Neuropsychological Deficits in Childhood Epilepsy Syndromes
- First Online:
- Cite this article as:
- MacAllister, W.S. & Schaffer, S.G. Neuropsychol Rev (2007) 17: 427. doi:10.1007/s11065-007-9048-4
- 1.1k Views
Seizure disorders are relatively common in childhood, and the International League Against Epilepsy (ILAE) provides a hierarchical classification system to define seizure types. At the final level of classification, specific epilepsy syndromes are defined that represent a complex of signs and symptoms unique to an epilepsy condition. The present review discusses the issues related to several of these epilepsy syndromes in childhood, including those classified as generalized idiopathic epilepsies (e.g., childhood absence epilepsy, juvenile absence epilepsy, juvenile myoclonic epilepsy), focal epilepsies (benign rolandic epilepsy, occipital epilepsy, temporal lobe epilepsy, frontal lobe epilepsy) and the “epileptic encephalopathies,” including Dravet’s Syndrome, West Syndrome, Lennox–Gastaut Syndrome, Myoclonic Astatic Epilepsy, and Landau–Kleffner Syndrome. For each syndrome, the epidemiology, clinical manifestations, treatments, and neuropsychological findings are discussed.
Estimates suggest that in the United States, approximately 5% of children will experience a seizure prior to the age of 20 (Hauser et al. 1996). About 25% of those who experience a single convulsive episode will go on to meet formal criteria for epilepsy, which requires recurrent unprovoked seizures. Accordingly, a large amount of effort and research has gone into understanding their etiology, associated features, appropriate intervention, and management. Children with epilepsy are known to have a range of cognitive deficits. An understanding of these deficits is essential to helping each child maximize their academic potential, as research has shown that school achievement in children with epilepsy is often lower than would be predicted based on global measures of cognitive function (Farwell et al. 1985).
Once a diagnosis of epilepsy is established in a child, syndrome classification is considered such that the best clinical management plan can be established. The International League Against Epilepsy (ILAE) classification involves a hierarchical system with three tiers. At the first level, the seizure type is described as generalized, localization related, or undetermined. The second level classifies seizures according to whether they are idiopathic, symptomatic, or cryptogenic. It should be noted that it has been proposed that the term cryptogenic be replaced with the term “probably symptomatic,” to indicate that there is a presumed underlying structural brain abnormality that has not yet been identified. At the final tier of classification, a specific syndrome is assigned. Specific epilepsy syndromes represent a complex of signs and symptoms that define a unique epilepsy condition. Historically, accurate syndrome classification has been challenging, and prior research documents some degree of inter-rater disagreement (Berg et al. 1999, 2000). In reviewing the available literature on the neuropsychological deficits in children with epilepsy syndromes, it is clear that terms are not used consistently across studies. As a result, it is not always known if different studies on a given syndrome are evaluating exactly the same types of children. Future studies should seek to use strict classification guidelines and standard terminology.
The aims of the present review are to draw together the relevant research on pediatric seizure syndromes. In doing so we will review the etiology of syndromes, the prevalence, treatment considerations, and the neuropsychological findings. An exhaustive review of all childhood syndromes would not be possible in the space available. Our review, therefore, will focus on syndromes that are commonly evaluated by pediatric neuropsychologists and have a strong empirical literature on the cognitive findings of such syndromes. Syndromes for which there is a paucity of cognitive data available will not be discussed. Specifically, the present review will discuss the generalized idiopathic epilepsies (e.g., childhood absence epilepsy, juvenile absence epilepsy, juvenile myoclonic epilepsy), focal epilepsies (benign rolandic epilepsy [BRE], occipital epilepsy, temporal lobe epilepsy [TLE], frontal lobe epilepsy [FLE]), and the “epileptic encephalopathies” (Dravet’s syndrome, West syndrome, Lennox–Gastaut syndrome, myoclonic astatic epilepsy, and Landau–Kleffner syndrome).
Generalized Idiopathic Epilepsies
The ILAE specifies three syndromes under the classification of generalized idiopathic epilepsies. These include childhood absence seizures, juvenile absence seizures, and juvenile myoclonic epilepsy. (There are other idiopathic generalized epilepsies that are less commonly seen, such as perioral myoclonia with absences and eyelid myoclonia with absences, but these syndromes have not yet been included in the ILAE classification and will not be addressed here due to limited information.) Absence seizures are further subclassified as typical absence seizures, atypical absence seizures, and absence status. Typical absence seizures may be either simple, with impairment of consciousness as the sole manifestation, or complex, where one may see mild clonic manifestations, changes in tone, and automatisms. An earlier study verified that automatisms during absences might involve repetitive movements. Moreover, memory for interactions during absences may vary from patient to patient and from seizure to seizure; some may be entirely anmestic for events that occur during absences, whereas others may have some memory for these events (Freemon et al. 1973). In complex typical absence seizures, differentiation from complex partial seizures can be difficult and requires EEG corroboration. Atypical absences are often characterized by a less abrupt onset and cessation, and a longer duration. Moreover, there are often more obvious changes in tone associated with these events (Commission on Classification and Terminology of the International League Against Epilepsy 1981). Atypical absence seizures commonly begin prior to the age of 5 years and are nearly always associated with other seizure types, such as generalized tonic-clonic seizures, myoclonic, atonic, and tonic seizures. Mental retardation is common.
Childhood Absence Epilepsy and Juvenile Absence Epilepsy
The term pyknolepsy is often used to describe typical absence seizures (both simple and complex) in children and is synonymous with the syndrome of childhood absence epilepsy. The term “petit mal” was previously used to describe absence seizures in school age children, and though this term is not used in the formal classification, it is still used colloquially. Onset is generally between age three and puberty, and girls are more frequently affected than boys. The absences in these children occur quite frequently (several times per day) and tend to occur in clusters (Wirrell 2003). In contrast to childhood absence epilepsy, juvenile absence epilepsy tends to develop later, typically with puberty. The two conditions can be challenging to differentiate, and many epidemiological and neuropsychological studies of absence epilepsies consider them together. Thus, the information available that is specific to this type is somewhat limited. In addition to absence seizures, this syndrome is often associated with generalized tonic-clonic seizures upon awakening in about 80% of cases (Loiseau et al. 1995). Moreover, myoclonic seizures are present in about 15% of cases, which can also make this syndrome difficult to distinguish from juvenile myoclonic epilepsy (Reutens and Berkovic 1995).
Absence seizures account for between two and 11 percent of all seizure types across all age ranges, with the highest rates seen in the first 10 years of life (Blume et al. 1973; Cavazzuti 1980; Dalby 1969; Heijbel et al. 1975; Livingston et al. 1965). Juvenile absence epilepsy accounts for approximately 11–20% of patients with absence epilepsy, and males and females are equally affected (Wirrell et al. 1996). In short, there are considerably fewer numbers of patients with juvenile absence epilepsy than with childhood absence epilepsy, though it is generally considered to be a more serious disorder. The presence of concomitant generalized tonic-clonic seizures is associated with a poorer prognosis (Tovia et al. 2006). With respect to the etiology, genetic factors are considered to be most salient. For example, it has been demonstrated that absence seizures and generalized spike and wave discharges are both inherited traits (Metrakos and Metrakos 1961). Given the fact that bilateral synchronous 3 Hz spike and wave activity is a defining feature of absence seizures, deep subcortical structures are implicated in the pathophysiology; studies have implicated the thalamocortical circuitry in absence seizures (Gloor and Fariello 1988).
