Effect of Focus Lateralization on Memory Assessment during the Intracarotid Amobarbital Procedure

BRAIN AND COGNITION 33, 224–241 (1997)ARTICLE NO. BR970894

Effect of Focus Lateralization on Memory Assessmentduring the Intracarotid Amobarbital Procedure

ISABELLE ROULEAU AND JANIK ROBIDOUX

Service de Neurologie, Hopital Notre-Dame, Montreal, Canada, and Laboratoire deNeuroscience de la Cognition, Universite du Quebec a Montreal, Montreal, Canada

AND

RAYMONDE LABRECQUE AND CAROLE DENAULT

Service de Neurologie, Hopital Notre-Dame, Montreal, Canada

Despite the use of stimuli that can be processed by both hemispheres, a numberof studies have reported lower memory scores after the left intracarotid amobarbitalprocedure (IAP) than after the right IAP. Because of that, failure after ipsilateralIAP is observed more often in patients with a left temporal seizure focus (LT) thanin right temporal patients (RT), possibly needlessly excluding some LT patientsfrom surgery. In order to overcome the deleterious effects of anesthetizing the domi-nant hemisphere, we designed an IAP protocol that did not promote verbal encodingof the stimuli. For this purpose, a large number of visual and tactile stimuli (coloredpictures and real objects) were presented to be recognized later. The effect of seizurefocus lateralization was examined in 82 temporal lobe epileptic patients who under-went IAP as part of their presurgical evaluation. As expected, for both RT and LTpatients, long-term recognition of pictures presented under the effect of amobarbitalwas highly sensitive to the presence of a contralateral epileptic focus. However,contrary to what is generally reported, LT patients performed better than RT patientswhen their left (ipsilateral) hemisphere was anesthetized. In RT patients, althoughmemory scores were lower after the left contralateral injection, the disparity in mem-ory scores between the right and left injection was not as marked as in LT patients.These results are discussed in terms of the influence of type of processing requiredduring the initial encoding on later recognition during IAP. 1997 Academic Press

Although the intracarotid amobarbital procedure (IAP) is commonly usedto assess memory functions prior to temporal lobectomy, there is no consen-

We are grateful to Jean-Marc Saint-Hilaire, Normand Giard, and Michele Sammaritano,neurologists, and Jacques Lesage and Daniel Roy, neurorodiologists, for their collaboration.We thank Francois Richer, Ph.D. for his critical comments on earlier drafts of the manuscript.This work was supported by a FRSQ clinical research fellowship to Dr. Rouleau. Addressreprint requests to Isabelle Rouleau, at the Lab. de Neuroscience de la Cognition, Universitedu Quebec a Montreal, Case postale 8888, Succursale Centre-ville, Montreal, Quebec H3C3P8, Canada.

2240278-2626/97 $25.00Copyright 1997 by Academic PressAll rights of reproduction in any form reserved.

FOCUS LATERALIZATION AND MEMORY 225

sus in the literature on the best procedure for assessing these functions (fora review, see Rausch, Silfvenius, Wieser, Dodrill, Meador, & Jones-Gotman,1993). An ideal test should be reliable enough to evaluate the risk of severepostoperative amnesia, as well as sensitive enough to indicate a lateralizeddysfunction. Despite differences in memory assessment, most authors agreethat performance is influenced both by the presence of contralateral dysfunc-tion and by the functional hemispherectomy produced by inactivation of onehemisphere. In clear-cut cases, injections contralateral to the focus shouldlead to very poor memory performance and injection ipsilateral to the focusshould produce good performance. This difference between performanceafter ipsilateral and contralateral injections is a fundamental aspect of thereliability of the IAP. In many centers, IAP memory disparity between ipsi-lateral and contralateral injection is successfully used to lateralize seizurefocus (Engel, Rausch, Lieb, Kuhl, & Crandall, 1981; Loring, Meador, Lee,Nichols, King, Gallagher, Murro, & Smith, 1995; Perrine, 1994; Perrine,Gershengorn, Brown, Choi, Luciano, & Devinsky, 1993; Perrine, West-erveld, Sass, Devinsky, Dogali, Spencer, Luciano, & Nelson, 1995; Powell,Polkey, & Canavan, 1987; Rausch, Babb, Engel, & Crandall, 1989; Sperling,O’Connor, Saykin, Phillips, Morrell, Bridgman, French, & Gonatas, 1992;Wyllie, Naugle, Chelune, Luders, Morris, & Skibinski, 1991a) and to predictmaterial-specific memory loss following unilateral temporal lobectomy(Chelune, 1995; Loring et al., 1995; Wyllie, Naugle, Awad, Chelune, Luders,Dinner, Skibinski, & Ahl, 1991b).

To assess memory functions during IAP, it is necessary to use materialthat can be processed adequately by the noninjected hemisphere. To comparethe competence of the two hemispheres, stimuli such as pictures or drawingsof common objects are included in most protocols; these stimuli are pre-sumed to be processed equally well by the dominant and the nondominanthemisphere (i.e., amenable to dual coding; Paivio & Csapo, 1972).

