Basic cognitive functions: general cognitive abilities, attention, and episodic memory
In patients with idiopathic generalized epilepsies, intellectual abilities are usually within the average range; however, they tend to be lower than in the general population (Hommet et al., 2006). In JME, most authors report average IQs with no significant group differences to healthy controls (Sonmez et al., 2004; Piazzini et al., 2008; Iqbal et al., 2009; Roebling et al., 2009; O’Muircheartaigh et al., 2011). Devinsky et al. (1997) highlighted one patient with an IQ of 69 who was excluded from their analysis. Despite group matching for educational exposure, lower verbal IQs (VIQs) have been reported in patients compared to healthy controls; however, with IQ levels remaining within normal limits (Pascalicchio et al., 2007; Wandschneider et al., 2010). Disease onset during early puberty hits a vulnerable educational phase (Wandschneider et al., 2010) and JME patients may not be able to benefit to the same extent as their peers, despite comparable educational exposure.
Sonmez et al. (2004) found no differences between JME patients and healthy controls using digit span to estimate global attention. Roebling et al. (2009) did not find any impairment of attention in JME using the Trail Making Test and the Stroop Test. In contrast, patients were affected on both tests in a different study population (Pascalicchio et al., 2007). Another study reported significantly slower reaction times in patients with JME compared with healthy controls on a measure of alertness (Wandschneider et al., 2010).
Functions of the temporal lobes, such as verbal and nonverbal episodic memory, are generally reported as spared (Roebling et al., 2009; Wandschneider et al., 2010). Two studies, however, have reported more widespread cognitive dysfunction. Sonmez et al. (2004) observed verbal and nonverbal learning deficits in addition to impairments on a range of frontal lobe tasks. In a study of 50 patients with JME and 50 controls, Pascalicchio et al. (2007) reported weak short- and long-term verbal and long-term nonverbal memory. The JME group was globally impaired in almost all other cognitive domains, including IQ, attention, processing speed, working memory, and executive functions. The findings of both these studies do not provide specific evidence for compromised temporal lobe functions.
Working memory in neuropsychological and imaging studies
Swartz et al. (1994) performed a visual working memory study, using patients with JME initially as a “patient control” group for comparisons of 15 patients with frontal lobe epilepsy (FLE), nine patients with JME, and 15 healthy controls. Pairs of abstract visual images were presented and subjects had to indicate by pressing a button whether the images matched. Defined by the delay between the two images, two conditions were created: The immediate match to sample task (IMS), with an image delay of 100 msec, controlled for attention, motivation, motor function, and habituation, and the delayed match to sample task (DMS), with an image delay of 8,000 msec, evaluated visual working memory. Two categories of errors (match-on-mismatch and mismatch-on-match) were recorded as well as reaction times. The FLE group presented with the weakest performance showing impairment on both tasks, particularly during the DMS condition. The JME patients’ performance was comparable to controls on the IMS task, but it was impaired during the DMS, the working memory condition, holding an intermediate position between the FLE group and the controls. Age and educational level were not found to be related to test performance.
In a subsequent 18-fluorodeoxyglucose–positron emission tomography (18FDG-PET) study, the same visual working memory paradigm was evaluated in nine patients with JME and 14 controls (Swartz et al., 1996). The JME group performed less well than the controls on the working memory task, whereas performance during the IMS task was again comparable in both groups. At resting state, 18FDG uptake in patients was decreased in the ventral premotor cortex, caudate, the dorsolateral prefrontal cortex (DLPFC) bilaterally, and the left premotor area, representing widespread frontal impairment. Controls activated areas that are thought to support working memory function, whereas patients presented with a “hypofrontality state” (Swartz et al., 1996) in keeping with their poorer performance on the DMS task. Increased metabolism of the lateral orbital and medial temporal regions in the JME group was interpreted as a compensatory mechanism for prefrontal dysfunction. The authors concluded that dysfunctional thalamo-frontocortical networks might account for both ictogenesis and poor working memory performance in JME.
