Cognitive outcome of parietooccipital resection in children with epilepsy
Address correspondence to Sarah Lippé, PhD, Professeure Adjointe/Assistant Professor, Département de psychologie/Department of Psychology, Université de Montréal/University of Montreal, C.P. 6128, Succursale Centre-Ville, Montréal, QC, H3C 3J7 Canada. E-mail: email@example.com
Purpose: We followed the neuropsychological development of five children who underwent unilateral neurosurgery of the occipitoparietal lobes as a treatment for epilepsy caused by a developmental lesion (cortical dysplasia).
Methods: The follow-up period ranged from 3–7 years postsurgery.
Results: Two participants had a verbal intelligence quotient (IQ) >100 and three had a verbal IQ between 65 and 80. All five children had abnormal nonverbal IQ and exhibited deficits related to visual attention, object recognition, and praxis. Nevertheless, our results suggest that brain plasticity after parietooccipital epilepsy surgery in young children allows for a schooling level of cognitive skills such as reading and arithmetic.
Discussion: Although recovery for visual perceptual cognition was more limited than for verbal functions, long-term neuropsychological outcomes showed that early surgery for epilepsy offers the possibility of optimizing cognitive outcomes in children with posterior intractable epilepsies.
Neurosurgery is an accepted treatment for intractable partial epilepsy in pediatric and adult populations. It has been proven effective, particularly for intractable temporal lobe epilepsy (Olivier, 1992; Wiebe et al., 2001; Jambaqué et al., 2007; Kim et al., 2008). In children, surgery for extratemporal lobe epilepsy represents approximately 50% of procedures and has been shown successful in one-half to two-thirds of the cases (Wyllie et al., 1998; Cossu et al., 2008). The most frequent cause of surgically treated extratemporal intractable epilepsy is cortical dysplasia, an early developmental malformation of the cortex (Wyllie et al., 1998; Janszky et al., 2000; Lortie et al., 2002; Lerner et al., 2009).
A number of studies have reported that surgical treatment can improve cognitive prognosis in children with epilepsy (Bayard & Lassonde, 2001; Helmstaedter, 2004). Scores on attention and working memory, episodic memory tasks, and naming tests have been shown to improve (Gleissner et al., 2005; Jambaqué et al., 2007), allowing for higher scholastic achievement.
However, very few studies have measured the neuropsychological outcomes for surgery in parietal and occipital epilepsies (Smith & Billingsley, 2001). Jambaqué et al. (1998) reported visual apperceptive agnosia with severe face recognition impairment, executive function deficits, and autistic symptoms in a 13-year-old girl who had undergone a right occipital lobectomy for cortical dysplasia at age 7 years. Remarkably, her academic abilities in spelling, reading, and arithmetic were higher than her functional intelligence level, perhaps as a result of special abilities often encountered in patients with autism. Cohen et al. (2004) demonstrated brain reorganization of the visual word form area (VWFA) toward the right hemisphere in a young epileptic child operated at the age of 4 years for a Sturge-Weber syndrome in the left occipitotemporal lobe. Sinclair et al. (2005) reviewed the epileptic outcomes of nine children with parietal resections and six with occipital resections. Although cognitive outcome was not addressed, occipital lobe patients showed visual field losses. Focusing on the possibility of generalized impairments, Gleissner et al. (2008) assessed 15 children with parietal lobe epilepsy. One year after surgery, these children were seizure-free with postoperative improvements in attention and behavior, but they showed a decrease in performance intelligence quotient (IQ).
In the present study, we assessed the long-term cognitive outcome—from 3–7 years postsurgery—of occipitoparietal epilepsy in five children who showed early onset due to Taylor-type focal cortical dysplasia. Neurosurgical resection involved the occipital lobe and temporal or parietal lobe junctions. Neuropsychological assessment specifically targeted the cognitive functions of the posterior regions of the brain. Although the occipital lobe is mainly responsible for visual perception and discrimination, the parietal lobe is important for visuospatial orientation, attention, saccades, numerical cognition, grapheme–phoneme transcoding, estimating time and quantities, and praxis (Leff et al., 2001; Hubbard et al., 2005; Pouthas et al., 2005; Cavanna & Trimble, 2006). Occipitotemporal areas are involved in color, object, and face recognition; reading; and other functions (Dehaene et al., 2001; Casarotto et al., 2008). In both children and adults, the impact of a lesion in the occipital and parietal lobes can result in visual agnosia, neglect, visuospatial disorientation, alexia, sensorial extinction, and Gerstmann syndrome (Laurent-Vannier et al., 2003). Given the early maturation of the posterior areas of the brain, we predicted reduced recovery of visuospatial abilities as compared to language development. Within the posterior regions, we also hypothesized that lesion type and location (right- or left-sided resection) would differentially impact cognitive function, given that the right and left hemispheres show distinct preferences for Gestalt versus analytical processing (Van Kleeck & Kosslyn, 1989). We posited that neurodevelopmental lesions, such as cortical dysplasia and early onset epilepsy, would entail reduced recovery of visuoperceptual cognition compared to verbal function. Because the literature reports of long-term follow-up in children with postoperative occipitoparietal epilepsy are scarce, our neuropsychological case descriptions from 3–7 years postsurgery focused on visuospatial abilities and scholastic skills such as reading and arithmetic.
