SEARCH

SEARCH BY CITATION

Keywords:

  • Epilepsy;
  • Memory;
  • Intracarotid amobarbital test;
  • Anterior temporal lobectomy;
  • Prediction

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Summary:  Purpose: The intracarotid amobarbital test (IAT) has been shown to predict verbal memory changes after anterior temporal lobectomy (ATL). Seeking to extend these findings, we examined two questions: (a) What is the relationship between material-specific aspects of IAT memory and material-specific memory changes after ATL? and (b) Which IAT memory score(s) optimally predict memory changes after surgery, the memory score after injection ipsilateral to the seizure focus, the memory score after injection contralateral to the seizure focus, or the IAT asymmetry score, comprising the ipsilateral minus contralateral injection scores?

Methods: Seventy left hemisphere language-dominant patients undergoing ATL for treatment of medically refractory seizures were administered a verbal and visuospatial recognition memory test before surgery and 3 weeks after surgery. IAT memory recognition scores for words and designs were used to predict verbal and visuospatial memory changes after surgery.

Results: After surgery, left ATL patients declined in verbal memory, whereas right ATL patients declined in visuospatial memory. IAT total recognition memory scores (collapsed across all types of materials) and IAT word memory scores were associated with postoperative verbal memory decline. This relationship was significant for the IAT ipsilateral injection memory scores and the IAT hemispheric asymmetry scores. IAT memory performances were not related to visuospatial memory changes.

Conclusions: Results indicate IAT memory measures to be related to postoperative verbal, but not visuospatial, memory change. A specific relationship was found between postoperative verbal memory change and IAT verbal memory after injection ipsilateral to the seizure focus, when relying primarily on the contralateral hemisphere. This finding is consistent with the functional reserve model of memory change in ATL.

Studies of patients undergoing anterior temporal lobectomy (ATL) for the treatment of medically refractory complex partial epilepsy suggest that memory processing for different types of material tends to be lateralized. After ATL in the left, language dominant, hemisphere (LATL), there is a significant decline in verbal memory, taking the form of poor recall of short stories (1–6), difficulty with word-list learning tasks (5–7), as well as difficulty learning and recalling unrelated word pairs (6,8,9). Although evidence is somewhat conflicting in regard to nonverbal memory, several studies have shown visuospatial memory deficits after ATL in the nondominant right temporal lobe (RATL) (10), apparent in disturbed learning of new, unfamiliar faces (11–13), musical compositions (14), and designs and patterns (15).

Not all patients undergoing ATL have equally severe memory changes after surgery. Age at onset of neurologic problems has been determined to be related to the degree of postoperative cognitive decline. Specifically, those patients with neurologic injury earlier in life have shown greater sparing of language and memory functions after unilateral ATL (16,17). This effect is presumed to be due to the reorganization of brain functioning after damage to critical temporal lobe structures early in life. The location and extent of surgery are also significant variables affecting postoperative cognitive outcome. Despite much early evidence supporting the idea that larger hippocampal resection results in greater decrement in memory functions (11,18), recent evidence indicates a more complicated interaction. Specifically, a larger resection may produce no more cognitive decrements than a smaller resection, when the hippocampal tissue removed is dysfunctional (19–21). Cognitive functioning is also related to the extent of seizure control after surgery. Patients who become seizure free after surgery tend to experience fewer material-specific memory deficits (22). The functional integrity of the contralateral nonresected tissue has been associated with postoperative cognitive outcome. Patients with extensive atrophy of the contralateral hippocampus experience more severe cognitive deficits after surgery, which is presumed to be due to the inability of contralateral brain structures to compensate for resected tissue (23–25).

Preoperative performance on the intracarotid amobarbital test (IAT) is another factor related to postoperative memory outcome. The IAT was developed initially to ascertain atypical hemispheric lateralization of language functions in patients who were being evaluated for resective surgery (26). Subsequently the test was adapted to examine the relative contribution of each hemisphere to long-term memory to identify those patients at risk for developing global amnesia after surgery (27). Global amnesia after ATL is a rare occurrence, however. Therefore, more recent attention has focused on the usefulness of the IAT in predicting the material-specific (verbal or visuospatial) memory changes that occur more commonly after ATL.

Several studies have examined the relationship between IAT memory and postoperative memory outcome after ATL. Wyllie et al. (10) compared two different IAT scoring methods for predicting material-specific decline after ATL. IAT memory stimuli included a maximum of 20 items, including four objects, two nonverbal drawings, five written words, one color, one number, three pictures, and four spoken phrases. Item presentation occurred in this order and was discontinued with the return of normal strength. On average, 16 items were presented after the left injection, and 18 items were presented after the right injection. The first scoring method involved the use of an absolute IAT memory retention score for each hemisphere. A separate score was computed for each hemisphere, comprising performance on all presented stimuli. The second IAT scoring method used a comparison score between the left and right hemispheres (i.e., failure in a given hemisphere was defined as a retention score that was at least 20% lower than that of the other hemisphere). This method also used both verbal and nonverbal stimuli. Wyllie et al. (10) did not find that IAT absolute retention scores (the first method described) predicted material-specific memory changes after either left or right ATL. When they used the comparative method, however, correlations were noted between IAT scores and postoperative material-specific memory change. Although differences on pre- to postoperative material-specific memory change scores between the group that passed the IAT and the group that failed the IAT were not significant, results suggested that the IAT interpreted comparatively may be helpful in the assessment of relative risk for material-specific memory decline after surgery.