Despite the use of dually encodable stimuli, many authors have reportedlower memory scores following dominant hemisphere injections than whenthe nondominant hemisphere was anesthetized (Aasly & Silfvenius, 1990;Jones-Gotman, McMackin, Cendes, Andermann, Evans, Olivier, & Peters,1993; Loring, Meador, & Lee, 1992; Silfvenius & Aasly, 1991). Poorerscores were observed following dominant hemisphere injection even thoughretention was tested when aphasia was no longer present, and recognitionrather than recall was used to test retention. Furthermore, Rausch, Fedio,Ary, Engel, and Crandall (1984) have observed deficits after dominant hemi-sphere injection in a number of nonverbal tasks such as matching-to-sampleof common objects and short-term recognition memory for pictures. In an-other study, Loring, Lee, and Meador (1989) did not observe any differencebetween right and left injection in the recognition of two objects. However,false-positive errors were more frequent following left injection regardlessof the site of epileptic focus.

226 ROULEAU ET AL.

When the effect of cerebral dominance for language (worse results follow-ing injection of the dominant hemisphere than after nondominant injection)is added to the effect of site (worse results following injection contralateralto the focus than after ipsilateral injection), the worst performance should beobserved in right temporal (RT) patients following dominant, contralateral,injection, whereas the best performance should be observed in RT after right,nondominant, ipsilateral injection. The performances of left temporal (LT)patients following ipsilateral and contralateral injection should be muchcloser.

This pattern was reported by Rausch et al. (1984). In RT patients, perfor-mance following right (ipsilateral) injection was near perfect while left (con-tralateral) injection induced significant impairment. In LT patients, perfor-mances following left (ipsilateral) and right (contralateral) injection wereequivalent. Thus, the expected dissociation (i.e., ipsilateral vs. contralateralinjection) in delayed recognition of pictures of common objects presentedunder the effect of amobarbital was seen only in RT patients. In their proto-col, Rausch et al. (1984) asked subjects to name the pictures that were pre-sented to be remembered.

A comparable pattern was reported by Smith, McGlone, and Fox (1993),who found that retention was particularly deficient in RT patients after leftcontralateral injection. Following the right ipsilateral injection retention wasalmost perfect. In the Smith et al. protocol, each stimulus (one object andtwo black-and-white line drawings) was named by the examiner. The patientwas instructed to look at, name, and remember these items. Thus, the subjectheard and repeated the name of each item. Retention of a nursery rhyme thatthe subject had to repeat was also examined.

Similarly, Perrine et al. (1993) reported lower scores for the recognitionof object drawings in RT patients following left injection than followingright injection, whereas in LT patients no differences were observed betweenright and left injections. In the study by Loring, Lee, Meador, Flanigin,Smith, Figueroa, and Martin (1990), the difference in performance betweenipsilateral and contralateral injection also appeared larger in RT patients (3.9vs. 1.8) than in LT patients (2.4 vs. 1.3), although the interaction was notstatistically significant. In their protocol, the eight common objects that werepresented to be remembered were named twice to the patient.

These findings may have important clinical consequences. Because oftheir low memory scores following ipsilateral left dominant injection, someLT patient might needlessly be excluded from surgery. For instance, in theWyllie et al. study (1991a), only LT patients failed ipsilateral IAP, althoughtheir ipsilateral performance was not as bad as that observed following con-tralateral injections. A similar pattern was reported by Novelly and William-son (1989). In their study, most instances of false positive memory impair-ment after ipsilateral injection were observed in LT patients.

In order to overcome the deleterious effects of anesthetizing the dominant


hemisphere, we designed an IAP protocol that did not promote verbal encod-ing of the stimuli. For this purpose, a large number of visual and tactilestimuli (colored pictures and real objects) were presented one by one in acontinuous recognition task. Because of the salient morphological attributesof the stimuli, and since naming was not encouraged, we expected the mem-ory scores to be unaffected by the language dominance of the injected hemi-sphere. We thus expected both RT and LT patients to show good retentionscores following ipsilateral injection and to demonstrate significant deficitsafter contralateral injections; that is, equivalent differences between ipsilat-eral and contralateral injections should be observed in RT and in LT patients.

The effect of focus lateralization was assessed in three different retentionconditions. Since the medial temporal lobe region appears to be involved inmemory consolidation, we designed a protocol that included, in addition tothe retrograde memory measure, two measures of anterograde memory: ashort-term memory measure assessed 2 min after presentation of the stimuli,when one hemisphere was still under the effect of amobarbital, and a long-term memory measure, assessed after a delay of about 15 min, when thehemisphere had completely recovered from the effect of anesthesia. Sincethe same stimuli were used to assess retention in both short-term and long-term retention conditions, this protocol enabled us to study the consolidationof information acquired under amobarbital effect (Rouleau, Labrecque,Saint-Hilaire, Cardu, & Giard, 1989).

