Neuropsychological impairments are common in many epilepsy syndromes and are related to clinical factors such as seizure frequency and severity, age of seizure onset, as well as the underlying pathologic substrate. It is beyond the scope of this chapter to provide a comprehensive review of all epilepsy syndromes and their patterns of neuropsychological impairment. However, we will highlight several consistent neuropsychological principles.

As described by Hughlings Jackson, there are significant and independent contributions of both static and dynamic factors that affect brain function, and by extension, neuropsychological abilities. Morphologic or structural lesions are associated with relatively nonmodifiable neuropsychological deficits. In contrast, electroencephalographic (EEG) discharges, seizures, and epilepsy treatment are associated with more dynamic brain changes that, to varying degrees, are modifiable and are under direct physician management. Depending on epilepsy type (idiopathic vs. symptomatic), the relative contributions of specific factors will differ.

Disentangling the stable and dynamic cognitive influences in epilepsy often poses a major challenge because the causes of impaired neuropsychological function are not fully independent of each other. Treatment effects, for example, act on and interact with morphology and epilepsy. Although altered brain structure and function may result in epilepsy, epilepsy and its underpinnings may also alter functional cerebral organization. Finally, at the highest level, epilepsy-related cognitive impairment must be evaluated within the patient’s developmental context. Certain seizure syndromes show peaks at specific developmental stages, and etiology is associated with age at seizure onset. Cognitive profiles vary depending on age of seizure onset, with differences apparent depending on whether epilepsy develops in the maturing brain versus mature brain versus aging brain. However, age of seizure onset may simply reflect the expression of dysfunctional brain maturation.

Epilepsy is often dichotomized according to whether the EEG abnormalities involve the entire cerebrum (generalized epilepsy) or begin focally (partial epilepsy). Generalized epilepsy includes tonic–clonic seizures, juvenile absence epilepsy, and myoclonic epilepsy. Partial epilepsy includes a variety of seizure types including the so called “benign” partial epilepsy (e.g., benign epilepsy with centrotemporal spikes, or BECTS) as well as symptomatic focal epilepsy (e.g., mesial temporal lobe epilepsy and neocortical epilepsy). In this chapter, we describe disease effects on cognition as a function of epilepsy syndromes, age of onset, and epilepsy course. We also discuss the complex issue of whether poorly controlled seizures are associated with progressive cognitive decline. For ease of discussion, we categorize epilepsy subtypes according to whether they are considered to be idiopathic or symptomatic.

Idiopathic epilepsy, including both generalized and partial epilepsy expression, is characterized by a genetic predisposition and the absence of readily identifiable brain lesions. Although not completely silent behaviorally or cognitively, idiopathic epilepsy is generally easy to treat and is associated with less severe cognitive impairments than are other seizure types. Idiopathic generalized epilepsy (IGE) is characterized by generalized EEG abnormalities involving the entire cerebral cortex, whereas idiopathic partial epilepsy (IPE) is associated with regional EEG abnormalities (e.g., centrotemporal EEG in rolandic epilepsy).

As would be predicted from generalized EEG abnormalities, diffuse and generalized cognitive impairments are present, including deficits in attention, psychomotor speed, visuospatial skills, and nonverbal memory. Language and verbal memory, in contrast, appear unaffected.^62,102,107

The epileptiform discharges and cognition are also closely related. Not only does cognitive impairment vary as a function of seizure activity, but cognition may also induce seizures and seizure discharges.⁹² Although this relationship has been described in patients with symptomatic temporal lobe epilepsy (TLE),⁵⁰ patients with IGE are particularly likely to show neuropsychological EEG activation. Negative effects of spike-and-wave bursts exist for sensory and executive functions. Therefore, tasks requiring sustained attention are best suited to detect the cognitive effects of EEG changes in IGE.¹⁰² Although cumulative attentional effects may ultimately result in diminished level of function when they occur over long periods, decreased IQ is not a primary feature of the disease, with developmental delay and retardation developing from interference with cognitive functions over a long period of time. Absence epilepsy developing in early childhood is generally associated with poorer outcome than juvenile absence epilepsy.

