During seizures, the patient consistently spoke Urdu/English to his mother but English only to staff, suggesting preserved pragmatic social awareness despite ongoing ictal activity.

Seizure semiology reflects dynamic interactions between the ictal discharge, the physical and mental state of the individual, and the surrounding environmental context. These interactions, conceptualized within cognitive neuroscience through the framework of embodiment, introduce important methodological challenges for the analysis and classification of seizure phenomena (1).

Although focal seizures in each patient are often stereotyped, seizure expression may vary substantially depending on internal state and external circumstances, even when certain ictal signs remain consistent, such as automatisms.

Importantly, many seizure manifestations result from the engagement of widely distributed neural networks, despite originating from a spatially limited epileptogenic focus. Complex seizure semiologies incorporating emotional behavior, social interaction, and language production cannot be adequately explained by activation of a restricted cortical region alone; instead, a network-level perspective is essential to account for their phenomenology.

We report the case of a 27-year-old right-handed man with focal epilepsy whose seizures are characterized by loud, fearful vocalizations containing linguistically appropriate and biographically accurate speech. During seizures, he addresses his bilingual mother in either English or Urdu, while speaking exclusively in English to staff. Responsiveness is preserved during events, followed by complete postictal amnesia. This case illustrates the context-dependent engagement of language and emotional networks during focal seizures and highlights the role of embodied and socially embedded processes in seizure semiology.

The patient is a 27-year-old right-handed bilingual man with equal proficiency in Urdu and English. He has a history of congenital cataracts resulting in blindness, who developed epilepsy at 16 years of age. There was no family history of epilepsy. His sister had congenital blindness and intellectual disability and died from an infection at age 20; her death was reported by the family to have been a major psychological stressor for the patient. He did not receive antiseizure medications until after his second focal-to-bilateral tonic–clonic seizure at age 21, at which time levetiracetam was initiated but discontinued due to behavioral adverse effects. He was subsequently started on oxcarbazepine 600 mg twice daily, which effectively controlled the convulsive seizures, while his focal impaired-consciousness seizures persisted. At age 23, he had a routine EEG done which was reportedly within normal limits. The patient reported good mood and healthy sleep. He denied panic attacks, functional seizures, and any other history of psychological disorders.

According to his mother, the patient experienced two seizure types. One consisted of generalized convulsive events described as “full-body shaking,” consistent with focal to bilateral tonic–clonic seizures. The other consisted of a subjective sensation of falling accompanied by intense fear, loud vocalizations, and urgent pleas to be held tightly to prevent perceived imminent death. These seizures typically lasted approximately 2 min and were preceded by an aura of intense fear, lightheadedness, and palpitations. Postictal symptoms included coughing, confusion, and complete amnesia for the events.

Brain MRI demonstrated an arachnoid cyst in the left middle cranial fossa abutting the medial temporal lobe (Figure 1). Routine scalp EEG performed while the patient was taking oxcarbazepine was unremarkable. He was therefore admitted to the epilepsy monitoring unit for seizure characterization and optimization of medical management.

He was admitted to the Epilepsy Monitoring Unit (EMU) for a total of 5 days, and the detailed medication regimen is outlined in Table 1. At admission his oxcarbazepine level was 28 ug/mL. At discharge, he was started on zonisamide 150 mg twice a day and his oxcarbazepine dosage was increased to 900 mg twice a day.

During admission, four stereotypical seizures were recorded, two of which occurred during sleep. All seizures began with panicked, loud, and comprehensible verbalizations accompanied by intermittent lip smacking. The content of the ictal speech was consistent across events, with language selection varying according to the social context. When his mother was present, he addressed her in either Urdu or English; when speaking to medical staff, he used English exclusively. During each seizure, he called for help and repeatedly expressed fear that he was in danger and about to die, pleading to be held or hugged to prevent his death. He referenced his deceased sister and warned his mother that she would find his “corpse” if she did not hold him.

In two seizures, the patient described a subjective sensation of diffuse bodily shaking and possible redness. He asked his mother and staff to confirm whether visible shaking or redness was present, although neither was observed by witnesses. In three seizures, the onset of verbalization preceded the electrographic onset by approximately 5–10 s.

All seizures demonstrated a consistent electrographic pattern. Seizure onset was characterized by arciform 7-Hz theta activity localized to the right frontotemporal region, with maximal involvement at F8/T2 > T4 > T6. Activity subsequently spread to F10 and T10 and evolved into high-amplitude rhythmic 4-Hz notched delta activity with superimposed fast frequencies. The ictal rhythm then propagated to the right parasagittal region before transitioning to high-amplitude 3-Hz delta activity. Seizures terminated with generalized 2-Hz delta activity of increasing amplitude. Figure 2 shows representative samples from the video-EEG recording; detailed descriptions of the recorded seizures are provided below.