METHOD

Subjects

The sample consisted of 82 patients (36 women, 46 men), all with a unilateral temporalepileptic focus and all candidates for surgery. The localization of the epileptic focus wasconfirmed by depth-electrode recordings of spontaneous seizures in most subjects. There were47 subjects with a left temporal lobe focus (LT) and 35 with a right temporal lobe focus (RT).Eleven patients were left-handed, 3 were ambidextrous, and 68 were right-handed. Seventypatients (85.4%) had a left cerebral dominance for language and 5 (6.1%) had a right domi-nance. Bilateral language was observed in 7 (8.5%) patients. A battery including standardizedtests of intellectual, language, memory, visuomotor, and perceptual abilities as well as testssensitive to frontal lobe dysfunction was administered to every subject pre- and postopera-tively.

Procedure

Right and left injections were performed in all patients on 2 consecutive days. The suspectednondominant hemisphere for language was usually injected first since the initial applicationof the protocol was easier in the absence of aphasia. An angiographic contrast study was doneprior to each test to ensure the appropriate localization of the catheter and to visualize anyanatomical variations of the arterial system. The sodium amobarbital (150 mg, 10% solution)was injected within a 5- to 10-sec interval into the internal carotid artery through a femoralcatheter. Continuous scalp EEG monitoring and repeated neurological examination were usedto evaluate the level of anesthesia following the amobarbital injection. In addition to memory

228 ROULEAU ET AL.

FIG. 1. Example of visual stimuli used. Each colored photograph of a single object or animalwas presented on a white background.

evaluation, language was also assessed during right and left hemisphere anesthesia in everypatient, but this procedure will not be detailed here.

To compare memory performance after left and right injections, dually encodable stimuliwere used in this study so that they should be adequately processed by both hemispheres(Paivio & Csapo, 1972). Visual stimuli were colored photographs of single objects or animalsshown on a white background (see Fig. 1). They were presented to the hemifield ipsilateralto the injection. Tactile stimuli were familiar objects with easily recognizable shapes. Theywere placed in the nonparalyzed hand and were never seen by the subject. For each memorycondition (retrograde, short-term, and long-term memory), visual stimuli were always pre-sented before tactile stimuli. To avoid effects related to cerebral dominance for language, thesubject was instructed not to name the stimuli and the examiner did not name them either.Retention was tested by a continuous yes/no recognition task. The same sensory modalitywas used for presentation and recognition. Different sets of stimuli were used for the twohemispheric injections. A practice session, following the same procedure but using differentstimuli, was always given the evening before the first test to ensure that the patients understoodthe instructions and the general design of the task.

Differences between protocols. Over the years, two different protocols were used to assessmemory. In the first protocol, both visual and tactile stimuli were presented, whereas onlyvisual stimuli were used in the second. The tactile stimuli were dropped from the protocolbecause it seemed that they had produced a significant increase in exposure time in somepatients who had tactile recognition difficulties. The memory conditions and the total numberof stimuli to be remembered remained the same in both protocols. An equal number of subjectsreceived protocol 1 (LT, 23; RT, 18) and protocol 2 (LT, 24; RT, 17). As three different


TABLE 1Order of Presentation of the Stimuli

Before the 20 Min afterinjection During IAP the injection

A (n) A (y) RETRO A (y)A (n) B (n) C (n)A (n) A (y) RETRO B (y) LTM23A (n) A (y) RETRO B′ (y) LTM13A (n) B (n) C (n)

B (n) B′ (y) LTM13A (y) RETRO C (n)B (n) B (y) LTM23B (n) C (n)A (y) RETRO A (y)B′ (n) B (y) LTM23B (y) STM C (n)B′ (n) C (n)B′ (n) A (y)B (y) STM C (n)B (y) STM B′ (y) LTM13B′ (n) C (n)B (y) STM C (n)B′ (n) B (y) LTM23B (y) STM C (n)

B′ (y) LTM13A (y)C (n)B (y) LTM23C (n)B′ (y) LTM13A (y)C (n)C (n)

Note. A, stimuli presented before the injection, usedto assess RETRO; B, stimuli presented during the IAP,used to assess STM and LTM 23; B′, stimuli used asdistractors for STM, used to assess LTM 13; C, stimuliused as distractors in LTM assessment; (y), yes is theexpected response; (n), no is the expected response;boldface indicates tactile stimuli (in protocol 1).

retention conditions were examined, the procedure used and the differences between the proto-cols will be described for each retention condition. The order of presentation of the stimulifor each retention condition is presented in Table 1.

Retrograde memory (RETRO). Five stimuli were presented before the injection, to be recog-nized later. In protocol 1, these visual and tactile stimuli were presented within 3 to 1 minbefore the injection. In the second protocol, the five visual stimuli were presented as close aspossible to the moment of injection (one stimulus every 15 sec, starting 1.5 min before injec-

230 ROULEAU ET AL.

tion). Retention of the stimuli was tested approximately 3 to 5 min after the injection whendrowsiness, if present, had disappeared but a high percentage of slow waves was still presentin the EEG on the injected side. Ten stimuli (five presented before the injection and fivedistractors) were presented one after the other and patients had to indicate which ones theyhad seen or touched before the injection (yes/no recognition).

