memory and intelligence outcome following surgery for intractable temporal lobe epilepsy:...

www.elsevier.com/locate/yebeh

Epilepsy & Behavior 6 (2005) 147–155

Memory and intelligence outcome following surgeryfor intractable temporal lobe epilepsy: relationship

to seizure outcome and evaluation using a customizedneuropsychological battery

Sujoy K. Sanyala, P. Sarat Chandraa,*, Surya Guptab, Manjari Tripathic, V.P. Singha,Satish Jainc, M.V. Padmac, V.S. Mehtaa

a Department of Neurosurgery, All India Institute of Medical Sciences, New Delhi 110 029, Indiab Department of Psychiatry, All India Institute of Medical Sciences, New Delhi 110 029, Indiac Department of Neurology, All India Institute of Medical Sciences, New Delhi 110 029, India

Received 12 October 2004; revised 21 November 2004; accepted 25 November 2004

Abstract

The main objectives of this prospective study were to (1) assess memory and intelligence outcome following surgery for intrac-table temporal lobe epilepsy, (2) correlate this with seizure outcome and side of surgery, and (3) perform (1) and (2) using an indig-enously developed battery customized to the Indian population. Prior to use in our epilepsy surgery program, the test–retest andinterexaminer variance reliability of this battery had been established in both normal and cognitively compromised populations.The memory scores were overall rather than material-specific. The battery was administered to right-handed adults undergoing sur-gery for intractable temporal lobe epilepsy without any evidence of opposite temporal lobe abnormality, both presurgery and post-surgery at a mean follow-up of 8 months. Twenty-five consecutive patients were included; 13 underwent right and 12 underwent lefttemporal surgery. Seizure outcome was assessed using Engel�s classification. Among 13 patients who underwent right temporal sur-gery, although 4 patients with poor seizure outcome had insignificant changes in scores, 7 of 9 patients with good seizure outcomeexhibited considerable (>20% over preoperative) improvement in their memory and intelligence scores. Statistical analysis using Stu-dent�s t test and the Mann–Whitney test revealed that the patients who underwent right temporal surgery with good seizure outcomehad significant improvement in both memory (P = 0.007) and intelligence (P = 0.043) scores compared with those with poor seizureoutcome. In contrast, patients who underwent left temporal surgery had no significant change in cognitive scores irrespective of sei-zure outcome. Cognitive improvement seems to occur in patients with good seizure outcome following nondominant temporal lobesurgery for intractable epilepsy with no evidence of pathology in the opposite temporal lobe. The same finding was not observed inpatients undergoing left temporal surgery.� 2004 Elsevier Inc. All rights reserved.

1. Introduction

Most neuropsychological batteries used in patientswith intractable temporal lobe epilepsy have focusedon evaluation of material-specific memory functions

1525-5050/$ - see front matter � 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.yebeh.2004.11.015

* Corresponding author. Fax: +91 011 265 88207.E-mail address: [email protected] (P. Sarat Chandra).

(learning and retention of verbal information for lefttemporal involvement vis-a-vis recall of nonverbal infor-mation for right temporal involvement) to assist in thedetermination of seizure lateralization/localization[1,2], as well as in the demonstration of postsurgicalchanges in these material-specific memory functions[11,15–30,32,35].

Now, however, attention is increasingly being fo-cused on the overall cognitive outcome following such

mailto:[email protected]

148 S.K. Sanyal et al. / Epilepsy & Behavior 6 (2005) 147–155

procedures, particularly because the temporal lobe is sointimately involved in the processes of memory. Littleliterature [42,43] exists on whether successful seizurecontrol, following surgery for intractable temporal lobeepilepsy, has a concomitant positive influence on overallcognition, especially on memory.

It is also important (particularly in Asian countries)to use a customized neuropsychological battery that ad-dresses the local customs and educational background,rather than the existing batteries designed for Westernpopulations.

Our study is unique for two reasons:

1. A customized neuropsychological battery in a locallanguage (Hindi), adequately validated and tested inthe normal population, was used in our epilepsy sur-gery program.

2. This study is one of the very few studies of its kindfrom the Asian subcontinent reporting cognitive out-come after surgery for intractable epilepsy from alarge tertiary referral center.

2. Aims and objectives

1. To assess cognitive outcome in the spheres of mem-ory and intelligence following surgery for intractabletemporal lobe epilepsy (using a validated customizedneuropsychological battery).

2. To correlate the cognitive outcome with seizure out-come and side of surgery (dominant/nondominant).

3. Materials and methods

This prospective study was conducted on patientsoperated on for medically intractable temporal lobe epi-lepsy from March 2001 to December 2002. The inclu-sion criteria for our study were: (1) nonmentallyretarded, right-handed adults (>15 years of age) under-going surgery for intractable temporal lobe epilepsy; and(2) no MRI/SPECT evidence of pathology in the oppo-site temporal lobe. We performed no test other thandetermination of right-handedness to establish left hemi-spheric dominance in our patients.

3.1. Presurgical evaluation

The presurgical evaluation of potential candidates in-cluded a detailed history and clinical examination, EEG,long-term video/EEG recording, and high-quality MRIscan (including T1-weighted inversion recovery se-quences and FLAIR sequences obtained in obliquecoronal planes 3 mm thick, perpendicular to the princi-

pal axis of the hippocampal formation and sensitive en-ough to detect subtle alterations in hippocampal size,morphology, and signal intensity). Gadoliniumenhancement was used if structural pathology was sus-pected. An interictal SPECT study was done using Tc-HMPAO. In doubtful cases, an ictal SPECT was doneusing Tc-ECD.

