Neuropsychological impairment is an important comorbidity of epilepsy (Dodrill, 2004) and affects up to 82% of children with epilepsy (Cormack et al., 2007; Berg et al., 2008; Hoie et al., 2008). Intractable localization-related epilepsy of childhood onset can result in more generalized neuropsychological effects such as greater reduction of levels of intellectual function, memory, and executive function than would be expected from a focal epileptogenic process (Hermann et al., 2002). Although there are numerous publications on the detrimental effects of chronic intractable epilepsy on neuropsychological performance in children and adolescents (O'Leary et al., 1981, 1983), to date, there is no report on the structural correlates of neuropsychological deficits beyond brain volumetric assessment in children with epilepsy (Hermann et al., 2006).
White matter (WM) tracts are thought to play a crucial role in linking the different components of the cortical processing networks necessary for cognition (Mesulam, 1990). Hence, injury to the WM may lead to loss of “connectivity” between these cortical regions and therefore neuropsychological impairment (O'Sullivan et al., 2001). Along these lines, there is evidence to suggest that WM is vulnerable to seizure-related injury (Jorgensen et al., 1980; Dwyer & Wasterlain, 1982). Therefore, impairment in WM integrity related to epilepsy may result in reduced neuropsychological function. The integrity of the WM can be assessed with diffusion tensor imaging (DTI).
Our hypotheses were first that nonlesional intractable localization-related epilepsy is characterized by diffuse bilateral WM changes, as manifested by reduced fractional anisotropy (FA) and elevated mean diffusivity (MD); and second, FA and MD demonstrate good correlation with neuropsychological performance. The aims of this study were to assess DTI measures of regional WM in children with nonlesional localization-related epilepsy compared to controls, and to assess the relation between lobar WM and neuropsychological performance.
- Top of page
- Supporting Information
We found that patients with nonlesional localization-related epilepsy performed worse than healthy subjects in multiple domains of neuropsychological performance including intelligence, language, and executive function. Patients had reduced FA in multiple lobes including bilateral frontal, bilateral temporal, right parietal, and right occipital WM relative to healthy controls. There was a significant correlation between reduced right temporal FA and impaired language and executive function, as well as between reduced body of corpus callosum FA and decreased intelligence and language in patients. Although there was no significant difference in MD between patients and controls, there was a significant correlation between elevated left parietal MD and impaired language in patients. Neither FA nor MD was correlated with neuropsychological function in controls.
Considerable neuropsychological impairment has been reported in children and adolescents with chronic epilepsy (Farwell et al., 1985; Schoenfeld et al., 1999; Smith et al., 2002; Germano et al., 2005). Impairment in a wide range of cognitive functions including intelligence, language, attention, executive function, and psychomotor speed was also evident in children with newly diagnosed epilepsy (Oostrom et al., 2003; Hermann et al., 2006), including children with localization-related epilepsy (Oostrom et al., 2003; Hermann et al., 2006). Childhood-onset localization-related epilepsy has been associated with more generalized cognitive effects than would be expected from a focal epileptogenic process (Hermann et al., 1997; Oyegbile et al., 2004a,b). A wide range of cognitive deficits including intelligence, language, executive function, and memory has been reported in childhood-onset temporal lobe epilepsy compared to more specific deficits seen in adult-onset temporal lobe epilepsy (Hermann et al., 2002). We have found similar findings in that, relative to controls, children with localization-related epilepsy have widespread cognitive impairment involving intelligence, language, and executive function, including the subgroup with FLE.
We found widespread abnormal WM in both cerebral hemispheres in children with localization-related epilepsy, including the subgroup with FLE. The WM may be vulnerable to seizure-related changes in the pediatric brain. Dwyer and Wasterlain (1982) have examined the effects of electroconvulsive-induced seizures in rats at different developmental stages. They reported that seizures occurring early in development selectively impaired myelin accumulation out of proportion to their overall effect on brain growth. Hermann et al. (2002) examined patients with childhood-onset chronic temporal lobe epilepsy and found greater reduction in WM volume compared to GM volume. During normal development in children and adolescents, there are greater age-related linear increases in WM volumes compared to GM volumes as demonstrated by in vivo quantitative MR (Pfefferbaum et al., 1994; Giedd et al., 1996; Courchesne et al., 2000; De Bellis et al., 2001; Paus et al., 2001). The rapid growth of the WM in childhood may predispose the WM to seizure-related changes.
The widespread WM changes in patients with localization-related epilepsy may be related to spread of seizures. The kainate model of epilepsy in rats has shown widespread DTI abnormalities (Sierra et al., 2011), involving dorsal endopiriform nucleus, external capsule, corpus callosum, dentate gyrus, thalamus, optic tract, horizontal limb of the diagonal band, stria medullaris, habenula, entorhinal cortex, and superior colliculus. Histology targeted to the abnormal FA showed altered myelination, neurodegeneration, and/or calcification of the tissue. DTI abnormalities have also been reported outside the hippocampus in temporal lobe epilepsy, including the internal capsule, external capsule, corpus callosum, fornix, and cingulum (Arfanakis et al., 2002; Concha et al., 2005; Thivard et al., 2005; Gross et al., 2006; Focke et al., 2008). These findings appear to implicate a larger dysfunctional network, reaching well beyond the seizure focus in chronic unilateral epilepsy (Thivard et al., 2005; Nilsson et al., 2008). The treatment implication of the widespread WM impairment remains to be elucidated. Further study is needed to clarify if this widespread impairment in WM is predictive of poor neuropsychological outcome following epilepsy surgery.
