An aura usually lasts seconds to minutes and immediately precedes the signs of an attack. On occasion, auras can be long-lasting, continuous, or recurrent, with short intervening breaks. Intracranial electroencephalographic (EEG) studies show that prolonged auras (aura continua) can represent continuous or recurrent seizures, a form of focal status epilepticus (4). More frequently, for hours to days before attacks, patients may experience prodromal symptoms of nervousness, anxiety, dizziness, and headache that should not be regarded as auras. The prodrome may be evident on awakening and signals a seizure that will occur later in the day. Sometimes a patient may not be conscious of anything untoward, but family members or friends may describe irritability or “a mean streak.” Such prodromal symptoms resemble those recounted by patients with migraine. Gowers’ speculation (1) that the prodrome is “indicative of slight disturbance of the nerve centres” has not been improved on.

Less commonly, some patients with generalized epilepsy may experience stereotyped sensations before the occurrence of a generalized seizure. Like an aura, these premonitory symptoms immediately precede the seizure but are brief and nonspecific, lacking the features that suggest activation of a circumscribed area of the cortex. Sensations include dizziness, warmth, cold, generalized tingling, anxiety, and a spaced-out, or confused, feeling. Rarely, illformed visual imagery and abdominal sensations have been reported. Some of the sensations likely correspond to a buildup of absence (dizziness, lightheadedness, and confusion) or myoclonic seizures (anxiety, restlessness, jumpiness, and jerking).

Although an aura reflects activation of functional cortex by a circumscribed seizure discharge, the seizure discharge frequently spreads. When the seizure discharge spreads along a single functional area, such as the postcentral gyrus, the sensory equivalent of focal motor jacksonian march is seen. An aura can also spread across different functional regions. A seizure that begins in the primary visual area of the occipital lobe and spreads to the temporal limbic structures may present with initial transient blindness, followed by other sensations referable to the temporal lobe (5).

Multiple sensations can occur even when seizure activity is relatively confined to one region, as at the start of a temporal lobe seizure. In some cases, the multiple auras can be dissected out along a sequence, implying spread of the seizure discharge. Anxiety, epigastric, and “indescribable” sensations commonly precede the more complex phenomena of déjà vu and other illusions of vision or sound (6), although the sequence is not always stable in different attacks. In other cases, a time series cannot be discerned, and the multiple sensations seem to occur simultaneously. An alternative explanation for multiple auras could be that they are secondary to activation of a system with access to more than one functional region. The temporal limbic system, with extensive connections to the septum, hypothalamus, temporal neocortex, insula, and parieto-occipital association cortex, is an example (7). In support of this hypothesis, electrical stimulation of temporal limbic structures by depth electrodes can produce different sets of sensations at different times, despite stimulation of the same contact (8).

The incidence of auras in large populations is imprecise. In the 32-year epidemiologic study from Rochester, Minnesota (9), epilepsy with focal sensory seizures was reported in 3.7% of all patients. An additional 26.4% were classified as having temporal lobe epilepsy, but the incidence of aura in this group was not separately reported. In the two large series of clinic- and office-based patients with epilepsy studied by Gowers (1) and by Lennox and Cobb (10) (Table 15.1), an aura was present in 56% of patients. Although notable discrepancies exist between the two series with respect to relative frequencies of unilateral somatosensory auras, bilateral general sensations, and visual auras, other categories are remarkably consistent. Differences are most likely explained on the basis of definitions of aura type.

The findings in other series were more variable. In patients with complex partial seizures or temporal lobe epilepsy, the incidence of auras ranged between 22.5% and 83% (6,11, 12, 13, 14, 15). During long-term scalp and sphenoidal or intracranial EEG monitoring of seizures, 46% to 70% of patients reported auras, either independent of or as part of their habitual seizures (16, 17, 18).