Eleven children with medically refractory epilepsy underwent surgical resection of the posterior region of the brain at the Fondation Ophtalmologique Rothschild de Paris (France) between 1999 and 2007 (Figs S1 and S2). Children were selected based on age at surgery (≤7 years of age), current age (between 8 and 12 years), and postoperative status (seizure-free or greatly improved). Five children fulfilled the inclusion criteria and agreed to participate in the study. The project was approved by the Paediatric Neurosurgery Committee of the Fondation Ophtalmologique de Rothschild. Both parents and children gave informed consent. We reviewed preoperative and postoperative clinical investigations, which included intracranial video-electroencephalography (video-EEG), magnetic resonance imaging (MRI), and IQ findings from Wechsler or developmental scales. Clinical information for each patient is detailed in Table 1 and in the individual case descriptions.
Table 1. Patients’ descriptions
|1/F||Birth||Partial||Daily||Right||Taylor-type/focal dysplasia||Right/Occ-Par||5.9||Right occipital pole, cuneus until postcentral sulcus, total lingual gyrus, fusiform gyrus, posterior part of the angular gyrus||1/3 anterior angular gyrus and lateral occipital gyrus||Yes||Carbamazepine||None||Regular school, with special teaching|
|2/M||3 months||Partial/IS||Weekly||Left||Taylor-type focal dysplasia||Left/Par-Occ||0.7||Left occipital, mainly cortical resection: lateral occipital gyrus, small part of the superior cuneus||Fusiform and lingual gyri||Subtle seizures||Vigabatrin||Carbamazepine||Regular school with support for caligraphy|
|3/M||20 months||Partial||Daily||Right||Taylor-type focal dysplasia||Left/Par-Occ||6.2||Left occipital lobe, precuneus except upper part of superior parietal lobule, lingual gyrus, fusiform gyrus||Angular and supramarginal gyri||Yes||Carbamazepine + Topiramate||None||Normal school with, support for writting and mathematics|
|4/F||8 months||Partial||Daily||Left||Taylor- type focal dysplasia||Left/Occ-Par-Temp||3.2||Left occipital pole, internal face, cuneus, inferior half of the precuneus, lateral occipital gyrus, posterior part of angular gyrus||Supramarginal gyrus, post central gyrus, superior half of the superior parietal gyrus||Yesa||Carbamazepine||Oxcarbazepine||2 years lag with special academic support|
|5/M||Birth||Partial||Daily||Right||Taylor-type focal dysplasia||Left/Occ||7||Left occipital pole, cuneus, precuneus except upper part of superior parietal lobule, lingual and fusiform gyri||1/3 anterior of the fusiform, angular and lateral occipital gyrus, superior part of superior parietal lobule||Yes||Carbamazepine + Clobazam + Stiripentol||None||Normal school with support for reading and writing|
The follow-up period ranged from 3–7 years. All children underwent a neuroophthalmologic and comprehensive neuropsychological evaluation, designed to assess cognitive functions related to the posterior region of the brain. Verbal and nonverbal intellectual functions were evaluated preoperatively and postoperatively on Wechsler or developmental scales (Table 2). Postoperatively, more extensive neuropsychological and ophthalmologic evaluations were performed. The neuroophthalmologic evaluation included visual acuity and the Goldmann visual field examination. We used standardized tests to assess the following five subdomains of nonverbal cognition: basic visual discrimination, visual attention, visual recognition, spatial perception, and praxis. In addition, we evaluated working memory, executive functions, and academic skills such as reading and numerical cognition.
Table 2. Pre- and postsurgery IQ scores
Screening tests from the Visual Object and Space Perception (VOSP) battery (Warrington & James, 1991) were used to assess the basic visual discrimination level of each child. We measured visual attention with cancellation tasks: the NEPSY Visual Attention Task (Korkman et al., 1998) and the Bell test from the Batterie d’évaluation de la négligence unilatérale du Geren (BEN) (Rousseaux et al., 2001). Subtests for eye and head deviation, extinction, and line bisection were used from the Batterie to assess symptoms of neglect. Visual recognition was tested through naming and designation of objects (Jambaqué & Dellatolas, 2000), discrimination of overlapping and embedded figures (Ghent, 1956), and recognition of degraded figures (Gestalt closure) (Kaufman & Kaufman, 1983). For facial recognition, we used the short form of the Benton Facial Recognition Test (Benton, 1983a).
