Polyspikes are characterized by a run of two or more spikes, whereas the polyspike-wave complex consists of polyspikes followed by slow waves (5). In GGE, polyspikes usually occur in the form of high-amplitude rhythmic bursts with synchronized and generalized distribution (Figure 1).

This is an abnormal response manifesting with the generation of spike-wave complexes, polyspikes, or polyspike-wave discharges during intermittent photic stimulation (5). The PPR is under the influence of several confounding variables including age, sex, ethnicity, genetics, antiepileptic medication use, state of alertness (sleep vs wakefulness), sleep deprivation, and the stimulation technique. There are three grades of PPR: (1) posterior stimulus dependent response, (2) posterior stimulus independent response, and (3) generalized response (23). The response to photic stimulation is defined as self-sustained when the epileptiform discharges outlast the stimulus by ≥100 ms (24). It is most frequently detected in JME (83%) followed by CAE (21%) and JAE (25%) (7). However, PPR can also be elicited in 0.3–4% adults without a history of epilepsy (25, 26). It is detected more frequently (14.2%) in asymptomatic children (27). The influence of various confounders including stimulation techniques may explain the wide range of results reported in the literature.

Epileptiform discharges characteristic of eye-closure sensitivity emerge within 1–3 s of eye closure and last for 1–4 s. However, the discharges do not persist for the total duration when eyes remain closed (Figure 2). Photosensitivity and eye-closure sensitivity are related phenomena (28).

Epileptiform discharges, generalized or occipital, triggered by the elimination of fixation and central vision are the hallmarks of FOS (29). This abnormality needs to be distinguished from photosensitivity and eye-closure sensitivity. In FOS, epileptiform discharges persist for the total duration of eye closure and disappear on eye opening (Figure 2) (29). To confirm FOS, central vision and fixation should be abolished with the application of spherical lenses, Frenzel lenses, or Ganzfeld stimulation technique (30). FOS has been described in GGE and occipital epilepsy (30). In some patients, photosensitivity and FOS coexist (31).

The overlap between generalized epileptiform discharges and K-complexes (epileptiform K-complexes) as well as sleep spindles (epileptiform sleep spindles) has been known to researchers for many decades (32, 33). This overlap generates complexes with a very characteristic morphology and topography (Figure 3) (33). A recent study has found this to be common in GGE with 65% of patients demonstrating epileptiform K-complexes and 10% epileptiform sleep spindles (34). These abnormalities probably indicate the link between microarousals and epileptiform discharges in GGE (34).

Occipital intermittent rhythmic delta activity is characterized by transient unilateral or bilateral occipital runs of 2–3 Hz, regular, rhythmic, and sinusoidal delta activity (5). Deep stages of sleep, as well as eye opening, typically attenuate OIRDA, whereas drowsiness and hyperventilation make it more prominent (35). It is detected in approximately one-third of patients diagnosed with CAE (36). Though often reported as an EEG abnormality of CAE, OIRDA is not specific to epilepsy. It is occasionally seen in encephalopathies, particularly in children (35).

High-amplitude, generalized, polyspike activity of 10–16 Hz with the frontocentral maximum is the EEG hallmark of myoclonic seizures (19, 37). These typical discharges may be preceded by irregular 2–5 Hz GSW activity and sometimes followed by irregular slow waves of 1–3 Hz (Figure 4) (19, 37, 38). The EEG seizure may be several seconds longer than the clinical seizure (37, 38).

Bilateral, regular, symmetrical, and synchronous 3-Hz spike-wave activity (range 2.5–4 Hz) sometimes admixed with polyspike-wave discharges on a normal background is the hallmark of a typical absence seizure (Figure 5) (39, 40). There are some differences among syndromes.

It is not unusual for the initial ictal discharge to be atypical. It could be non-generalized, spike-wave, polyspike-wave, or irregular discharges, with typical GSW activity appearing after an average of 0.7 s (7).