Short-term retention (STM). Under the effect of amobarbital, the 5 distractors that had justbeen used to test retrograde memory were presented a second time, interspersed among 5 newstimuli. The number of items correctly recognized was used as a measure of short-term mem-ory. An interval of 1 to 2 min separated the initial presentation of an item from its secondpresentation, for recognition. In the first protocol, visual stimuli were shown before tactilestimuli in each retention condition. To assess retrograde memory, the patients were first shown6 visual stimuli (3 target items that had been presented before the injection and 3 distractors)and then 4 tactile stimuli (2 from before the injection and 2 distractors). They were then shown6 visual stimuli and 4 tactile stimuli for the assessment of STM. Altogether, 20 items wereshown in protocol 1, 12 pictures and 8 objects. In contrast, as only visual stimuli were usedin protocol 2, the total stimuli consisted of 20 pictures, presented one after the other foryes/no recognition. The order of presentation of the stimuli is shown in Table 1.

Long-term retention (LTM). Twenty minutes after injection, after the slow-wave EEG andneurological signs of anesthesia had cleared, subjects were asked to recognize all stimuli seenor touched before and after the injection. A series of stimuli was presented for yes/no recogni-tion, including the 5 presented before injection and the 5 presented twice under the effect ofamobarbital (LTM 23). These stimuli were the same that were used to assess short-termretention. In the second protocol, the stimuli presented only once during the effect of amobarbi-tal (i.e., the ones used as distractors in the short-term recognition test) were also used as anadditional measure of long-term retention (LTM 13). Twelve distractors were used in thefirst protocol and 15 in the second.

To take false recognitions into account, we computed a score for each retention conditionby adding the number of stimuli correctly recognized to the number of distractors correctlyrejected (maximum score, 10). The total score was multiplied by 100 to obtain the percentagecorrect. For the two long-term retention measures (LTM 13 and LTM 23), the score wascalculated with the following formula: percentage correct 5 [true recognition 1 (5 2 falserecognitions /2) 3 100]. Thus, the total number of false recognitions was assigned equally toboth measures of long-term retention. The percentage correct obtained in each retention condi-tion (RETRO, STM, LTM 13, LTM 23) was compared for left and right injections for eachsubject, as a function of site (ipsilateral vs. contralateral to the epileptic focus), lateralizationof the focus (RT vs. LT), and the protocol used.

For clinical purpose, a pass–fail score of #70% correct was used. This criterion was selectedon the basis of the performances observed in our previous study (Rouleau et al., 1989).

RESULTS

Results obtained in the various retention conditions were analyzed withANOVAs with repeated measures on site (ipsilateral vs. contralateral to sei-zure focus) and conditions (RETRO, STM, LTM 13, LTM 23). Between-subjects variables were focus lateralization (RT vs. LT) and protocol (1 vs.2). When indicated, analyses of simple effects were carried out to breakdown interaction effects.

1. Effect of Site (Ipsilateral vs. Contralateral Injections)

A first ANOVA was performed on the entire sample (N 5 82), regardlessof the cerebral language representation. The scores obtained on the various


retention conditions (RETRO, STM, LTM 23) were analyzed as a functionof site (ipsilateral vs. contralateral to seizure focus). This ANOVA yieldeda significant main effect of condition, F(2, 152) 5 27.57, p , .0001, of site,F(1, 76) 5 10.62, p 5 .002, and a significant Condition 3 Site interaction,F(2, 152) 5 11.47, p , .0001. Results are presented in Fig. 2a.

A second analysis, including the long-term retention of stimuli presentedonly once under the effect of amobarbital (LTM 13), was performed onlywith subjects who were administered protocol 2 (see Fig. 2b). Significantmain effects of site, F(1, 38) 5 6.35, p 5 .016, and condition, F(3, 114) 521.40, p , .0001, were noted, as well as a significant Site 3 Conditioninteraction, F(3, 114) 5 4.79, p 5 .004.

Analysis of the simple effects revealed that the Site 3 Condition interac-tion was due to the difference between ipsilateral and contralateral injectionon long-term retention of the stimuli presented twice during IAP (LTM 23),F(1, 76) 5 32.83, p , .0001.

2. Effect of Protocol

Results obtained with protocol 1 and 2 were compared for the variousretention conditions. This ANOVA with repeated measures on conditionsyielded a significant effect of protocol, F(1, 75) 5 14.01, p , .0001, and asignificant Protocol 3 Condition interaction, F(2, 150) 5 3.89, p 5 .022.Analysis of simple effects showed that this interaction was due to lowerscores observed with protocol 2 (visual stimuli only) on the short-term mem-ory measure, F(1, 77) 5 17.34, p , .0001, and, to a lesser extent, to thelower scores noted on the retrograde memory measure, F(1, 75) 5 6.59, p5 .012.

Timing differences between the protocols. The differences in timing be-tween the protocols are shown in Table 2. In general, memory testing wascarried out later in the second protocol (3.37 min vs. 5.52 min postinjection,F(1, 79) 5 31.38, p , .0001), but the total time required to present the 20stimuli was equivalent in both protocols (approximately 2 min/average 5 secper stimulus). Thus, contrary to our expectations, the use of visual stimulionly, instead of both visual and tactile stimuli, did not change the time re-quired for presentation of the material. In both protocols, material was pre-sented later when the dominant hemisphere was anesthetized, F(1, 67) 535.24, p , .0001.

