splitting brain from mind is clearly absurd. And yet, in medicine, we persist: in textbooks, in curricula, and in our clinics we continue to categorize illnesses as being of psychiatry or of neurology. In many ways, psychosis and epilepsy remain archetypes of each. Yet with a modern lens, a more elaborate picture is beginning to take shape.

First, at a basic, epidemiological level, the two conditions co-occur at rates far higher than could occur by chance: the incidence of psychosis in patients with epilepsy is 5 times higher than in the general population; the incidence of epilepsy is twice as high in patients with psychosis relative to normal control subjects (4,5). Given these data, it is clear that some connection exists.

Digging deeper, the next clue comes from location. It is now well established that individuals whose seizure focus is located in the temporal lobe are more likely to have psychosis of epilepsy (POE) (5). Moreover, the positive symptoms seen with POE—auditory hallucinations, paranoia, and grandiosity—are qualitatively similar to other presentations of psychosis. Is it possible that they share an underlying mechanism?

One of the most cutting-edge areas of research in psychosis is exploring how we decide what in the environment is important—a concept known as salience (6). The amygdala, deep within the temporal lobe, is thought to play a crucial role. Under ordinary circumstances, when someone encounters an unexpected threat or reward (e.g., a gorgon or a Golden Apple), the amygdala signals to the prefrontal cortex—pay attention! Essentially, the pathway serves as an environmental alert system.

Now, imagine that this mechanism becomes dysregulated—all of a sudden, routine things could be perceived as holding excessive significance. A song on the radio or a sideways glance from a neighbor become the nidus for paranoid ideations or other delusions. Considerable data have bolstered this hypothesis, including numerous functional imaging studies showing increased limbic activation in individuals with positive-symptom predominant schizophrenia (6). [Of note: there’s far more to the model than this. A key separate aspect is the role of the prefrontal cortex in regulating limbic signaling and contextualizing the flagged stimulus. Negative-symptom predominant schizophrenia is hypothesized to be related to decreased activity in this regulatory system (7).]

Around the same time the salience model of psychosis was emerging, Michael Trimble and a team of researchers were interested in exploring the connection between temporal lobe functioning and POE. In temporal lobe epilepsy (TLE), limbic structures are often reduced in size. This finding is sufficiently common that it is sometimes used as a diagnostic hallmark to localize the seizure focus—but it isn’t found in all TLE patients. Dr Trimble’s team hypothesized that the presence of psychosis may be related to this variability. So, they took two groups of patients with TLE, one with a history of psychosis and one without. Remarkably, they found the amygdalae were larger bilaterally in the POE group (8). Follow-up studies of medial temporal lobe function confirmed a relationship with psychosis in TLE and hint at a tantalizing possible connection: perhaps these individuals’ psychotic symptoms can be explained by overactivity in the salience system (9).

Now, hold on to your toga, because the next part of this story is wild. Long before current standards were established for research, a group of scientists was interested in exploring the relationship between epilepsy and psychosis. These scientists were aware that scalp EEGs tend to be normal in psychosis. But they were also aware of the profound limitations of the methodology—scalp EEGs fail to identify or localize signals from deep structures. So, they took a bold (if massively unethical) approach: they implanted intracranial electrodes throughout the brains of asylum patients, including groups with TLE with POE and individuals with schizophrenia who were experiencing active psychotic symptoms (10). The expectation, according to the long-established dogma of the field, was that only individuals with epilepsy—a disease of the brain—would show abnormalities. What they found was astonishing: both groups showed abnormal spike and wave activity in deep intracranial leads (9,10). Nosology be damned! Decades before the modern neuroscience revolution, they showed that schizophrenia is indeed brain based.

The story of Hercules is powerful in many ways, including how it reveals our historical struggles to integrate mind and brain, madness and motor symptoms. What was once a single entity was later divided in two. But, through the lens of modern neuroscience, we can start to reconcile the common biology of epilepsy, POE, and idiopathic psychosis. All three are brain-based syndromes that can be evaluated with a comprehensive neuropsychiatric approach. The EEG was the first useful biomarker for understanding a clinical cause of psychosis. Many more examples have followed: neurosyphilis, Lewy body dementia, the 22q11.2 deletion syndrome, and autoimmune encephalitis are all well-described diseases that are known to cause psychosis. Schizophrenia, by contrast, remains a syndrome: a clinical diagnosis of exclusion, to be used only when we are unable to identify the biological cause of the illness. This type of syndromic label is essential for selecting evidence-based treatments, but we should be careful to not infer any notion of a unifying pathophysiology.

In many ways, the work of our field has circled back to that of our pioneers: to continue the herculean task of parsing complex clinical presentations, such as epilepsy and psychosis, into discrete entities. The hope is that ongoing work can reduce the number of cases to which we apply the outdated label of “schizophrenia” and replace them instead with more precise diagnoses that, in turn, allow for disease-modifying treatments.