BECTS is a common epilepsy syndrome (10%–15%), beginning between 5 and 9 years of age and extending into adolescence. It has a favorable prognosis, and most patients become seizure-free after puberty. Its neuropsychological prognosis, however, is less benign. During its active phase, neuropsychological deficits may include attention, motor functions, short-term memory, visual and perceptive abilities. Language difficulty relating to the interictal dysfunction of the perisylvian language areas, however, is a major characteristic of BECTS.¹⁰² Learning disabilities are common in BECTS,¹⁴³ although they are not progressive in nature. Although children rapidly improve in most areas following seizure remission, minor problems in executive functions and verbal comprehension persist.^80,103 Complete seizure remission is generally needed for a favorable cognitive outcome.

Juvenile myoclonus epilepsy (JME) generally begins between 12 and 18 years of age, and is characterized by neuropsychological and behavioral features associated with
frontal dysexecutive impairment such as reasoning difficulty, poor concept formation, and decreased mental speed and flexibility.^{22,64,65,102,130} Of course, frontal lobe dysfunction is not specific for JME. Whether frontal lobe cognitive dysfunction together with personality change (e.g., limited self-control, suggestibility, indifference, rapid mood changes) form a syndrome characteristic of JME merits further study.⁶⁵ The presence of more focal impairments in addition to generalized slowing is consistent with the view that IGE should no longer be considered purely a “generalized” epilepsy. EEG, histologic, structural and functional imaging studies suggest a specific involvement of frontal lobes, thalamus, and thalamocortical loops in IGE.^{1,75,100,118,147}

In conclusion, a wide range of rather mild impairments may be associated with idiopathic generalized or idiopathic partial epilepsy. Mild generalized impairment and learning difficultly have been observed. These are best understood from the close relationship between active epileptic processes interfering with cognitive networks of lower-order functions, on perceptive and executive functions, and on the interference of epilepsy with critical periods of cognitive development (i.e., before, during, or after language acquisition, or at the time before or during frontal executive function development). Frontal/executive functions are the last to fully develop and therefore may represent a common endpoint for the impairments seen in idiopathic epilepsy. Following epilepsy remission, neuropsychological recovery from active epilepsy-driven impairment can be observed. However, some long-term residual deficits may persist, particularly when epilepsy has significantly interfered with cognitive development.

In contrast to idiopathic epilepsy, the cognitive profiles of symptomatic epilepsy are more strongly related to epilepsy location and etiology. The temporal lobes and temporomesial structures are particularly vulnerable to seizure development, and TLE accounts for approximately 70% of chronic symptomatic epilepsy. Approximately half of TLE patients have hippocampal sclerosis (HS) or hippocampal atrophy, although whether mesial TLE represents a distinct nosological entity or a syndrome is still a matter of debate.¹⁴⁶ Mesial TLE is characterized by impaired declarative memory.⁵⁷ Patients with earlier seizure onset tend to have lower IQs, reflecting the interference of seizures (and perhaps their treatment with antiepileptic medications) with normal cognitive and brain development.⁵⁴ Accompanying the IQ with earlier seizure onset is a reduction of total brain volume, including both gray and white matter.⁵⁴ Memory impairment occurs independent of the age of seizure onset, although the nature of the memory impairment depends on when seizures begin. A more generalized memory impairment occurs with earlier seizure onset, whereas a more focal and material-specific memory impairment that varies according to seizure onset laterality is seen with later seizure onset.⁴³

With later seizure development, left temporal/left temporomesial epilepsy is associated with material-specific impairment of verbal learning and memory. Mesial and neocortical structures differentially contribute to verbal memory, with mesial structures subserving consolidation and retrieval, and neocortical structures being more associated with content processing. Thus, impaired delayed recall is more indicative of mesial rather than neocortical temporal lobe damage.⁴⁸ Impairment of verbal learning, short-term memory, and naming (i.e., semantic memory) are less specific but also may reflect left inferotemporal or temporolateral lesions.^{39,40,42,125,126} Naming impairment is associated with hippocampal volume,¹⁴ and also related to functional activity reflected by spectroscopy.¹²³ Like memory, the magnitude of naming impairment is strongly associated with seizure onset age.