Valproic acid, ethosuximide, and clonazepam have all proven effective in decreasing seizure frequency (Browne et al. 1975; Browne 1976, 1978; Bruni et al. 1980; Lund and Trolle 1973; Sherard et al. 1980). Generally speaking, cognitive side effects are low with ethosuximide and valproic acid and a recent study suggests that children with absence epilepsy being treated with these medications experience improved attention, visual memory, and fine motor skills (Siren et al. 2007). Whereas many patients with childhood absence epilepsy will eventually enter remission, those with juvenile absence epilepsy will generally require lifelong treatment. Some suggest remission rates in approximately 80% of childhood absence epilepsy cases. However, a number of patients, 44% in one study, with unremitting childhood absence epilepsy will develop juvenile myoclonic epilepsy (Wirrell et al. 1996).
Despite remission, however, the psychosocial outcome in children with childhood absence epilepsy may be poor. One study followed such children into adulthood and compared them to a group of controls with a history of juvenile rheumatoid arthritis. In comparison to the arthritis group, those with a history of absence epilepsy were more likely to drop out of high school, have unplanned pregnancies, and abuse substances (Wirrell et al. 1997).
In 2001, a sample of children with childhood absence epilepsy was studied neuropsychologically (Pavone et al. 2001). The sample included six boys and ten girls that ranged in age from 6 to 16 years, with a median age of onset of 5.3 years. These children were compared to a control group matched for gender and had comparable age and socioeconomic status (Pavone et al. 2001). To assess cognitive function, patients and controls were administered the Italian version of the Wechsler Intelligence Scale for Children—Revised, as well as the Raven’s Progressive Matrices, Rey Complex Figure, and the Test of Memory and Learning. The results showed that 81% of the patient group had IQ scores in the average range, with three patients having scores in the borderline range. Overall, however, patients had lower full scale IQ’s than did the matched controls (Pavone et al. 2001). To assess more specific neuropsychological domains, composite indices were created by calculating norm-referenced scores for each task, and percentile ranks were averaged for each domain. Domains included general cognition, language, visual spatial skills, and memory. Overall, the patients showed a lower level of general cognition, visual spatial skills, nonverbal memory, and delayed recall. The authors concluded that childhood absence epilepsy patients show slight, but statistically significant, deficits in global cognitive function, visual spatial functions, and visual memory. Verbal skills and verbal memory were less affected (Pavone et al. 2001). The finding that general intellectual functioning is lowered in children with generalized idiopathic epilepsy syndromes is consistent with the results of later work (Nolan et al. 2003). However, the latter study considered children with generalized idiopathic epilepsy as a group and did not consider childhood absence epilepsy specifically. Thus, direct comparisons between these studies are difficult.
Likewise, Jambaque et al. (1993) studied a large group of children with various seizure types. These included 18 patients with generalized idiopathic epilepsies, 12 of whom had childhood absence epilepsy. As with the results reported above, the group of children with generalized idiopathic epilepsy performed poorly relative to controls on tasks of visual memory, whereas verbal memory was intact (Jambaque et al. 1993). Again, however, this study did not consider children with childhood absence epilepsy alone; the generalized idiopathic epilepsy group also contained three children with juvenile absence epilepsy and three children with generalized tonic-clonic seizures on awakening.
Nolan et al. (2004) studied memory function in children with mixed seizure syndromes, including childhood absence epilepsy, FLE, and TLE. According to the authors, this was the first prospective study of memory functioning in children with epilepsy where cases were strictly classified according to the criteria of the ILAE (Nolan et al. 2004). The mean age of onset for the group of 13 with childhood absence epilepsy was 5.5 years, and they averaged 9.5 years at assessment. To assess memory, the children were administered five subtests from the Wide Range Assessment of Memory and Learning (WRAML), as well as the Rey Complex Figure. Though children with absence epilepsy showed the least degree of memory impairment in comparison to the other seizure types, they did show impairment relative to normative data. Specific weaknesses were seen on the finger windows subtest of the WRAML, a test requiring sustained visual attention and working memory, as well as on the Rey Complex Figure, which suggests fairly specific visual-spatial skill deficits consistent with prior research.
Another investigation studied children with absence epilepsy whose onset occurred prior to the age of 3 years (Chaix et al. 2003). Ten cases were identified, including seven girls and three boys: five cases presented with simple absences, and five cases with complex absences (i.e., with automatisms or eyelid myoclonia, or myoclonic jerks). The authors concluded that the overall prognosis for children with early onset absence epilepsy is relatively poor in comparison to what is seen in childhood absence epilepsy more generally. However, this investigation suffers from limitations that make firm conclusions difficult. First and foremost, the assessment was not standardized across participants, with different measures used in different cases. Moreover, two cases received no formal testing. Equally important is that global IQ measures were used, rather than tests of specific neuropsychological functions; tests employed may not have had sufficient sensitivity to detect subtle deficits.
Juvenile Myclonic Epilepsy
Juvenile myoclonic epilepsy, which has also been termed “impulsive petit mal,” generally begins between ages 12 and 18 and requires lifelong treatment. It is the most common primary generalized epilepsy syndrome in adolescence, accounting for between 4 and 10% of all epilepsies. This syndrome is often associated with myoclonic jerks of the neck, shoulders, and arms (Janz 1985). Most also experience generalized tonic-clonic seizures, and as many as 40% experience absences (Asconape and Penry 1984). Seizures are often brought on by sleep deprivation, stress, alcohol, or illicit drug use. Photosensitivity is common in this syndrome as well, as it occurs in about 30% of patients (Wolf and Goosses 1986). The characteristic EEG findings in this syndrome include high amplitude generalized symmetrical and synchronous 4–6 Hz polyspike-wave complexes. There is a genetic contribution to this syndrome and it has been mapped to chromosome 6p21.3 (Delgado-Escueta et al. 1989).
Despite the ubiquity of juvenile myoclonic epilepsy, the diagnosis is often missed. Several studies documented unusually long period between initial seizures and correct diagnosis, ranging from 8.3 to 15 years across studies (Grunewald et al. 1992; Panayiotopoulos et al. 1991; Vazquez et al. 1993). Delayed diagnosis has been attributed to factors including a patient’s failure to report myoclonic jerks to their physicians, or a physician’s failure to specifically inquire about myoclonic jerks.
As with other generalized idiopathic epilepsies of childhood, valproic acid, ethosuximide, and clonazepam are all used to treat juvenile myoclonic epilepsy (Browne et al. 1975; Browne 1976, 1978; Bruni et al. 1980; Lund and Trolle 1973; Sherard et al. 1980), though valproic acid is considered a better choice in patients with both absences and generalized tonic-clonic seizures, as seen in this syndrome (Benbadis 2005). Lamotrigine and topiramate have shown some success, but are not yet specifically approved for use in juvenile myoclonic epilepsy (Beran et al. 1998; Cross 2002; Frank et al. 1999). Interestingly, one study showed greater attention, short-term memory, and processing speed deficits in juvenile myoclonic epilepsy patients that were treated with topiramate versus valproic acid (de Araujo Filho et al. 2006).
