Brain and Language 102 (2007) 1321 www.elsevier.

com/locate/b&l

Verbal memory compensation: Application to left and right temporal lobe epileptic patients
ronique Lespinet-Najib a, Alain Rougier b, Christel Bresson a,*, Ve Bernard Claverie a, Bernard NKaoua a
b

Victor Segalen Bordeaux 2, Laboratoire de Sciences Cognitives (UPRES EA-487), Institut de Cognitique, Universite t. 2A 146 Rue Le o Saignat, 33076 Bordeaux cedex, France Domaine Universitaire de Carreire, Zone Nord, Ba rimentale et Clinique (EA 2967), Universite Victor Segalen Bordeaux 2, Domaine Universitaire de Carreire, Laboratoire dEpileptologie Expe te 78 146 Rue Le o Saignat, 33076 Bordeaux cedex, France Bo Accepted 2 June 2006 Available online 12 July 2006

Abstract This study investigates the compensatory impact of cognitive aids on left and right temporal lobe epileptic patients suering from verbal memory disorders, who were candidates for surgery. Cognitive aids are dened in the levels-of-processing framework and deal with the depth of encoding, the elaboration of information, and the use of retrieval cues. Results indicate dierential compensatory impact for left and right epileptic patients and are discussed according to the HERA model and the compensation framework. 2006 Elsevier Inc. All rights reserved.
Keywords: Cognitive compensation; Verbal memory; Temporal lobe epilepsy; Hemispheric specialization; Levels-of-processing

1. Introduction Temporal lobe epilepsy is the most important focal epilepsy due to its high prevalence, drug resistance, and commonly disabling eects on memory functions (Engel, 1996). Seizures concern hippocampal and parahippocampal structures, which mediate declarative memory (Cohen & Squire, 1980; Eichenbaum, 2001; Nadel & Moscovitch, 2001). Thus, poor memory is a primary complaint of patients with temporal lobe epilepsy (TLE). Indeed, a large body of literature is devoted to memory impairment in TLE and studies have frequently focused on the potential dierences that occur when seizures are generated from left versus right temporal lobe (Andersson-Roswall, Engman, Samuelsson, Sjo berg-Larsson, & Malmgren, 2004; Grith, Pyzalski, Seidenberg, & Hermann, 2004; Helmstaedter, 2002; Reminger et al., 2004; Schwarcz & Witter, 2002).

Corresponding author. Fax: +33 0 5 57 57 46 07. E-mail address: bresson@scico.u-bordeaux2.fr (C. Bresson).

One way of approaching memory decits is the levels-ofprocessing framework (LOP) developed by Craik & Lockhart in 1972. In this framework, the durability and distinctiveness of memory traces are direct functions of the nature and quality of the encoding operations. Depth of processing is seen as a processing continuum from shallow (sensory) to deep (semantic) processes, with some types of processing (typically sensory analyses) preceding others (typically conceptual analyses). So, the deeper the initial processing, the better subsequent memory performance is expected to be. To complete this theoretical approach (Lockhart & Craik, 1990), two other main concepts have been added: elaboration and encoding specicity. Therefore, memory performance depends upon three dimensions, depth of processing, elaboration, and encoding specicity (kind of retrieval) of the information operative at the two memory stages, encoding and retrieval. The rst dimension in the LOP theory, depth of processing, states that deeper processing is associated with higher levels of subsequent remembering. Encoding information at a shallow level (e.g., phonetic) leads to worse

0093-934X/$ - see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bandl.2006.06.005

14

C. Bresson et al. / Brain and Language 102 (2007) 1321

performance than encoding information at a deeper level (e.g., semantic). For example, memorizing brain, train, rain leads to worse performances than memorizing bear, horse, dog. These results have been obtained in numerous experiments with healthy subjects (e.g., Bentin, Moscovitch, & Nirhod, 1998; Craik & Tulving, 1975; Fujii et al., 2002) and are explained by the fact that encoding semantic information activates semantic associates in memory that are more ecient for the retrieval of information than less meaningful phonetic associates. The second dimension, elaboration, demonstrated by Craik and Tulving (1975), states that an elaborated memory trace is remembered more readily. Information is elaborated by paying attention to the specic meaning i.e., the semantic aspect of associated information (Symons & Johnson, 1997). The richness or extensiveness of processing (Lockhart, Craik, & Jacoby, 1976) makes it possible to retrieve elaborated information via several pathways (Einstein & Hunt, 1980; Lockhart et al., 1976). Among the ways to handle elaboration, the self-reference eect (Kuiper & Rogers, 1979; Maki & McCaul, 1985; Rogers, Kuiper, & Kirker, 1977) describes the fact that asking the subject if the word describes him/her produces a better memory trace. In addition, the self-generated cuing paradigm, introduced by Ma ntyla and Nilsson (1983), involves the subject in elaborating the information to be memorized by making him/her producing the cue to be matched with the information. Finally, the third dimension concerns the retrieval stage and refers to encoding specicity (Tulving & Thomson, 1973). Given a specic type of encoding, retrieval is optimized when the retrieval test is designed to utilize the same type of information (Fisher & Craik, 1977; Marmurek, 1995). So, cued recall or recognition provides better retrieval performance than free recall, with no context. These three dimensions derived from the levels-of-processing theory can be studied together to determine the impact of their interaction on memory performance. Interaction between depth of processing and encoding specicity shows that retrieval (cued recall) is more eective for a deep encoding than for a shallow encoding (Fisher & Craik, 1977). Secondly, interaction between elaboration and encoding specicity has been observed. Ma ntyla and Nilsson (1988) demonstrated that self-generated cueing is much more eective in cued than free recall. This indicates that elaboration improves the eciency of encoding specicity. Thus, interaction between depth of processing and elaboration demonstrated that depth was boosted by elaboration. The superiority of semantic over phonetic encoding is enhanced by elaboration. Finally, the triple interaction has been obtained. Indeed a contextualized recall of an elaborated and deeply processed information results in better performance (Ma ntyla & Nilsson, 1988). In the levels-of-processing theory, the three dimensions could be seen as aids that enhance performance and so could be proposed to compensate cognitive decits.