Some patients who do not describe auras may have had them early in their illness. Anecdotal experience suggests that auras may disappear as the disease progresses, and as seizures cause increasingly profound loss of awareness and postictal confusion. As Lennox and Cobb (10) stated: “It is more accurate to speak of the recollection of aura(s) rather than of their presence.” Young children may lack the verbal capacity to describe the sensations that herald a seizure, even though their actions indicate some awareness of the impending event. Similarly, adults who deny having any warning, may nevertheless press the seizure alarm button during video-EEG monitoring but have no recollection of doing so. The seizure either induces an amnesia so immediate that there is no memory of a warning or causes retrograde amnesia. This is supported by a study that showed amnesia for auras depended on the severity of the seizure (19). An isolated aura is nearly always recollected and associated with either no ictal discharge or a unilateral EEG ictal discharge. The aura is more likely to be forgotten if the seizure becomes secondarily generalized and involves bilateral EEG ictal discharge.

The complete cessation of all seizures—the desired goal of successful epilepsy surgery—cannot always be achieved. Auras can persist as isolated phenomena following epilepsy surgery, when complex partial or secondarily generalized seizures no longer occur, and even after discontinuation of antiepileptic drugs. Isolated postoperative auras are often ignored and classified among the “seizure-free” outcomes. In a few studies, isolated postoperative auras occurred in 20% (20,21) to 35% (22) of patients following surgery for temporal lobe epilepsy and in 22% of patients after focal resective surgery unselected for location (23). Residual auras seem particularly common following temporal lobe surgery, and may be related to incomplete removal of the mesial temporal structures comprising the amygdala, hippocampus, and parahippocampal gyrus. The persistence of epigastric auras after functional hemispherectomy, in which the insula is the only cortical structure still functionally connected on the side of surgery, suggests that continuing seizure activity in that structure may be another mechanism. Postoperative auras commonly recurred within the
first 6 months of surgery and tended to persist (22). Although isolated postoperative auras are widely regarded as being of little significance, they may accompany an increased risk for recurrence of complex partial seizures (22) and reduced quality of life on self-assessment (23).

A small number of patients may lose their auras following temporal lobectomy, even as they continue to have postoperative complex partial seizures; others may experience a different aura. Such alterations occurred in 55% of patients who had residual postoperative seizures (20).

A long duration of epilepsy correlates with an increased incidence of auras, which Lennox and Cobb (10) thought “presumably due to the greater total number of seizures and the greater likelihood of experiencing aura.” Early onset epilepsy, lower intelligence quotient (IQ), male gender, and right temporal lobe focus are all associated with a higher incidence of “simple primitive” auras, whereas complex “intellectual” auras with illusions or hallucinations accompany male gender and a verbal IQ greater than 100 (24).

Aura content may be related to a patient’s psychological makeup. Stimulation of various mesial limbic structures elicited auras with features that were intimately related to ongoing psychopathologic processes (25). Emotional responses and hallucinations produced by electrical stimulation were reported to depend on the background affective state (26,27). Similarly, patients who experienced anxiety or fear during temporal lobe electrical stimulation scored higher on the Psychasthenia scale of the Minnesota Multiphasic Personality Inventory, whereas those experiencing dreamlike or memory-like hallucinations scored higher on the Schizophrenia scale (8). The aura phenomena shown to be sensitive to personality factors are precisely those that comprise an individual’s personality. Thus, the memory flashback that may be recalled in an aura is not a generic item, but rather an experience specific to a patient.

An aura provides evidence of focal seizure onset. The nature of the symptoms may localize the epileptogenic zone. All sensations near the onset of seizures are not necessarily auras, however. It is important to differentiate auras from prodromes and from nonspecific premonitions before generalized seizures. Auras may vary in the same patient or occur in combination but should show a certain stereotypy and consistency. It may be particularly difficult to classify a first seizure based on the report of a preceding sensation. One study (28) noted poor interobserver agreement about the nature of such preceding sensations. In the same study, the incidence of generalized versus focal epileptiform EEG abnormalities and of structural abnormalities on computed tomography was similar in the 67 patients with and the 82 patients without sensations preceding a generalized convulsion. At 1-year follow-up, seizures had recurred in 22 of the 67 patients with preceding sensations, but only 11 of these had clinical indications that the recurrences were of focal onset. Thus, self-report of a preceding sensation in an isolated first convulsion may not be a reliable indicator of focal epilepsy.