3. Effect of Cerebral Dominance for Language

The effect of cerebral dominance for language was examined in the fourdifferent retention conditions. Only patients with a well-lateralized cerebraldominance for language were included in this analysis (N 5 75). The resultsare shown in Fig. 3. An ANOVA with repeated measures on dominance(dominant vs. nondominant injection) and on condition (RETRO, STM,

232 ROULEAU ET AL.

FIG. 2. Mean percentage correct as a function of site (ipsilateral vs. contralateral to thefocus) and retention conditions. (a) Protocol 1: Both visual and tactile stimuli were presented.(b) Protocol 2: Only visual stimuli were presented. RETRO, retrograde memory; STM, short-term retention; LTM 23, long-term retention of the stimuli presented twice under the effectof amobarbital; LTM 13, long-term retention of the stimuli presented once under the effectof amobarbital.


TABLE 2Differences in Timing between the Two Protocols

Timing (sec)

Protocol 1 Protocol 2

BeginningL injection 284.1 379.3R injection 202.1 319.8

EndL injection 410.1 497.2R injection 319.4 439.1

DurationL injection 126.0 117.9R injection 117.3 119.3

Presentation rate Approx. 5 sec/stimulus

LTM) did not show a significant main effect of dominance, F(1, 69) 5 .24,p 5 .624. However, there was a significant main effect of retention condition,F(2, 138) 5 26.34, p , .0001, and a significant Dominance 3 Conditioninteraction, F(2, 138) 5 9.5, p , .0001.

Analysis of the simple effects of this interaction showed a significant effectof dominance on the measure of retrograde memory, F(1, 69) 5 8.26, p 5.005. Lower scores were noted after dominant hemisphere anesthesia, possi-bly related to language disruption (e.g., misunderstanding of the task andperseveration) in the first few minutes after injection, which may have inter-fered with completion of the task. In contrast, the long-term retention of thestimuli presented twice (LTM 23) during IAP was significantly more im-paired when the nondominant hemisphere was anesthetized than when thedominant hemisphere was injected, F(1, 69) 5 13.39, p , .0001. A similartrend was observed for long-term retention of the stimuli presented once(LTM 13) during the effect of amobarbital, F(1, 32) 5 2.91, p 5 .097.These results suggest that functional inactivation of the nondominant hemi-sphere results in more deficits in long-term retention than injection of thedominant hemisphere, even if the stimuli presented were chosen to be duallyencodable.

4. Effect of Focus Lateralization (RT vs. LT)

A final ANOVA was performed on the effect of lateralization of the epi-leptic focus (right vs. left temporal lobe focus) on the various retention condi-tions. Only patients with left cerebral dominance for language were includedin this analysis (N 5 70). The results are presented in Fig. 4.

There was no overall main effect of focus lateralization on the variousretention conditions, F(1, 63) 5 1.30, p 5 .259. However, there was a sig-nificant Focus 3 Condition 3 Site interaction, F(2, 122) 5 9.09, p , .0001.

234 ROULEAU ET AL.

FIG. 3. Effect of cerebral dominance for language. Mean percentage correct as a function ofhemisphere injected (dominant versus nondominant) for various retention conditions. RETRO,retrograde memory; STM; short-term retention; LTM 23, long-term retention of the stimulipresented twice under the effect of amobarbital; LTM 13, long-term retention of the stimulipresented once under the effect of amobarbital.

To interpret this interaction, additional ANOVAs with repeated measureson site (ipsilateral vs. contralateral) were performed for each retention condi-tion separately as a function of focus lateralization (RT vs. LT).

Retrograde memory. the ANOVA revealed a significant Focus 3 Site in-teraction for the retrograde memory measure, F(1, 61) 5 11.19, p 5 .001. Ascan be seen in Fig. 4a, scores were generally lower following left, dominantinjection, especially in RT patients (contralateral to the seizure focus).

Short-term memory. Results are shown in Fig. 4b. There was no significantFocus 3 Site interaction for the measure of short-term retention, F(1, 63)5 0.01, p 5 .918.

Long-term memory. There was a significant Focus 3 Site interaction forlong-term retention of stimuli presented twice (LTM 23) under the effect ofamobarbital, F(1, 66) 5 9.09, p 5 .004. As depicted in Fig. 4c, the differencebetween ipsilateral and contralateral memory performance was considerablylarger in LT patients than in RT patients. In LT patients, the left, dominant,ipsilateral injection resulted in a near-perfect performance, whereas the right,nondominant, contralateral injection produced a marked deficit. In RT pa-tients, the difference in performance between the right, nondominant, ipsilat-


eral injection and the left, dominant, contralateral injection was not as appar-ent. A significant Focus 3 Site interaction, F(1, 31) 5 4.90, p 5 .034, wasalso observed for the long-term retention of the pictures presented once(LTM 13) under the effect of amobarbital (protocol 2 only; see Fig. 4d),although the overall performances were lower, probably due to the fact thatstimuli were presented only once.