Myoclonic absence seizures are semiologically characterized by absences in association with tonic contractions resulting in progressive upper limb elevation and superimposed rhythmic myoclonic jerks (2). The impairment of awareness is less pronounced in comparison to typical absence seizures. Both myoclonic absence and typical absence seizure have similar ictal EEG patterns (Figure 6) (2, 41). Polygraphic recordings are used to demonstrate the correlation between the ictal EEG and tonic as well as myoclonic activity. Often triggered by hyperventilation, myoclonic absence seizures are less frequently (14%) induced by intermittent photic stimulation (41).

Eyelid myoclonia with absences, EEG paroxysms/seizures triggered by eye closure, and photosensitivity are the main features of Jeavons syndrome (42). Sometimes, eyelid myoclonia may not be associated with an absence seizure (42). The ictal EEG typically shows generalized high-amplitude polyspikes and polyspike-wave discharges of 3–6 Hz lasting 1.5–6 s (Figure 7) (42). The ictal discharges occur with or before the onset of eyelid myoclonia (42). The EEG abnormalities are usually triggered by eye closure, intermittent photic stimulation, and hyperventilation (42). FOS may coexist (42).

Muscle and movement artifacts mask the EEG during GTCS unless muscle relaxants are used to paralyze the subject. Generalized polyspike-wave bursts usually mark the ictal onset. Generalized amplitude attenuation follows, with or without low voltage, generalized, 20–40 Hz fast activity superimposing for a few seconds. The onset of the tonic phase coincides with the voltage attenuation. Then, generalized rhythmic alpha activity (10–12 Hz) evolves with increasing amplitude and decreasing frequency accompanied by the ongoing tonic phase. When the decreasing frequency reaches 4 Hz, repetitive polyspike-wave complexes emerge accompanied by myoclonic and clonic jerking semiologically. With the progression of the seizure, periodic bursts of polyspike-wave discharges appear with background suppression in between. Generalized EEG suppression is seen for a variable period with the termination of clonic jerking. The gradual recovery is marked by the restoration of the background rhythm from irregular generalized delta slowing proceeding to theta, and finally alpha rhythm (38).

There are circadian variations in seizures and epileptiform discharges in GGE. Generalized spike-wave (GSW) activity is seen more often in non-rapid eye movement (NREM) sleep, but rare in rapid eye movement sleep (49). In GGE, sleep deprivation significantly increases the density of spike-wave discharges in both sleep and wakefulness (50). In JME, routine EEGs (without sleep deprivation) done in the morning are more often abnormal than those done in the afternoon (51). In JME, sleep EEG always shows epileptiform discharges (52).

Epileptiform discharges in GGE appear to be closely related to sleep–wake cycle. A retrospective study based on 24-h ambulatory EEGs found that 4.6% of patients had epileptiform discharges correlating with awakening. All patients who had epileptiform discharges on awakening were diagnosed with GGE. The epileptiform discharges were detected between 20 and 50 min following awakening in JME (53).

The interaction between circadian rhythmicity and the sleep–wake cycle in the generation of epileptiform discharges in GGE has been evaluated in a recent study (54). Epileptiform discharges are significantly shorter in duration and more frequent during the NREM sleep compared with wakefulness. When quantified, 67% of epileptiform discharges are detected in NREM sleep whereas 33% occurs in wakefulness. The distribution of epileptiform discharges demonstrates two peaks (11 p.m. to 7 a.m. and 12 noon to 4 p.m.) and two troughs (6 p.m. to 8 p.m. and 9 a.m. to 11 a.m.) (54). These findings highlight the variability in the diagnostic yield in relation to the time-of-day and sleep–wake cycle. The best time for the optimal yield of EEG abnormalities is from 11 p.m. to 7 a.m. Similarly, capturing natural sleep during the EEG recording significantly increases the diagnostic yield (54). Hence, 24-h ambulatory EEG is a very useful diagnostic tool in GGE.

Hyperventilation is routinely used as an activation method in EEG. Hyperventilation-induced EEG abnormalities seem to depend on the severity of hypocapnia and the reduction in cerebral blood flow (55).