We evaluated spatial perception using the NEPSY Arrows subtest (Korkman et al., 1998) and the Benton Judgement of Line Orientation Test (JLOT) (Benton, 1983a; Benton et al., 1983b), with children’s norms (Lindgren & Benton, 1980). To measure visuoconstructive abilities and praxis, we used the Rey Complex Figure Test (Rey, 1959) and NEPSY Hand Movements and Visuomotor Precision subtests.
Reading was assessed by subtests from two batteries: the Test de l’Alouette (Lefavrais, 1967) and the Outil de Dépistage des Dyslexies (ODEDYS) (Neurocognition, 2002). In both cases, accuracy and speed were measured.
Numerical cognition was evaluated by the Zareki-R battery of scores (2006), which includes assessment of visual counting, mental counting, number writing, number reading, oral mental calculation, visual quantity estimation, and verbal contextual estimation.
Visual span and working memory were tested with the Wechsler Intelligence Scale for Children – Fourth Edition (WISC-IV) Integrated Spatial Span subtest. Auditory attention and working memory were tested further by comparison to visual tasks, using the NEPSY Auditory Attention and the WISC-IV Integrated Digit Span tests. Executive function was assessed through the NEPSY Tower subtest, and behavior was evaluated on the Child Behaviour Check List (CBCL) questionnaire (Achenbach, 1991) filled out by the parents.
Standardized neuropsychological test scores were classified into two categories, according to the mean and standard deviation (SD) of each test: 1.5 SD below the mean was qualified as a “deficit,” whereas 1.0 SD above or below the mean were qualified as “average.” With some exceptions, a pathological threshold was used: for the Benton Facial Recognition Test, this was more than two points from the mean (Benton), and for the embedded figures task, two omissions (Ghent, 1956). For the scoring of the BEN subtests (Bell test, Eye and Head Deviation, Extinction and Line Bisection), young adult norms were used as a guideline (Rousseaux et al., 2001). Finally, T scores above 60 on the CBCL questionnaire (Achenbach, 1991) were rated pathological.
Preoperative data and type of surgery
All five children experienced seizure onset in the early months of life, characterized by various types of clinical ocular manifestations. The seizures were stereotypic partial seizures beginning with a conjugate gaze deviation or a strabismus followed by ocular clonus; one child (Patient 2) presented infantile spasms, which were well controlled by Vigabatrin. MRI lesions were visible in the occipital lobe in all children (Patients 1–5) as well as in the parietal lobe in three (Patients 1–3) and in the temporoparietal junction in one (Patient 4). Only one child, Patient 1, had a right hemisphere lesion. In four children, between 20 and more than 100 seizures were recorded during the period of intracranial video-EEG monitoring, whereas one patient, Patient 2, presented with only two seizures. The preoperative IQ scores revealed a range from mild deficit (Patient 1) to mental delay (Patients 3 and 5), and autistic features in one patient (Patient 4). All children exhibited poor visual perception and recognition in everyday life. The posterior part of the occipitotemporal gyrus was included in the occipital and posterior parietal lobe resections (Figs S1 and S2) in all participants but one (Patient 2) (Tables 1 and 2) .
Postoperatively, all children improved their intellectual abilities, but with significant differences between verbal and nonverbal components (see Table 2 and detailed description). Although all nonverbal IQs were abnormal and three children had a verbal IQ below 80, two had verbal IQs above 100. Parental rating of behavior (on the CBCL questionnaire) was average for four of the five children. Nevertheless, behavioral difficulties were observed during our evaluations and are reported in the case descriptions.
All participants displayed normal visual discrimination (VOSP). However, all performed at least 1.5 SD below the mean on object naming (DEN-48), visuomotor praxis (hand movements), and regular word reading (ODEDYS-time) tasks (Table 3). Four of the five children performed at least 1.5 SD below the mean on visual attention tests (Bell and NEPSY), visual perceptual integration skills (degraded forms Kaufman Assessment Battery for Children [K-ABC]), embedded figure discrimination (Ghent), face recognition (Benton), visuospatial orientation (Benton JLOT and NEPSY Arrows), visuoconstructive praxis (Rey figure), visuomotor precision (NEPSY), and irregular word reading and pseudo-word reading time (ODEDYS) (Table 1).