Several studies have documented the cognitive and psychological difficulties in this syndrome. For example, Devinsky et al. (1997) evaluated 15 juvenile myoclonic epilepsy patients and compared the results of neuropsychological tests of executive functions to those obtained by a group of TLE patients matched for IQ. Variable performances were seen across juvenile myoclonic epilepsy patients, with some showing no cognitive difficulties and others showing significant impairment. Overall, high frequencies of impaired performances were seen on tests of concept-formation/abstract reasoning, speed of cognition, planning, and organization. Significant differences between patient groups were seen on tests of mental flexibility and concept formation (Devinsky et al. 1997). A later study compared the neuropsychological performances of 50 patients with juvenile myoclonic epilepsy to 50 age, education, and gender matched controls. This study showed more widespread deficits than earlier work. Patients had significantly poorer performances on tests of attention, inhibition, working memory, processing speed, and mental flexibility, in addition to deficits in verbal and visual memory, naming, and verbal fluency. Additionally, duration of epilepsy was associated with greater cognitive impairment, but not for patients with more education (Pascalicchio et al. 2007).
Trinka et al. (2006) administered the Structured Clinical Interview for DSM-IV Axis I (SCID-I) and the Structured Clinical Interview for DSM-IV Axis II (SCID-II) to 21 males and 22 females with juvenile myoclonic epilepsy. Thirty-five percent had a comorbid Axis I psychiatric disorder, most commonly with anxious features, and 23% had comorbid personality disorders. These numbers are slightly higher than community estimates, which show 27% incidence rate for Axis I disorders and 13.4% rate of personality disorders (Trinka et al. 2006). However, the rates of mood and personality disorders in juvenile myoclonic epilepsy appear to be lower than that seen in other seizure types, such as TLE (Perini et al. 1996).
To summarize, of the three generalized idiopathic epilepsy syndromes discussed here, childhood absence epilepsy tends to have the most favorable outcome. A fair number of these patients will experience remission after a period of time, but some may go on to experience other seizure types. From a cognitive perspective, children with childhood absence epilepsy may have subtle deficits in attention and visual memory skills. Limited information is available regarding the outcomes of juvenile absence epilepsy, but it is generally considered to be a less benign form of epilepsy. In juvenile myoclonic epilepsy, frontal deficits may predominate, but there is clear evidence of memory and language deficits in these individuals as well.
The focal epilepsies are characterized by seizures in which onset of the initial electrical activation occurs in a specific (i.e., focal) area of the brain. Based on their etiology, partial seizures are distinguished as (1) idiopathic; (2) symptomatic; or (3) cryptogenic (i.e., probably symptomatic). Partial seizures predominate in these patients. Overall, partial seizures account for approximately 40–60% of children with childhood epilepsy, and complex partial seizures appear to be more prevalent in children than simple partial seizures (Cowan 2002).
Seizure syndromes classified under the heading of focal idiopathic epilepsies of childhood, also termed the “benign” focal epilepsies, are among the most common epilepsies in children between ages three and 12. The two most commonly recognized and studied focal idiopathic epilepsies are benign epilepsy with centrotemporal spikes (also termed Benign Rolandic Epilepsy) and childhood epilepsy with occipital paroxysms. Focal symptomatic (or presumably symptomatic) epilepsy refers to epilepsies in which there is a known or suspected underlying lesion, although it should be noted that in many cases no underlying pathology is ever identified. TLE and FLE are the primary syndromes included under the classification of symptomatic or probably symptomatic epilepsy. Little information is available regarding incidence and prevalence rates specific to temporal and frontal lobe epilepsy, as these patients are often grouped together in epidemiological studies. The total prevalence rates for simple partial seizures, complex partial seizures, and secondary generalized seizures in children have been reported as 2–12, 8–31, and 7–29% respectively (Cowan 2002).
The treatment of partial seizures in childhood involves the use of first generation and recently introduced antiepileptic drugs, as well as nonpharmacologic options including the ketogenic diet, vagus nerve stimulation, and surgical intervention. First line antiepileptic therapies for partial seizures include carbamazepine and valproic acid, whereas phenobarbital and phenytoin are usually considered as last choice drugs because of their adverse event profiles; phenobarbital, topiramate, and zonisamide have been associated with the most cognitive side effects (Coppola 2004).
Benign Epilepsy of Childhood with Centrotemporal Spikes (Benign Rolandic Epilepsy)
Benign Rolandic Epilepsy is the most common epilepsy syndrome of childhood, representing about 15% of all childhood seizure disorders (Sidenvall et al. 1993). Onset is typically between ages 3 and 13 years, with a peak around ages 8–10. Males are more frequently affected. A genetic basis for BRE has been suggested by twin studies (Eeg-Olofsson et al. 1982), and it has been linked to chromosome 15q14 (Neubauer et al. 1998). Though there was an early suggestion of an autosomal inheritance pattern, later studies suggest a more multifactorial inheritance, and more recent work has shown that non-inherited factors are more important than once believed (Vadlamudi et al. 2006). The characteristic EEG findings involve high voltage centrotemporal spikes followed by slow waves. These tend to be activated by sleep and may shift from side to side. Seizures seen in rolandic epilepsy frequently affect the oropharyngeal muscles and are often described as brief, simple partial, hemifacial motor seizures, but they may evolve into generalized tonic-clonic seizures. Treatment with antiepileptic medications is not always warranted because the seizures tend to be infrequent, typically occur at night, and often spontaneously remit by the mid-teenage years. That being said, there are certain circumstances under which treatment is favored, including a young age of onset, a short period of time between the first three seizures, and daytime seizures (Bourgeois 2000). The typical treatment, when warranted, is valproic acid or carbamazepine, although it should be noted that carbamazepine may adversely affect verbal memory in children treated for BRE (Seidel and Mitchell 1999). Levitiracetam has also been approved for use in children with BRE (Bello-Espinosa and Roberts 2003).
Despite its status as a benign syndrome, there is increasing evidence to suggest that these children have neuropsychological impairments that may be overlooked. A 1999 study in Turkey suggested deficits in vocabulary, prosody, motor skills, and frontal functions. In this investigation, 20 children diagnosed with BRE were compared to 15 controls of comparable age and gender distribution. The patient group was more frequently impaired on tests of vocabulary, dyspraxia in the leg to imitation, dysprosody, response inhibition on a go/no-go task, and motor deficits of the upper extremities. Although this study suggested that deficits in children with BRE might indeed exist, as acknowledged by the authors, a major weakness of this study is the lack of standardized neuropsychological measures in Turkish (Gunduz et al. 1999).
Using a more comprehensive neuropsychological battery, Croona et al. studied 17 children with BRE and compared performances to age, gender, and estimated intelligence matched controls. Intellectual ability was assessed by Raven’s Progressive Matrices. A variety of tasks were used to examine attention span, verbal memory, visual memory, and executive functioning. Moreover, parents and teachers completed questionnaires regarding the children’s cognitive function and academic achievement abilities. Patients and controls performed equally well on numerous tasks, including digit span, block span, Trailmaking Test, the Rey Complex Figure delayed recall, and the Spatial Learning Task. However, the patient group performed more poorly than controls on tests of verbal fluency, verbal memory, and planning. Moreover, the parents of patients reported more difficulties with respect to distractibility, concentration, mood swings, impulsivity, understanding instructions, etc. (Croona et al. 1999).