3.2. Surgical procedures

Formesial temporal sclerosis, a standard anteromedialtemporal lobectomy (3.5 cm of lateral neocortex from thepole if dominant and 5.5 cm if nondominant) with asubpial amygdalohippocampectomy was performed.For lesional temporal lobe epilepsy, lesionectomy underelectrocorticographic guidance was performed.

3.3. Seizure outcome assessment

After discharge, follow-up visits were scheduled at 2weeks, 1 month, 3 months, 6 months, and every 6months thereafter. Antiepileptic drugs were continuedfor at least a year and then gradually tapered.

Seizure outcome was assessed using Engel�s grading[3].

3.4. Battery used for cognitive evaluation

For neuropsychological evaluation, we used a cus-tomized battery drafted in Hindi, the local languagefor a large portion of the Indian population. We call thisbattery the AIIMS Comprehensive NeuropsychologicalBattery in Hindi (Adult Form) (AIIMS = All IndiaInstitute of Medical Sciences).

This battery comprises 160 items spread over two ba-sic scales of memory and intelligence, with a 5-pointcredit scale (from 0 to 4) for each item for evaluatingperformance, to enhance the precision of the battery.The items were allotted credits along the dimensionsof accuracy, frequency, speed, quality, and appropriate-ness. A rating of 0 indicates normal performance, 4indicates abnormal performance and 1–3 suggest inter-mediate performance.

The memory scale included recognition of birds andanimals shown on cards, items assessing remote and re-cent memory, immediate verbal recall of words, immedi-ate memory for digits both forward and backward,verbal recall with homogeneous interference, verbal re-call with heterogeneous interference, immediate recallof visually presented designs, nonverbal recall withhomogeneous interference, and nonverbal recall withheterogeneous interference. Memory scores are there-fore not material-specific, but rather are overall memoryscores.

The intelligence scale included items assessing generalknowledge, comprehension of social situations, ability

S.K. Sanyal et al. / Epilepsy & Behavior 6 (2005) 147–155 149

to explain popular proverbs, ability to determine oppo-sites, ability to find similarities and differences, ability toform analogies, logical reasoning for numbers, ability toidentify missing elements in a complex geometric config-uration, and a maze learning test.

Some of the items had to be answered within a partic-ular time frame, which meant that credits on these itemswere allotted along the dimension of speed. Immediatememory for digits was scored 1 to 5 depending on whatwas the maximum number of digits recalled as a string.Verbal recall of sentences was scored depending onaccuracy of recall. The appropriateness of reply alsodetermined the score for some items, for example, itemsassessing ability to explain common proverbs.

The test items had to be customized for the Indianperspective starting from the language in which theywere administered. Our proverbs are different, our geog-raphy is different, our calendar year is different, our so-cial situations are different. Also, although a significantportion of the population are illiterate, almost all areintelligent regarding matters concerning their immediateenvironment and profession. For example, a farmer,while illiterate, is quite aware of the kinds of crops tobe planted during a particular season. This obviouslynecessitated construction and utilization of a customizedbattery.

3.4.1. Development of norms for the battery

Prior to use in our epilepsy surgery program, this bat-tery was tested on a group of 175 normal controls. Themeans and standard deviations of scores attained by thisgroup were used to develop formulas for conversion ofraw memory and intelligence scores into t scores. Thegroup of normal controls (90 males, 85 females) con-tained individuals with no known neurological/psychiat-ric illnesses spread over all age groups (>15 years of age)and diverse educational backgrounds. As the impor-tance of the variables age and education in influencingtest scores in neuropsychological studies is well known,formulas were evolved to determine ‘‘expected t scores’’based on age and education (using regression analysison the same population of 175 normal controls). If thet score obtained for any patient was higher than the ex-pected t score, the performance was termed abnormal.Lower scores therefore denote better performance.

3.4.2. Reliability tests of the battery

The reliability of the battery in terms of test–retestand interexaminer variance had also been establishedprior to use in our epilepsy surgery program. On bothof these reliability tests, the reliability coefficients(0.926 and 0.967, respectively) were significant at theP = 0.001 level, which means that scores are stable overtime and there is negligible interexaminer variance inscoring the items. The time frame for the test–retest reli-ability determination was 4 months.

Thus constructed and validated, this battery was usedin our epilepsy surgery program.

3.5. Patient population

Twenty-five consecutive patients satisfying the inclu-sion criteria were included in this analysis: 13 underwentright temporal surgery and 12 underwent left temporalsurgery. Twenty-two underwent a formal anteromedialtemporal lobectomy (12 right, 10 left) and 3 underwentlesionectomy (1 right, 2 left).

3.6. Use of the battery in our program

The memory and intelligence scales of our customizedbattery were administered to these patients preopera-tively, as well as following surgery at a mean follow-up of 8 months. The tests were administered twice andat a minimum interval of 6 months from the time of pre-operative testing to ensure that there were no practice ef-fects. Moreover, because the test–retest reliability of thisbattery had been shown to be good at 4 months, wecould reasonably expect scores to be stable, in the ab-sence of any interventions, at 6 months.

