Postictal But Not Interictal Hemispatial Neglect in Patients with Seizures of Lateralized Onset


Address correspondence and reprint requests to Dr. O. Prilipko at Sleep Clinic, Stanford University, 401 Quarry Road, Stanford, CA 94305-5730, U.S.A. E-mail:


Summary: Purpose: Unilateral spatial neglect, defined as a failure to report, respond, or orient to stimuli that are presented contralaterally, has been widely documented after brain damage to right, and to a lesser degree, left frontotemporoparietal networks. Group studies involving patients with seizures with a lateralized focus have demonstrated transient dysfunctions in memory and language; however, so far, only two case reports have described transient neglect after an epileptic seizure.

Methods: To assess the existence and consistency of this phenomenon, we evaluated 33 epilepsy patients on a line-bisection task in interictal and postictal states as compared with an age- and sex-matched control group.

Results: Spatial neglect, as determined by this test, was found in the postictal but not interictal examination in patients with right parietal epileptic foci and was maximal for the left-positioned lines, whereas no neglect was found in other groups.

Conclusions: Our findings indicate that patients with right parietal foci can present a transient neglect phenomenon on the line-bisection task in the postictal period, even in the absence of overt clinical neglect signs. These findings might be useful in establishing the laterality and even localization of epileptic foci based on the postictal neuropsychological evaluation.

Classically, when patients with pharmacoresistant epilepsy undergo presurgical evaluation of their disease before surgical intervention, a complete neuropsychological examination is carried out. This generally serves as a baseline for postsurgical follow-up. In the last decade, neuropsychology has proved useful in contributing to the lateralization, and even localization, of the epileptic focus, provided that a postictal evaluation is performed (Pegna et al., 1998).

The postictal course of cognitive deficits after seizures has been found transiently to follow the pattern of impairments observed in patients with permanent brain lesions. For example, temporal epileptic foci have been found to produce severe memory defects that are not observed in frontal lobe seizures (Helmstaedter et al., 1994) and that disappear within about 1 h. In the verbal domain, left hemispheric seizures have further been seen to produce longer-lasting language deficits than seizures of right hemisphere origin (Privitera et al., 1991).

Applying a similar line of reasoning, one would expect posterior right hemispheric foci to exhibit transient impairments akin to the longer-lasting deficits observed when these regions sustain permanent damage. Thus postictal deficits after seizures from right frontal and parietal foci would plausibly produce unilateral spatial neglect.

Cortical damage to right frontal, temporal, or parietal areas are known to produce a symptom termed unilateral spatial neglect (USN) (Chamorro et al., 1997; Karnath et al., 2001; Maguire and Ogden, 2002).

USN is defined as a failure to report, respond, or orient to stimuli that are presented contralaterally to a brain lesion, usually right parietal and frontal lobe lesions surrounding the frontal eye field, a failure that is not due to elementary sensory or motor disorders (Heilman et al., 1993). Emerging evidence indicates that neglect is a multicomponent syndrome, which can combine one or several features, probably depending on the extent and type of impaired networks (Heilman et al., 1997; Rafal, 1997; Mesulam, 1999; Bartolomeo and Chokron, 2001; Kerkhoff, 2001). The two major subtypes of neglect to be distinguished are the sensory or perceptual neglect and motor or intentional neglect. Different aspects of USN can be present simultaneously or in isolation (i.e., multisensory neglect or neglect of one sensory modality). It has also been shown that USN can affect near and far space differentially (Halligan and Marshall, 1991; Vuilleumier et al., 1998) and that the boundary between “near” and “far” space can vary with the use of tools (Berti and Frassinetti, 2000; Pegna et al., 2001). Finally, neglect can be present to a greater or lesser degree ranging from massive multimodal neglect to discrete unimodal extinction.