Previous studies in adults with epilepsy have found a relation between impaired WM integrity and neuropsychological function. In one study of temporal lobe epilepsy, tractography of the parahippocampal gyrus demonstrated that reduced FA in left parahippocampal gyrus, and FA in the left and right parahippocampal gyri correlated with verbal and nonverbal memory, respectively (Yogarajah et al., 2008). McDonald et al. (2008) found that reduced FA and elevated MD of multiple fiber tracts (including uncinate fasciculus, arcuate fasciculus, parahippocampal cingulum, and inferior frontooccipital fasciculus) were related to verbal memory and naming performances but not with nonverbal memory or fluency. Diehl et al. (2008) evaluated the uncinate fasciculus of patients with temporal lobe epilepsy and found reduced FA of right uncinate fasciculus was correlated with visual delayed memory, whereas MD of left uncinate fasciculus was negatively correlated with auditory immediate and delayed memory. O'Muircheartaigh et al. (2011) found that reduced FA in the WM of supplementary motor area predicted word naming tasks and expression scores, and reduced FA in posterior cingulate predicted cognitive inhibitions scores on mental flexibility task in adults with juvenile myoclonic epilepsy. A more recent study in adults with nonlesional localization-related epilepsy found that IQ and cognitive impairment were strongly associated with clustering and path lengths as assessed using WM connectivity network (Vaessen et al., 2012). To our knowledge, the present study is the first to assess the relation between WM integrity in children with nonlesional localization-related epilepsy and neuropsychological function. Children with nonlesional localization-related epilepsy demonstrated a significant correlation between reduced right temporal FA and impaired language and executive function, between reduced body of corpus callosum FA and decreased intelligence and language, and between elevated left parietal MD and impaired language. The impaired WM integrity did not appear to be tightly linked to specific cognitive abilities, for instance poor language function was related to impaired WM integrity in right temporal, body of corpus callosum, and right parietal WM. This may be related to using lobar WM rather than specific white matter tracts, and using composite scores of neuropsychological function rather than specific neuropsychological abilities. The association between impaired language and reduced right temporal FA could be related to the role of the right anterior temporal lobe in language. Early in life, both left and right hemispheres contribute to language processing, with an increase in left hemisphere contribution with increasing age (Holland et al., 2001; Brown et al., 2005; Ressel et al., 2008; Kadis et al., 2011). It may be that the disturbance in networks for language processing in children with epilepsy alters the timing or degree of the contribution of right hemisphere structures for language. There is also evidence that the right temporal lobe plays a role in certain aspects of language such as semantic processing, object naming (Lambon Ralph et al., 2001), and language comprehension (Crinion & Price, 2005).
The severity of abnormal WM was related to age at seizure onset, with greater reduction in FA associated with earlier age at seizure onset. Riley et al. (2010) have also found that lower FA in the posterior corpus callosum was significantly correlated to earlier age at seizure onset but not epilepsy duration. In addition, animal studies have demonstrated repeated electroconvulsive-induced seizures in the immature rat resulted in reduced brain growth, an effect found to be dependent on the developmental stage of the animals when seizures were induced, being more severe when seizures occurred at younger ages (Wasterlain & Plum, 1973; Jorgensen et al., 1980). These data suggest that the immature brain may be more susceptible to changes related to recurrent seizures. The relation between abnormal WM and duration of epilepsy was weaker.
We have examined children with a variety of seizures. The heterogeneity in the seizure semiology of our patient population reflects the heterogeneity of the clinical population of pediatric patients being evaluated for epilepsy surgery. However, within our patient population, there was a greater predominance of FLE; hence, the subgroup analyses were conducted. We found that patients with FLE have widespread reduced FA involving the frontal and extrafrontal WM, including bilateral frontal, right temporal, right parietal, and genu, body, and splenium of corpus callosum. We also found a trend for correlation between FA of the body of corpus callosum and executive function, and between MD of the left frontal, left temporal, and left parietal WM with language in patients with FLE. These findings confirm that there are widespread effects of a focal epileptogenic process on structural and cognitive function in pediatric epilepsy.
In summary, we found widespread abnormal WM in children with nonlesional localization-related epilepsy, which was associated with impaired neuropsychological function. Impaired WM integrity may be related to the direct effect of seizures. Alternatively, the data may reflect an underlying structural abnormality that is responsible for both the seizures and the neuropsychological deficits, or a combination of these factors. The impairment in WM integrity observed in our cohort of children with nonlesional localization-related epilepsy may reflect disruption in the connectivity for cortical processing networks, which is necessary for the development of cognition.