4. Results

Statistical analysis was done using Student�s t test anda nonparametric test of significance, the Mann–Whitneytest, on SPSS software. P 6 0.05 was taken as statisti-cally significant.

As postoperative cognitive outcome was determinedwithin 1 year of surgery, there was no possibility of drugwithdrawal confounding the results because we startedwithdrawing drugs 1 year after surgery.

4.1. Patients undergoing right temporal surgery

Thirteen patients (Table 1) underwent right temporalsurgery. Twelve anteromedial temporal lobectomy (8 formesial temporal sclerosis (MTS), 1 for medial temporaloligodendroglioma, 2 for medial temporal gliosis, and 1for medial temporal cortical dysplasia). The other pa-tient underwent a lesionectomy for a right posteriortemporal cavernoma (patient 6).

Among the 8 patients with right MTS, 6 had good sei-zure outcome (5 Class I, 1 Class II) and 2 had poor sei-zure outcome (Class IV). Among the 6 patients withgood seizure outcome, 5 (patients 1–4 and 9) made con-siderable improvements in memory and intelligencescores (>20% over the preoperative score). Two of themhad preoperative abnormal (compared to the expectedscores) memory and intelligence scores, which normal-ized postoperatively (patients 2 and 9). Among thosewhose memory improved, the average improvement over

Table 1Memory and intelligence evaluation, right temporal lobe (n = 13)

Memory Intelligence Time of postoptesting (min)Preop t

scorePostop t

scoreExpected t

scoreDifference(preop–postop)

Preop t

scorePostop t

scoreExpected t


Class I seizure outcome (good outcome)1. MTSa,b 64 46 67 18 65 42 69 23 82. MTSb 79(abn) 59(norm) 67 20 71(abn) 47(norm) 66 24 83. MTSb 62 40 66 22 64 42 68 22 74. MTSb 60 42 65 18 60 44 67 16 95. MTS 47 41 69 6 39 39 72 0 76. Lesb 78(abn) 41(norm) 72 37 63 39 72 24 97. Lobb 60 41 67 19 62 45 68 17 78. Lobc 49 40 65 9 37 37 66 0 9

Class II seizure outcome (good outcome)9. MTSb 83(abn) 49(norm) 69 34 72(abn) 56(norm) 71 16 8Mean (SD) 20.33 (10.08) 14 (1.5)

Class IV seizure outcome (poor outcome)10. MTS 40 34 64 6 39 40 63 �1 711. MTS 41 36 66 5 40 40 64 0 812. Lob 53 56 71 �3 52 46 74 6 813. Lob 43 41 72 2 39 38 72 1 9

Mean (SD) 2.5 (4.04) 1.5 (3.1)

a MTS, mesial temporal sclerosis; Les, lesionectomy; Lob, lobectomy (AMTL); norm, normal score; abn, abnormal score.b Considerable improvement (>20% over preop score).c <20% improvement over preop score.


preoperative memory scores was 31% (range, 28–41%).Among the patients whose intelligence scores improved,the average improvement over preoperative intelligencescores was 30% (range, 22–35%). Both patients with classIV seizure outcome (Nos. 10 and 11) showed noimprovement in memory and intelligence scores.

Among the 5 who underwent surgery for lesionaltemporal lobe epilepsy (4 anteromedial temporal lobec-tomies [AMTLs] and 1 lesionectomy), 3 had Class I and2 had Class IV seizure outcome. Of the 3 with good sei-zure outcome, 2 (patients 6 and 7) improved consider-ably in memory and intelligence scores (>20% over thepreoperative score), while the third (patient 8) had an18% improvement in memory scores. Patient 6 had pre-operative abnormal memory scores that normalizedpostoperatively. The memory scores of patients 6 and7 improved considerably, 32 and 47% over preoperativescores, and their intelligence scores improved 27 and38%, respectively. (The patient who improved the most47 and 38% in memory and intelligence scores, respec-tively, had undergone only lesionectomy [patient 6].)In both patients with Class IV seizure outcome (patients12 and 13), memory and intelligence scores did not im-prove. Summarizing, among the 13 patients who under-went right temporal surgery, the changes in scores of 4patients with poor seizure outcome (Engel�s Class III/IV) were insignificant, whereas the memory and intelli-gence scores of 7 of 9 patients with good seizure out-come (Engel�s Class I/II) improved considerably(>20% over the preoperative score).

To demonstrate the relationship, if any, betweengood seizure outcome (Engel�s Class I/II) and improve-ment in memory and intelligence scores, we calculatedthe difference between the preoperative score and post-operative score (lower scores are better scores) for eachpatient with respect to good and poor seizure outcome.Thereafter, we calculated the mean difference (andstandard deviation) between the preoperative and post-operative memory and intelligence scores for patientsin both categories. We did not include absolute scoresin our analysis because they are influenced by otherfactors like education. Rather, we applied statisticaltests of significance on SPSS software to see whetherthe difference between preoperative and postoperativememory and intelligence scores differed significantlybetween those who had good seizure outcome andthose who had poor seizure outcome. This differencein scores is unlikely to be influenced by any factorother than seizure outcome, surgical procedure, andthe possibility of practice effects (which we feel areminimal, as explained earlier) being the same in bothgroups.