Northcott et al. studied 16 girls and 26 boys with BRE (Northcott et al. 2005). The mean age at evaluation was 8.5 years, with seizure onset ranging from 3 to 11. The patients were administered a comprehensive battery of neuropsychological tasks, and one sample z-tests were used to compare performances to normative populations. EEG features were also assessed, including spike frequency, trains (i.e., runs of discharges of 5s or more), and laterality. Correlational analyses assessed the relations between EEG variables and cognitive functions (Northcott et al. 2005). The sample had higher than expected performances on tests of general intellectual functioning and general language. Further, academic achievement (i.e., basic reading, spelling, and mathematics) was not impaired relative to normative values. However, both verbal and visual memory indices were below expected levels, as were more specific language functions (e.g., phonological awareness). Thus, in contrast to prior work (Croona et al. 1999), this study showed more generalized memory deficits, possibly due to improved statistical power. EEG features were minimally associated with cognitive variables, but there was no relation between spike burden and laterality (Northcott et al. 2005). Interestingly, a subset of this sample was followed longitudinally. Twenty-eight participants underwent re-evaluation, and differences were evaluated via t-tests. Improvement was observed in verbal memory and receptive language, whereas visual memory performance and phonological awareness did not improve (Northcott et al. 2006).
Thus, despite being considered an epilepsy syndrome with a benign course, individuals with BRE may present with clear neuropsychological deficits despite overall intact intelligence. These include deficits in aspects of language and memory, which is not surprising given the characteristic epileptiform discharges in the centrotemporal regions. There may be some degree of motor and executive function impairment in this group as well.
Childhood Epilepsy with Occipital Paroxysms
Occipital lobe epilepsy accounts for 6–8% of individuals with focal epilepsy (Manford et al. 1992) and is more common in children than in adults (Sveinbjornsdottir and Duncan 1993). There are two types of occipital lobe epilepsy currently recognized by the ILAE, including early onset benign childhood occipital lobe epilepsy (Panayiotopoulos type) and late onset childhood occipital lobe epilepsy (Gastaut type; Engel 2006). The former generally has a more favorable outcome. Though there are few studies on this syndrome, several neuropsychological studies have found cognitive impairments in children with benign childhood occipital epilepsy (Chilosi et al. 2006; Germano et al. 2005; Gulgonen et al. 2000).
In one study, children with occipital lobe epilepsy performed significantly lower than controls (matched for age, sex, and socioeconomic status) on measures of intellectual functioning, particularly with respect to performance IQ (Gulgonen et al. 2000); however, another study found verbal IQ to be differentially impacted relative to controls, with no difference found between groups on performance IQ (Germano et al. 2005). In both studies, the children with epilepsy showed significantly reduced performances across a wide variety of cognitive tasks, including attention, memory, visuospatial skills, language skills, and motor skills, and in the Gulgonen et al. (2000) study, group differences remained significant (or reached significance) when intellectual functioning was controlled for. Gulgonen et al. (2000) did not find differences between groups with respect to academic achievement, visuomotor skills, or executive functioning, whereas Germano et al. (2005) found reduced reading, writing, and calculation abilities in their epilepsy group; in the latter study, performances on several lexical tasks were associated with visual memory and graphomotor skills.
Temporal Lobe Epilepsy
Temporal Lobe Epilepsy is among the most common types of epilepsy in both adults and children, and the associated cognitive profile is undoubtedly the most studied of the epilepsy syndromes (Deonna et al. 1986). Complex partial seizures are the most common type of seizure associated with TLE, although simple partial seizures do occur, and complex partial seizures often lead to secondarily generalization (Cowan 2002). Auras often precede temporal lobe seizures and may involve somatosensory symptoms, including epigastric rising sensation, olfactory symptoms, and emotional symptoms (e.g., feeling of fear); many patients report experiencing a “funny feeling” prior to the onset of the seizure. The actual complex partial seizure generally consists of behavioral arrest, unresponsiveness and staring, stereotyped automatisms, and tonic motor phenomena (Bourgeois 1998). Auras and seizures can present very differently in infants and young children.
Temporal Lobe Epilepsy in children usually begins during the school years, and when temporal lobe seizures are seen in young children there is often an underlying lesion (Jambaque 2001). Onset in childhood is generally associated with a poorer cognitive outcome than in adult onset cases. Further, these children are at a greater risk of developing learning disabilities, which may have implications for success later in life (Jambaque et al. 1993; Williams et al. 1996). Lower age of onset has been associated with lower performance IQ, and patients with later onset performed better when copying the Rey Complex Figure (Jambaque et al. 2007). Persistent seizures in childhood might slow down the rate of cognitive and psychological development (Oguni et al. 2000). Children with longer duration of epilepsy have been shown to have lower IQs than children with a shorter seizure history (Robinson et al. 2000), and long-term effects of intractable seizures may be more deleterious for children than adults (Bjornaes et al. 2001). Not surprisingly, the most prominent deficits pertain to memory, which may be seen in the context of average intellectual functioning. However, impairments are not always limited to memory, and several factors contribute to the progression of broader deficits, including age at seizure onset and level of seizure control (Hermann et al. 2002).
In a cross-sectional investigation of patients with TLE (early versus late onset) and controls, the early onset group (onset before age 14) performed more poorly than the late onset group and controls on all measures of intelligence and memory (Hermann et al. 2002). There were fewer differences between the late onset group and controls, despite the fact that this group had epilepsy for an average of 16 years. Further, neuroimaging showed that the early onset group had smaller total cerebral tissue volume and smaller hippocampal volume relative to late onset patients and controls. Within the early onset patients, duration of epilepsy was directly related to cognition when age of onset, gender, and education were covaried. The authors concluded that onset of TLE in childhood adversely affects the development of brain structures and function, which often extends outside the primary epileptogenic region. Childhood onset TLE has also been associated with reduction in corpus callosum volume when compared to both late onset patients and healthy controls (Hermann et al. 2003). Smaller corpus callosum volume was linked to poorer performances on nonverbal problem solving tasks, immediate memory, speeded complex motor ability, and fine motor dexterity.
Whereas several adult studies have shown dissociation between verbal and visual memory impairment as a function of side of focus, child studies have been less consistent. Several child studies have found either no relationship between the side of seizure focus and material specific memory performances (Camfield et al. 1984; Lendt et al. 1999), equivalent deficits in verbal memory performance regardless of seizure side (Igarashi et al. 1995), or lower verbal memory performances in left-TLE patients in the absence of concomitant visual memory impairments in right-TLE patients (Clusmann et al. 2004). However, some have indeed shown modality specific impairments (Cohen 1992; Jambaque et al. 1993, 2007). Beardsworth and Zaidel found children with right-TLE to be differentially impaired in memory for faces relative to left-TLE patients and controls (1994). In a study that used proton magnetic resonance spectroscopy in children with intractable TLE, patients with left temporal pathology showed more impaired verbal learning than patients with right-sided pathology (Gadian et al. 1996); the authors did not administer a visual memory task. This investigation also showed poorer general verbal abilities associated with left temporal pathology, while right-sided pathology was associated with a loss of nonverbal cognitive functions. Forty-five percent had bilateral pathology; the authors noted that without sophisticated imaging, undetected pathology could account for contradictions in the literature.
In contrast, an earlier study did not find differences between cognitive profiles of children with “pure right” versus “pure left” temporal lobe seizure foci, and few participants showed evidence of cognitive dysfunction at all (Camfield et al. 1984). However, the majority had excellent seizure control, and very few had severe protracted TLE. Further, they did not account for handedness or language lateralization. Subsequent studies have suggested a moderating effect of language lateralization to overall intellectual functioning. Children with atypical language dominance may perform more poorly on both verbal and nonverbal IQ subtests (Billingsley and Smith 2000). Gleissner et al. (2003) compared neuropsychological performances of children with left-sided focal epilepsy and either atypical language representation (i.e., right or bilateral) or left-hemisphere language representation patients matched for age. Among the atypical language group, there were more patients with atypical hand dominance, extratemporal lesions, and earlier onset of epilepsy. Consistent with a “crowding effect,” patients with atypical language dominance demonstrated significantly lower performance in visual memory, yet there was a trend in this group to show better verbal memory performance than seen in the typical language dominance group. These findings highlight the importance of considering the possibility of language shift in presurgical neuropsychological workups in patients with early left-hemisphere damage. In cases where atypical language dominance is established, an extratemporal focus should be considered. In the absence of data on language lateralization, handedness should be considered when interpreting neuropsychological profiles of these patients.