As TLE patients main complaint concerns their memory decits, the mismatch between their memory skills and the demands of everyday life (Fisher et al., 2000; Guerreiro, Jones-Gotman, Andermann, Bastos, & Cendes, 2001; Smith, 1989), these patients are candidates for compensation studies to assess the positive impact of the three LOP dimensions (seen as compensation aids) on performance. In the epileptic research eld, some studies have shown that a decit in memorizing information at a shallow level disappears when patients are helped by deep processing of information, although authors do not agree on the relationship between decits and laterality of the foci (Helmstaedter & Kurthen, 2001). For some, this compensatory impact of depth of processing is observed only in left temporal lobe epileptic (L-TLE) (Helmstaedter, Gleissner, Di Perna, & Elger, 1997; Jokeit, Okujava, & Woermann, 2001) or right temporal lobe epileptic (R-TLE) patients (Christianson, Nilsson, Sa isa , & Silfvenius, 1992; Mungas, Ehlers, Walton, & McCutchen, 1985). Finally, some works do not nd any impact of depth of processing, irrespective of the laterality of the epileptic foci (Gleissner & Elger, 2001; NKaoua, Lespinet, Barsse, Rougier, & Claverie, 2001; Tro ster et al., 1995). Furthermore, it could be noted that in patients who suer from a Korsako syndrome, authors have not found benets from deep processing whereas patients performance levels after shallow processing were nearly normal (Cermak, 1979; Cermak, Reale, & Baker, 1978). Few studies have investigated elaboration. In a memorization task using self-generated cueing, Lespinet-Najib et al. (2004) demonstrated that elaboration did not help L-TLE patients in phonetic (shallow) or semantic (deep) processing. On the contrary, the R-TLE group beneted from elaboration in phonetic processing, as well as in semantic processing. The third dimension, retrieval, was tested by comparing free and cued, or coping recall. Numerous authors have used this paradigm (Christianson, Silfvenius, & Nilsson, 1987, 1989; Savage, Saling, Davis, & Berkovic, 2002). Mungas et al. (1985) found that cued recall only had an impact in L-TLE patients after deep processing. According to Lespinet-Najib and collaborators (2004) L-TLE patients did not benet from cued recall in deep processing, whereas R-TLE patients did. These ndings indicate that the interaction between depth of processing and retrieval does not provide cognitive aid in left TLE, whereas it does in right TLE. Although verbal memory decits have been widely studied, there is no consensus as to the prole of left and right temporal lobe epileptic patients memory disorders or the compensatory power of cognitive aids, which is crucial from a rehabilitation perspective. While individual cognitive aids may be useful, the impact of their interaction is of real interest. We therefore studied the impact of each dimension of the LOP theory and examined how their interactions potentiated performance. In our approach,

C. Bresson et al. / Brain and Language 102 (2007) 1321

15

three compensation outcomes were identied by comparing performance of left- and right-TLE patients to healthy subjects. In a no compensation outcome, healthy subjects improved their performance when at least one aid was used, whereas the TLE groups did not. In a partial compensation outcome, the TLE groups improved their performance using the aid, but did not achieve the same performance as healthy subjects. Finally, in a full compensation outcome, the TLE groups achieved the same performance level as healthy subjects when the aid was provided. The aim of our study is to investigate the compensation eect of the three cognitive aids, separately and in interaction, on L- and R-TLE patients. 2. Materials and methods 2.1. Participants Thirty patients with intractable partial epilepsy due to temporal lobe epilepsy, who had been referred for evaluation for epilepsy surgery, were included in this study. Fourteen patients suered from left temporal lobe epilepsy and 16 from right temporal lobe epilepsy. Seizure origin in all patients retained for analysis was assessed by standard preoperative diagnostic procedures, including videotape/electroencephalographic monitoring and MRI. Both groups were right handed and matched for gender, current age, age at seizure onset, seizure frequency, seizure duration, number of antiepileptic drugs (AEDs), and educational level. A normal control group (NC) consisted of sixteen persons matched with the TLE groups for age, educational level, and gender. As medication and epilepsy can reduce processing speed (Shulman & Barr, 2002), patients completed standard preTable 1 Group characteristics Parameters Age, in years M (SD) Number of male/female Educational level, in years M (SD) Age of seizure onset, in years M (SD) Epilepsy duration, in years M (SD) Seizure frequency per month Under 5 510 Over 10 No. of AEDsa, M (SD) Coding subtestbM (SD) TMT AcM (SD) TMT BcM (SD) Logical Memory IdM (SD) Logical Memory IIdM (SD)
a b c d