A more recent functional imaging study examined 19 patients with JME using a verbal and nonverbal functional MRI (fMRI) working memory paradigm (Roebling et al., 2009). During the visual-spatial paradigm, subjects were assessed with a modified version of the Sternberg Item Recognition Test. A visual grid was presented, holding either a triangle or a square, and subjects were asked to memorize the positions of the items within the grid. After an interval, the grid was presented again, containing either a triangle or a square. Participants had to decide whether one of the symbols had been in the same position in the previous grid, irrespective of shape. During the verbal memory task, phonologically similar letters were shown, either in upper or lower case. During the response condition, a single letter was presented, and subjects had to decide whether it had been shown in the previous condition, irrespective of case. Both groups performed equally well on these tasks and no significant group differences were detected on fMRI-activation patterns. The authors suggested that these inconsistent findings may indicate a heterogeneous epilepsy syndrome, in which frontal lobe dysfunction is present only in a JME subgroup. However, this failure to detect differences between groups might be caused by an insufficiently challenging working memory task.
Vollmar et al. (2011) investigated a larger JME population (n = 30) with a different, and more challenging, working memory fMRI paradigm, the dot back task (Kumari et al., 2009). During the task, dots were randomly presented on a spatial grid. There were three different response conditions: During the “0-back” task, participants were instructed to move a joystick toward the current position of the dot; in the “1-back” condition to the position of the dot in the previous grid and in the “2-back” to its position two stimuli back. Patients and controls performed equally well on all three tasks and showed significant fMRI activation of working memory networks, after subtracting “0-back” from “1-back” and “2-back” in order to control for the motor component. FMRI cortical activation patterns, however, differed significantly with increasing task demand. During the “2-back” condition, the motor cortex and supplementary motor area (SMA) increasingly coactivated with working memory networks in patients (Fig. 1). The authors also described increased functional connectivity between the motor system and areas of higher cognitive functions within the frontal and parietal lobes. This increased functional connectivity in the JME group was interpreted as a possible mechanism for seizures triggered by cognitive effort, in line with clinical observations that myoclonic jerks are often precipitated by cognitive tasks (Inoue & Kubota, 2000). These studies underscore the importance of imaging studies to detect differences in cognitive networks, whereas group differences are too subtle to be detected during conventional neuropsychological testing.
Figure 1. Functional MRI findings in JME, rendered on Montreal Neurological Institute (MNI) template. Red shows the bilateral frontal and parietal working memory network, activated in the “2-back minus 0-back” condition. Orange shows the increased activation during this cognitive effort in patients with JME compared to healthy controls. Both functional maps are thresholded at T-scores > 3.
Download figure to PowerPoint
Executive functions in neuropsychological and imaging studies
Executive functions in JME are controversially discussed (see Table S2).
Devinsky et al. (1997) assessed 15 patients with JME and 15 with TLE, with tests sensitive to executive functions: inhibition and psychomotor speed (Trailmaking Test, Stroop Test); abstract reasoning, concept formation and, mental flexibility (Wisconsin Card Sorting Test [WCST], Booklet Categories Test); planning (Mazes from the WISC-R); and verbal fluency (Controlled Oral Word Association Test). In the JME group, performance was variable. Some patients were impaired on only 3 of 11 tests, whereas others were impaired on at least 4 tests. The poorest results were reported for concept formation and mental flexibility (60–64% of patients impaired), followed by cognitive speed (47–53% of patients) and planning (33%). In comparison with the TLE group, patients with JME performed less well on tests requiring mental flexibility and concept formation. No specific pattern of cognitive dysfunction was elicited in JME, although the authors concluded that their findings were supportive of frontal lobe dysfunction, which may lead to maladaptive behavior with social consequences.
Other studies have confirmed impairments of distinct executive functions, namely response inhibition (Sonmez et al., 2004; Pascalicchio et al., 2007; Piazzini et al., 2008; Wandschneider et al., 2010; O’Muircheartaigh et al., 2011; Kim et al., 2012), mental flexibility and concept formation (Pascalicchio et al., 2007; Piazzini et al., 2008; O’Muircheartaigh et al., 2011), and verbal fluency (Swartz et al., 1996; Sonmez et al., 2004; Pascalicchio et al., 2007; Roebling et al., 2009; Wandschneider et al., 2010; O’Muircheartaigh et al., 2011; Kim et al., 2012).
To date, few imaging studies evaluated structural and functional brain correlates of executive dysfunction in JME.