Table 3. Patients showing deficit or average score on the different scales
| Visual attention task (NEPSY, Bell test)||1, 2, 4, 5||3|
| Symptoms of neglect (BEN)|
| Extinction||2, 5||1, 3, 4|
| Line bisection (BEN)||1||2, 3, 4, 5|
| Object recognition (Jambaqué et Dellatolas)||1, 2, 3, 4, 5|| |
| Embedded figures (Ghent)||1, 2, 3, 4||5|
| Gestalt closure (K-ABC)||1, 2, 3, 4||5|
| Face recognition (Benton)||1, 3, 4, 5||2|
|Spatial perception and praxies|
| Visuospatial orientation (Arrows and Benton JLOT)||1, 3, 4||2, 5|
| Visuoconstructive (copy of the Rey complex figure)||1, 3, 4, 5||2|
| Visuomotor precision (NEPSY)||1, 2, 4, 5||3|
| Imitation of hand movements (NEPSY)||1, 2, 3, 4, 5|| |
| Text reading (Lefavrais)a||1, 4, 5||2, 3|
| Word and pseudo-word reading (ODEDYS) regular word (errors)||5||1, 2, 3, 4|
| Irregular word (errors)||4, 5||1, 2, 3|
| Regular word (time)||1, 2, 3, 4, 5|| |
| Irregular word (time)||2, 3, 4, 5||1|
| Pseudo-word (errors)||4||1, 2, 3, 5|
| Pseudo-word (time)||1, 2, 4, 5||3|
| Calculation (Zareki-R) number reading||3||1, 2, 4, 5|
| Counting||3||1, 2, 4, 5|
| Number writing||3||1, 2, 4, 5|
| Visual estimation||4||1, 2, 3, 5|
| Verbal estimation|| ||1, 2, 3, 4, 5|
On the other hand, only Patient 4 did not show normal auditory attention and planning abilities (NEPSY Auditory Attention and NEPSY Tower, SD < −1.5). The two children with normal verbal IQ (Patients 1 and 2) performed above the mean on the planning task. Numerical abilities such as number reading and writing, counting, and verbal and visual estimation were within average range for four of the five children (exception, Patient 3).
Case descriptions of participants with normal verbal IQs
Patient 1. Experienced seizures at birth with ictal and interictal right occipital EEG activity. Seizures were characterized by upward left elevations of the eye globes and ocular clonus. After the second neurosurgery, at age 5½ years, the child became seizure-free without need for medication, although upward left ocular movements were still present at random instances. The visual field test showed left hemianopsia. This girl was 9.7 years old at the time of the neuropsychological evaluation.
Visual attention and recognition
The patient exhibited left hemianopsia with upward left eye deviation irregularly present. Behaviorally, she explored her visual field mainly with her right eye and showed visual attention deficits. On the line-bisection task, she exhibited left visual neglect, and was also impaired on object recognition, with poor visual perceptual integration performance on object naming and Gestalt closure tests. The patient was highly impaired when face recognition involved lighting and angle variation between the target and responder, contrasting her perfect performance on identical front-view photographs.
Spatial perception and praxis
Spatial perception was markedly impaired (Benton JLOT and the NEPSY Arrows task). On the copy of the Rey figure, she could not easily estimate shape dimensions or distance between elements. She showed also dyspraxia, with marked difficulty in imitating hand movement positions, and poor visuomotor precision abilities.
This patient was significantly slow at reading single words and text, and made more mistakes reading pseudo-words and regular words than irregular words. On the other hand, she was not significantly impaired on numerical cognition. She could count items, count backwards mentally, read and write numbers, do oral mental arithmetic (addition, subtraction, multiplication), and estimate visually and verbally contextualized quantities.
Patient 2. Seizures were characterized by left eye and head deviations and ocular clonus, associated with infantile spasms and psychomotor regression one month after the initial presentation. The preoperative neurologic examination showed mild axial hypotonia, poor socialization skills with impaired visual contact, if any, and an inability to catch objects. The right eye was affected with strabismus. After neurosurgery and introduction of low-dose carbamazepine, the patient showed amazing development of language and interaction skills.
Visual attention and recognition
At 8 years old, this boy showed right inferior quadranopsia, hypermetropia, astigmatism, and strabismus. On hemineglect tasks, he showed extinction signs in the left ear, whereas the line-bisection test was significantly skewed to the right. On the Bell cancellation task, he omitted more stimuli in the left hemispace than the right. Furthermore, he exhibited visual inattention with disorganized visual searching abilities. Object recognition was impaired and even more so when the forms were degraded or embedded. On the Benton Facial Recognition Test, he was slow and made mistakes on the matching of identical front-view photographs.