operative neuropsychological testing: the Coding (WAIS-R subtest, Wechsler, 2000) and Trail Making Test (TMT), A and B (Reitan & Wolfson, 1985). Finally, the Wechsler memory Scale-III (WMS-III) verbal subtests Logical Memory I (short-term verbal memory) and Logical Memory II (long-term verbal memory) (Wechsler, 1997, 2001) were performed by the three groups in order to determine the verbal memory decits of the Land R-TLE patients in comparison to normal control group. (Demographic and clinical and neuropsychological characteristics of the groups are summarized in Table 1). 2.1.1. Demographic characteristics There was no dierence between the three groups in terms of age [F(2, 43) = 0.4; p > .05] and educational level [F(2, 43) = 1.1; p > .05]. 2.1.2. Clinical characteristics L-TLE and R-TLE groups did not dier in terms of seizure frequency [v2 = 0.4; p > .05], age at seizure onset [t = 0.5; p > .05], seizure duration [t = 0.2; p > .05], or number of AEDs [t = 0.3; p > .05] (Table 1). 2.1.3. Neuropsychological characteristics There was no dierence between the two groups of patients in terms of processing speed: Coding [t = 0.8 p > .05], TMT A [t = 0.8; p > .05], or TMT B [t = 1.1; p > .05]. Concerning the Wechsler memory Scale, a twoway (Groups * Logical Memory) analysis of variance (ANOVA), with repeated measures for the last factor, was used to assess verbal memory decits in the two groups of patients. The ANOVA revealed single eects. Groups eect [F(2, 43) = 20.7; p 6 .0001]: patients are impaired in comparison to NC group as revealed by the Fisher post hoc analysis [p 6 .0001]. Logical Memory eect

L-TLE (n = 14) 31 (13) 4/10 11 (3) 13 (12) 18 (12) 5 5 4 2.6 (0.9) 53.1 (15.4) 32.3 (18.4) 75.7 (40.5) 39.1 (14.7) 23.8 (16.6)

R-TLE (n = 16) 35 (10) 9/7 12 (2) 15 (11) 19 (14) 7 4 5 2.7 (0.8) 48.2 (17.1) 37.7 (16.9) 95.8 (56.1) 44.5 (16.8) 36.8 (17.6)

NC (n = 16) 32 (10) 10/6 12 (2)

67.5 (13.4) 63.6 (13.1)

Antiepileptic drug. Scores. Times in seconds. Percent of recall.

16

C. Bresson et al. / Brain and Language 102 (2007) 1321

[F(1, 43) = 86.5; p 6 .0001]: Logical Memory I (short term memory) is better performed than Logical Memory II (long term memory). Finally interaction between Groups and Logical Memory is shown [F(2, 43) = 11.8; p 6 .0001]. For Logical Memory I, L-, and R-TLE patients performed worth than NC group. For Logical Memory II, L-TLE performed worth than R-TLE patients who performed worth than NC group. 2.2. Apparatus Four lists were developed. Two lists required shallow processing (phonetic lists) and the two other, deep processing (semantic lists). For each level of processing one list is constituted of word pairs and the other list of single words. The word pairs list (targetcue) corresponds to the experimenter provided cue condition (named E-provided condition) that enables low elaboration, the cue is given to the subject by the experimenter. The single words list corresponds to the subject produced cue condition (named S-produced condition) that enables deep elaboration because the subject produces his cue. For phonetic processing: One list consisted of 16 isolated words (self-generated cueing condition) to enable the subject to produce the rhyme cue (e.g., Brainsubject produced cue); One list consisted of 16 targetcue rhyming-pairs, ending with 16 dierent phonemes (e.g., BrainTrain). For semantic processing: One list consisted of 16 isolated words (self-generated cueing condition) to enable the subject to produce the semantic cue (e.g., Brainsubject produced cue); One list is of 16 targetcue pairs representing examples of dierent categories coupled with the category per se. (e.g., BrainHead). The presentation order of each list was counterbalanced by subject. The verbal stimuli were drawn from the Brulex database (Content, Mousty, & Radeau, 1990), their lexical frequency mean was 1885 occurrences per 100,000,000. Stimuli concreteness was evaluated in a preliminary study, in which students had to score 96 words on a seven-point scale (mean = 5.8 0.8). 2.3. Procedure Each subject had to memorize the four word-lists, one at a time. At the encoding stage, words were presented orally at the rate of one word every 5 s. For each word, either the experimenter (E-provided condition) or the subject (S-produced condition) gave the cue. At the retrieval stage a free recall was followed by a cued recall, where the experiment-