Hyperventilation often induces ictal and interictal abnormalities in children diagnosed with absence seizures (56). Hyperventilation triggered absence seizures in 67% of patients in a pediatric cohort (mean age 9.3 years) diagnosed with JAE and CAE (55). In untreated children, hyperventilation induces absence seizures more often in CAE and JAE (87% each) in comparison to JME (33%) (7). In contrast, another study involving a predominantly adult cohort, during hyperventilation, no one with generalized epilepsy had seizures and only 12.2% had an increase in interictal epileptiform discharges (57). Hyperventilation-induced GSW paroxysms were found in only 12.3% of adult patients with GGE on treatment (13). These studies suggest that during hyperventilation absence seizures are more likely to occur in the younger age group with untreated CAE and JAE.

Intermittent photic stimulation is a routine induction technique during EEG recordings. The PPR is more often seen in generalized epilepsy than in focal epilepsy. It is under the influence of several variables including age, sex, antiepileptic drug therapy, level of arousal, sleep deprivation, and the stimulation technique (4).

Reflex seizures on exposure to specific stimuli are sometimes encountered in GGE. The use of such stimuli as activating procedures during the EEG recording in selected patients is an option to improve the yield of EEG abnormalities.

Reflex seizures involving visual stimulation have been reported in several epilepsy syndromes including GGE symptomatic generalized epilepsy, and occipital epilepsy (58). Flickering lights, patterns, video games, and television are among the common visual triggers. Both photosensitivity and pattern sensitivity are implicated in television and video game induced seizures. Around 90% of patients with electrographic pattern sensitivity also demonstrate PPR (59). 3D television and movies do not pose a higher risk of reflex seizures than 2D television and movies (60).

Non-verbal cognitive stimuli such as thinking and praxis may induce reflex seizures in GGE. In a study involving reflex epilepsy triggered by spatial tasks, card or board games, and calculation, 96% experienced generalized tonic-clonic seizures often preceded by myoclonic jerks, whereas 68% demonstrated generalized epileptiform discharges on EEG (61). Another study involving 480 patients found that neuropsychological tasks provoked epileptiform discharges in 38 patients and 36 of those patients were diagnosed with GGE (62). Mental arithmetic and decision-making may trigger “noogenic” (thinking-associated) seizures among susceptible individuals. Cognitive activity in conjunction with planned motor tasks usually with hands is implicated in praxis-induced seizures (63). Praxis-induced reflex seizures are particularly common in JME (45, 62). Reading, talking, and writing are examples of verbal cognitive stimuli that may trigger reflex seizures. Both generalized and focal epilepsies have been reported under this category (63).

Several electroclinical syndromes such as CAE, JAE, JME, and epilepsy with generalized epilepsy with tonic-clonic seizures alone (GTCSA) have been described in GGE. In this review, we will focus on the four main syndromes: CAE, JAE, JME, and GTCSA. It should be noted that apart from the electroclinical syndrome, epileptiform abnormalities in GGE are under the influence of many variables including sex, age, the state of alertness, activation procedures, techniques of EEG recording, and antiepileptic drug therapy (4).

Childhood absence epilepsy is typically seen in children and the EEG signature is “generalized, bisynchronous, and symmetrical 3-Hz spike-wave discharges emerging from a normal background” (2). Fragments of GSW discharges are seen in >90% of cases, predominantly in drowsiness and sleep (7). Interictal polyspikes usually occur in drowsiness and sleep (7). Polyspike-wave discharges were detected in 26% of patients in a different series (64). Among untreated children with CAE, only 21% demonstrate PPR, whereas hyperventilation-induced absence seizures are seen in 87% (7).

Occipital intermittent rhythmic delta activity is seen in 20–30% of CAE subjects (8, 36), and 40% of those have a notched appearance (36).