Loring et al. (7) also made use of IAT hemispheric asymmetry scores, in the prediction of postoperative memory functioning. This study used the difference between left hemisphere injection and right hemisphere injection IAT recognition memory scores for dually encoded real objects to predict material-specific memory decline after ATL. Results indicated that greater verbal memory decline after left ATL is associated with less asymmetry in hemispheric IAT memory scores (7), which can be taken to indicate either that medial temporal lobe structures ipsilateral to the resection are less sclerosed, or that medial temporal lobe structures contralateral to the resection are more impaired.

Kneebone et al. (4) examined the relationship between an IAT dichotomous pass/fail memory score after injection of the contralateral, nonsurgical hemisphere and material-specific memory decline after ATL. In this study, IAT stimuli consisting of objects, designs, words, colors, numbers, and spoken phrases were combined into one total score for each hemisphere. These authors found that patients with a left-sided seizure focus who “passed” the IAT after the contralateral, right hemisphere injection (when they were relying primarily on the epileptic left hemisphere) experienced greater verbal memory decline after left ATL than did those who “failed” the IAT after the contralateral, right hemisphere injection.

In summary, findings to date suggest that IAT memory performances may provide useful information regarding the extent of verbal memory decline after surgery. However, this recent research is unclear in terms of the relative utility of various IAT memory scores in this prediction. Two of the studies previously cited indicate IAT hemispheric differences to be most useful for predicting postoperative memory decline (7,10). Conversely, Kneebone et al. (4) found that memory performance after injection of the nonepileptic hemisphere was related to postoperative memory change.

In an attempt to provide a framework for the interpretation of such conflicting data, Chelune (28) presented two, not mutually exclusive ways to understand the relationship between presurgical memory functioning and postsurgical memory performance. The functional reserve model posits that postsurgical memory deficits are related to the ability of the medial temporal lobe structures contralateral to the seizure focus to support memory functions. It assumes that memory capacity after unilateral temporal lobe resection is a direct function of functional integrity of the contralateral temporal tissue. Therefore, measures demonstrating greater impairment of function in the hemisphere contralateral to seizure focus would relate to worse memory outcome. Alternatively, the functional adequacy model states that postsurgical memory deficits are the result of the adequacy of the ipsilateral temporal lobe (that which is resected) to support memory functions. This model assumes that the removal of compromised brain tissue will not adversely affect memory functioning, whereas the removal of functionally healthy tissue will compromise memory abilities.

The purpose of this study was to clarify and extend this line of investigation through the examination of two specific questions:

  • 1
    What is the relationship between material-specific aspects of IAT memory performance and material-specific memory changes after ATL? Material-specific aspects of memory functions refer to memory capabilities for only one type of material (e.g., words or designs). Given the evidence supporting the hemispheric lateralization of memory functions, it would follow that material-specific aspects of the IAT might better predict material-specific memory changes after ATL. To our knowledge, the relationship between material-specific IAT memory measures and memory after ATL has not been examined to date. We expected that IAT measures of verbal memory would be specifically related to postoperative verbal memory changes, whereas IAT measures of nonverbal visuospatial memory would be related to postoperative changes in visuospatial memory functions.
  • 2
    Another question addressed in this study concerns whether changes in memory performance after ATL can be explained best by the model of functional adequacy or functional reserve or both. Studies to date provide contradictory results as to which IAT score is most useful in predicting material-specific memory changes after surgery: IAT memory scores after injection of the hemisphere ipsilateral to the seizure focus (ipsilateral injection), IAT memory scores after injection contralateral to the seizure focus (contralateral injection), or IAT memory asymmetry scores (IAT memory performance after ipsilateral minus contralateral injection)? According to the functional reserve hypothesis, we would expect that memory performance after IAT injection ipsilateral to the seizure focus would be related most strongly to postsurgical changes in memory function. According to the functional adequacy hypothesis, memory performance after IAT injection contralateral to the seizure focus would be expected to be most strongly related to postsurgical changes in memory function. Alternatively, it may be that the relative capacities of the two hemispheres, captured in an asymmetry score, best predict memory change after ATL.

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Participants

Participants were 70 patients with intractable partial epilepsy who underwent ATL (right ATL, n = 42; left ATL, n = 28) following the protocol described by Sperling et al. (29). All patients included in the study were determined to be left hemisphere language dominant, based on the IAT (30). There were no significant differences between the right and left ATL groups in terms of age at surgery, age at first neurologic risk, or age at seizure onset (Table 1). WAIS-R Full-Scale IQ was significantly higher in the right ATL group as compared with the left ATL group. The two groups also differed significantly in terms of mean years of education, with the right ATL group having a slightly higher mean education than the left ATL group.