Attention impairments are widespread in children with epilepsy, regardless of seizure focus, and such deficits may be a leading cause of learning difficulties in these patients (Dunn and Kronenberger 2005; Schubert 2005). Though improvements in attention are often seen after temporal lobe surgery (Clusmann et al. 2004; Jambaque et al. 2007; Lendt et al. 1999), children who undergo resective surgery may not improve to levels seen in normally developing children (Billingsley et al. 2000). With respect to treatment, evidence clearly indicates that stimulants are safe and effective in treating the attention deficits seen in children with epilepsy (Dunn and Kronenberger 2005).
Executive function impairments have also been seen in children with TLE that were administered a modified version of the Wisconsin Card Sorting Test (WCST; Igarashi et al. 2002). Children with TLE and hippocampal atrophy performed more poorly (i.e., fewer categories achieved and more perseverative errors) than those without visible hippocampal damage, though impairment was less severe than that seen in children with FLE. Interestingly, seizure onset prior to 40 months of age was associated with better performance, which was attributed to greater brain plasticity in younger children.
Language deficits, such as deficits in naming and vocabulary, have been seen in children with left-TLE (Jambaque et al. 1993), and earlier onset may be associated with developmental language problems (Jambaque 2001). Jambaque and colleagues found post-operative naming to be higher in children with later epilepsy onset, despite the fact that post-operative gains in Verbal IQ were associated with earlier age of onset, a finding that that may point to a specific risk to the development of semantic knowledge in younger patients (Jambaque et al. 2007). Reading development may be problematic in children with TLE (Williams et al. 1996), and these patients show greater reading difficulties than do children with idiopathic generalized epilepsy.
The role of the temporal lobes (especially the right temporal lobe) in social and emotional perception is well established in the literature, but few studies have examined this in children with TLE. One study showed that right-TLE patients with early-onset seizures (before age five) were selectively impaired in recognizing fearful faces when compared to right-TLE patients with late-onset seizures and controls. Those with early onset seizures were also impaired in their ability to recognize expressions of sadness and disgust relative to controls, whereas there was not a significant difference between controls and the late-onset seizure group (Meletti et al. 2003). In another study that compared the ability to perceive emotional gesturing and prosody in children with right- and left-TLE, only the right-TLE group was significantly impaired relative to healthy controls (Cohen et al. 1990). There was no significant difference between the left- and right-TLE groups, or between the left-TLE group and controls.
Surgery is recognized as a safe treatment option for adults with intractable TLE, but neurologists tend to be more conservative in considering this in children, despite mounting evidence of positive outcomes in children who received surgery after poor management with medication. Patients that received resective surgery as a child have been shown to have better long-term socioeconomic outcome and quality of life (Keene et al. 1998a), as well as overall satisfaction with the surgery (Keene et al. 1998b), when seizures were reduced by 50% or more (Engel Class I–III). Further, studies have shown that cognitive functioning tends to remain stable or improve after surgery, especially when it is undertaken early in the disease course (Jambaque et al. 2007). The risk for decline in intelligence level after temporal lobe surgery is low (Kuehn et al. 2002; Westerveld et al. 2000; Williams et al. 1998), but there is evidence that higher preoperative functioning is a risk factor for postoperative decline (Szabo et al. 1998). Risk factors for decline may also include older age at time of surgery and the presence of a structural lesion other than mesial temporal sclerosis on imaging.
Children who underwent left temporal lobectomy not only maintained presurgical verbal intellectual functioning but also were found to improve significantly in nonverbal intellectual functioning (Jambaque et al. 2007; Westerveld et al. 2000). Children who have undergone right temporal lobectomies typically do not show significant changes in intellectual functioning (Jambaque et al. 2007; Westerveld et al. 2000). In comparing adults and children who have undergone temporal lobectomies in the context of similar pathology, adults tend to show less recovery of memory skills (Gleissner et al. 2005). Generally speaking, post-operative declines may be seen at shorter retest intervals, but improvement is generally observed at longer follow-up (Adams et al. 1990; Szabo et al. 1998; Williams et al. 1998). In an investigation of children and adolescents who received a selective surgical procedure that spares the lateral temporal cortex (transparahippocampal selective amygdalohippocampectomy), no changes in cognitive functioning were seen post-operatively, with the exception of significant improvement in rote verbal memory scores among patients who underwent right-sided surgery (Robinson et al. 2000). In this study, high premorbid functioning was not associated with post-operative decline, as has been seen after anterior temporal lobectomy (Chelune et al. 1998; Hermann et al. 1995; Szabo et al. 1998); therefore this may be a better surgical option for children, particularly in children who show strong verbal memory skills prior to surgery.
The intracarotid amobarbital procedure (Wada test), which can establish language dominance and lateralized memory functions, has shown utility in predicting post-surgical memory improvement in children. In one study, verbal memory was improved for children who showed asymmetries in the predicted direction (i.e., memory after the injection on the side ipsilateral to surgery better than memory after contralateral injection), whereas children who did not show such memory asymmetries showed a significant decline in verbal memory post-surgically (Lee et al. 2005).
To summarize, there are both similarities and differences between the profiles of children and adults with TLE. While onset later in life is linked primarily with memory deficits, an earlier onset appears to have a more widespread impact on brain structure and function. With respect to memory, although several adult studies have found modality specific memory deficits depending on the side of pathology, results from child studies are far from conclusive in this regard, perhaps due to the greater plasticity in children. Language lateralization further complicates this issue. Regarding treatment of TLE in childhood, the preponderance of evidence suggests that not only is surgery a safe and effective treatment for reducing seizures in children, but it also appears to be associated with positive cognitive outcomes. However, high pre-operative verbal memory skills have been associated with greater post-operative declines, though it is worth noting that the effect of prolonged seizures may ultimately be more devastating in the long run. A better outcome can generally be expected when surgery is undertaken early in the disease course. Neuropsychologists consulted during pre-surgical workups should be aware of issues contributing to post-operative outcomes.
Frontal Lobe Epilepsy
Frontal lobe epilepsy accounts for approximately 30% of partial epilepsy, and it represents the largest subgroup of extratemporal lobe epilepsy. Frontal lobe seizures often have a bizarre clinical presentation, in the context of little or no interictal and ictal EEG abnormalities, and they are often mistaken for nonepileptic seizures (Arunkumar et al. 2001). Most neuropsychological studies of FLE in children have been single case studies (Hernandez et al. 2001). In a study that compared neuropsychological performances of children with symptomatic TLE and FLE, a differential impairment in concept formation on a modified (i.e., simpler) WCST was seen in the latter (Igarashi et al. 1995), a finding that was confirmed in a subsequent study (Igarashi et al. 2002). However, a study that used the original version of the WCST did not find children with asymptomatic FLE to be impaired relative to controls with respect to number of categories completed or in the number of perseverative responses (Riva et al. 2005), despite deficits on other executive measures, including phonemic (but not semantic) fluency and design fluency. Mean Full Scale IQ was average in the latter study, while lower IQ scores were reported in the Igarashi studies, and this may account for the discrepant findings. Riva et al. administered the California Verbal Learning Test in addition to traditional measures of executive functioning, and though this sample was not impaired relative to controls, age at seizure onset and duration of epilepsy were related to the total number of words recalled, use of semantic clustering strategies, short delay free recall, and long delay free and cued recall. In another study that compared children with idiopathic FLE and TLE on measures of attention, executive functioning, memory (verbal and nonverbal), and adaptive functioning, FLE patients showed greater deficits in planning and executive functions in the context of intact memory performances; the opposite pattern was seen in the TLE group (Culhane-Shelburne et al. 2002). Furthermore, 50% of the variance in the Vineland Adaptive Behavior Composite was accounted for by performances on two measures of executive functioning (Stroop—Color/Word and Tower of London—Rule Violations).