Spatial perception and praxis
Spatial perception was relatively preserved but the child was greatly impaired on visuomotor precision, calligraphy, and the imitation of hand positions.
Reading assessment showed that word reading was slower than expected, but error rates were normal or better than average. Furthermore, text reading performance was within normal range. The child showed normal numerical cognition.
Case descriptions of participants with abnormal verbal IQ
Patient 3. This patient was 19 months old at epilepsy onset. The EEG showed ictal and interictal left occipitoparietal abnormal activity. At 5 years of age, this left-handed boy had permanent deviation of the right eye and tended not to sufficiently use the right side of his body. He exhibited language delay but was characterized by evident visual agnosia and dyspraxia. At 8 years 8 months of age, the patient still showed a lack of behavioral control.
Visual attention and recognition
Ophthalmologic examination showed a right eye deviation, which was irregularly present. Assessed through cancellation tasks, he exhibited visual attention difficulties with a similar number of omissions on each side of the hemispace. In addition, he showed significant symptoms of apperceptive agnosia. He could recognize identical figures, but object recognition, especially for degraded and embedded figures, was altered. He made mistakes on both semanticolexical (e.g., stairs for ladder) and visual (e.g., caterpillar for snail) levels. On the Benton Facial Recognition Test, the child’s performance was also highly impaired.
Spatial perception and praxis
Visuospatial orientation abilities as well as constructive abilities were markedly defective.
Reading abilities were relatively preserved for text. The boy showed only a 1-year lag in reading speed for regular, irregular, and pseudo-words. Unlike other children in our sample, he was significantly impaired on numerical cognition, although, especially number reading and writing. Many mistakes were visuospatial, for example, reading 305 as 135, but other errors involved a misrepresentation of numerical knowledge (reading 138 as one thirteen eight; reading 6,485 as 74,895). Visual and verbal contextual estimation abilities, however, were normal.
Patient 4. The first neurologic examination found psychomotor delay and abnormal eye movements. Seizures occurred with upward right elevations of the eye globes and body, and the EEG showed ictal and interictal left occipital activity. Postoperative MRI revealed blurring of the white–gray matter junction over the left posterior temporal lobe, which extended anteriorly to the resection. The patient was 11 years of age at the time of the neuropsychological evaluation.
Visual attention and recognition
Ophthalmologic examination confirmed strabismus, hypermetropia, and right hemianopsia. On cancellation tasks, the patient omitted a large number of items on the left and right fields but was still able to explore the entire visual field. She showed extinction phenomena only in the tactile modality. This girl is the only child in our sample displaying aural–verbal as well as visual attention deficits. Visuospatial memory and verbal working memory were both impaired. Object recognition was altered (naming and designation), mainly for degraded forms and embedded figures. Object-naming errors were mainly semantic (e.g., gripper named as screwdriver). On the other hand, she could perfectly recognize front-view photographs of the face but was very much impaired when skew and lighting conditions changed.
Spatial perception and praxis
The patient showed impaired visuospatial abilities on the Benton JLOT and the NEPSY Arrows task. The copy of the Rey figure was segmented and she could not estimate the dimensions of shapes or the distances between elements. The child exhibited marked impairment during hand movement and visuomotor precision tasks.
This girl showed the greatest reading deficits of all the participants, as she was significantly slow at reading single words and text. Furthermore, the number of mistakes was similar whether she was reading regular words, irregular words, or pseudo-words. She made visual, phonologic, and regularization mistakes. On the other hand, she was not significantly impaired on numerical cognition. She could count items, read and write numbers, and estimate verbally contextualized quantities. However, the visual quantities estimation task was riddled with attention difficulties; in fact, this patient showed hyperactivity and lack of behavioral control.
Patient 5. Preoperatively, the EEG revealed ictal and interictal left parietooccipital abnormal activity. Seizures occurred with left upward eye movements and ocular clonus. At the age of 2 years, the child had mild psychomotor delay and interacted poorly with his mother. Behavioral difficulties related to food, with signs of anorexia. Two neurosurgeries were needed to obtain freedom from seizures. The boy was 10 years and 8 months old at the time of the neuropsychological evaluation.
Visual attention and recognition
The ophthalmologic examination showed right hemianopsia, but the patient compensated well. Although he showed right visual extinction when stimulated bilaterally, he had a lesser level of visual attention deficit than the other children in the sample. Visuospatial working memory was within average. However, this boy could not name objects properly despite preserved visual perceptual integration abilities. He could recognize front-view photographs of faces, but was impaired when lighting and skew varied between the target and the responder.