In contrast to left TLE, right TLE tends affect performance on figural or nonverbal memory tasks.³⁴ However, this relationship is less consistent than that between left TLE and verbal memory,⁶ an effect that has been attributed to nonverbal memory networks being more bilaterally distributed than verbal memory, covert verbalization during task performance, or the type of test and test materials (abstractness, complexity) used. Consequently, using figural memory tests to infer mesial temporal dysfunction will often falsely lateralize seizure onset. However, false lateralizing figural memory impairment in left TLE may also reflect atypical language dominance or sex differences.⁴³

Even though the area of seizure onset in focal TLE is limited, neuropsychological impairment often extends beyond the seizure onset zone.^59,89 These “frontal” deficits imply impaired functional connectivity that is disrupted with a temporal lobe focus, and may be considered to reflect “nociferous cortex” effects, in which the negative effects associated with ongoing seizure discharge impair brain function at some distance from the active seizure focus.¹³⁴ However, magnetic resonance imaging (MRI) volumetrics have demonstrated prominent disruption in ipsilateral hippocampus and neural connectivity (i.e., white matter volume loss) that extends beyond the temporal lobe, affecting both ipsilateral and contralateral hemispheres.¹²⁸ TLE patients with secondary generalized seizures are at higher risk of additional general neuropsychological impairment.⁷⁰

Frontal lobe epilepsy (FLE) is seen in approximately 20% of patients with partial onset seizures, and is associated with a less consistent neuropsychological profile than TLE. In contrast to TLE, in which HS is the predominant morphologic feature, frontal lobe epilepsy is associated with a more heterogeneous array of etiologic factors. Moreover, executive functions mediated by the frontal lobe contribute to most other cognitive functions, resulting in diffuse and nonspecific neuropsychological impairments. Patients suffer from attention problems, problems with working memory, mental flexibility, response inhibition, or planning. Tests of motor coordination appear particularly sensitive to frontal lobe epilepsy. At the highest level, a dysexecutive syndrome may comprise problems with response selection, initiation, execution, and inhibition. No consistent lateralized impairment has been associated with focal left versus right FLE.^{25,41,129,139}

The neuropsychological characteristics of parietal lobe epilepsy and occipital lobe epilepsy have rarely been described in a series using adequate sample sizes. Acute parietal or occipital neuropsychological symptoms become evident in seizure semiology, but in chronic epilepsy (most often those patients exhibiting early lesions or malformations), the classic posterior symptoms of aphasia, alexia, agraphia, acalculia, agnosia, and neglect are very uncommon. Primary or secondary perceptive and sensory problems that may be evident at the beginning of epilepsy are often well compensated for behaviorally. Impairments are diffuse, and, as described with seizure semiology and EEG, often mimic frontal or temporal lobe dysfunction.^52,72 Nevertheless, tests of stereognosis or haptic search may be sensitive to parietal lobe epilepsy.^71,117

Given the many potential influences on cognition for patients with epilepsy, such as age of onset, disease substrate, or seizure frequency and severity, antiepilepsy drugs (AEDs) occupy a unique position because they are under the direct control of the treating physician and his patients. Although the choice of specific AED is guided by seizure type and epilepsy syndrome,⁶¹ within seizure/syndrome categories, AED selection is typically based on clinical experience rather than evidence-based practice. Most major AEDs used to treat partial epilepsy have comparable efficacy,⁷⁶ although many recently introduced AEDs are associated with more favorable tolerability profiles that includes less neuropsychological impairment.⁷³