Profiles of children with FLE, TLE, and generalized epilepsy (typical absence) were compared on tests of motor ability and executive functioning (Hernandez et al. 2002). Groups were matched for age, IQ, age at seizure onset, and duration of epilepsy. The groups did not differ with respect to IQ (full scale, verbal, or performance), but children with FLE showed deficits on tasks of motor coordination, response generation, impulse control, and planning ability relative to the other seizure groups. With regard to motor skills, deficits were particularly apparent during tasks involving bimanual coordination and asymmetrical movements, with younger children (ages 8–12) showing more pronounced deficits. In a related study on the same cohort of children, performances on measures of attention, processing speed, and memory (verbal and nonverbal), as well as parent ratings of behavior were compared (Hernandez et al. 2003). All three groups were impaired relative to controls on attention and memory tasks, but attention problems were more prominent in the FLE group, particularly with respect to sustained attention and complex working memory on a continuous performance test. This group also displayed differential impairments in perceptual organization and self-regulation of behavior across several tasks. On a list-learning task, more intrusion errors were made and these patients were more prone to interference than the other patient groups, despite comparable overall learning and recall between groups; all three groups were deficient relative to healthy controls. The FLE and TLE groups both had difficulty reproducing a complex figure relative the generalized epilepsy group and controls. Recall difficulties in the FLE group were due to organizational difficulties when copying the design, whereas the TLE group’s difficulties appeared to be due to impaired encoding. Thought problems, behavior problems, and social adjustment problems were reported more frequently in children with FLE than in the other two groups on behavior rating scales. The authors noted that the number of antiepileptic drugs taken by children did not appear to affect attention, memory, and behavior significantly.
Another group examined performances of children with TLE, FLE, and controls on a measure of preparatory attention (Auclair et al. 2005). They found that the FLE group showed a higher mean slope of response time to a target as a function of distractor probability (i.e., a distractor was occasionally presented prior to the target) compared to children with TLE and control subjects. Children with TLE did not differ from controls in their response time slope. A high incidence of attention deficit and hyperactivity, learning difficulties, and behavioral problems has been observed in children with FLE, even when seizures are well-controlled (Prevost et al. 2006).
Cohen and Le Normand (1998) conducted yearly evaluations of receptive and expressive language skills in a small sample of young children with left frontal simple partial seizures and compared results to controls. Analyses of individual language trajectories revealed a clear dissociation in linguistic performance between comprehension and production. Linguistic comprehension gradually improved to reach normal performance levels by age seven, whereas production, even at later stages, remained quite poor (Cohen and Le Normand 1998). Children with FLE exhibited significant deficits on phonological processing tasks when compared to children with TLE and generalized absence seizures (Vanasse et al. 2005).
Surgical intervention is also a viable option for patients with FLE, though the likelihood of becoming seizure-free post-surgically is lower in this group in comparison to temporal lobe surgery patients, and there is increased risk of damage to eloquent cortex. Lendt and colleagues investigated the neuropsychological function of children with FLE before and 1 year after surgery (Lendt et al. 2002), and compared their performance to children who underwent temporal lobe surgery for seizures. The FLE group demonstrated higher intellectual functioning prior to surgery but were more impaired in manual motor coordination than the TLE group. Post-surgically, both groups improved in attention, processing speed, short- and long-term memory, and manual motor coordination (although the latter was not significant). There was no improvement in executive functions in either group. Post-operative improvements did not depend on complete seizure relief after surgery. Age at surgery and side of resection did not moderate the effects that were seen. Only two patients showed a decline in language, executive, or motor functions; no factors could be identified that accounted for these changes. Overall, outcome was favorable in this group, but the small sample size and associated poor statistical power may explain the lack of group differences in the cognitive outcome of FLE and TLE patients. The authors point out an interesting caveat to these results: future deficits may emerge as the frontal lobes continue to develop into adolescence and adulthood (a phenomenon that has been reported after frontal lesions), and they underscored the need for longitudinal studies that utilize longer retest intervals.
In summary, FLE in children appears to be associated primarily with deficits in executive functioning, attention, and motor skills, often in the context of normal intellectual function. Executive deficits are believed to be the primary cause of memory impairments in this group. Further, social and behavioral problems are common. While surgery has been shown to improve attention, processing speed, and memory in children with FLE, the benefit to overall executive functioning and/or behavior is less clear. However, few studies have examined this issue and results are limited by small sample sizes.
Epileptic encephalopathies are a group of syndromes characterized by deterioration of sensory functions, motor functions, and/or cognition as a result of epileptic activity, including frequent seizures and/or prominent interictal paroxysmal activity (Nabbout and Dulac 2003). In most, there is either a regression of cognitive development or a failure to attain developmental milestones. The long-term cognitive and behavioral outcome is often poor. Despite commonalities between syndromes, studies have attempted to differentiate the cognitive profiles of disorders; some have distinct patterns of strengths and weaknesses. Several syndromes present with such profound developmental delay that they are rarely seen by pediatric neuropsychologists for assessment, and neuropsychological studies are virtually absent. Only those syndromes for which neuropsychological studies were available will be discussed below.
Dravet’s Syndrome (Severe Myoclonic Epilepsy of Infants)
The incidence of Dravet’s syndrome is estimated to be one in 40,000 children under the age of seven (Hurst 1990). Onset occurs within the first year of life in normally developing children. A family history of epilepsy or febrile convulsions have been noted in up to 64% of cases (Ohtsuka et al. 1991) and recent investigations have found mutations in the SCNA1 gene in approximately 35% of children with Dravet’s syndrome (Oguni et al. 2005). There is molecular evidence to suggest Dravet’s syndrome may be on the spectrum of generalized epilepsy with febrile seizures “plus,” along with myoclonic astatic epilepsy (discussed below; Scheffer et al. 2001). Dravet’s syndrome is characterized by frequent prolonged seizures that often lead to recurrent status epilepticus. Seizure focus tends to switch sides from one event to the next, and they are refractory to antiepileptic medications (Wirrell et al. 2005). Myoclonic seizures appear between 1 and 4 years and are often accompanied by absence and focal seizures. A recent study that assessed the intellectual functioning of 53 patients with Dravet’s syndrome found mild developmental delay in 18 children (34%), moderate delay in 22 children (41.5%), and severe delay in 14 children (26%; Caraballo and Fejerman 2006). Hyperactivity was observed in the majority (85%). Interpersonal relationships rarely exceed the developmental level of 2 years of age, and these patients sometimes exhibit autistic traits (Casse-Perot et al. 2001).