In a quantitative MR spectroscopy (MRS) study, Savic et al. (2004) compared patients with either JME or idiopathic generalized epilepsy (IGE) with generalized tonic–clonic seizures (GTCS) to a group of healthy controls. N-Acetyl aspartate (NAA) is a neuron-specific metabolite. Reduced levels can be associated with neuronal dysfunction or damage (Savic et al., 2000). The authors reported reduced frontal lobe NAA concentrations in the JME group (n = 25) in relation to both healthy controls (n = 10) and GTCS patients (n = 20). In addition, JME patients with low frontal NAA concentrations showed frontal lobe dysfunction on a brief neuropsychological assessment. In contrast, frontal lobe functions were spared in JME patients with normal NAA concentrations, GTCS patients, and controls. Hence, although a prefrontal neuronal lesion may be present in some JME patients, JME seems to be a heterogeneous condition. Low NAA levels did not correlate with any other clinical parameter, such as current seizure frequency or seizure number over lifetime.
Pulsipher et al. (2009) aimed to investigate the integrity of thalamo-frontocortical networks in relation to executive function in recent-onset JME. Twenty JME patients within 12 months of diagnosis were compared to an epilepsy control group of 12 patients with recent-onset benign childhood epilepsy with centrotemporal spikes (BECTS) and 51 healthy controls (first-degree cousins). Groups were matched for gender, duration of epilepsy, and IQ. JME patients were significantly older than both the BECTS and healthy control groups, yet age was not found to be related to test performance. Participants were assessed with three subtests of the Delis-Kaplan Executive Function System (D-KEFS) and a questionnaire for parents, the Behavior Rating Inventory of Executive Function (BRIEF). JME patients performed less well than controls on D-KEFS Inhibition. Behavioral Regulation and Metacognition scores of the BRIEF were also significantly lower in the JME group. Quantitative MRI measurements revealed smaller thalamic volumes and greater frontal cerebrospinal fluid (CSF) in JME patients than in healthy controls and BECTS patients. Thalamic and frontal volumes predicted D-KEFS performance only for the JME group. Of interest, JME patients were showing volumetric abnormalities already within 12 months of seizure onset, suggesting a clinically significant disruption of the thalamo-frontocortical circuitry, leading to both seizures and neurocognitive deficits. It remains uncertain how early structural abnormalities present in disease evolution, although the distinct volumetric abnormalities seem unlikely to be the result of chronic seizures.
In a resting FDG-PET study (McDonald et al., 2006), regional cerebral rates of glucose uptake values (rCMRGlc) were regressed on various executive function test scores in patients with frontal lobe epilepsy (FLE; n = 18), JME (n = 10) and healthy controls (n = 14). The executive function battery included measures of cognitive flexibility, fluency, response inhibition, working memory, and sustained attention. In the JME group, frontal hypometabolic values predicted impairment on measures of figural fluency and cognitive flexibility.
O’Muircheartaigh et al. (2011) reported subtle dysfunctions in verbal fluency, comprehension and expression, mental flexibility, and response inhibition in a cohort of 28 JME patients. In a structural and diffusion tensor MRI (DTI) study, voxel-based morphometry revealed reductions in gray matter volume in the SMA and posterior cingulate cortex. Fractional anisotropy (FA) in the SMA predicted performance in tasks of word naming and expression. Gray matter volumes of the posterior cingulate cortex and FA correlated with scores on the mental flexibility task. The authors describe their JME cohort as relatively high functioning but with neuropsychological evaluation revealing subtle cognitive impairments.
The finding of reduced frontal FA was recently replicated in another JME cohort (Kim et al., 2012) and correlated with disease severity but not with the reduced neuropsychological performance observed in JME.
A diffusion tensor tractography study in JME revealed increased structural connectivity between the prefrontal cortex and the motor cortex and SMA, which correlated with the functional connectivity of adjacent areas depicted by fMRI, supporting the hypothesis of network alterations facilitating the cognitive triggering of seizures. The authors discussed that an increased proportion of fibers crossing the SMA could be an explanation for reduced FA findings in the mesial frontal region, rather than neuronal loss (Vollmar et al., 2012).
In summary, imaging studies support dysfunction within thalamo-frontocortical networks as possible neuronal correlate of executive function impairment in JME.