We examined postoperative changes in the memory function of 68 people with unilateral hippocampal sclerosis (HS) who underwent epilepsy surgery and had no previous clinical memory impairments. One in four with right HS (RHS) and one in five with left HS (LHS) in our sample of 323 people with unilateral HS performed within normal limits on memory tests that are sensitive to hippocampal pathology in group studies. People with intact memory function prior to surgery were significantly younger and had higher IQs than those with memory impairments. The majority of those with intact memory functions prior to surgery had significant postoperative declines on the memory measures at the 1-year follow-up and no longer functioned within the average range. People with RHS and left LHS were equally at risk of a postoperative decline. Postoperative deterioration was not related to seizure outcome. A higher IQ appears to protect against postoperative memory decline in this group. This finding does not support the notion that intact memory function in people with unilateral HS represents a migration of memory function to the contralateral structures, and thus protection from surgical insult. People with good preoperative memory function in the context of unilateral HS should therefore be counseled regarding the likelihood of a significant postoperative decline in memory function following a temporal lobe resection.
Unilateral hippocampal sclerosis (HS) is associated with memory deficits in people with epilepsy. These deficits have been correlated with the extent of hippocampal pathology, with greater pathology associated with more extensive memory deficits demonstrated by magnetic resonance imaging (MRI) studies of hippocampal volume, and cell counting studies using resected tissue (Baxendale, 1995). In terms of surgical outcome, people with greater pathology tend to show less memory decline following surgery than those who have less damaged structures removed. This is thought to be due to the fact that most people with extensive hippocampal pathology have already “lost” much of the memory function associated with the structure prior to surgery (Helmstaedter et al., 2011).
There is a small subset of people who have clear unilateral HS but perform within normal limits on tests of learning and memory that are usually sensitive to such damage. It is often difficult to know how to counsel these people prior to epilepsy surgery regarding any likely neuropsychological outcomes, since removal of extensive pathology is not generally associated with significant decline, but good postoperative function is. It is sometimes assumed that the “good” memory function seen on baseline preoperative testing must have migrated to the contralateral side, and thus will be preserved following surgery; however, this is difficult to establish. A “pass” on the Wada test is not sufficient to establish this, as the test is only designed to screen for amnesic risk and is methodologically unsuited to reliably predict how good someone's memory will be after surgery (Baxendale, 2009). Functional MRI memory paradigms show much promise in this area (Bonelli et al., 2010; Binder, 2011), but at present they remain largely in the research domain with respect to predicting the extent of memory integrity for individuals following surgery.
The aim of this study was to examine postoperative changes in the memory function of people with clear unilateral HS and function within the normal range prior to surgery. We aimed to answer the following questions.
What proportion of people with unilateral HS function in the normal range on memory tests that are usually sensitive to hippocampal pathology?
What happens to their memory function following surgery?
What factors differentiate those who decline from those who maintain memory functions following epilepsy surgery in this subset of people?
The participants were selected from a database of >750 people with epilepsy who have undergone epilepsy surgery at our center over the past two decades and who have been followed up at 1 year postsurgically. People with unilateral HS evident on their MRI scan were included, according to standard qualitative and quantitative criteria (Woermann et al., 1998). MRI scans were obtained on a 1.5 T GE scanner prior to June 2004 and on a GE 3T scanner since. People with bilateral HS or dual pathologies—that is, HS plus other abnormalities—evident on the scan were excluded from the study (with the exception of patients with HS plus cerebellar atrophy or nonspecific white matter lesions) A total of 323 participants, 146 people with right HS (RHS) and 177 people with left HS (LHS) fulfilled these inclusion criteria.
All participants underwent a neuropsychological assessment as part of their presurgical assessment. Measures of cognitive function included in this study were the Verbal IQ and Performance IQ indices of intellectual function from the Wechsler Adult Intelligence Scales. Four measures of memory from the UK Adult Memory & Information Processing Battery (AMIPB) and its successor the Birt Memory & Information Processing Battery were included.
On the list learning task the subject is read a list of 15 words and asked to recall as many as possible over five consecutive trials, leading to a possible maximum of 75 items. We used this as our index of Verbal Learning. On the design learning task the person is shown a design with nine features on a structured grid for 10 s and then asked to reproduce it on a blank grid. Again there are five learning trials leading to a maximum of 45 points. We used this as our index of Visual Learning. The immediate and delayed scores (40 min) from a prose recall task were also used in this study. We used these scores as our indices of Initial Registration and Verbal Retention, respectively. All of these tests have been described previously and have been reported to be sensitive to hippocampal pathology (Baxendale et al., 1998, 2006).
Participants were deemed to have intact memory functions if their scores on all four of these indices of memory function (Verbal Learning, Visual Learning, Initial Registration, Verbal Retention) fell at or above the 25th percentile on the standardized norms for each test. Clinically significant declines in function were determined using cutoffs based on reliable change indices at 80% (RCIs), in a normal healthy population for each task (Table 1).