Spatial perception and praxis
The patient showed well-preserved visuospatial orientation abilities but poor spatial construction abilities. He also was affected with dyspraxia, poor imitation of hand positions, and visuomotor imprecision.
This patient was one of the most impaired in reading skills, being significantly slow at single words and text. Errors were made while reading pseudo-words, regular words, and irregular words, making visual mistakes 28% of the time. On the other hand, numerical cognition tasks were easily performed. Assessed by NEPSY tests, auditory attention was within average, but digit span and verbal working memory were diminished.
All children experienced seizure onset before 2 years of age. After surgery, three children were seizure-free and off medication, whereas the remaining two showed infrequent seizure reoccurrence and received one antiepileptic drug. Four of five children attained normal schooling level. Verbal IQ was improved by up to 35 points after the surgery. As in children with Sturge-Weber syndrome with drug-resistant epilepsy, the control of epilepsy and reduction of pharmacotherapy may have prevented severe cognitive impairments (Sujansky & Conradi, 1995). Three to seven years postoperatively, all children showed significant discrepancies between verbal and nonverbal IQs. Although two of five children displayed normal verbal IQ, all were moderately impaired on nonverbal IQ. Our observations are in concordance with the significant postsurgery difference between verbal and nonverbal IQ reported in adults with posterior resections, regardless of side of surgery or visual field loss (Luerding et al., 2004).
Our neuropsychological evaluation suggests that a history of early severe epilepsy with a focal cortical dysplasia followed by surgical resection will have an impact on the corresponding visual, spatial, and academic cognitive abilities. All five children showed cognitive deficits related to visual attention, object recognition, praxis, and reading speed.
Visual attention and recognition
The posterior parietal lobe is considered to have a special role in spatial functions (Nachev & Husain, 2006). Indeed, visual spatial attention was impaired in all five children. Subregions such as the temporoparietal junctions are involved in the spatial attention network (Mesulam, 1981; Vallar et al., 1993). They are strongly associated with neglect syndrome, an attention deficit disorder that causes patients to demonstrate pathologic bias for ipsilesional items, neglecting the contralateral ones (Mattingley et al., 1997). The one child with a right-sided lesion (Patient 1) was the most impaired on spatial attention and showed signs of neglect. Our observations are consistent with those of recent studies reporting that the right temporoparietal junction, part of which had been resected in Patient 1, is related to the network involved in reorienting behaviors following attentional capture by other stimuli (Corbetta & Shulman, 2002; Mitchell, 2008). In addition, neglect has recently been described in children with left and right parietal acquired lesions (Laurent-Vannier et al., 2003). In our sample, only Patient 1 with a right occipitoparietal lesion showed neglect of the contralateral hemispace. With regard to developmental lesions, our results do not confirm the previously prevailing hypothesis of more frequent spatial neglect of the right hemispace in very young children (Laurent-Vannier et al., 2003; Trauner, 2003).
Most children were impaired on degraded object recognition or object naming. Object recognition is known to involve the lateral occipital and the fusiform cortex bilaterally (Eger et al., 2004, 2007; Pegna et al., 2004) and to be modulated by top-down processes by the frontoparietal network. Children were unable to recognize or name an object without contextual cues, which suggests that recovered brain processes related to seizure freedom and/or brain plasticity failed to completely compensate for the unilateral posterior lesions. Interestingly, the one child who could accurately identify degraded objects had the lateral occipital cortex (LOC) preserved out of his left lesion (Patient 5), whereas the analogous LOC-sparing surgery in a right hemisphere lesion still impaired the child on the same task (Patient 1). The right hemisphere is typically associated with spatial transformation of a mental Gestalt, a process likely to be involved in degraded object recognition tasks (Warrington & James, 1991; Humphreys et al., 1992). Our observations, therefore, confirm the bilateral LOC role in object recognition.
Facial recognition, tested using the Benton Facial Recognition Test, was impaired in four of the five children. The task is separated into two parts. Only the first part does not require visuospatial skills, as it shows front-view photographs; the second part shows the angles and lighting in a different view. The lesions and cognitive difficulties experienced by the subject are reflected in corresponding deficits on the first or second part of the task. For example, the only child showing normal performance on orientation tasks (a patient with a small lateral resection in the left hemisphere), made mistakes on the first section but very few on the second. In contrast, the most impaired child on orientation tasks, with a median right hemisphere resection, showed a perfect performance on the first part, but a worse performance on the second part of the test. Other reports on facial perception have revealed limited postlesional plasticity in the pediatric population resulting from both right and left lesions (Mancini et al., 1994; de Schonen et al., 2005). Our study supports the heterogeneity of the mechanisms related to developmental prosopagnosia (Duchaine & Nakayama, 2006).