Statistical analysis using Student�s t test and theMann–Whitney test revealed that patients withgood seizure outcome after right temporal surgeryexhibited significant improvement in both memory(P = 0.007, Mann–Whitney U = 0.5) and intelligence(P = 0.043, Mann–Whitney U = 5) scores (Figs. 1and 2) compared with patients with poor seizureoutcome.

Fig. 1. Memory outcome following right temporal surgery. On the Y

axis are the memory t scores, and on the X axis, the number ofpatients. Both patients with good seizure outcome (left) and patientswith poor seizure outcome (right) are represented. Lower memory t

scores are better scores. The area marked in gray indicates theimprovement in memory t scores. As can be seen, memory scores ofpatients with good seizure outcome improved significantly (P = 0.007).Good seizure outcome = Engel�s Class I/II. Poor seizure out-come = Engel�s Class III/IV.

Fig. 2. Intelligence outcome following right temporal surgery. On theY axis are intelligence t scores, and on the X axis, the number ofpatients. Both patients with good seizure outcome (left) and patientswith poor seizure outcome (right) are represented. Lower intelligence tscores are better scores. The area marked in gray indicates theimprovement in intelligence t scores. As can be seen, intelligence scoresof patients with good seizure outcome improved significantly(P = 0.043). Good seizure outcome = Engel�s Class I/II. Poor seizureoutcome = Engel�s Class III/IV.


4.2. Patients undergoing left temporal surgery

Memory and intelligence were evaluated in 12 pa-tients (Table 2) who underwent left temporal surgery.Seven underwent left AMTL for MTS, and three under-went left AMTL for lesions including medial temporalgliosis, subependymoma, and oligodendroglioma. Twounderwent left temporal lesionectomies, one for a pilo-cytic astrocytoma and the other for a dysembryonic neu-roepithelial tumor (DNET).

Among the 7 patients with left MTS, 2 (patients 11and 12) had Class IV seizure outcomes, 1 (patient 8)had a Class III outcome, and the remaining 4 (patients1–4) had Class I outcomes. Among the 4 patients withClass I outcomes, only one patient (patient 1) had a30% improvement in memory score (this patient�s intel-ligence scores remained the same). The other 3 patientswith left MTS with Class I outcomes (patients 2–4)exhibited only modest improvements in memory and

intelligence scores ranging from 10 to 15%. One patientwith Class IV seizure outcome (though his seizures be-came less bothersome, occurring only at night) showedlittle improvement in an abnormal memory score, whichremained abnormal (patient 11), though his intelligencescore improved 25% over his preoperative score. Theother patient with Class IV outcome (patient 12) hadabout the same scores on follow-up. The patient whohad a Class III outcome (patient 8) had modestimprovements of 17 and 12% in memory and intelligencescores. Therefore, on the whole, the memory and intelli-gence profile of left MTS patients, like that of right MTSpatients, seems to improve with seizure control, thoughto a much more modest degree.

Among the five who underwent surgery for left tempo-ral lesional epilepsy, 3 had Class I seizure outcomes (pa-tients 5–7), and 2 had Class III outcomes (patients 9and 10). Among the 3 patients with Class I outcomes,memory and intelligence scores improved in 2 (patients5 and 7) and remained the same in 1 (patient 6). Themem-ory scores of patients 5 and 7 improved by 22 and 18%,while their intelligence scores improved by 6 and 5% (pa-tient 7 had undergone only temporal lesionectomy).Interestingly, the initially abnormal memory scores ofboth patients (5 and 7) normalized. Patients 9 and 10 withClass III seizure outcomes had almost the same memoryscores postoperatively (though patient 9 made an 8%improvement, which normalized his memory score). Thissame patient had a 34% improvement in intelligencescores. Therefore, similar to patients with left MTS, pa-tients operated on for left temporal lobe lesions also seemto improve their memory and intelligence profile with sei-zure control, though to a much more modest degree thanpatients undergoing right temporal surgery.

We again applied the same method as described ear-lier to examine whether there was any difference in thechange in memory and intelligence scores postopera-tively between those who had a good seizure outcomeand those who had a poor seizure outcome. Analysis re-vealed that there was no statistical difference betweenthese two groups (good and poor seizure outcome) asfar as changes in memory scores (P = 0.251) and intelli-gence scores (P = 0.12) are concerned (Figs. 3 and 4).

Statistically significant improvements in memory andintelligence scores with good seizure outcome were ob-served only in patients undergoing right temporal sur-gery. It is necessary to reiterate at this stage thatmemory scores in our series were overall memory scoresand not material-specific memory scores aimed at evalu-ating solely the memory functions of the lobe resected.