er gave the imposed or produced cue to enable the subject to retrieve the target. 3. Results A four-way (Groups * Depth of processing * Elaboration * Retrieval) analysis of variance (ANOVA), with repeated measures for all factors except Groups, was used to compare the total scores. The single eects, partial interactions, and four-way interaction revealed by the ANOVA are presented below. The data showed the single eect of each of the three dimensions: Depth of processing [F(1, 43) = 288.9; p 6 .0001]; Elaboration [F(1, 43) = 20.2; p 6 .0001]; and Retrieval [F(1, 43) = 152.4; p 6 .0001]. These results indicated respectively enhanced performances for semantic processing (48.7 23.9) in comparison to phonetic one (25.3 17.1), for elaborated information (41.5 25.8) in comparison to non elaborated one (32.5 20.9), and for cued recall (45.2 28.1) in comparison to free recall (28.8 14.7). These results are irrespective of the group. The ANOVA also revealed a Groups eect [F(2, 43) = 11.3; p 6 .001], where left (31.1 21.7) and right (33.1 22.4) epileptic patients had lower performance scores than normal control subjects (46 24.6), as revealed by the Fisher post hoc analysis [p 6 .005]. Results also showed two-by-two interactions between the aids: Depth of processing and Elaboration: [F(1, 43) = 6.7; p 6 .01]; Depth of processing and Retrieval: [F(1, 43) = 107.8; p < .0001]; Elaboration and Retrieval: [F(1, 43) = 22.5; p 6 .0001]. Adding an aid in a situation where another aid was already present enhanced performance. Elaboration had much more eect on deep processing than on shallow one. Cued recall enhanced performance for deep processing in comparison to shallow processing and also for elaborated processing relative to less elaborated. In addition, the triple interaction between the aids is observed. Each group beneted from the interactions of all three aids: Depth of processing, Elaboration and Retrieval: [F(1, 43) = 5.2; p 6 .05]. Cued recall of semantic (deep processing) and elaborated information induced the best performance. Concerning Groups, rst an interaction Groups by Retrieval and second the four-way interaction between the three aids and groups were observed. The two-way interaction Groups by Retrieval [F(2, 43) = 9.4; p 6 .001] revealed that the NC and RTLE groups enhanced cued recall performance (condition with an aid) in comparison to free recall, although both groups of patients were still impaired in comparison to the NC group. The four-way interaction Groups, Depth of processing, Elaboration and Retrieval [F(2, 43) = 3.3; p 6 .05] showed that the interaction between the three aids was dierent in the three groups and could be interpreted as below. Graph 1 presents means of performance for the fourway interaction which are enumerated in Table 2. Finally,

C. Bresson et al. / Brain and Language 102 (2007) 1321

% of recall 100 80 60 40 20 0

17

NC L-TLE R-TLE

ESESESESprovided produced provided produced provided produced provided produced Shallow Free recall Deep Shallow Cued recall Deep

Graph 1. Means of performance for the three groups.

we collapse the results and the dierences of the three variables between their two modalities in Table 3. To do so, mean of the four conditions that implicate each variable (two conditions by modality) have been calculated. For the phonetic processing the L-TLE patients did not enhance performance whatever the aid whereas R-TLE patients improved performance when information is elaborated. For a semantic processing in an E-provided condition and free recall task, the L-TLE patients were better than R-TLE. Moreover, all the groups achieved better performance when the three aids were given, but the two groups of patients did not match NC performance (compensation was partial). Finally, Table 3 shows that L-TLE patients benet globally from deep processing whereas R-TLE patients enhance performance in cued recall conditions.
Table 2 Means and standard deviations for the four-way interaction Shallow processing (phonetic) E-provided Free recall L-TLE M (SD) R-TLE M (SD) NC M (SD) 18.7 (16) 18 (11) 29.7 (15.7) Cued recall 21 (14) 17.2 (11.9) 32 (15.4) S-produced Free recall 19.2 (8.3) 24.2 (12) 29.7 (14)

4. Discussion When somebody suers from a disease, the rst step is to dene and quantify the disorder and, second, to determine how to cure the disease or minimize its eects. Analysis of the literature on temporal lobe epilepsy and memory showed that many experiments had investigated the extent of verbal memory decits but very few had explored ways of reducing them (Engelberts et al., 2002). The aim of our study was to introduce a compensatory approach by presenting cognitive aids to epileptic patients, in order to investigate a possible alleviation of their memory disorders. The term compensation is used in a wide range of studies concerning neuronal plasticity and/or aging research (e.g., Becker et al., 1996; Helmstaedter & Elger, 1998). In our work, compensation refers to a psychological framework developed by Ba ckman and Dixon (Ba ckman, 1985; Ba ckman & Dixon, 1992), and is dened as a

Deep processing (semantic) E-provided Cued recall 15.6 (14.2) 27.7 (20.5) 46.8 (23.3) Free recall 32.1 (15.4) 26.6 (15.5) 34.4 (10.9) Cued recall 42.4 (24) 51.5 (22.8) 64.1 (15.2) S-produced Free recall 35.3 (7.9) 33.2 (9.6) 43.7 (16.1) Cued recall 64.3 (19.6) 66.8 (17.6) 87.9 (12.3)

Table 3 Dierences between means of performances for the three LOP dimensions Depth Shallow NC L-TLE R-TLE 35 19 22 Deep 58 44 45 Dierence 23 25 23 Elaboration E-provided 40 29 28 S-provided 52 34 38 Dierence 12 5 10 Retrieval Free recall 34 26 26 Cued recall 58 36 41 Dierence 23 10 15