Wolff et al. (2006) conducted a prospective longitudinal neuropsychological study in 14 patients with Dravet’s syndrome in order to elucidate the nature of the cognitive and behavioral profiles associated with this syndrome. The authors administered a battery of 50 neuropsychological tests, and scores were used to calculate a global developmental quotient, as well as a quotient for each domain. Before the onset of seizures, psychomotor development and behavior were reportedly normal in all of the children. The age at first assessment ranged from 11 months to 12 years. Four children underwent neuropsychological testing in the first 2 years of life, and their global developmental quotients were found to be only “slightly deficient” at that time. Longitudinal data indicated that after the age of four, the developmental quotient tended to decrease until age six, after which it remained relatively stable at a low level. Severe mental retardation, hyperactivity, poor relational capacity, and gestural stereotypies were evident in most children even before the age of six. Subtest results revealed poor visuomotor skills in all children over 4 years. Language results were more heterogeneous, yet language skills were better than visuospatial skills in only three children; these three patients showed a more favorable outcome overall. A trend was noted whereby children with fewer convulsive seizures (<5 per month) performed better than those with higher seizure frequency.
Also known as infantile spasms, West syndrome is diagnosed in children who exhibit the triad of infantile spasms, psychomotor deterioration, and hypsarrhythmia. Onset of seizures generally occurs between three and 12 months of age. Causative factors include neurological insult (e.g., infection or hypoxia-ischemia), cortical malformations, neurocutaneous syndrome (e.g., tuberous sclerosis complex and Sturge–Weber syndrome), chromosomal or single gene disorder, or an inborn error of metabolism (Wirrell et al. 2005). The incidence rates of West syndrome have been estimated at 2–4.5 per 10,000 live births per year, while the prevalence rates for children 10 years of age or older are about 1.5–2 per 10,000 (Cowan 2002). Approximately 50–70% of children with infantile spasms will develop other seizure types (Cowan 2002), with Lennox–Gastaut syndrome comprising approximately 20–50% of these patients (Appleton 2001). At least a third of the cases of West syndrome are idiopathic (Nabbout and Dulac 2003), and these patients generally show a more favorable outcome (Guzzetta 2006).
Early signs of cognitive impairment include deterioration of responsiveness and sensory abilities, as well as poor social contact, which is often the reason that parents seek medical attention (Guzzetta 2006). Impaired eye-hand coordination was found to be deficient when infants were evaluated, even in idiopathic patients who appear to be functioning normally otherwise (Rando et al. 2005). Impaired visual functioning involving acuity, ocular motility, visual field, visual attention, and visual scanning skills have been found in several studies with infants (Guzzetta et al. 2002; Jambaque et al. 1993; Rando et al. 2004). Guzzetta (2006) reported an associated deficit in auditory attention, and visual and auditory functions were both correlated with cognitive competence on the Griffith Scales of Infant Development. In general, long-term cognitive outcome appears to depend on the underlying causes (e.g., lesion characteristics); however, about 80% of individuals with the syndrome present with mental retardation (Guzzetta 2006; Koo et al. 1993; Matsumoto et al. 1981). Patients with the idiopathic form of the disease whose spasms had completely resolved within the first year were evaluated immediately after spasms ceased and subsequently underwent two follow-up evaluations (4–6 years later and 6–8 years later; Gaily et al. 1999). Deficits were seen in attention, learning, and memory, even in cases where intellectual functioning was normal. Perception and fine motor skills at the first assessment (8–15 months after cessation of infantile spasms) strongly predicted later cognitive functioning. In an earlier study that also focused on the idiopathic form, preserved visual tracking during infancy was associated with a more favorable outcome (Dulac et al. 1993). Results from a longitudinal study that followed 214 patients with West Syndrome for 20–35 years (or until death) showed normal or only slightly impaired intellectual outcome in a quarter of the patients (Riikonen 2001), although they too noted that specific cognitive deficits were seen in some patients with normal intelligence. Interestingly, in contrast to what is generally reported in the literature, only a third of these patients were found to have the idiopathic form of the syndrome, but this difference might represent diagnostic errors. In most cases there was no change in intellectual level as compared to earlier evaluations at a mean age of 8 years, and those that were ultimately placed in lower IQ categories all had severe drug-resistant epilepsy. Behavioral disorders, often accompanied by autistic symptomology, have been noted as a potential outcome in approximately 15% of patients with West syndrome, although the incidence rate increases to 70% in patients with tuberous sclerosis complex (Guzzetta 2006).
Lennox–Gastaut syndrome is very rare, with an estimated incidence rate of 1–2 per 100,000 in children from birth to 14 years of age; the estimated prevalence rate is much higher, at 1.3–2.6 per 10,000 (Cowan 2002). The peak age of onset of seizures ranges from 2 to 8 years of age (Wirrell et al. 2005), and approximately 20–60% of children with Lennox–Gastaut syndrome have a history of infantile spasms. The syndrome usually results from focal, multifocal, or diffuse brain damage, although idiopathic cases are reported and usually present with a later age of onset (Nabbout and Dulac 2003). Children with Lennox–Gastaut syndrome almost always present with both tonic and akinetic seizures, and slow generalized spike-and-wave discharges on EEG. The outcome for these children is generally poor, with an estimated 91% exhibiting mental retardation, often times profound (Cowan 2002). Further, there is strong evidence to suggest that intellectual functioning continues to deteriorate with age (Oguni et al. 1996), and one study reported that of patients who had normal intelligence at baseline, only 26% of these patients (33% of idiopathic, 17% indeterminate, and 20% of symptomatic) had normal or borderline normal intelligence at long-term follow-up (Goldsmith et al. 2000). Additionally, many patients with Lennox–Gastaut syndrome develop psychiatric and behavioral problems over time (Besag 2006), and long-term follow-up of patients has revealed perseverative behavior, psychomotor slowing, and apathy (Kieffer-Renaux et al. 2001). Poorer outcomes are predicted when onset occurs prior to 3 years of age and there is a prior history of West syndrome (Wirrell et al. 2005). One retrospective study found that age of onset was a better predictor of outcome than whether or not the disorder was idiopathic in nature (Goldsmith et al. 2000); however, late onset patients rarely had a history of infantile spasms, and normal or borderline normal outcomes were seen in a greater proportion of patients with idiopathic Lennox–Gastaut syndrome than in symptomatic or indeterminate forms. This effect disappeared when patients with mental retardation at the onset were excluded from analyses. Vagus nerve stimulation has been successful in improving seizure frequency in some patients and may also lead to slight improvements in cognition, behavior, and mood. However, there is insufficient evidence to determine whether such gains are maintained over time (Majoie et al. 2001, 2005).