Table 1. Average scores on the memory tests for the intact versus impaired groups and published norms for the test (controls)
Functional and structural incongruities prior to surgery
Twenty-one percent of the whole sample (n = 68) performed at an average level or above on all four indices of memory function prior to surgery indicating intact memory function. Thirty-four had LHS (representing 19% of those with LHS), and 34 had RHS (23% of those with RHS).
The intact group did not differ from the impaired group in terms of the extent of unilateral volume loss measured on MRI (Intact group, mean smallest-to-largest hippocampal ratio [Hvol ratio] = 71% [standard deviation, SD = 15.6]; Impaired group Hvol ratio = 70% [SD = 14.2], p > 0.05). Males and females were equally represented in both groups (χ2 = 0.05, d.f. = 1, p > 0.05).
Those with intact preoperative memory function had significantly higher Verbal IQs (t = −6.5, d.f. = 292, p < 0.001), and Performance IQs (t = −6.0, d.f. = 287, p < 0.001) than people with memory impairments prior to surgery and were younger at the time of surgery (t = 2.8, d.f. = 298, p < 0.01). The two groups did not differ in the age of onset of their seizures (t = −0.77, d.f. = 271, p > 0.05). Scores on the memory tests for both groups are presented in Table 1.
Of the 68 people who had intact memory function prior to surgery, 24 (n = 12 LHS; n = 12 RHS) maintained memory function in the average range or above at the 1-year postoperative follow-up. Patterns of postoperative memory function and deterioration, based on the number of tests on which the patients no longer score within the average range, are shown in Figure 1.
Participants who maintained memory functions in the average range did not differ from those who declined in their age at the time of surgery (t = 0.27, d.f. = 66, p > 0.05), age of onset of seizures (t = 1.15, d.f. = 66, p > 0.05), extent of HS (Declined group – Hvol ratio = 72% (SD = 16.8); Maintained Group – Hvol ratio = 69% (SD = 13.5), p > 0.05), or gender distribution (χ2 = 0.02, d.f. = 1, p > 0.05). People who maintained memory functions in the average range postoperatively, however, had significantly higher IQ scores (VIQ t = −1.78, d.f. = 64, p < 0.05; PIQ t = −1.89, d.f. = 63, p > 0.05).
Seventy percent of the whole patient group were classified as seizure-free at the 1-year postoperative assessment (International League Against Epilepsy [ILAE] Outcome Classification 1 and 2; Wieser et al., 2001). In the group as a whole, people with intact memory were not more likely to become seizure-free than those with memory impairments prior to surgery (χ2 = 0.14, d.f. = 1, p > 0.05).
In the group who had intact memory prior to surgery (n = 68), those who maintained memory function in the average range were not more likely to be seizure-free at the 1-year follow-up than those whose memory declined (χ2 = 0.06, d.f. = 1, p > 0.05).
Extent of memory decline
Forty-seven participants (69%) with intact memory functions prior to surgery experienced a significant postoperative deterioration of function, based on RCI cutoffs on one or more of the measures of memory function at the 1-year postoperative assessment. Fig. 2 illustrates the number of tests (maximum 4) on which the patients scored below the RCI at the 1-year postoperative assessment. Two individuals experienced a global postoperative decline in function, with a significant deterioration on all four measures of memory (n = 1 RHS; n = 1 LHS). Both were seizure-free following surgery.
Forty-seven percent of participants with impaired memory function experienced a significant postoperative deterioration of function on one or more of the measures of memory function at the 1-year postoperative assessment.
Between one fourth and one fifth of people with unilateral HS perform within normal limits on memory tests that have been shown to be sensitive to hippocampal pathology in group studies. These individuals tend to be younger and have higher IQs than those with memory impairments prior to surgery.
The majority of people with intact memory functions prior to surgery demonstrate significant postoperative declines on one or more of these memory measures following surgery and no longer function within the average range. People with RHS and LHS are equally at risk.
This finding does not support the notion that intact memory function in people with unilateral HS represents a migration of function to the contralateral structures in most. Our results suggest that intact memory functions prior to surgery are typically disrupted by the unilateral resection of a pathologic hippocampus. It is possible that preoperatively intact functions are subserved by adjacent ipsilateral cortex that is damaged during the surgery. Functional MRI (fMRI) studies are beginning to reveal the distribution of memory networks pre- and postoperatively, and efforts are underway to develop the clinical applications of this technology for the prediction of individual outcomes following surgery (Bonelli et al., 2010).
The finding that a higher IQ protects against postoperative memory decline, however, suggests that the maintenance of intact memory function pre- and postsurgically may also be the result of these people developing effective, intellectual compensatory strategies. More work is needed to investigate these possibilities. In the meantime, people with good preoperative memory function in the context of unilateral HS should be counseled regarding the high likelihood of a significant postoperative decline in memory function following a temporal lobe resection.
We are grateful to Prof. Matthew Walker and Prof John Duncan from UCL Institute of Neurology, for their support. JWS is supported by the Marvin Weil Epilepsy Research Fund. This work was undertaken at UCLH/UCL, which received a proportion of funding from the Department of Health's National Institute for Health Research Biomedical Research Centres' funding scheme.
None of the authors has any conflict of interest in relation to this work. 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.