Epilepsy

I distinguish an epileptic event from the enduring predisposition that makes it epilepsy. In Choi’s lecture, that distinction connects clinical classification, hippocampal networks, social stigma, and the electrophysiology used to translate between patients and experimental models.

Clinical overview

Terminology and operational definitions

The lecture separates seizure, convulsion, and epilepsy. “Seizure” can broadly describe a sudden event; an epileptic seizure specifically involves transient symptoms or signs from abnormally excessive or synchronous neuronal activity. Its onset and termination delimit a paroxysm, with subjective experiences, observable manifestations, or both. A convulsion describes abnormal excessive muscle contraction, usually bilateral, and does not establish epilepsy. Epileptic seizures need not be convulsive.

The lecture moves from recurrent seizures to the 2005 conceptual definition: an enduring predisposition with neurobiological, cognitive, psychological, and social consequences, requiring at least one seizure. This permits diagnosis after one event but leaves “enduring predisposition” without an operational threshold. Earlier epidemiological definitions required two unprovoked seizures at least 24 hours apart, excluding acute symptomatic events and a single cluster within one day.

The ILAE 2014 practical definition supplies three alternatives:

1. At least two unprovoked or reflex seizures more than 24 hours apart.
2. One unprovoked or reflex seizure with an estimated recurrence probability of at least 60% over the next ten years, comparable to the risk after two unprovoked seizures.
3. Diagnosis of an epilepsy syndrome.

Epidemiology and age-specific causes

The lecture describes common disease with U-shaped incidence: high in early life, lower through much of adulthood, and rising again in older people. It reports higher incidence in developing-country populations and an increasing elderly burden. I retain these as the lecture’s epidemiological framing rather than universal rates.

Its risk factors include head trauma, CNS infection, stroke, Alzheimer’s disease, cerebral palsy, intellectual disability, and previous febrile seizures. Perinatal injury, malformation, and infection dominate the youngest age groups; febrile seizures feature in infancy and toddlerhood. Idiopathic epilepsies become prominent in school-age children and adolescents. Infection, tumors, and especially cerebrovascular disease matter in adults. A risk history does not itself establish epilepsy, and a febrile or acute symptomatic seizure is not interchangeable with an enduring seizure disorder.

Anatomy and pathophysiology

Networks and hippocampal sclerosis

The lecture summarizes epilepsy through hyperexcitability and hypersynchrony. Focal seizures involve networks within one hemisphere; generalized seizures engage bilateral networks. Manifestations depend on which structures are recruited and how activity spreads.

Mesial temporal lobe epilepsy with hippocampal sclerosis, mTLE-HS, is the lecture’s main structural example. Its proposed sequence begins with hippocampal injury, followed by neuronal loss involving the dentate gyrus and CA regions, aberrant synapse formation, and compensatory excitatory circuits that support an epileptic network. I read this as network reorganization in which structural damage and circuit activity describe different parts of the mechanism.

Translating between bedside and bench

The closing slides connect animal seizure models with human and rodent ictal recordings, rodent status epilepticus, Fourier analysis, compressed spectral arrays, hippocampal-slice multielectrode arrays, and cell recordings. The lecturer insists on electrophysiology alongside behavior.

In the slide’s pentylenetetrazole example, Racine convulsion scores are compared with electrocorticography, and the first spike coincides with a myoclonic jerk. The caption cautions that spike frequency and seizure expression do not have a settled correspondence.

Beyond the slide caption, the cited Aygun and Ayyildiz study found different vortioxetine effects across induced-seizure and genetic absence models. I read this as a reason to specify which mechanism and measurement an experiment represents. Behavioral severity, electrical activity, and a human epilepsy syndrome require separate validation.

Diagnostic workflow and imaging findings

Classify the seizure and then the epilepsy

I retain the lecture’s 2017 framework. Fisher et al. distinguish focal, generalized, and unknown onset, with motor and nonmotor categories and awareness specified for focal seizures. “Focal impaired awareness” replaces the older complex partial terminology; “focal to bilateral tonic-clonic” replaces secondary generalization.

The separate Scheffer et al. framework proceeds from seizure type to epilepsy type and, where possible, syndrome. Epilepsy types are focal, generalized, combined generalized and focal, or unknown. Structural, genetic, infectious, metabolic, immune, and unknown etiologies are considered throughout.

The lecturer cautions that these classifications do not map one to one. Absence seizures do not establish absence epilepsy; they also occur in JME and other syndromes. Bilateral tonic-clonic activity can follow either generalized onset or focal spread. An aura or initial focal motor sign can change that interpretation.

Recognize juvenile myoclonic epilepsy

The lecture presents JME as an age-related idiopathic generalized epilepsy, typically recognized around adolescence. Brief arm jerks shortly after awakening may go unreported until a tonic-clonic seizure prompts assessment. Sleep deprivation and awakening are characteristic contexts. Absences occur in approximately one third of cases, a lecture figure also supported by the ILAE syndrome resource. The same resource confirms that photic stimulation can facilitate seizures, where the slide places a question mark.

Reconstruct the mesial temporal sequence

The lecture’s characteristic sequence is an aura, impaired awareness with motion arrest, automatisms with possible language disturbance, and sometimes secondary generalization. Auras include epigastric rising, fear, déjà vu, jamais vu, olfactory hallucination, and depersonalization. The slide distinguishes these from well-formed hallucinations, which it treats as atypical for mesial temporal onset.