4.3. Overall seizure outcome

In our epilepsy surgery program, we achieved overallClass I seizure outcome rates in 76% of patients withMTS, 73.5% of patients with lesional temporal lobe

Table 2Memory and intelligence evaluation (left temporal) (n = 12)

Memory Intelligence Time of postoperativetesting (min)Preop t

scorePostop t

scoreExpected t


Preop t

scorePostop t

scoreExpected t


Class I seizure outcome (good outcome)1. MTSa,b 57 40 69 17 44 46 72 �2 92. MTSc 67 59 67 8 77 67 68 10 83. MTSc 67 60 68 7 68 60 70 8 74. MTSc 48 42 66 6 49 42 68 7 95. Lobc 82(abn) 64(norm) 70 18 70 66 72 4 76. Les 44 40 70 4 43 41 73 2 87. Lesc 78(abn) 64(norm) 69 14 63 60 72 3 8Mean (SD) 10.57 (5.65) 4.57 (2.3)

Class III seizure outcome (poor outcome)8. MTSc 59 49 73 10 69 61 75 8 99. Lobd 73(abn) 67(norm) 70 6 71 47 73 24 710. Lob 60 56 70 4 62 59 71 3 8

Class IV seizure outcome (poor outcome)11. MTSd# 80(abn) 75(abn) 71 5 67 50 72 17 812. MTS 52 50 69 2 54 51 70 3 8

Mean (SD) 5.4 (2.96) 11 (5.6)

a MTS, mesial temporal sclerosis; Les, lesionectomy; Lob, lobectomy (AMTL); norm, normal score; abn, abnormal score.b Considerable improvement in memory (>20% over preop score) without improvement in intelligence score.c <20% improvement over preop score.d Improvement in intelligence score without improvement in memory score.

Fig. 3. Memory outcome following left temporal surgery. On theY axisare memory t scores, and on the X axis, the number of patients. Bothpatients with good seizure outcome (left) and patients with poor seizureoutcome (right) are represented. Lower memory t scores are betterscores. The area marked in gray indicates the improvement in memory tscores. As can be seen, memory scores of patients undergoing lefttemporal surgery exhibit no significant postoperative change irrespec-tive of seizure outcome. Good seizure outcome = Engel�s Class I/II.Poor seizure outcome = Engel�s Class III/IV.

Fig. 4. Intelligence outcome following left temporal surgery. On the Yaxis are intelligence t scores, and on the X axis, the number of patients.Both patients with good seizure outcome (left) and patients with poorseizure outcome (right) are represented. Lower intelligence t scores arebetter scores. The area marked in gray indicates the improvement inintelligence t scores. As can be seen, intelligence scores of patientsundergoing left temporal surgery exhibit no significant postoperativechange irrespective of seizure outcome. Good seizure out-come = Engel�s Class I/II. Poor seizure outcome = Engel�s Class III/IV.


epilepsy undergoing AMTL, and 79% of patients withlesional temporal lobe epilepsy undergoing onlylesionectomies.

5. Discussion

An important issue raised by most studies evaluatingcognitive outcome following surgery for intractable tem-poral lobe epilepsy relates to the functional adequacy of

the lobe being resected, and it has been established thatfunctional inadequacy of the lobe to be resected in con-trast to functional adequacy of the lobe not to be re-sected is crucial to a good cognitive outcome [4–14].The functional adequacy of the temporal lobes can beevaluated directly with the Wada test [4,33]. Studies[5–10,34] using structural/functional neuroimaging forindirect evaluation of temporal lobe function have con-clusively established that patients with imaging evidenceof pathology in the unoperated temporal lobe, as well as


patients with lesser evidence of pathology in the oper-ated temporal lobe, are at risk of memory deterioration,thereby also proving that imaging evidence of pathologyis a good indirect indicator of the functional status ofthe temporal lobes.

Davies et al. [11] showed that patients who hadundergone left AMTL without significant sclerosis inthe resected hippocampus had a greater risk of deterio-ration in the ability to encode new verbal information,probably because the resected hippocampus was func-tionally reasonably adequate. Hermann [12], Seidenberg[13], and Westerveld [14] and their co-workers also dem-onstrated similar findings.

For the same reasons as mentioned above, we in-cluded only those patients in our study with no obviouspathology in the opposite temporal lobe.

Overall, many of the literature reports regardingpostoperative cognitive outcome after temporal lobec-tomy have tended to demonstrate objective declines inmaterial-specific verbal memory scores followingAMTL of the dominant lobe [11,17–25]. However, moreinconsistent, material-specific visuospatial memory de-cline after a nondominant temporal lobe resection hasbeen also demonstrated [18,20–23,26–28].

With respect to intelligence, Dlugos et al. [29] re-ported significant language-related cognitive declinesafter left temporal lobectomy, whereas no decline wasobserved in patients undergoing right temporal lobec-tomy. Ivnik et al. [30] demonstrated that although sum-mary IQ scores were maintained in the average range,left AMTL patients were less effective in language-de-pendent cognitive tasks postoperatively, whereas rightAMTL patients were either unchanged or slightly bettercognitively after the operation.

Though the above studies seem to suggest that tem-poral lobe surgery for intractable temporal lobe epilepsymay carry a high risk of cognitive morbidity, studiesincluding ours evaluating overall cognitive functionsrather than material-specific functions have demon-strated a contrary finding.