This case illustrates an unusual seizure semiology characterized by intense fear of death, loud and coherent vocalizations, contextually appropriate bilingual language use, accurate autobiographical references, and complete postictal amnesia, features that collectively imply engagement of a broad, integrative network. The patient consistently addressed his mother in Urdu or English while speaking exclusively in English to staff, demonstrating preserved pragmatic awareness of his social environment. Notably, he referenced his deceased sister during seizures despite complete amnesia afterward, highlighting a dissociation between access to autobiographical information during the seizure and subsequent consolidation failure.

Ictal fear is a recognized but relatively uncommon manifestation, most often linked to the amygdala and mesial temporal networks (2–5). In typical cases, fear is brief and nonverbal (6); extended, coherent speech with preserved syntax, pragmatic appropriateness, and autobiographical specificity is rare. The sustained verbalization of impending death in this patient resembles the case reported by Biraben et al., but differs in its linguistic richness and social contingency, suggesting concurrent engagement of limbic and higher-order associative networks (7).

Complete postictal amnesia with preserved behavioral responsiveness and language output aligns with “pure amnestic seizures,” thought to reflect transient disruption of mesial temporal memory encoding—particularly bilateral hippocampal structures—while sparing online retrieval and expression (8–10). This dissociation underscores the separation between real-time cognitive access and durable consolidation.

The phenomenology of this patient’s seizures shares partial overlap with neural systems engaged by mortality-related cognition (existential threat). Experimental work implicates a distributed network including the amygdala, anterior cingulate cortex (ACC), posterior cingulate cortex, caudate, and parietal association regions in self-referential and regulatory processing (1, 11). The right amygdala, in particular, appears to show heightened sensitivity to threat detection and emotional arousal during mortality-related cognition, even in the absence of conscious fear (1). The rostral ACC is similarly implicated—consistent with anticipatory anxiety, conflict monitoring, and processing of socially relevant threat—with several studies demonstrating greater activation to death-related cues than to other aversive stimuli (1). Of note, the ictal discharge was observed on the right in our subject, but a left hemispheric onset cannot be ruled out since he has a left temporal arachnoid cyst and exhibited a delay between clinical onset and ictal EEG changes. Scalp EEG has known limitations, particularly for deep or mesial temporal generators, and may preferentially capture propagated activity rather than the seizure onset. Early low-amplitude ictal discharges may remain below the detection threshold. Accordingly, we cannot confirm whether the arachnoid cyst is a mere incidental finding.

Epilepsy studies support overlapping circuitry for ictal fear and existential threat: amygdala activation can induce profound dread and autonomic changes; the insula contributes to interoception and subjective threat; ACC–amygdala interactions modulate fear regulation; and mesial temporal structures (including the hippocampus) can allow autobiographical content to surface during seizures. Subcortical modulators (thalamus, brainstem) may amplify autonomic sensations relevant to seizure-related fear and SUDEP risk (12). Conceptually, seizure-associated fear of death can be viewed as a pathological, temporally compressed overactivation of circuits that subserve existential threat in non-pathological conditions.

A motivational distinction offers a heuristic: existential threats preferentially recruit the Behavioral Inhibition System (BIS), centered on medial frontal/ACC circuits, promoting vigilance and sustained anxiety, whereas immediate physical threats more prominently engage fight–flight–freeze systems (13). Mortality salience is associated with prolonged self-referential processing and heightened vigilance (e.g., ACC activation, late positive potentials), unlike the typically brief affective surge of nonexistential threats (14, 15). This framework may help explain the unusually prolonged, narratively elaborated ictal fear observed here.

Individual differences modulate mortality-related neural responses: secure attachment, social-coalitional traits (e.g., leadership), and high self-esteem appear to buffer engagement of self-referential and parietal networks, whereas insecure attachment, low self-esteem, or pre-existing fear of death amplify such responses (16–18). While not directly measured in our patient, these factors may shape vulnerability to elaborated fear content during seizures.

In summary, convergent evidence from epilepsy and functional neuroimaging indicates that fear of death recruits overlapping—though not identical—substrates across physiological and pathological contexts. The amygdala, insula, mesial temporal regions, and ACC constitute a core network for both existential fear and ictal dread, with differences in context, temporal dynamics, and phenomenology. Seizures provide a unique, paroxysmal window into these systems.

Recognizing this overlap enriches interpretation of seizure semiology and cautions against misdiagnosis as primary psychiatric conditions, while highlighting how death-related threat processing is embedded in fundamental human cognition and emotion (19).