Spatial perception and praxis
All children showed dyspraxia, ranging from movement imitations to constructions. Movement imitation is thought to involve a complex network in the frontal and parietal regions (Iacoboni et al., 1999; Shmuelof & Zohary, 2007). Authors have proposed that the anterior intraparietal area is involved in a kinesthetic copy of the movements to be used during action execution (Iacoboni et al., 1999), whereas the inferior parietal lobule is associated with body schema knowledge (Berlucchi & Aglioti, 1997). Although the anterior intraparietal area was not altered in most of our children, the epilepsy and the resections, with the inferior parietal lobe partially resected in most children, must have produced a disconnect in the neural network that prevented them from performing accurate gesture imitations. As expected, visuomotor precision and complex figure copying were impaired, being associated with a number of brain areas in the frontal, parietal, and occipital lobes bilaterally, such as the cuneus (Tzagarakis et al., 2009), which had been resected in most children in our sample.
All children were impaired with regard to reading speed. Reading is a complex task involving an extended network. Word recognition has recently been strongly associated with the ventral stream, where a functional specialization for recurrent spelling patterns was found in the left occipitotemporal sulcus, named the visual word form area (VWFA) (Cohen et al., 2003; Marinkovic et al., 2003; McCandliss et al., 2003; Binder et al., 2006; Gaillard et al., 2006). The dorsal stream is thought to be involved in word recognition, when serial reading ensues from degraded special properties of words or spatial rearrangement of letters (Gitelman et al., 2005; Vinckier et al., 2006; Cohen et al., 2008). Grapheme–phoneme conversion and semantic processing of words involve left temporal and frontal regions of the brain (Simos et al., 2002; Mechelli et al., 2005).
The two most impaired children had lesions spread over the left ventral stream (Patients 3 and 5). Patient 3 had an extensive left dysplasia and Patient 5 had a resection of the left fusiform gyrus (including the VWFA) at 7 years of age. Indeed, plasticity may have been limited in Patient 5 by virtue of age and the fact that he had acquired a level of reading expertise before the surgery. The two children with the least impaired reading speed had very distinct lesions. Patient 4 had extensive left dysplasia combined with a resected fusiform gyrus, whereas Patient 2 had a small neurosurgery with preserved left fusiform gyrus. Cohen et al. (2004) reported hemispheric shift for reading but not for language in a child of preschool age at the time of left occipital resection. In Patient 4, the large left hemispheric dysplasia, the early resection (6 years of age), and early seizure freedom and reduction of pharmacotherapy could have resulted in brain reorganization toward the right hemisphere allowing for a certain level of reading. In all cases, it is possible that visual field loss may have altered the saccade movements, causing a reduction in reading speed. Inefficient fixation during reading has been described in hemianoptic patients 6 months after the acquired lesion (McDonald, 2006). Although it should be considered for the two least impaired children, who mainly showed alteration in reading speed instead of accuracy, this alternative explanation cannot hold true in the most impaired children because both accuracy and reading speed were impaired.
Although the impact of developmental lesions and neurosurgery did not create alexia in any of the children, it did impair all of them on word reading speed. Parietal regions of the brain are involved in spatial attention processes, orienting and decoding letters and words, and consequently in early learning of lexical functions. Epelbaum et al. (2008) showed that disconnecting the occipital cortex from the VWFA (by altering the inferior longitudinal fasciculus) was sufficient to provoke reading deficits. Therefore, it would seem that up to 7 years postsurgery, a right or left posterior lesion of the extent seen in our subjects prevents the child from performing at normal reading capacity at school age, but does not prevent the individual from learning how to read.
Numerical cognition refers to processes such as subitizing (estimation for quantities lower than four), estimating, counting, calculating, and comparing quantities or numbers. Studies have demonstrated the role of a frontoparietal network for numerical cognition (Roland & Friberg, 1985; Dehaene et al., 1996). Within this network, the intraparietal sulcus regions, bilaterally, are specifically involved in most numerical cognition processes. In children, comparison tasks of Arabic numerals and color nonnumerical symbols induced bilateral intraparietal sulcus activity in normal children but not in children showing developmental dyscalculia (Mussolin et al., 2010). Only one child in our sample was impaired on numerical cognition, the one in whom the largest part of the superior parietal lobule was resected (Patient 3). The child showed deficits in number reading, number writing, and dot counting. He did not show deficits on visual estimation, whereas the one child demonstrating difficulties on the estimation tasks had an occipitotemporal lesion (Patient 4). A plausible explanation is that the estimation task was modulated by spatial working memory (Culham & Kanwisher, 2001), which was altered in the latter (Patient 4) but not the former (Patient 3).