Kubu et al. [31] described 10 patients who underwenttemporal lobectomy after failing the Wada test bilater-ally, despite having only a unilateral onset of their sei-zures. Contrary to what would be expected from thebilateral Wada failure, the majority of postoperativememory scores were consistent with or better than thoseobtained preoperatively, though some material-specificdeclines were noted consistent with the expected effectsof the temporal lobectomy. To explain this result, Kubuet al. furthered the hypothesis that it is possible that anepileptogenic mesial temporal lobe can give rise to revers-ible cognitive dysfunction involving the contralateral me-sial temporal structures (possibly due to the fact thatthere is an obvious direct connection between the twotemporal lobes through the anterior commissure andthe corpus callosum) [36], leading to failure of the Wada

bilaterally. However, over a longer period after removalof the primary epileptogenic lesion (and the constant neg-ative electrophysiological influence of the primary epilep-togenic lesion), functioning of the contralateral mesialtemporal structures could improve, possibly to the nor-mal level. Such a phenomenon has support in other clin-ical observations [37–41], including improved cognitivefunctioning after hemispherectomy in infantile hemiple-gia. Cavazutti et al. [39] noted improvement in cognitivefunctioning of the contralateral temporal lobe after ipsi-lateral removal of temporal lobe tumors presenting withepilepsy. We also feel that the removal of the deleteriousinfluence of seizures may lead to cognitive improvement,as two of the three patients in our series who underwentonly temporal lesionectomy exhibited significantimprovement in their overall memory and intelligencescores following Class I seizure outcomes.

Cavazutti et al. [39], Meier and French [44], Augus-tine and Novelly [45], Blakemore and Falconer [46],Olivier [47], Rausch and Crandall [21], Selwa et al.[25], and Mayanagi et al. [48] demonstrated that intelli-gence scores increase after AMTL, especially after non-dominant resection, although they did not demonstrateany relationship to seizure control. However, Dodrillet al. [49], Rausch and Crandall [21], Wachi et al. [42],Leonard [43], Martin et al. [52], Robinson et al. [51],and Bjornaes et al. [50] demonstrated a relationship be-tween postoperative improvement in intelligence scoresand seizure control following AMTL.

The literature seems to be more noncommittal regard-ing improvement of memory in contrast to intelligence,following good seizure outcome after AMTL. Robinsonet al. [51], Rausch and Crandall [21], and Martin et al.[52] found that while patients with good seizure controlshowed significant improvements in intelligence quo-tients, they had memory deficits of the type associatedwith side of temporal lobe resection. However, theyfound that right-sided surgery patients experiencedimprovement in verbal memory functions, which wasmost likely due to the relief from the continuous electricalstimuli of intractable epilepsy. But they could not demon-strate any improvement in overall memory followinggood seizure outcome. Wachi et al. [42], however, foundthat the postoperative improvement of seizure-free pa-tients on the Wechsler�s Adult Intelligence Scale—Re-vised (WAIS-R) and Wechsler Memory Scale—Revised(WMS-R) was statistically significant. Leonard [43] dem-onstrated a similar effect of good seizure outcome onintelligence and delayed verbal memory.

If we combine the results of all these studies with ourfindings, what emerges is:

1. Functional inadequacy of the operated temporal lobeand functional adequacy of the nonoperated tempo-ral lobe are crucial to a good cognitive outcome.The direct way of evaluating temporal lobe function


is the Wada test [4]. However, preoperative structuraland functional neuroimaging can provide a fair ideaof the same [5–10].

2. Demonstration of objective decline in material-spe-cific memory scores, especially episodic verbal mem-ory scores, and also of language-dependentcognitive functions, especially after dominant AMTL[11,17–30], have raised the specter of cognitive mor-bidity associated with temporal lobectomy. However,these studies demonstrating deterioration in memoryhave evaluated mainly material-specific episodicmemory after AMTL and thereby cannot be com-pared with our study, which measures overall mem-ory scores, which provide a better idea of a patient�soverall functional status.

3. There is in all probability a role for continuing epi-lepsy in reversibly compromising cognitive function-ing of patients with intractable epilepsy [31,37–41].We feel that seizure control may lead to cognitiveimprovement, because two of the three patients inour series who underwent only lesionectomies andno removal of surrounding brain exhibited significantimprovement in their overall memory scores follow-ing good seizure outcomes.

4. A good seizure outcome following surgery for intrac-table temporal lobe epilepsy has been demonstratedto translate into improved IQ scores in quite a fewstudies [21,42,43,49–52], including ours, though somestudies have demonstrated a definite decline in lan-guage-dependent cognitive scores after left AMTL.

5. Improvement in overall memory scores (as opposedto material-specific scores) has been demonstratedless frequently and has even been less correlated witha good seizure outcome [42,43]. On the other hand,the deterioration in material-specific memory scoresobserved after AMTL is likely to raise skepticismabout memory outcome following this procedure.We have, however, demonstrated that good seizureoutcome may lead to significant memory improve-ment among right-handed adults undergoing non-dominant temporal lobe surgery (with no evidenceof pathology in the opposite temporal lobe). We feelthat the presumably functional contralateral tempo-ral lobe, when freed from the deleterious influenceof intractable epilepsy, probably functions better,leading to cognitive improvement. However, goodseizure outcome did not translate into cognitiveimprovement for patients undergoing left temporalsurgery (though they had no deterioration), probablybecause the functional right temporal lobe in thesepatients has less reserve than the left, in terms ofmemory processing. Because we measured postopera-tive cognitive outcome within 1 year of surgery, thereis no possibility of antiepileptic drug withdrawal com-pounding the results as we started to withdraw drugs1 year after surgery.

6. Conclusions

Significant cognitive improvement seems to be associ-ated with good seizure outcome in right-handed adultsundergoing nondominant surgery for intractable tempo-ral lobe epilepsy with no evidence of pathology in theopposite temporal lobe. Good seizure outcome did nottranslate into cognitive improvement for patients under-going dominant temporal surgery (though they demon-strated no deterioration).