18

C. Bresson et al. / Brain and Language 102 (2007) 1321

process that makes it possible to overcome or reduce verbal memory decits. These processes were explored via the LOP theory (Craik & Lockhart, 1972) by manipulating the depth of processing of information, elaboration, and kind of retrieval. 4.1. Healthy subjects and cognitive aids In the healthy subjects, each cognitive aid was eective, alone and in interaction with another aid, as shown by the fact that their performance improved consistently from the condition with no aid to the one with three aids. Deep (semantic) processing led to better subsequent memory performance than shallow perceptual or physical processing (Craik & Lockhart, 1972; Craik & Tulving, 1975). An elaborated processing enhanced performance (Ma ntyla & Nilsson, 1983). Finally, encoding specicity also had an eect (Tulving & Thomson, 1973). Given a specic type of encoding, retrieval was optimized when the test was designed to use the same type of information. For example, to benet from phonetic encoding, a rhyme-cued recall is eective (Fisher & Craik, 1977; Morris, Bransford, & Franks, 1977). By combining depth of processing and retrieval aids, we found that semantic processing was enhanced by semanticcued recall. Giving these two aids together boosts memory performances. This result is in agreement with previous studies, which showed that rhyme encoding coupled with a rhyme retrieval task resulted in poorer performance than a combination of semantic encoding and semantic retrieval (Fisher & Craik, 1977; Lespinet-Najib et al., 2004; Morris et al., 1977) The second interaction in our study concerned elaboration and depth of processing. We demonstrated that elaboration aid coupled with depth of processing aid in a memorization task improved performance (Kuiper & Rogers, 1979; Maki & McCaul, 1985; Rogers et al., 1977). Finally, the interaction between all three dimensions conrmed that each individual eect was enhanced by the presence of the others, as previously shown by Ma ntyla and Nilsson (1983) or Lespinet-Najib et al. (2004). Our results conrmed that LOP dimensions facilitated memorization and could be used as cognitive aids in the eld of compensation. 4.2. Epileptic patients and cognitive compensation The two groups of patients present verbal memory deficits revealed by the Wechsler memory test. Our results show that both groups beneted from cognitive aids such as deep processing, elaboration, or cued recall. The ANOVA revealed the global eect of each of these factors. Furthermore, the Retrieval-Groups interaction revealed that the eect of the retrieval aid diered from one group to another. L-TLE beneted relatively little from cued recall; whereas R-TLE and NC performance were enhanced, even if R-TLE do not achieve NC performances (compensation for R-TLE was only partial). Furthermore,

the interaction data indicate that the combined eects of the three aids diered from group to group with the following ndings: For phonetic processing: L-TLE did not enhance performance whatever the aid whereas R-TLE improved performances when information is elaborated. So, L-TLE patients suer from a memory decit in encoding phonetic information. When both of the other aids were given, the performance of this group did not improve. These results reect those of NKaoua et al. (2001) and Rains (1987), who reported on decits in an L-TLE group, and support the fact that phonetic processing selectively engages the left cortex (Berman, Mandelkern, Phan, & Zaidel, 2003; Coney, 2002; Demonet et al., 1992, Demonet, Price, Wise, & Frackowiak, 1994). The R-TLE group, however, did benet from cognitive aids at the phonetic processing stage, particularly from elaboration. These results are in agreement with those of Lespinet-Najib et al. (2004) and Mungas et al. (1985), who showed verbal memory decits for shallow levels of processing in an L-TLE group but fewer (Mungas et al., 1985) or none (Lespinet-Najib et al., 2004) in an R-TLE group. For semantic processing: L-TLE patients presented better performance than R-TLE. So, when asked to memorize semantic information, L-TLE patients did improve their performance. This result may be interpreted as reecting the fact that their diculty in engaging semantic processing spontaneously is attenuated by an appropriate semantic aid. The memory performance of R-TLE patients was improved by cued recall, which ndings may also be interpreted as reecting their diculty in engaging retrieval processing, which is overcome by providing a retrieval aid. Finally, all the groups achieved better performance when the three aids were given, but the two groups of patients did not match NC performance (compensation was partial). At this stage, the two groups of patients presented their most important decit in comparison to healthy subjects but at the same time they presented their better performance. So, they did not benet as healthy subjects of the conjunction of the three aids, as a function of the cognitive load of the task. Indeed, L-TLE patients performed as healthy subjects when only one aid (deep processing) was proposed but did not reach their performance when the three aids were given. Another main result of our study is that L-TLE patients had diculty engaging spontaneously in encoding processes (deep processing as to be proposed to enable the improvement of their performance), whereas R-TLE patients had retrieval process impairments (a cue is necessary to enable memory performance improvement). This nding may be interpreted in the light of functional imaging studies. It has often been found that left lateralization