Several neuropsychological tests are used to determine the presence of USN. Among these, line bisection has been used extensively because of its high sensitivity (Schenkenberg et al., 1980). In this task, the patients are required to mark the center of a series of horizontal different-sized lines drawn at different locations on a sheet of paper. Patients with right hemisphere damage show a typical deviation of the subjective center to the ipsilesional side. Multiple studies have also been conducted on normal subjects on visual line-bisection tasks (Jewell and McCourt, 2000). A certain number of these studies have reported an interesting phenomenon of leftward bisection error, which was called “pseudoneglect” as opposed to the rightward bisection error seen in patients with left neglect. However, considerable between-study variability exists, because of factors such as subject age, hand used, line length, hemispace line position, and the direction in which subjects initiate motor and visual scanning. In particular, the direction of the line-bisection error seems to be influenced by the direction of visual-scanning initiation (i.e., subjects forced to scan visually from left to right tend to deviate to the left, whereas subjects scanning from right to left tend to deviate to the right) (Chokron et al., 1998).

In this study, we investigated the extent to which USN might be observed in the postictal neuropsychological examination of 33 patients with unilateral focal seizures by using a line-bisection task. Patients assessed at our presurgical epilepsy unit with unambiguously localized and unilateral focus were compared on their inter- and postictal productions on a line-bisection task. The results were further compared with a matched control population. We hypothesized that right hemispheric foci would produce greater deviations to the right in line bisection due to USN, whereas left hemispheric foci would produce only slight (or no) leftward deviation.



In total, 33 patients with unifocal pharmacoresistant epilepsy (15 left and 18 right; see Table 1 for details) were examined in our study. The patient group included 12 female and 21 male patients with a mean age of 30 years (range, 8–43 years). The younger subjects were not removed to maintain a sample representative of the population commonly seen in the clinical setting. All patients except one (ambidextrous) were right-handed. The focus was determined on the basis of long-term video-EEG monitoring, high-resolution structural MRI, interictal PET, interictal and ictal SPECT, as well as interictal and postictal neuropsychological evaluation. Each patient had undergone one interictal and one to four postictal line-bisection tasks, depending on the number of seizures (Table 1). Seventy-seven postictal evaluations were analyzed.

Table 1. Data of 33 patients
Patient numberSexAge (yr)Epileptic focusType of seizureLesion typeAge at onsetNumber of analyzed seizures
  1. CPS, complex partial seizure; SGTCs, secondarily generalized tonic–clonic seizures; SPS, simple partial seizure; HS, hippocampal sclerosis; DNET, dysembryoplastic neuroepithelial tumor; postop, after operation.

 1M40Temp RCPS, SGTCSHS 84
 2M33Temp RCPS, SGTCSHS 72
 3F43Temp RCPSLesional314
 4M27Temp RCPSHS 54
 5M28Temp RCPSHS211
 6M42Temp RCPS, SGTCSCryptogenic271
 7M41Temp RCPS, SGTCSPosttraumatic271
 8F43Temp RCPSHippocampal atrophy 51
 9F16Temp RCPSDNET121
10M 8Temp RCPS, SGTCSPachygyria, white matter abnormalities0,151
11M35Temp RCPS, SGTCSNonlesional301
12M40Temp RCPS, SGTCSBilateral HS304
14M39Temp LCPS, SGTCSCavernoma374
15M32Temp LCPS, SGTCSCryptogenic153
16M39Temp LCPS, SGTCSHS153
17M29Temp LCPS, SGTCSHS 71
18M26Temp LCPS, SGTCSPostop gliosis211
19M28Temp LCPSHS41
20F20Temp LCPSDNET182
21F26Temp LCPS, SGTCSPosttraumatic142
22F22Front RCPSDysplasia103
23F15Front LCPS, SGTCSCryptogenic104
24F29Front LCPSCryptogenic214
25F14Front LCPS, SGTCSCryptogenic113
26F16Pariet RCPS, absences, spasmsDysplasia 72
27F40Pariet RCPSDNET164
28M34Pariet RCPSDNET172
29M32Pariet RSPSPosttraumatic151
30M30Pariet RCPSCortical malformation123
31M41Pariet LCPS, SGTCSCryptogenic173
32F43Pariet LSPS, CPSPosttraumatic 81
33M12Pariet LCPS, SGTCSDysplasia 94

An age- and sex-matched group of 30 controls was selected by using the Edinburgh handedness questionnaire and a health questionnaire, addressing the presence or absence of neurologic, psychiatric, and other diseases; history of head trauma; learning difficulties; and actual medication. The controls were then evaluated with a line-bisection task in the same conditions as the patient's group.