References

[1] Scoville W, Milner B. Loss of recent memory after bilateralhippocampal lesions. J Neurol Neurosurg Psychiatry1957;20:11–21.

[2] Milner B. Psychological aspects of focal epilepsy and its neuro-surgical management. Adv Neurol 1975;8:299–321.

[3] Engel Jr J. Outcome with respect to epileptic seizures. In: Engel JrJ, editor. Surgical treatment of the epilepsies. New York: RavenPress; 1987.

[4] Sabsevitz DS, Swanson SJ, Morris GL, Mueller WM, SeidenbergM. Memory outcome after left anterior temporal lobectomy inpatients with expected and reversed Wada memory asymmetryscores. Epilepsia 2001;42:1401–8.

[5] Pauli E, Eberhardt KW, Schafer I, Tomandl B, Huk WJ, StefanH. Chemical shift imaging spectroscopy and memory function intemporal lobe epilepsy. Epilepsia 2000;41:282–9.

[6] Trenerry MR, Jack Jr CR, Ivnik RJ, Sharbrough FW, CascinoGD, Hirschorn KA. MRI hippocampal volumes and memoryfunction before and after temporal lobectomy. Neurology1993;43:1800–5.

[7] Wendel JD, Trenerry MR, Xu YC, et al. The relationshipbetween quantitative T2 relaxometry and memory in nonlesionaltemporal lobe epilepsy. Epilepsia 2001;42:863–8.

[8] Rocchetta ID, Gadian DG, Connelly A, et al. Verbal memoryimpairment after right temporal lobe surgery: role of contralateral damage as revealed by hydrogen-1 magnetic resonancespectroscopy and T2 relaxometry. Neurology 1995;45:797–802.

[9] Rausch R, Henry TR, Ary CM, Engel Jr J, Mazziotta J.Asymmetric interictal glucose hypo metabolism and cognitiveperformance in epileptic patients. Arch Neurol 1994;51:139–44.

[10] Griffith HR, Perlman SB, Woodard AR, Rutecki PA, Jones JC,Ramirez LF. Preoperative FDG-PET temporal lobe hypo metab-olism and verbal memory after temporal lobectomy. Neurology2000;54:1161–5.

[11] Davies KG, Bell BD, Bush AJ, Wyler AR. Prediction of verbalmemory loss in individuals after anterior temporal lobectomy.Epilepsia 1998;39:820–8.

[12] Hermann BP, Wyler AR, Somes G, Berry 3rd AD, Dohan Jr FC.Pathological status of the mesial temporal lobe predicts memoryoutcome from left anterior temporal lobectomy. Neurosurgery1992;31:652–6.

[13] Seidenberg M, Hermann B, Wyler AR, Davies K, Dohan Jr FC,Leveroni C. Neuropsychological outcome following anteriortemporal lobectomy in patients with and without the syndromeof mesial temporal lobe epilepsy. Neuropsychology1998;12:303–16.

[14] Westerveld M, Sass KJ, Chelune GJ, et al. Temporal lobectomyin children: cognitive outcome. J Neurosurg 2000;92:24–30.

[15] Rausch R, Babb TL. Hippocampal neuron loss and memoryscores before and after temporal lobe surgery for epilepsy. ArchNeurol 1993;50:812–7.


[16] Hermann BP, Seidenberg M, Haltiner A, Wyler AR. Relationshipof age at onset, chronologic age and adequacy of preoperativeperformance to verbal memory change after anterior temporallobectomy. Epilepsia 1995;36:137–45.

[17] Martin RC, Sawrie SM, Roth DL, et al. Individual memorychange after anterior temporal lobectomy: a base rate analysisusing regression-based outcome methodology. Epilepsia1998;39:1075–82.

[18] Chelune GJ, Naugle RI, Luders H, Sedlak J, Awad IA. Individualchange after epilepsy surgery: practice effects and base-rateinformation. Neuropsychology 1993;7:41–52.

[19] Ojemann G, Dodrill C. Verbal memory deficits after left temporallobectomy for epilepsy: mechanism and prediction. J Neurosurg1985;62:101–7.

[20] Novelly RA, Augustine EA, Mattson RH, et al. Selectivememory improvement and impairment in temporal lobectomyfor epilepsy. Ann Neurol 1984;15:64–7.

[21] Rausch R, Crandall PH. Psychological status related to surgicalcontrol of temporal lobe seizures. Epilepsia 1982;23:191–202.

[22] Ivnik RJ, Sharbrough FW, Laws ER. Anterior temporal lobec-tomy for the control of partial complex seizures: information forcouncil patients. Mayo Clin Proc 1988;63:783–93.

[23] Lee GP, Loring DW, Thompson JL. Construct validity ofmaterial specific memory measures following unilateral temporallobectomy. Psychol Assess 1989;1:192–7.

[24] Szabo CA, Wyllie E, Stanford LD, et al. Neuropsychologicaleffect of temporal lobe resection in preadolescent children withepilepsy. Epilepsia 1998;39:814–9.

[25] Selwa LM, Berent S, Giordani B, Henry TR, Buchtel HA, RossDA. Serial cognitive testing in temporal lobe epilepsy: longitudi-nal changes with medical and surgical therapies. Epilepsia1994;35:743–9.

[26] Saykin AJ, Gur RC, Sussman NM, O�Connor MJ, Gur RE.Memory deficits before and after temporal lobectomy: effects oflaterality and age of onset. Brain Cogn 1989;9:191–200.