C. Bresson et al. / Brain and Language 102 (2007) 1321

19

of cerebral activity is correlated with the encoding stage, whereas activity on the right is correlated with the retrieval stage (Golby et al., 2001), as in the HERA (Hemispheric Encoding Retrieval Asymmetry) model developed by Tulving, Kapur, Craik, Moscovitch, and Houle (1994). In their article, Tulving and collaborators (1994) reviewed numerous publications and demonstrated that frontal lobes are implicated in encoding and retrieval. But when we examine those publications we can observe that temporal lobes are also implicated (e.g., Squire et al., 1992; Wise et al., 1991). So, the HERA model could also be used in temporal lobe pathology like epilepsy. The asymmetry reported by the HERA model in processing verbal material has been also demonstrated to be independent of material modality (Habib, Nyberg, & Tulving, 2003; Nyberg, Cabeza, & Tulving, 1996) as it has been observed for verbal and non-verbal material. However, these results are not consensual (Blanchet et al., 2001). Campo et al. (2005), using MEG, found right as well as left medial temporal lobe activation during verbal encoding. Finally, Greicius et al. (2003) tested the HIPER (Hippocampus Encoding Retrieval) model of Lepage, Habib, and Tulving (1998), developed using PET studies, which states that the anterior hippocampus is activated by encoding and the posterior hippocampus by retrieval. These authors refuted the HIPER model by nding left and right activation of the entire hippocampus for encoding and retrieval stages using fMRI and a visual word recognition task. However, the results of our study are clearly aligned with the HERA ndings, as patients suering from a left hemisphere injury develop an encoding decit, whereas patients with a right hemisphere epileptic focus develop a retrieval decit. Therefore, the fact that the ability of the L-TLE to compensate depends mainly on the deep processing aid which is given at the encoding stage (they achieved healthy subjects performance), while the R-TLE need retrieval aid given by the cue at the retrieval stage (R-TLE patients always enhanced performance when a cue is provided), may be explained by the use of the residual ability of their damaged hemisphere. Indeed, the left hemisphere is largely implicated at encoding stage and the right hemisphere at retrieval stage (HERA model). So, our ndings that L-TLE patients benet from the encoding aid and R-TLE patients from the retrieval aid are in favor of the hypothesis that they use residual hippocampal ability. This assumption is in agreement with the hippocampal functional adequacy model of the damaged hippocampus described by Chelune (1995), where functional adequacy is dened as the residual capacity of the epileptic hippocampus to perform memory functions. When the three aids are given, they are proposed at encoding and retrieval stages so our results suggest that patients rely on the residual ability of their damaged hippocampus and on their contralateral hemisphere reserves: encoding aids solicit the damaged hippocampus of L-TLE patients and retrieval aids involve their undamaged hemisphere, and conversely for R-TLE patients.

To conclude, from a perspective of rehabilitation, this study is a rst step in understanding the type of cognitive tool to use in neuropsychological sessions. To improve L-TLE memory performance, patients would be helped to use depth of processing more eectively, e.g., by self-reference or self-generated cues. The performance of R-TLE patients is enhanced by all the cognitive aids, provided that a retrieval aid is presented. A cognitive tool dealing with retrieval cues is likely to be eective. Further research is required to observe the impact of dierent cognitive aids on compensation, in order to develop cognitive tools specially adapted to each type of epileptic patients decit, depending on several factors, including laterality, seizure frequency, number of antiepileptic drugs, and age of onset. References
Andersson-Roswall, L., Engman, E., Samuelsson, H., Sjo berg-Larsson, C., & Malmgren, K. (2004). Verbal memory decline and adverse eects on cognition in adult patients with pharmacoresistant partial epilepsy: a longitudinal controlled study of 36 patients. Epilepsy & Behavior, 5, 677686. Ba ckman, L. (1985). Compensation and recoding: a framework for aging and memory research. Scandinavian Journal of Psychology, 26, 193207. Ba ckman, L., & Dixon, RA. (1992). Psychological compensation: a theoretical framework. Psychological Bulletin, 112, 259283. Becker, JT., Mintun, MA., Aleva, BA., Wiseman, MB., Nichols, BS., & DeKosky, ST. (1996). Compensatory reallocation of brain resources supporting verbal episodic memory in Alzheimers disease. Neurology, 46, 692700. Bentin, S., Moscovitch, M., & Nirhod, O. (1998). Levels of processing and selective attention eects on encoding in memory. Acta Psychologica, 98, 311341. Berman, S. M., Mandelkern, M. A., Phan, H., & Zaidel, E. (2003). Complementary hemispheric specialization for word and accent detection. Neuroimage, 19, 319331. Blanchet, S., Desgranges, B., Denise, P., Lechevalier, B., Eustache, F., & Faure, S. (2001). New questions on the hemispheric encoding/retrieval asymmetry (HERA) model assessed by divided visual-eld tachistoscopy in normal subjects. Neuropsychologia, 39(5), 502509. ndez, S., & , F., Ortiz, T., Capilla, A., Ferna Campo, P., Maestu ndez, A. (2005). Is medial temporal lobe activation specic for Ferna encoding long-term memories? NeuroImage, 25(1), 3442. Cermak, L. S. (1979). Amnesic patients level of processing. In L. S. Cermak & F. I. M. Craik (Eds.), Levels of processing in human memory. Hillsdale, NJ: Erlbaum. Cermak, L. S., Reale, L., & Baker, E. (1978). Alcoholic Korsako patients retrieval from semantic memory. Brain and Language, 5(2), 215226. Chelune, G. J. (1995). Hippocampal adequacy versus functional reserve: predicting memory functions following temporal lobectomy. Archives of Clinical Neuropsychology, 10, 413432. Christianson, S. A., Nilsson, L. G., Sa isa , J., & Silfvenius, H. (1992). Visual half-eld testing of memory functions in patients considered for surgical treatment of intractable complex partial epilepsy. Acta Neurologica Scandinavica, 86, 545554. Christianson, S. A., Nilsson, L. G., & Silfvenius, H. (1989). Pre- and postoperative memory of dichotically presented words in patients with complex partial seizures. Neuropsychologia, 27, 427436. Christianson, SA., Silfvenius, H., & Nilsson, L. G. (1987). Hemisphere memory of concrete and abstract information determined with the intracarotid Sodium Amytal test. Epilepsy Research, 1, 185193.