The onset of JAE is in teenage years (12–17 years). Absence seizures are less frequent but myoclonus is more common in JAE compared with CAE. In comparison to CAE, generalized tonic-clonic seizures more frequently precede the onset absence seizures in JAE (2). Fragmented discharges and polyspikes are seen in all patients, mostly in drowsiness and sleep (7).

Patients with JME typically experience their first seizure at puberty (12–18 years). The typical semiologic feature is myoclonic seizures predominantly involving arms. Generalized tonic-clonic seizures occur more frequently than absences (2). Sleep deprivation and alcohol are potent seizure triggers. Seizures, particularly myoclonus, frequently occur after awakening from sleep (19).

The classic EEG abnormalities in JME are generalized polyspikes and polyspike-wave discharges (19, 38). The interictal EEG is characterized by 3–6 Hz spike and polyspike-wave discharges in an irregular mix (28). Focal EEG abnormalities are common (45). PPR is seen in the majority (7). Both eye-closure sensitivity and FOS have been reported in JME (45).

This condition is characterized by GTCS occurring on awakening or at random times. The median age of onset is (18 years) significantly older than JME and JAE (65). The interictal EEG demonstrates generalized polyspikes, polyspike-waves, and spike-wave discharges similar to other GGE syndromes. The mean spike-wave frequency is 3.6 Hz. The density of epileptiform discharges is significantly lower than CAE, JAE, and JME (66).

In all GGE syndromes, the initial frequency of GSW activity is faster. In the next phase, the discharges become more regular and slower in frequency by 0.4–0.6 Hz. The frequency decreases again in the terminal phase of CAE and JAE (20). The highest median frequency of GSW during the first second of an absence seizure is in JME (3.5 Hz). It is marginally slower in JAE (3.25 Hz) and CAE (3 Hz) (7). In JME, the GSW activity often tends to be faster (>3.5 Hz) (19, 38, 67). A more recent study based on 24-h EEGs found median GSW frequencies of 3.3 (CAE), 3.1 (JAE), 3.8 (JME), and 3.5 (GTCSA). But the differences were not statistically significant (66).

Childhood absence epilepsy and JAE demonstrate similar morphologies of GSW discharges. Multiple spikes preceding or overlapping slow waves give rise to an appearance of compressed “W”s in absence seizures of JME (20). The polyspike-wave activity is seen more often in JME and JAE than CAE (7). CAE and JAE have longer EEG seizure durations than JME (20, 68). The longest EEG absence seizure is seen in JAE, whereas the shortest is in GTCSA (66).

Absence seizures typically demonstrate well organized regular and rhythmic ictal EEG pattern. In disorganized discharges, regular rhythmic activity is interrupted by, (a) brief (<1 s) and transient interruptions in ictal rhythm, or (b) waveforms of different frequency and/or morphology (Figure 13) (7). Disorganized ictal discharges are 110 times more likely to occur in JME than CAE and eight times more likely in JAE than CAE (7). It is also influenced by provoking techniques, the state of arousal, and the age (7). Irregular and disorganized paroxysms are also seen in GTCSA though less frequently (66).

Table 1 summarizes key EEG differences among the four main GGE syndromes.

A study based on dense array EEG in absence seizures has demonstrated spike-wave discharge onset in the dorsolateral frontal and orbital frontal regions followed by rapid and stereotypic propagation (71). Electrical source analysis of dense array EEG data has revealed frontotemporal networks involving the slow wave and spike propagation through ventromedial frontal networks during absence seizures (72).

A combined MEG/EEG study has described a prefrontal–insular–thalamic network in absence epilepsy (73). In JAE, the spike-wave discharge onset is in focal cortical regions with subsequent involvement of the default mode network as demonstrated by synchronous MEG/EEG data (74).

Electroencephalographic and functional MRI is a non-invasive technique to measure regional brain activation during epileptiform discharges using blood oxygenation level-dependent contrast (75). A recent critical review has elicited three key features among EEG-fMRI findings in GGE: (a) activation of the thalamus, (b) activation of cortical regions, particularly frontal, and (c) deactivation of default mode areas (47).