Interictal and postictal evaluations in patients were carried out under electroencephalographic control. Postictal assessment was performed within 15 min to 1 h after the seizure and was initiated upon spatiotemporal reorientation.


The line-bisection task consisted of seven to 20 lines (seven for the control group) of different length distributed to the left, right, or center of the A4 paper sheet, which was placed in front of the participant, centered on the body axis. The line length varied from 98 to 198 mm. Patients and controls were instructed to bisect all the lines in their center as correctly as possible. No time limit, hand indication, or scanning direction was imposed. The deviation from the objective center of the line was measured in millimeters and assigned a positive value for rightward deviations and a negative value for leftward deviations. To minimize the effect of the line length, the deviation value was expressed relative to the total line length with a positive or negative sign (relative deviation value). Correctly bisected lines were thus given a zero value. The values obtained were classified according to the location (temporal, parietal, and frontal) and side (right and left) of the epileptic focus, time of testing (inter- and postictal) and spatial position of the line on the sheet (left, center, and right).

Statistical analysis:

We used a general mixed-model analysis of variance to take into account the unbalanced nature of the data and all the sources of variation of the data. Fixed effects were side, location, and position of the lines on the sheet; random effects included patients and seizures; and covariables were constituted by the interictal performance of each patient. This analysis was performed by using program 3V of the BMDP statistical software (BMDP Statistical Software, University of California. Release 7.0, 1990). First, patients in the interictal phase and controls were compared for deviations according to lines position on the sheet. Second, patients in the interictal phase were compared according to location and side of the epileptic focus and to spatial position of the lines on the sheet. Third, in the postictal phase, the performance of the epilepsy patients in all evaluations were compared according to the side and location of lesion, and for lines position on the sheet; deviation for each line was adjusted for corresponding mean deviation in the interictal phase (the covariate). Patients and seizures were considered random effects. When interactions were significant, subsequent analyses were performed separately for each location of the lesion (temporal, parietal, and frontal), and also separately for each side (left and right lesions).


Interictal performance

Patients versus controls

No significant differences were seen between controls and the whole group of epilepsy patients, nor was any interaction observed between these two groups and the position of the lines on the sheet. The only significant effect for both groups concerns the position of the lines on the sheet, with negative deviations for left and center lines and positive deviations for right ones (for control and all epilepsy patients, respectively: −9.5 and −3.9 for left lines, −3.2 and −4.7 for central lines, and 1.7 and 3.8 for right lines) (p < 0.0001).

Patients only

Significant interactions were seen between factors (location and side of focus and lines position); see Table 2B. A significant interaction (p = 0.0012) was seen for parietal group between side of foci and position of lines: with comparable negative deviations for right foci, but an important negative deviation for left lines and positive deviations for right lines in left foci.

Table 2. Statistical analysis of deviation values for the patients and controls in the interictal condition
A. Average of average values of relative deviations in controls and epileptic patients (interictal condition)
Lines positionControls N = 30Focus location
Left N = 3Right N = 5(4)‡Left N = 9Right N = 12(10)‡Left N = 3Right N = 1
  1. N = 5 (4)‡:5 Patients for central lines, but only 4 for left or right lines.