20

C. Bresson et al. / Brain and Language 102 (2007) 1321 Helmstaedter, C. (2002). Eects of chronic epilepsy on declarative memory systems. Progress in Brain Research, 135, 439452. Helmstaedter, C., & Elger, CE. (1998). Functional plasticity after left anterior temporal lobectomy: reconstitution and compensation of verbal memory functions. Epilepsia, 39, 399406. Helmstaedter, C., Gleissner, U., Di Perna, M., & Elger, CE. (1997). Relational verbal memory processing in patients with temporal lobe epilepsy. Cortex, 33, 667678. Helmstaedter, C., & Kurthen, M. (2001). Memory and epilepsy: characteristics, course, and inuence of drugs and surgery. Current Opinion in Neurology, 14, 211216. Jokeit, H., Okujava, M., & Woermann, FG. (2001). Memory fIMR lateralizes temporal lobe epilepsy. Neurology, 57, 17861793. Kuiper, N. A., & Rogers, T. B. (1979). Encoding of personal information: self-other dierences. Journal of Personality and Social Psychology, 37, 499514. Lepage, M., Habib, R., & Tulving, E. (1998). Hippocampal PET activations of memory encoding and retrieval: the HIPER model. Hippocampus, 8, 313322. on, H., Bresson, C., Rougier, A., Lespinet-Najib, V., NKaoua, B., Sauze & Claverie, B. (2004). Levels of processing with free and cued recall and unilateral temporal lobe epilepsy. Brain and Language, 89, 8390. Lockhart, R. S., & Craik, F. I. M. (1990). Levels of processing: a retrospective commentary on a framework for memory research. Canadian Journal of psychology, 44, 87112. Lockhart, R. S., Craik, F. I. M., & Jacoby, L. L. (1976). Depth of processing, recognition and recall. In J. Brown (Ed.), Recall and recognition. New York: Wiley. Maki, R. H., & McCaul, K. D. (1985). The eects of self-reference versus other reference on the recall of traits and nouns. Bulletin of the Psychonomic Society, 23, 169172. Ma ntyla , T., & Nilsson, L. G. (1983). Are my cues better than yours: uniqueness and reconstruction as prerequisites for optimal recall of verbal materials? Scandinavian Journal of Psychology, 24, 303312. Ma ntyla, T., & Nilsson, L. G. (1988). Cue distinctiveness and forgetting: eectiveness of self-generated retrieval cues in delayed recall. Journal of Experimental Psychology: Learning, Memory, and Cognition, 14(3), 502509. Marmurek, H. C. (1995). Encoding, retrieval, main eects and interactions: were Lockhart and Craik (1990) on the level? Canadian Journal of Experimental Psychology, 49(2), 174190. Morris, C. D., Bransford, J. D., & Franks, J. J. (1977). Levels of processing versus tranfert appropriate processing. Journal of Verbal Learning and Verbal Behavior, 16, 519533. Mungas, D., Ehlers, C., Walton, N., & McCutchen, CB. (1985). Verbal learning dierences in epileptic patients with left and right temporal lobe foci. Epilepsia, 26, 340345. Nadel, L., & Moscovitch, M. (2001). The hippocampal complex and longterm memory revisited. Trends in Cognitive Sciences, 228230. NKaoua, B., Lespinet, V., Barsse, A., Rougier, A., & Claverie, B. (2001). Exploration of hemispheric specialization and lexico-semantic processing in unilateral temporal lobe epilepsy with verbal uency tasks. Neuropsychologia, 39, 635642. Nyberg, L., Cabeza, R., & Tulving, E. (1996). PET studies of encoding and retrieval: the HERA model. Psychonomic Bulletin & Review, 3(2), 14135. Rains, G. D. (1987). Incidental verbal memory as a function of depth of encoding in patients with temporal lobe lesions. Journal of Clinical and Experimental Neuropsychology, 9, 18. Reitan, R. R. M., & Wolfson, D. (1985). The HalsteadReitan neuropsychological test battery. Tucson: Neuropsychology Press. Reminger, SL., Kaszniak, AW., Labiner, DM., Littrell, L. D., David, B. T., Ryan, L., et al. (2004). Bilateral hippocampal volume predicts verbal memory function in temporal lobe epilepsy. Epilepsy & Behavior, 2004(5), 687695. Rogers, T. B., Kuiper, N. A., & Kirker, W. S. (1977). Self-reference and the encoding of personal information. Journal of Personality and Social Psychology, 35, 677688.