DISCUSSION

Contrary to what is generally reported, LT patients did not show poorerperformances than RT patients in the majority of the retention conditionsassessed. LT patients were actually better than RT patients in the long-termrecognition of pictures and tactile objects, showing near-perfect performancewhen the left hemisphere was anesthetized. This finding does not support thehypothesis of a baseline difference in memory efficiency that was provided toexplain the lower scores observed in LT patients after left ipsilateral injection(e.g., Wyllie et al., 1991b).

As expected, for both RT and LT patients, recognition of pictures pre-sented under the effect of amobarbital was highly sensitive to the presenceof a contralateral epileptic focus. These findings replicate the ‘‘site’’ effect(ipsilateral vs. contralateral to the epileptic focus) that was reported in ourprevious study (Rouleau et al., 1989), and is in agreement with a number ofstudies that confirm the value of IAP in the prediction of focus lateralization(Loring et al., 1990, 1992; Loring, Meador, Lee, King, Gallagher, Murro, &Smith, 1994; Perrine, 1994; Perrine et al., 1993, 1995; Rausch et al., 1984;Wyllie et al., 1991). However, contrary to what was reported by a numberof authors (e.g., Perrine et al., 1993; Rausch et al., 1984), the differencebetween ipsilateral and contralateral injection in long-term recognition ofpictures and objects was smaller in RT than in LT patients. In our study,failure to remember material presented after ipsilateral injection was seenalmost exclusively in RT patients. This was especially true with the firstprotocol; no LT patient failed the test after the left ipsilateral injection,whereas 3 of 18 patients failed the test following the right ipsilateral injec-tion. An opposite tendency was reported by Wyllie et al. (1991a); in theirstudy, no RT patients failed the test after right injection, whereas 4 of 20LT patients failed after the left ipsilateral injection. In the Wyllie et al. study,the failure observed following left ipsilateral injection might be due in partto the use of verbal material (written words, numbers, and sentences). Itmight also be due to the presentation of pictures, objects, and colors that wereprobably presented to be named, although what the patient was specificallyinstructed to do with the stimuli is not explicitly mentioned in the article.Asking the patient to name the stimuli might limit retention when the domi-nant hemisphere is anesthetized while facilitating performance after rightnondominant injection.

236 ROULEAU ET AL.

FIG. 4. Comparison of performance (percentage correct) observed after right and left injec-tion as a function of focus lateralization (right temporal [RT] versus left temporal [LT]) forvarious retention conditions. (a) RETRO, retrograde memory; (b) STM, short-term retention;(c) LTM 23, long-term retention of the stimuli presented twice under the effect of amobarbital;(d) LTM 13, long-term retention of the stimuli presented once under the effect of amobarbital.


238 ROULEAU ET AL.

Our results are similar to those reported by Christianson, Saisa, and Silf-venius (1990). In their protocol, three pictures were presented along withwords, faces, and nonsense drawings. For retention of pictures, the lowestperformance was noted in LT patients after right contralateral injection (20%correct) and the best performance was observed, also in LT patients, afterleft ipsilateral injection (63% correct). The differences between ipsilateral(38% correct) and contralateral (25% correct) injections were much smallerin RT patients. In the Christianson et al. protocol, subjects were neither askedto name the pictures nor were the names of the pictures provided. Similarly,in our protocol, subjects were instructed not to name the pictures and objectssince we wanted the total presentation time to be equivalent after right andleft injection, and additional time would have been required for naming afterdominant hemisphere injection. By attempting to proscribe verbal encodingof the material and by presenting a large number of stimuli at a relativelyfast pace (approximately 5 sec/stimulus) within a short interval of time, wemay have encouraged the retention of salient morphological attributes of ourstimuli such as color and shape. In this respect, the lower retention scoresobserved in RT patients after right ipsilateral injection could be the conse-quence of the right functional hemispherectomy produced by the injection.This hypothesis is supported by a number of studies that have shown aright hemispheric dominance for the retention of visuospatial information(Boller & DeRenzi, 1967; Pigott & Milner, 1993; Warrington & Rabin,1978) and tactile information (Bottini, Cappa, Sterzi, & Vignolo, 1993; Mil-ner & Taylor, 1972). More recently, Kaplan et al. (1994) have shown, withinjection of posterior cerebral artery, that the right temporal lobe is moreefficient than the left in the retention of specific visual attributes of objects.

The negative effect of anesthetizing the nondominant hemisphere for theretention of a large number of rapidly presented stimuli was even more pro-nounced in the second protocol despite the fact that testing was performedlater. The fact that only pictures were presented probably increased thebuildup of proactive interference (PI). Absence of change in sensory modal-ity (visual, tactile) may have prevented release from PI in the second proto-col. Although semantic shift is recognized as being highly powerful for PIrelease, a change in presentation modality can be as effective (Wickens,1970).

A similar buildup of PI could also explain, as least partially, the poorerperformance observed with the second protocol in the measure of short-termrecognition. Unlike protocol 1, there was no change in sensory modalitybetween the first and the second series of 10 stimuli. Since STM was mea-sured with the last 10 pictures, the subject had already seen 10 pictures whenthe evaluation of STM started; this may have led to an increase in PI.