Combined TMS and EEG is an emerging non-invasive technique with a potential to study the functional connectivity of the brain (76). A protocol to study GGE patients with TMS-EEG has been recently described (77). However, to date, changes in network connectivity have mostly been studied with TMS-EEG in focal epilepsy (78), while evaluation in GGE remaining in its infancy (79).

Graph theory is a mathematical concept to study brain connectivity. It describes networks in terms of interrelationship between nodes (brain regions) and edges (connections) (80). Graph theory is increasingly being used as a tool to analyze epileptic networks. A recent study has reported increased local connectivity in the frontal regions with spike-wave discharges in JME (81). Another study based on graph theory using EEG data found similarities in network topology between patient with GGE and their unaffected relatives (82).

Conclusions should be drawn with care from these studies due to various limitations. There is wide variability in the methodology among studies. In particular, EEG-fMRI studies vary in terms of study paradigms, and methods of data acquisition as well as analysis. Additionally, most studies are based on GSW activity. In GGE, there are other EEG abnormalities and underpinning network mechanisms may be different in those. Despite such limitations, there is growing support for the hypothesis that spike-wave discharges originate from a cortical focus with rapid spread to the thalamus followed by entrainment of the cortico-thalamo-cortical loop resulting in the classic GSW activity observed in GGE (83).

Paroxysmal disorders ranging from syncope to psychogenic non-epileptic seizures can be misdiagnosed as epilepsy. The rate of misdiagnosis is as high as 20–30% in general practice and outpatient clinics (94, 95). Misdiagnosis is likely to happen when an individual presenting with a non-epileptic disorder undergoes an EEG test yielding epileptiform abnormalities. It has been shown that 0.5% of healthy adults in the general population have epileptiform abnormalities in the EEG (25). Among healthy school children, the prevalence of GSW activity in the EEG is 0.9% (96). Epileptiform discharges are more frequently (37%) detected among the offspring of patients with epilepsy. 6% of healthy first-degree relatives of JME probands demonstrate typical generalized epileptiform discharges (97, 98). Hence, we wish to emphasize the importance of clinical correlation of EEG abnormalities in establishing the diagnosis of epilepsy.

Atypical features, including focal epileptiform discharges, can potentially result in delayed diagnosis and misdiagnosis of GGE. The rate of misdiagnosis can be as high as 91% and the mean delay to the diagnosis ranges from 6 to 15 years in studies (47). As a result, many patients receive inappropriate antiepileptic drugs such as carbamazepine leading to paradoxical worsening of some seizures (47).

The 6-Hz spike-wave (phantom spike-wave) pattern consists of bursts of generalized symmetric spike-wave discharges with a very low amplitude spike component (106). The bursts are typically very brief but can last up to 4 s on rare occasions (106). The amplitude maxima can be anterior or posterior (107). This variant, particularly the type with posterior maximum, usually emerges from drowsiness and disappears during deep sleep. Though the typical frequency is 6 Hz, it can range from 4 to 7.5 Hz (107). The spike component is <25µV in the majority and >75 µV in 5% (107). This is a benign variant of no clinical significance.

The 14- and 6-Hz positive burst pattern (14 and 6 Hz positive spikes or ctenoids) is a benign variant most often seen in early teens, which then becomes infrequent with advancing age. These spikes are surface positive in polarity, occurring in bursts of <1 s, with unilateral or bilateral distribution and posterior dominance during drowsiness and light sleep (106). This can be mistaken for polyspikes, but careful analysis of the polarity, frequency, and distribution should help clarify the diagnosis.

Small sharp spikes (benign sporadic sleep spikes) are seen in adults during drowsiness and light sleep. It disappears in deep sleep. The sharp waves are usually diphasic with a low amplitude (<50 µV) and brief duration (<50 ms) without an after-going slow wave. The spikes occur in the form of isolated transients with a unilateral or bilateral widespread field most prominent in the temporal regions (106).