Myoclonic Astatic Epilepsy
Myoclonic astatic epilepsy is a syndrome that begins between the ages of 2 and 5 years in otherwise normally developing children (Nabbout and Dulac 2003). Genetic factors are believed to play a role in its development, and there is molecular evidence to suggest that it may be on the spectrum of generalized epilepsy with febrile seizures “plus,” along with Dravet syndrome (Scheffer et al. 2001). Seizures begin as generalized tonic-clonic, but are eventually followed by myoclonic astatic seizures of increasing frequency. Neuropsychological outcome in children with myoclonic astatic epilepsy is quite variable and depends on the course of the seizure disorder (Kaminska et al. 1999; Oguni et al. 2002). Patients who present with developmental delay are sometimes misdiagnosed, because they share features of patients with idiopathic Lennox–Gastaut syndrome. However, several studies have argued that these are indeed two separate entities. Kaminska et al. (1999) examined the clinical profiles and outcome of 72 children with idiopathic generalized epilepsy and found that they were able to classify them into three distinct groups: (1) myoclonic astatic epilepsy group with favorable outcome; (2) myoclonic astatic epilepsy group with unfavorable outcome; and (3) Lennox–Gastaut syndrome group. The myoclonic astatic epilepsy groups both presented initially with similar cognitive abilities (mean IQ of 85 during the first 2 years). However, the second group showed a deterioration in cognitive functioning (IQ <50 in 83% of patients), apparently due to persisting seizures that were distinct from the pattern seen in the first group. In contrast, seizures remitted entirely in the first group and cognitive functioning was found to be only mildly impacted at least 1 year after the last myoclonic event. Unlike either of the myoclonic astatic epilepsy groups, the Lennox–Gastaut group presented with a different seizure profile and had mental retardation from the outset. Other studies have failed to find a specific neuropsychological pattern associated with myoclonic astatic epilepsy but suggest that cognitive and behavioral functioning are primarily affected by the duration and frequency of epileptiform activity (Filippini et al. 2006; Stephani 2006). Filippini and colleagues described a representative patient (from a group of seven) with the transitory form of myoclonic astatic epilepsy who showed cognitive and behavioral disturbances that remitted entirely with successful pharmacological treatment of seizures.
Continuous Spike-and-waves During Slow-wave Sleep/Landau–Kleffner Syndrome
Continuous spike-and-waves during slow-wave sleep accounts for less than 1% of childhood onset epilepsies (Kramer et al. 1998), yet it has received considerable attention in the literature due to its dramatic effects on cognition and behavior. Peak age of onset is between 5 and 7 years, and there is a slight male preponderance (McVicar and Shinnar 2004). Although most patients are reported to have normal development prior to onset of symptoms, pre-existing neurological abnormalities are reported in approximately one third of patients. Patients with continuous spike-and-waves during slow-wave sleep often have partial and generalized seizures at the onset of the epilepsy syndrome, yet the onset of cognitive deterioration leads to a more comprehensive workup that reveals generalized spike-wave discharges that occupy >85% of slow-wave sleep, which is referred to as electrical status epilepticus in slow-wave sleep (Camfield and Camfield 2002). Global intellectual functioning, language, temporo-spatial disorientation, motor function, and behavior are all affected (Tassinari et al. 2000). In their review of literature on electrical status epilepticus in slow-wave sleep, Tassinari et al. (2000) hypothesized that long-term and stable cognitive impairment is the direct consequence of the status of continuous spike-and-waves during sleep. They cite several factors that are supported in the literature, including (1) a close temporal association between the presence of status epilepticus during sleep and neuropsychological regression; (2) a parallel between duration of electrical status epilepticus in slow-wave sleep and the final neuropsychological outcome; and (3) the strict association between the pattern of neuropsychological deficits and the location of the interictal focus. Seizures are not easily controlled with medication, although most patients’ seizures remit spontaneously in adolescence (McVicar and Shinnar 2004; Scholtes et al. 2005).
The most well studied disorder associated with continuous spike-and-waves during slow-wave sleep is Landau–Kleffner syndrome, which is characterized by a dramatic regression of language functions in children who have already developed normal speech, and relative sparing of other cognitive functions (Landau and Kleffner 1957; Rotenberg and Pearl 2003). The peak onset is between 3 and 8 years of age, while language regression in children under 2 or 3 years usually occurs in the context of a more global autistic regression (McVicar and Shinnar 2004). Clinical seizures are present in approximately 80% of children with Landau–Kleffner syndrome, although they are not necessary to make the diagnosis; in addition to electrical status epilepticus in slow-wave sleep, these patients often show prominent epileptiform activity in the temporal lobes, usually bilaterally (Deonna 1991; McVicar and Shinnar 2004). The incidence of Landau–Kleffner syndrome is unclear, but it is thought to be quite rare. Its pathology also remains unclear; however, one recent imaging study that compared the volumes of superior temporal areas in a small sample of these patients without specific anatomic abnormalities to controls found bilateral volume reduction in the patient group (Takeoka et al. 2004).
The language disorder of Landau–Kleffner syndrome is characterized by a severe deficit in comprehension, often referred to as a verbal auditory agnosia, though it often progresses to include language production as well. A deficit in phonemic discrimination has been described as the basic disorder underlying the deterioration of receptive language (Van Hout 2001). There has been considerable dispute in the literature regarding whether Landau–Kleffner syndrome is but one manifestation of continuous spike-and-waves during slow-wave sleep, or whether it should be included as an independent syndrome. While language deficits in continuous spike-and-waves during slow-wave sleep are often quite severe and long-standing, there is evidence to suggest that these patients show a profile that is distinct from Landau–Kleffner patients, which involves impairments in lexical and syntactic skills in the context of spared comprehension abilities (Debiais et al. 2007).
Unlike the poor outcomes typically seen in continuous spike-and-waves during slow-wave sleep, the prognosis of Landau–Kleffner is much more variable (Bishop 1985; Deonna et al. 1989; Veggiotti et al. 2002). Although the underlying seizure disorder is almost always treated successfully with antiepileptic medication, many patients continue to show language deficits (often severe) into adulthood, despite treatment (Deonna 1991). Studies have shown that a younger age at onset of the language regression is associated with a poorer language outcome (Bishop 1985), though one study found duration of electrical status epilepticus in slow-wave sleep to be a better predictor of outcome than age, with full recovery of language functions occurring exclusively in patients who had electrical status epilepticus in slow-wave sleep for less than 3 years (Robinson et al. 2001). The results of the latter study are limited by a small sample. Plaza et al. (2001) described a child who was diagnosed with Landau–Kleffner syndrome at 28 months (a right temporal focus was identified on EEG) who ultimately recovered language skills and acquired reading and spelling abilities. He showed complete left extinction in dichotic listening and a pattern of performance on memory tasks whereby very poor recall was observed when verbal information was presented auditorily, yet performances were intact when such information was presented visually. The authors suggested that early language regression might have been due to impaired auditory perception, which he overcame by using compensatory strategies that allowed the development of phonological skills from predominantly visual input.
Most studies of Landau–Kleffner syndrome have excluded patients whose receptive language was compromised before the diagnosis was made and have instead focused on the relationship of seizure control to language recovery (Klein et al. 2000). Klein and colleagues investigated the effect of premorbid language skills and behavior, along with clinical variables, on long-term language recovery. They found that premorbid language and behavior were more predictive of language recovery than persistence or absence of seizures. Results from this study underscored the importance of assessing premorbid language and behavior when making predictions regarding long-term outcome.
The neuropsychological literature with respect to seizure disorders is expansive, and it is well established that epilepsy in childhood is often associated with lowered general intellectual functioning and specific cognitive impairments. However, many studies have addressed seizure disorders in general, and do not identify specific epilepsy syndromes as outlined by the ILAE. The existing research on specific syndromes suggests that each may have very different cognitive outcomes, but more information is needed. Better characterization of the deficits unique to a given syndrome will aid clinicians in tailoring neuropsychological batteries to assess functioning in a given child. Future studies should strive to utilize well-characterized groups using strict ILAE classification guidelines. Given the low base rates of some syndromes, this may require collaborative multicenter efforts. Moreover, longitudinal studies that assess cognitive changes over time are warranted. In such studies, investigators should examine not only the natural progression of the cognitive profiles relevant to each syndrome, but also the pattern of changes subsequent to specific treatments (e.g., surgery, vagus nerve stimulator, etc.). As newer antiepileptic agents become available, studies should seek to elucidate the cognitive side effects or clarify their role in improving cognition.