Because AEDs decrease membrane excitability, increase postsynaptic inhibition, or alter the synchronization of neural networks, they are often associated with neuropsychological side effects including decreased motor/psychomotor speed and attention.⁹⁶ Adverse AED effects are a significant component of treatment effectiveness. The landmark VA Cooperative study reported that standard AEDs including carbamazepine are associated with significant adverse effects that contribute to initial treatment failure in more than 40% of patients,⁹³ and a separate European trial reported that tiredness was described by more than 50% and sleepiness by more than 35% of patients on phenytoin or carbamazepine monotherapy.⁴ Adverse AED effects are strongly associated with poor health reported by patients³⁰ and with decreased health-related quality of life.³² After seizure control, the most important aspect of AED treatment is the side effect profile, including problems with cognition, energy level, school performance, childbearing, coordination, and sexual function. Because of side effects, 20% of patients adjusted their AED dosing.²⁷

In young adults, neuropsychological AED profiles are generally comparable for the older-generation AEDs carbamazepine, phenytoin, and valproate, with each AED associated with modest psychomotor slowing accompanied by decreased attention and memory.⁹⁶ Neuropsychological side effects generally emerge according to a dose-dependent relationship⁹⁶; however, both quality of life³¹ and memory may be affected, even when AED blood levels are within standard therapeutic ranges. Central nervous system (CNS) effects of AEDs are reflected by EEG slowing that not only is correlated with short-term neuropsychological decline,^119,120 but is also related to poorer neuropsychological outcome following 1 year of treatment.²⁸ With the exception of topiramate^{87,96,122,135} and possibly zonisamide,^2,10 most newer-generation AEDs have more favorable tolerability and neuropsychological profiles than their predecessors.^86,97,98,99

Although direct head-to-head comparisons examining the neuropsychological profiles of newer AEDs have not typically included medications thought to have favorable neuropsychological outcomes, some data suggest differences in this regard. For example, in one study, oxcarbazepine was associated with both neuropsychological impairment and EEG slowing in healthy volunteers.¹²¹ Thus, some data suggest that important differences may exist among AEDs, even across newer agents considered to have favorable neuropsychological side effect profiles. Several recent Class I healthy-volunteer studies suggested increased risk of cognitive impairment associated with topiramate^97,121 Because there may be individuals who are at greater risk for developing cognitive impairment, it may be possible to ultimately predict individuals at increased risk for developing treatment-emergent side effects based on pharmacogenetic or pharmacokinetic patient characteristics.

In addition to poor performance on memory tests and other neuropsychological measures, epilepsy patients often complain of poor memory.¹⁹ Although both subjective and objective memory findings indicate decreased memory, subjective memory ratings and objective memory performance are poorly correlated.^{8,29,46,108,142} In studies with sufficient sample sizes, statistically significant relationships between objective and subjective performances have been reported, although these correlations are generally small and account for a small portion of the variance. In contrast, subjective memory correlates much more highly with mood.^{19,23,26,108,109} Depressed or anxious patients tend to rate their memory as poor, whereas patients less burdened by poor mood states rate their memory more favorably. Correlations generally account for approximately half of the variance,^5,26 with mood being the single best predictor of subjective memory functioning.^23,33,108

The association between subjective memory and mood is informative, yet a large portion of the variance remains to be explained. Most studies show no significant relationship between subjective memory and clinical factors such as sex/gender, chronologic age, seizure-onset age, seizure type, seizure frequency, region of seizure onset, and number of AEDs.^23,109 However, memory “complainers” may have a later age of
seizure onset,¹⁹ and a small inverse relationship between age and subjective memory reports has been described.^26,33 There is a tendency for patients on polytherapy to report greater cognitive difficulty than do patients on monotherapy,³³ although this relationship is well-established using formal neuropsychological measures.⁹⁶ Although most studies are restricted to TLE patients, those that included both temporal and extratemporal patients document greater reports of memory impairment than in TLE.¹⁹ Although some investigators report no influence of seizure laterality,⁵ others have found significant associations between perceived memory and objective verbal memory in left TLE patients, and with objective nonverbal memory in right TLE patients.¹² Although reports exist of a relationship between perceived and objective language performance,¹² others have not observed this relationship.⁵