  2. N.B.: An average deviation of 10.0 corresponds to 1.0mm in a 100mm line, 2.0mm in a 200mm line.

Left−9.5−17.1−5.1 −10.8−7.5−6.3 12.8
Central−3.2 −2.0−8.0−8.9−3.5−8.9 8.0
Right 1.7 5.1−4.2−0.1 9.1 16.4−8.6
B. Mixed model ANOVA (interictal condition)
Subgroup includedMain effectsFirst order interactions2nd -order F.S.L
FocusSideLinePosF SF.LS.L
RightNANA  .0018
ParietalNANA  .0012

For the temporal group, no interaction was seen, but significant differences were seen concerning position of the lines: right lines were either not deviated or deviated to the right, whereas left and center lines deviated to the left.

For the frontal group, a significant interaction was seen (p = 0.037), left foci deviated to the left for left and center lines and to the right for right lines, whereas the right frontal patient deviated to the right for left and center lines and to the left for right lines (see Table 2A).

Postictal performance

Taking all postictal data into account, we found a highly significant third-order interaction (p < 0.0001, Table 3B) between side, location of the focus, and position of the lines in the sheet.

Table 3. Statistical analysis of deviation values for patients in the postictal condition
A. Average of average values of relative deviations in epileptic patients (postictal condition)
Lines positionFocus location
Left (3.8)‡Right (5.12)‡Left (9.18)‡Right (12.25)‡Left (3.12)‡Right (1.3)‡
Left−19.237.4−5.9−8.0−2.4 19.5
Central−13.9 5.1−6.4−2.45.3−7.2
B. Mixed model ANOVA(post ictal condition)
Subgroup includedCovar InterictalMain effectsFirst order interactions2nd -order F. S. L
FocusSideLinePosF. SF. LS. L
  1. (3.8)‡: 3 patients with a total of 8 crises.

  2. N.B.: An average deviation of 10.0 corresponds to 1.0mm in 100mm line, 2.0mm in a 200mm line.

Right  .0085NANA<.0001 
Temporal <.0001 .49.25.042 

For parietal foci, a significant interaction between the side of the focus and position of the lines was present (p = 0.009) (Table 3B). Left parietal patients had comparable negative deviations (−19.2, −13.9, −10.9 for left, center, and right lines, respectively), whereas right parietal patients had very different, usually small deviations (5.1 and −2.5) for center and right lines, but large positive deviation (37.4) for left lines (Table 3A).

For temporal foci, the interaction between side of the focus and position of the lines was marginally significant (p = 0.042). Again, left temporal patients had negative deviations (−5.9, −6.4, and −4.4 for left, central and right lines, respectively), whereas right temporal patients varied in the direction of their deviations (−8.0, −2.4, and 10.6 for left, central, and right lines).

For frontal patients, a highly significant interaction between the side of the focus and position of the lines was seen (p = 0.0004). For left frontal foci, we found a small deviation for all lines (−2.4, 5.3, and 1.4 for left, central, and right lines), whereas the right frontal patient deviated to the right for left and center lines, to the left for right lines in a similar manner to that in the interictal condition (19.5, −7.2, and −12.9 for left, central, and right lines) (Table 3A).


In this study, we investigated the presence of USN after seizures of right and left hemispheric onset. Line bisection was thus performed interictally and postictally in patients with temporal, parietal, or frontal foci in the left or right hemisphere.

We found no difference in line bisection between controls and the whole group of 33 patients (right and left hemisphere foci) in the interictal phase. However, a significant effect was found in subsequent subgroup analysis in patients with left-hemispheric foci, with a trend for leftward deviation for left and central lines, and a rightward deviation for right-positioned lines. The right temporal group tended to demonstrate the same pattern (Table 2A). In contrast, a significant interaction was seen between right and left parietal and frontal groups for focus localization and line position. The rationale for subdividing left, central, and right lines was made in light of previous studies that had suggested an influence of the role of the line position in healthy subjects (Jewell and McCourt, 2000). This meta-analysis performed on 73 studies of line-bisection tasks revealed that subjects displace the subjective center in the direction of the hemispace in which the line is situated and in the direction of the hand used to perform the bisections. However, other studies failed to replicate these findings (Reuter-Lorenz and Posner, 1990). Our results are in concordance with several previous studies on line bisection in healthy participants that describe a “centrifugal” pattern of error (Reuter-Lorenz et al., 1990; Milner et al., 1992; McCourt and Jewell, 1999). Results of previous studies suggest that the size of the line to bisect may have an influence on the performance in this task (Jewell and McCourt, 2000). Here we used lines of different lengths and computed the relative deviations of values to minimize an eventual effect of the line length.