Another explanation for the lower STM scores observed in protocol 2could be due to a faulty understanding of the task; instead of using all thestimuli presented previously as a reference set, some patients may have based


their judgments only on the five stimuli presented before the injection andnot on the new stimuli presented under the effect of amobarbital. It is possiblethat in the first protocol, the change in modality increased the likelihood ofchanging reference to include the stimuli that had just been presented. Sucha misunderstanding of the task is supported by the larger proportion of pa-tients in protocol 2 who did not recognize a single picture in the short-termmemory condition. However, even when excluding these patients from theanalyses, STM scores remain lower with protocol 2 than with protocol 1.Thus, misunderstanding of the task alone is not sufficient to explain the lowerscores observed with the second protocol. Since neither the site of injection(ipsilateral vs. contralateral) nor the hemisphere injected (dominant vs. non-dominant) significantly influenced the short-term retention of the materialpresented under the effect of amobarbital, the buildup in PI due to the succes-sive presentation of a large number of pictures remains the best mechanismto explain the lower STM scores observed with the second protocol.

There was no retrograde amnesia produced by injection to either side,although there was a tendency for lower scores to occur after contralateralinjection, especially in LT patients. The differences between contralateraland ipsilateral memory scores were more pronounced with the second proto-col, possibly reflecting the effects of shortening the delay between presenta-tion of the stimuli and the injection.

In summary, the significant effect of focus lateralization that was observedin the long-term recognition of stimuli presented under the effect of amobar-bital appears to be related to the effect of cerebral dominance for languagerather than to baseline differences in memory efficiency between RT andLT patients. A significant effect of cerebral dominance for language wasobserved in our study even though dually encodable stimuli were used. Themain differences among the various IAP protocols that are commonly useddo not rely on the type of stimulus, but rather on the type of processingrequired when subjects are shown the items to be remembered. Whether thetask is verbal (such as naming a picture) or nonverbal (such as matching-to-sample) will directly influence which hemisphere will preferentially beinvolved. Obviously, we cannot rule out that our subjects named the picturesand objects subvocally. However, most of the subjects complied with theinstruction not to name the stimuli.

A comprehensive evaluation of memory functions during IAP should in-clude the presentation of dually encodable stimuli as well as stimuli that arepreferentially processed by one hemisphere. Dually encodable stimuli canbe used to compare the effect of right and left hemisphere injection and toestablish a pass–fail criterion. For this purpose, it would be important tohave the patient directly encode both the verbal and the nonverbal aspectsof the material. Because some studies have shown that it is possible to useIAP results to predict the postoperative material-specific memory decline,stimuli that are preferentially processed by one or the other hemisphere, such

240 ROULEAU ET AL.

as abstract designs or abstract words, could also be used to examine withmore refinement the level of functioning of each temporal lobe.

REFERENCES

Aasly, J., and Silfvenius, H. 1990. Evaluation of early and late presented tasks in the intra-carotid Amytal test for epileptic patients. Epilepsy Research, 7, 155–164.

Boller, F., & De Renzi, E. 1967. Relationship between visual memory defects and hemisphericlocus of lesion. Neurology, 17, 1052–1058.

Bottini, G., Cappa, S. F., Sterzi, R., & Vignolo, L. A. 1995. Intramodal somaesthetic recogni-tion disorders following right and left hemisphere damage. Brain, 118, 395–399.

Chelune, G. J. 1995. Hippocampal adequacy versus functional reserve: Predicting memoryfunctions following temporal lobectomy. Archives of Clinical Neuropsychology, 10(55),413–432.

Christianson, S. A., Saisa, J., & Silfvenius, H. 1990. Hemisphere memory differences in so-dium Amytal testing of epileptic patients. Journal of Clinical and Experimental Neuropsy-chology, 12(5), 681–694.

Engel, J., Jr., Rausch, R., Lieb, J. P., Kuhl, D. E., & Crandall, P. H. 1981. Correlation ofcriteria used for localizing epileptic foci in patients considered for surgical therapy ofepilepsy. Annals of Neurology, 9, 215–224.

Jones-Gotman, M., McMackin, D., Cendes, F., Andermann, F., Evans, A., Olivier, A., &Peters, T. 1993. Performance on intracarotid sodium amobarbital memory tests: Relation-ship to hippocampal atrophy as estimated by volumetric MRI. Epilepsia, 34, 94.

Kaplan, R. F., Meadows, M.-E., Verfaellie, M., Kwan, E., Ehrenberg, B. L., Bromfield, E. B. &Cohen, R. A. 1994. Lateralization of memory for the visual attributes of objects: Evidencefrom the posterior cerebral artery amobarbital test. Neurology, 44, 1069–1073.

Loring, D. W., Lee, G. P., & Meador, K. J. 1989. The intracarotid amobarbital sodium proce-dure: False-positive errors during recognition memory assessment. Archives of Neurology,46, 285–287.

Loring, D. W., Lee, G. P., Meador, K. J., Flanigin, H. F., Smith, J. R., Figueroa, R. E., andMartin, R. C. 1990. The intracarotid amobarbital procedure as a predictor of memoryfailure following unilateral temporal lobectomy. Neurology, 40, 605–610.