Cohen, N. J., & Squire, L. R. (1980). Preserved learning and retention of a pattern-analyzing skill in amnesia: dissociation of knowing how and knowing that. Science, 210, 207210. Coney, J. (2002). Lateral asymmetry in phonological processing: relating behavioral measures to neuroimaged structures. Brain and Language, 80, 355365. Content, A., Mousty, P., & Radeau, M. (1990). Brulex : Une base de es lexicales informatise e pour le franc crit et parle . LAnne e donne ais e Psychologique, 90, 551566. Craik, F. I. M., & Lockhart, RS. (1972). Levels of processing: a framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671684. Craik, F. I. M., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General, 104, 268294. Demonet, J. F., Chollet, F., Ramsay, S., Caedebat, D., Nespoulous, J. L., Wise, R., et al. (1992). The anatomy of phonological and semantic processing in normal subjects. Brain, 115, 17531768. Demonet, J. F., Price, C., Wise, R., & Frackowiak, R. S. (1994). A PET study of cognitive strategies in normal subjects during language tasks. Inuence of phonetic ambiguity and sequence processing on phoneme monitoring. Brain, 117, 671682. Eichenbaum, H. (2001). The hippocampus and declarative memory: cognitive mechanisms and neural codes. Behavioural Brain Research, 127, 199207. Einstein, G. O., & Hunt, R. R. (1980). Level of processing and origination: additive eects of individual item and relational processing. Journal of Experimental Psychology: Human Learning and Memory, 6, 588598. Engel, J. (1996). Introduction to temporal lobe epilepsy. Epilepsy Research, 26, 141150. Engelberts, N. H. J., Klein, M., Herman, J. A., Heimans, J. J., , D. G. A., & van der Ploeg, H. M. (2002). Kasteleijn-Nolst Trenite The eectiveness of cognitive rehabilitation for attention decits in focal seizures: a randomized controlled study. Epilepsia, 43(6), 587595. Fisher, R. P., & Craik, F. I. M. (1977). The interaction between encoding and retrieval operations in cued recall. Journal of Experimental Psychology: Human Learning and Memory, 3, 701711. Fisher, RS., Vickrey, BG., Gibson, P., Hermann, B., Penovich, P., Scherer, A., et al. (2000). The impact of epilepsy from patients perspective. I Descriptions and subjective perceptions. Epilepsy Research, 41, 3951. Fujii, T., Okuda, J., Tsukiura, T., Ohtake, H., Suzuki, M., Kawashima, R., et al. (2002). Encoding-related brain activity during deep processing of verbal materials: a PET study. Neuroscience Research, 44, 429438. Gleissner, U., & Elger, CE. (2001). The hippocampal contribution to verbal uency in patients with temporal lobe epilepsy. Cortex, 37, 5563. Golby, A. J., Poldrack, R. A., Brewer, J. B., Spencer, D., Desmond, J. E., Aron, A. P., et al. (2001). Material-specic lateralization in the medial temporal lobe and prefrontal cortex during memory encoding. Brain, 124, 18411854. Greicius, M. D., Krasnow, B., Boyett-Anderson, J. M., Eliez, S., Schatzberg, A. F., Reiss, A. L., et al. (2003). Regional analysis of hippocampal activation during memory encoding and retrieval: fMRI study. Hippocampus, 13(1), 164174. Grith, HR., Pyzalski, RW., Seidenberg, M., & Hermann, B. P. (2004). Memory relationships between MRI volumes and resting PET metabolism of medial temporal lobe structures. Epilepsy & Behavior, 5, 669676. Guerreiro, C. A. M., Jones-Gotman, M., Andermann, F., Bastos, A., & Cendes, F. (2001). Severe amnesia in epilepsy: causes, anatomopsychological considerations, and treatment. Epilepsy & Behavior, 2, 224246. Habib, R., Nyberg, L., & Tulving, E. (2003). Hemispheric asymmetries of memory: the HERA model revisited. Trends in Cognitive Sciences, 7(6), 241245.

C. Bresson et al. / Brain and Language 102 (2007) 1321 Savage, GR., Saling, M. M., Davis, CW., & Berkovic, SF. (2002). Direct and indirect measures of verbal relational memory following anterior temporal lobectomy. Neuropsychologia, 40, 302316. Schwarcz, R., & Witter, MP. (2002). Memory impairment in temporal lobe epilepsy: the role of entorhinal lesions. Epilepsy Research, 50, 161177. Shulman, M. B., & Barr, W. (2002). Treatment of memory disorders in epilepsy. Epilepsy & Behavior, 3, S30S34. Smith, ML. (1989). Memory disorders associated with temporal lobe lesions. In L. Squire & G. Gainotti (Eds.), Handbook of neuropsychology (pp. 93106). New York: Elsevier. Squire, L. R., Ojemann, J. G., Miezin, F. M., Petersen, S. E., Videen, T. O., & Raichle, M. E. (1992). Activation of the hippocampus in normal humans: functional anatomical study of memory. Proceeding of the National Academy of Sciences United States of the America, 89, 18371841. Symons, C. S., & Johnson, B. T. (1997). The self-reference eect in memory: a meta-analysis. Psychological Bulletin, 121(3), 371394.

21

Tro ster, A. I., Warmash, V., Osorio, I., Paolo, AM., Alexander, LJ., & Barr, WB. (1995). The role of semantic networks and search eciency in verbal uency performance in intractable temporal lobe epilepsy. Epilepsy Research, 21, 1926. Tulving, E., Kapur, S., Craik, F. I. M., Moscovitch, M., & Houle, S. (1994). Hemispheric encoding/retrieval asymmetry in episodic memory: position emission tomography ndings. Proceeding of the National Academy of Sciences United States of the America, 91, 20162020. Tulving, E., & Thomson, DM. (1973). Encoding specicity and retrieval processes in episodic memory. Psychological Review, 80, 352373. Wechsler, D. (1997, 2001). Wechsler memory scale (3rd edition), Psychological Corporation, New York. eme Wechsler, D. (2000). Echelle dintelligence de Wechsler pour Adultes, 3 dition. Les Editions du centre de Psychologie Applique e. e Wise, R., Chollet, F., Hadar, U., Friston, K., Honer, E., & Frackowiak, R. (1991). Distribution of cortical neural networks involved in word comprehension and word retrieval. Brain, 114, 18031817.