Furthermore, the minimal line length in our study is too big to observe a “crossover effect” (i.e., neglect patients err too far to the right in a long line-bisection task but bisect short lines too far to the left), as this phenomenon has been described for lines of less than 3 cm (Rueckert et al., 2002).

In the analysis of the postictal condition, we considered the side of the focus, the location of the focus, and the position of the lines on the paper as we hypothesized that the parietal focus, particularly in the right hemisphere, should induce the most important line-bisection deviations. As expected, we found a highly significant interaction between line position and side of the focus for parietal and frontal groups, whereas this interaction was of borderline significance for the temporal group. Patients in the right parietal group and, to a lesser degree, the right frontal patient showed an impressive rightward deviation on left-positioned lines and almost no deviation on centered and right-positioned lines. The right frontal patient displayed a similar pattern on the interictal evaluation, whereas right parietal patients had a completely different interictal pattern with slightly negative deviations on all lines, suggesting a direct role of seizures on the line-bisection pattern change in the right parietal group.

No effect of focus location or line position was seen for patients with left hemisphere foci.

Our finding of transient USN in seizures of right parietal onset was maximal for lines situated on the left side of the page. This pattern has already been described by Schenkenburg et al. (1980) in a study on patients with permanent cerebral damage. In this study, patients with right hemispheric damage displayed a strong tendency to neglect lines on the left side of the page as compared with lines situated at the center or on the right.

Transient deficits in visuospatial function have been reported in a previous investigation involving patients with right and left temporal foci. Visuospatial but not verbal memory was found to decrease after right temporal seizures, whereas the reverse was observed after seizures of left temporal origin (Andrewes et al., 1990). However, this study did not seek specifically for USN. By contrast, and in line with our observations, transient USN was reported in two cases of patients with epileptic seizures. Heilman and Howell reported a case of a patient with right parietooccipital seizures who exhibited ictal tactile extinction for left-hand stimuli and had a leftward ictal line-bisection error (in the presence of a tonic head and eye deviation to the left) followed by a rightward postictal bisection error, which was present to a lesser degree in the interictal condition (Heilman and Howell, 1980). Meador and Moser reported a second case of an epilepsy patient with right centroparietal seizures who demonstrated ictal visual and tactile extinction, rightward line-bisection error, hemi-inattention on cancellation tasks, hemidyslexia, and hemispatial neglect in drawings, with no signs of interictal neglect (Meador and Moser, 2000).

These reports demonstrate that at least some epilepsy patients with right parietal foci exhibit transitory signs of neglect in the immediate postictal period.

In conclusion, our results confirm the existence of functional neglect in epilepsy patients induced by seizures originating in the right parietal region, which can be demonstrated with a line-bisection task and is most evident on left-positioned lines. This observation provides further evidence that seizure activity affects cognitive functions that depend on cerebral structures situated close to the epileptic focus. Thus transient postictal neuropsychological deficits can be observed that are not necessarily present in the interictal phase, which can be useful in determining the lateralization and localization of the epileptic focus (Regard et al., 1994).

Interestingly, in most of our patients, transient USN could be demonstrated only in the line-bisection task and was not clinically obvious, suggesting that neglect should be actively sought in these patients even in the absence of clinical signs or complaints.


Acknowledgment:  This study was supported by SNF grants 32-68105, 32-104146, and 32-109928. We thank Dr. Laurent Spinelli for technical support as well as Dr. Eugene Mayer and Dr. Stephanie Ortigue for their advice and kind support.