Objective features include staring and arrest, lip smacking, chewing, tooth grinding, fumbling, and picking. Vocalization, spitting, and piloerection are also illustrated. The sequence can stop before bilateral convulsion; the early subjective component remains diagnostically relevant.

The lecture gives specific lateralizing clues. Forced head and eye version and unilateral tonic or dystonic posturing point contralaterally; unilateral automatisms point ipsilaterally. Ictal aphasia or speech arrest suggests the language-dominant hemisphere, usually the left. Preserved ictal speech, or automatisms with preserved awareness, suggests the nondominant hemisphere. Ipsilateral unilateral blinking is discussed with a caution: it more often implicates occipital or temporoparieto-occipital regions than mesial temporal onset. I would interpret these clues together with their timing and EEG.

Combine EEG and imaging

Neurological examination is usually normal in the lecture’s mTLE description. Interictal EEG may show anterior temporal sharp waves, spikes, and slowing; the ictal example is unilateral rhythmic activity at 5–7 Hz. As a literature addition, NICE explicitly advises against excluding epilepsy with EEG. A normal interictal recording cannot settle the diagnosis.

The lecture associates hippocampal sclerosis with hippocampal atrophy and increased T2/FLAIR signal on MRI. FDG-PET shows hypometabolism. Temporal perfusion may be reduced interictally and increased ictally on SPECT, making acquisition timing essential. Its SISCOM illustration combines subtraction of ictal and interictal SPECT with MRI coregistration. I read these modalities as complementary observations of structure, metabolism, and perfusion.

Differential diagnosis and management context

Separate provocation, mimics, and treatment context

The lecture distinguishes epilepsy from seizures accompanying acute metabolic, toxic, systemic, or CNS insults. As an addition beyond its brief workup material, NICE recommends history, eyewitness accounts or video, ECG for cardiac mimics, and assessment of metabolic disturbance. These help distinguish a seizure disorder from an acute precipitant or another transient event.

For mTLE-HS, the lecture emphasizes frequent medication resistance and lists carbamazepine, oxcarbazepine, and anterior temporal lobectomy. I retain these as its treatment context, without treating the list as a prescribing algorithm or surgery as appropriate from MRI alone.

Social stigma and the Korean name

The lecture explicitly treats epilepsy as a social label. It recounts stereotypes of aggression, hyperreligiosity, and hypersexuality, alongside educational, occupational, and relationship disadvantage. These are presented as stigma and social burden, not defining patient traits.

Its Korean comparison is 간질 versus 뇌전증. As supporting literature beyond the slides, Kim et al. describe replacing a stigmatizing name with a term meaning cerebroelectric disorder. The intended change places epilepsy within neurological disease and reduces associations with madness. I read this section as relevant to diagnostic communication: assigning a label can influence disclosure and how others interpret a person’s behavior. The naming intervention itself does not establish that discrimination has disappeared.

Implications for medical AI

I would separate discharge detection, recorded-seizure detection, and patient-level epilepsy diagnosis. The diagnostic event may never occur during recording. For clinical faithfulness auditing, I would preserve recording duration and ascertainment method, distinguish unobserved events from negative findings, and evaluate false alarms over continuous recording time.

I read the 2014 single-seizure criterion as embedding a risk model within the diagnosis. A dataset using it contains an estimated future quantity. I would record which diagnostic criterion established each label and assess recurrence prediction separately, with its ten-year horizon and follow-up limitations.

I would also preserve the classification version and original wording. Labels assigned before and after 2017 are not automatically equivalent. Mapping “complex partial” into a newer ontology requires supporting information, and missing onset evidence should remain visible.

For video semiology, I would audit whether lateralization predictions follow head version, dystonia, and automatisms or instead exploit camera position, bed orientation, electrode equipment, and site conventions. Patient and site separation, expert-timed signs, and comparisons that alter background while preserving the sign would make this a specific extension of my shortcut-learning research.

References

• Choi JY. Epilepsy & Seizure. Ajou University School of Medicine, lecture slides supplied for this note.
• Fisher et al. ILAE official report: a practical clinical definition of epilepsy. Epilepsia. 2014;55:475–482.
• Fisher et al. Operational classification of seizure types by the International League Against Epilepsy: Position Paper of the ILAE Commission for Classification and Terminology. Epilepsia. 2017;58:522–530.
• Scheffer et al. ILAE classification of the epilepsies: Position paper of the ILAE Commission for Classification and Terminology. Epilepsia. 2017;58:512–521.
• Kim et al. Changing name of epilepsy in Korea; cerebroelectric disorder (noi-jeon-jeung,뇌전증,): my epilepsy story. Epilepsia. 2014;55:384–386.
• Aygun and Ayyildiz. Vortioxetine increases absence-like seizures in WAG/Rij rats but decreases penicillin- and pentylenetetrazole-induced seizures in Wistar rats. Epilepsy & Behavior. 2021;116:107797.
• ILAE. Juvenile myoclonic epilepsy: seizures.
• NICE. Epilepsies in children, young people and adults: diagnosis and assessment. NG217.