Loring, D. W., Meador, K. J., and Lee, G. P. 1992. Amobarbital dose effects on Wada memorytesting. Journal of Epilepsy 5(3), 171–174.

Loring, D. W., Meador, K., Lee, G. P., Nichols, M., King, D., Gallagher, B., Murro, A., &Smith, J. 1995. Wada memory asymetries predict verbal memory decline after anteriortemporal lobectomy. Neurology, 45, 1329–1333.

Loring, D. W., Meador, K. J., Lee, G. P., King, M. D., Gallagher, B. B., Murro, A. M., & Smith,J. R. (1994). Stimulus timing effects on Wada memory testing. Archives of Neurology, 51,806–810.

Milner, B., & Taylor, L. 1972. Right hemisphere superiority in tactile pattern-recognition aftercerebral commissurotomy. Neuropsychologia, 10, 1–15.

Novelly, R. A., & Williamson, P. D. 1989. Incidence of false-positive memory impairmentin the intracarotid Amytal procedure. Epilepsia, 30(5), 711.

Paivio, A., & Csapo, K. 1972. Picture superiority in free recall: Imagery or dual coding?(Research Bulletin No. 243). London, Canada: University of Western Ontario.

Perrine, K. 1994. Future directions for functional mapping. Epilepsia, 35 (Suppl. 6), S90–S102.

Perrine, K., Gershengorn, J., Brown, E. R., Choi, L. S., Luciano, D. J., and Devinsky, O.1993. Material-specific memory in the intracarotid amobarbital procedure. Neurology, 43,706–711.

Perrine, K., Westersveld, M., Sass, K., Devinsky, O., Dogali, M., Spencer, D., Luciano, D., &


Nelson, P. 1995. Wada memory disparities predict seizure laterality and postoperativeseizure control. Epilepsia, 36, 851–856.

Pigott, S., & Milner, B. 1993. Memory for different aspects of complex visual scenes afterunilateral temporal or frontal lobe resection. Neuropsychologia, 31(1), 1–15.

Powell, G., Polkey, C., & Canavan, A. 1987. Lateralization of memory functions in epilepticpatients by use of the sodium Amytal (Wada) technique. Journal of Neurology, Neurosur-gery, and Psychiatry, 50, 665–672.

Raush, R., Babb, T. L., Engel, J., Jr., & Crandall, P. H. 1989. Memory following intracarotidamobarbital injection contralateral to hippocampal damage. Archives of Neurology, 46,783–788.

Raush, R., Fedio, P., Ary, C. M., Engel, J., & Crandall, P. H. 1984. Resumption of behaviorfollowing intracarotid sodium amobarbital injection. Annals of Neurology, 15(1), 31–35.

Rausch, R., Silfvenius, H., Wieser, H.-G., Dodrill, C., Meador, K., & Jones-Gotman, M. 1993.Intraarterial amobarbital procedures. In J. Engel, Jr. (Ed.), Surgical treatment of the epilep-sies. New York: Raven Press. 2nd ed., pp. 341–357.

Rausch, R. and MacDonald, K. 1997. Effects of hemisphere speech dominance and seizurefocus on patterns of behavioral response errors for three types of stimuli. Brain and Cogni-tion, 57, 161–177.

Rouleau, I., Labrecque, R., Saint-Hilaire, J. M. Cardu, B., & Giard, N. 1989. Short-term andlong-term memory deficit following Amytal injection: Further support for the memoryconsolidation hypothesis. Brain and Cognition, 11, 167–185.

Silfvenius, H., and Aasly, J. 1991. Cerebral hemisphere memory for manual object handling:An intracarotid Amytal study. Epilepsy Research, 8, 49–57.

Smith, I. M., McGlone, J., & Fox, A. J. 1993. Intracarotid amobarbital memory protocol:Muteness, dysphasia, and variations in arterial distribution of the drug do not affect recog-nition results. Journal of Epilepsy, 6, 75–84.

Sperling, J., O’Connor, M. J., Saykin, A. J., Phillips, C. A., Morrel, M. J., Bridgman, P. A.,French, J. A., & Gonatas, N. 1992. A noninvasive protocol for anterior temporal lobec-tomy. Neurology, 42, 416–422.

Warrington, E. K., & Rabin, P. 1970. A preliminary investigation of the relation betweenvisual perception and visual memory. Cortex, 6, 87–96.

Wickens, D. D. 1970. Encoding categories of words: An empirical approach to meaning.Psychological Review, 77, 1–15.

Wyllie, E., Naugle, R., Chelune, G., Luders, H., Morris, H., & Skibinski, C. (1991a). Intra-carotid amobarbital procedure. II. Lateralizing value in evaluation for temporal lobec-tomy. Epilepsia, 32, 865–869.

Wyllie, E., Naugle, R., Awad, I., Chelune, G., Luders, H., Dinner, D., Skibinski, C., & Ahl,J. 1991b. Intracarotid amobarbital procedure. I. Prediction of decreased modality-specificmemory scores after temporal lobectomy. Epilepsia, 32(6), 857–864.

Documents

Effect of Focus Lateralization on Memory Assessment during the Intracarotid Amobarbital Procedure