The sensation before an ictal event can be misleading in another situation. Sometimes, albeit rarely, patients have pseudoseizures beginning with an epileptic aura (29). The epileptic seizures often are well controlled except for the auras. Whether the pseudoseizure that follows the aura represents a learned response or occurs from other psychogenic mechanisms cannot be determined.

Current concepts of the localizing value of auras rely heavily on the pioneering studies of Penfield and Jasper (14), who correlated sensations and signs obtained through electrical stimulation of the awake patient with those of the patient’s spontaneous seizures. Subsequent studies with long-term intracranial electrodes for the recording of spontaneous seizures and extraoperative electrical brain stimulation have extended early observations (30, 31, 32, 33, 34).

Although an aura may help to localize the epileptogenic zone, an important point must be kept in mind: The initial sensation of an aura is related to the first functional brain area activated by the seizure that has access to consciousness, but this may not be the site of seizure origin. A seizure starting in the posterior parietal region may be initially asymptomatic, until ictal activity spreads to adjacent functional areas. Spread to the postcentral gyrus may elicit a somatosensory sensation as the first warning; propagation to the parietooccipital association cortex may give rise to initial visual illusions or hallucinations. Furthermore, it remains unclear whether experience of an aura is contingent on direct ictal involvement of the cortical areas subserving those functions or whether an aura sensation may also be evoked by excitation at a distance, provided a pathway of projection or facilitation exists between the site of excitation and an eloquent cortical structure. Both mechanisms are probably operative in human epilepsy. A sensory jacksonian march cannot be explained other than by ictal spread along the somatosensory cortex. The indistinguishable auras found in patients with hippocampal sclerosis and temporal neocortical pathology underlie the distributed network that functionally links the limbic and neocortical structures in the temporal lobe. Cortical stimulation in patients with extratemporal epilepsy also showed that sites at which an aura is reproduced can extend well beyond the expected functional map for those sensations (33).

The localizing value of auras has been studied in a number of ways. Penfield and Kristiansen (35) recorded the initial seizure phenomenon in 222 patients with focal epilepsy and commented on the likely localization of different auras. Auras reported in patients with well-defined epileptogenic foci in different brain regions can be compared
from different series (Table 15.2) or, better yet, prospectively (36) (Table 15.3). Data from patients (37,38) who become seizure free after localized brain resections are particularly important because their surgical outcome is absolute proof of the correct localization of the epileptogenic zone. Making comparisons from different series in the literature is hampered by several problems: Definitions of aura type are not uniform, data on different auras are often grouped in dissimilar ways, and classification rules may differ when multiple sensations occur in the same aura. In spite of the different approaches, however, retrospective and prospective series yielded a remarkably similar conclusion: Auras have localizing significance. Patients with temporal lobe epilepsy have the highest incidence of epigastric, emotional, and psychic auras (36,37). Frontal lobe epilepsy is distinguished by frequent reports of no aura (36,38). When an aura is present in a patient with frontal lobe epilepsy, cephalic and general body sensations predominate (36). Perirolandic epilepsy with centroparietal foci is most likely to involve somatosensory aura (39). Unsurprisingly, occipital lobe epilepsy is associated with the highest incidence of visual aura (36,40). No single aura sensation is necessarily restricted to a single lobe, however.

Except for unilateral somatosensory and visual auras contralateral to the site of seizure onset, the nature of an aura provides no reliable lateralizing information. Penfield and colleagues (14,41) reported that psychic illusions were lateralized mainly to the nondominant temporal lobe. Subsequently, these findings have been confirmed by some researchers (42) but refuted by others (6,12,16).