Side of surgery
Our sample allowed us to directly compare two children with very early epilepsy onset and lesions of the occipital-temporoparietal region, one in the left hemisphere (Patient 5) and one in the right (Patient 1). Both children were impaired as to reading, but showed a double dissociation in two aspects. First, although Patient 5 was significantly impaired on irregular word reading, Patient 1 showed normal performance. The lexicosemantic path necessary for irregular word reading is thought to involve the anterior part of the inferior frontal gyrus and of the fusiform gyrus (Mechelli et al., 2005), which had been resected in Patient 5 but not in Patient 1. Lesioned in the left hemisphere, Patient 5 was greatly impaired in lexical knowledge. In contrast, Patient 1 was able to directly read with the left hemisphere, where lexical knowledge is well developed. Second, although Patient 1 showed significant pseudo-word reading errors, Patient 5 had normal performance. Words are first processed within each hemisphere in the V4 area and then transferred through the corpus callosum to the left hemisphere, in the VWFA, around the temporooccipital sulcus (Cohen et al., 2000). Pseudo-word reading may further involve the dorsal premotor cortex (Mechelli et al., 2005). Our results suggest that pseudo-word reading may in fact involve other networks, including the right hemisphere. Both the right and left hemisphere fusiform gyri process newly presented words (Dehaene et al., 2001), and the posterior fusiform area has been shown to correlate with the frontal region involved in pseudo-word reading (dorsal premotor cortex) (Mechelli et al., 2005). We posit a right hemisphere contribution in the early detection of pseudo-words that could mimic serial reading processes known to activate biparietal regions (Vinckier et al., 2006; Cohen et al., 2008).
In summary, our sample was too small to quantify cognitive outcome with respect to dysplastic tissue, localization, resection size, time of surgery, seizure freedom, and pharmacotherapy reduction. However, the visuospatial impairment profile of our sample can be related to the neuroanatomic framework. We observed that the one child (Patient 2) showing the highest nonverbal IQ score was also the one with the smallest dysplastic tissue as visualized on MRI, had the least amount of resection, and was operated on at the youngest age. Indeed, this patient showed a very good outcome considering the classic long-term prognosis of infantile spasms (Jambaqué et al., 1993). Furthermore, the child showing the worst overall results (Patient 4), including deficits in oral–verbal attention, also showed the most extended dysplasia, involving the left temporal lobe. The other four patients may also show residual nonepileptogenic dysplastic tissue not obvious on MRI. In these cases, the dysplastic tissue may or may not be cognitively functional (Janszky et al., 2000). Patient 1, the only child with a right occipitoparietal resection, was the most impaired as concerns visual attention, visuospatial orientation, simultagnosia, and visuospatial construction abilities. In contrast, Patient 5, with a very similar resection of the left occipitoparietal region, showed the least visual attention and visuospatial working memory difficulties, but was the most impaired on reading. Finally, the only patient in whom the resection involved the left occipital-temporoparietal region (Patient 3), including the superior parietal lobe, was the only one to show numerical cognition impairments and had relatively preserved reading abilities.
From our small series, we can only convey that a number of variables including dysplasia size, resection size, and location are important in predicting verbal and nonverbal cognitive outcomes. However, given that verbal IQ improvements occurred after surgery, we can ascertain a beneficial role for seizure freedom and antiepileptic drug reduction in verbal brain processing. The mechanisms responsible for the visuospatial outcomes remain unclear. Whereas performance IQ did not significantly change after the surgery, the level of competence of each child is still striking when one considers the lesion sizes in some of the children. Possible recovery factors in visuospatial tasks may include distinct cognitive strategies, brain reorganization, and normalized brain activity resulting from seizure freedom and reduced medication.
One limitation of this study includes the fact that the more extensive neuropsychological evaluation was conducted only postoperatively. This limitation comes from the young age of the patients included in this study. For future work, more comprehensive preoperative assessment could help in better characterizing their improvements or worsening after surgery.
In conclusion, early surgical intervention in children with posterior intractable epilepsies offers the possibility of optimizing cognitive outcome and quality of life. We believe that, in future, more attention should be paid to neuropsychological assessment of these patients to better understand their visual perceptive deficiencies and provide better clinical and psychological management.
This work was supported by the Canadian Institutes of Health Research (Lippe S) and Unité de Neurochirurgie Pédiatrique, Fondation Ophtalmologique A. de Rothschild, Paris, France Inserm UMR663. Epilepsies de l’enfant et plasticité cérébrale, Université Paris Descartes, Paris, France.
We confirm that we have read the journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. None of the authors has any conflict of interest to disclose.