[27] Barr WV, Chelune GJ, Hermann BP, et al. The use of figuralreproduction tests as measures of nonverbal memory in epilepsysurgery candidates. J Int Neuropsychol Soc 1997;3:435–43.

[28] Phillips NA, McGlone J. Grouped data do not tell the wholestory: individual analysis of cognitive change. J Clin ExpNeuropsychol 1995;17:713–24.

[29] Dlugos DJ, Moss EM, Duhaime AC, Brooks-Kayal AR.Language related cognitive declines after left temporal lobectomyin children. Pediatr Neurol 1999;20:444–9.

[30] Ivnik RJ, Sharbrough FW, Laws Jr ER. Effects of anteriortemporal lobectomy on cognitive function. J Clin Psychol1987;43:128–37.

[31] Kubu CS, Girvin JP, McLachlan RS, Pavol M, Harnadek MCS.Does the intracarotid amobarbital procedure predict globalamnesia after temporal lobectomy? Epilepsia 2000;41:1321–9.

[32] Jokeit H, Ebner A, Holthausen H, Markowitsch HJ, Tuxhorn I.Reorganization of memory functions after human temporal lobedamage. NeuroReport 1996;7:1627–30.

[33] Dade IA, Jones-Gotman M. Sodium amobarbital memory tests:what do they predict? Brain Cogn 1997;33:189–209.

[34] Goldman-Rakic PS, Funahashi S, Bruce CJ. Neocortical memorycircuits. Cold Spring Harb Symp Quant Biol 1990;55:1025–38.

[35] Ojemann G, Dodrill CB. Verbal memory deficits after lefttemporal lobectomy for epilepsy. J Neurosurg 1985;62:101–7.

[36] Zola-Morgan S, Squire LR. Neuroanatomy of memory. AnnuRev Neurosci 1993;16:547–63.

[37] McFie J. The effects of hemispherectomy on intellectual func-tioning in cases of infantile hemiplegia. J Neurol NeurosurgPsychiatry 1961;24:240–9.

[38] Smith A, Sugar O. Development of above normal language andintelligence 21 years after left hemispherectomy. Neurology1975;25:813–8.

[39] Cavazutti V, Winston R, Baker R, Welch K. Psychologicalchanges following surgery for tumors in the temporal lobe. JNeurosurg 1980;53:618–26.

[40] Glosser G, Saykin AJ, Deutsch GK, O�Connor MJ, Sperling MR.Neural organization in material specific memory functions intemporal lobe epilepsy patients as assessed by the intracarotidamobarbital test. Neuropsychology 1995;9:449–56.

[41] Perrine K, Gershengorn J, Brown ET, Choi IS, Luciano DJ,Devinsky O. Material specific memory in the intracranial amo-barbital procedure. Neurology 1993;43:706–11.

[42] Wachi M, Tomikawa M, Fukuda M, et al. Neuropsychologicalchanges after surgical treatment for temporal lobe epilepsy.Epilepsia 2001;42(Suppl. 6):4–8.

[43] Leonard G. Temporal lobe surgery for epilepsy: neuropsycholog-ical variables related to surgical outcome. Can J Neurol Sci1991;18(4, Suppl.):593–7.

[44] Meier MJ, French LA. Longitudinal assessment of intellectualfunctioning following unilateral temporal lobectomy. J ClinPsychol 1966;22:22–7.

[45] Augustine EA, Novelly RA. Memory and IQ correlates of seizurecontrol following ATL. Epilepsia 1981;48:317–26.

[46] Blakemore CB, Falconer MA. Long-term effects of ATL oncertain cognitive functions. J Neurol Neurosurg Psychiatry1967;30:364–7.

[47] Oliver A. Risk and benefit in the surgery of epilepsy: complica-tions and positive results on seizure tendency and intellectualfunction. Acta Scand 1988;72(Suppl. 117):114–21.

[48] MayanagiY,WatanabeE,NagahoriY,NankaiM. Psychiatric andneuropsychological problems in epilepsy surgery: analysis of 100cases that underwent surgery. Epilepsia 2001;42(Suppl. 6):19–23.

[49] Dodrill CB, Hermann BP, Rausch R, et al. Neuropsychologicaltesting for assessing prognosis following surgery for epilepsy. In:Engel Jr J, editor. Surgical treatment of the epilepsies. NewYork: Raven Press; 1993. p. 263–71.

[50] Bjornaes H, Stabell KE, Heminghyt E, Roste GK, Bakke SJ.Resective surgery for intractable focal epilepsy in patients withlow IQ: predictors for seizure control and outcome with respect toseizures and neuropsychological and psychosocial functioning.Epilepsia 2004;45:131–9.

[51] Robinson S, Park TS, Blackburn LB, Bourgeois BF, Arnold ST,Dodson WE. Transparahippocampal selective amygdalohippo-campectomy in children and adolescents: efficacy of the procedureand cognitive morbidity in patients. J Neurosurg 2000;93:402–9.

[52] Martin P, Maestu F, Sola RG. Effects of surgical treatment onintellectual performance and memory in a Spanish sample ofdrug-resistant partial onset-temporal lobe epilepsy patients. Sei-zure 2002;11:151–6.

memory and intelligence outcome following surgery for intractable temporal lobe epilepsy:...

Documents