Table 1.  Demographic characteristics for epilepsy patients after right and left anterior temporal lobectomy
 Right ATL (n = 42)Left ATL (n = 28)t
MSDMSD
  • ATL, anterior temporal lobectomy; FSIQ, full-scale IQ.

  • a

    p < 0.05.

Age (yr)34.8110.8631.298.811.43
Education (yr)13.382.2112.461.432.11a
Age at first risk (yr)11.5911.977.539.721.49
Age at seizure onset  (yr)16.111.6710.9410.061.90
WAIS-R FSIQ94.0712.7187.7512.742.04a

IAT procedure

Each patient underwent the IAT prior to surgery, following the protocol described by Glosser et al. (30). A catheter is placed into an internal carotid artery to the junction with the middle cerebral artery via a transfemoral approach. Vascular anatomy is defined with a contrast injection (arteriogram). Sodium amobarbital (100–125 mg, diluted in 5 ml saline) is then injected by hand over a 5-s interval to produce a contralateral hemiparesis. The hemisphere ipsilateral to the suspected seizure focus is injected first. Approximately 45 min later, after an arteriogram of the hemisphere contralateral to the suspected seizure focus, an equivalent amount of amobarbital is injected into the contralateral hemisphere.

Behavioral testing begins immediately after verification of the hemiparesis. Language functioning is tested first. Approximately 90 s after injection, presentation of to-be-remembered memory stimuli is initiated. Memory stimuli consist of three common objects, which can be dually encoded using verbal or visual representations, presented for naming, three low-imagery words presented for reading aloud, and three line drawings of unfamiliar nonverbalizable abstract designs that are presented for visual inspection. This noted order of stimulus presentation was uniform across subjects to maintain consistency for clinical interpretation of the results. After return to baseline neurologic functioning, ∼10 min after injection, memory retention is assessed. A recognition memory paradigm is used with the nine targets presented among 18 foils in a standard order for yes/no forced-choice recognition. To account for guessing, a corrected memory recognition score is used, wherein the proportion of false-alarm errors (affirmative response when presented with a distractor) is subtracted from the proportion of hits (correct identifications of target items). Thus the corrected total memory recognition score is computed as follows: (number of hits/9) – (number of false alarms/18). Material-specific memory recognition scores are also computed separately for objects, words, and designs using a similar formula: (number of hits/3) – (number of false alarms/6).

Alternate test forms are used for the two injections. Order of administration of the forms is randomized over side of injection and side of focus across patients. Previous analyses indicate no significant differences between the two forms (31).

Dependent IAT measures include the corrected memory recognition scores after injection ipsilateral to the seizure focus, the corrected memory recognition scores after injection contralateral to the seizure focus, and the asymmetry scores consisting of the difference between corrected recognition memory scores after injection of the epileptic hemisphere minus corrected memory recognition score after injection of the nonepileptic hemisphere. Four hemispheric asymmetry scores were computed: (a) total recognition asymmetry score, summed over objects, words, and designs; (b) object recognition asymmetry score; (c) word recognition asymmetry score; and (d) design recognition asymmetry score. These asymmetry scores quantify the relative memory capabilities of the two hemispheres, with a positive number indicative of better memory capabilities after ipsilateral injection, when patients were relying primarily on the nonepileptic hemisphere, and a negative number indicative of better memory capabilities when relying primarily on the epileptic hemisphere.

Pre- and postoperative neuropsychological testing

The California Verbal Learning Test (32) and the Graduate Hospital Facial Memory Test (33) were administered once during a comprehensive preoperative neuropsychological evaluation that took place 2–6 months before surgery, and then a second time ∼3 weeks after surgery, when patients were neurologically stable.

The California Verbal Learning Test (CVLT) assesses learning of a supraspan list of 16 words, drawn from four semantic categories, which are presented over repeated trials. Retention of the words after different periods of filled delay is also assessed. A score from the delayed recognition condition was used as a measure of verbal memory in the analyses reported later, because it is most comparable to the measure of visuospatial memory, which is also derived from a recognition memory paradigm. A memory recognition score that corrects for false-positive errors was computed by using the hit rate (true positives/number of targets) minus false-alarm rate (number of false positives/number of distractors) (34).

The Graduate Hospital Facial Memory Test (FMT) (33) requires the patient to examine a page consisting of 20 photocopies of unfamiliar faces for 1 min. After a 30-min delay, the patient is asked to identify previously studied faces from an array of 40 faces (20 targets and 20 distractors). A recognition memory score is computed by subtracting the number of omission and commission errors from the number of correct identifications. This task is used as a measure of visuospatial memory, and it is similar to that used by Milner (5). However, difficulty is increased by the use of partially degraded stimuli. Saykin et al. (35) determined delayed recognition scores for the normative group to be 27.59 ± 7.00 (M ± SD).

To estimate the changes in memory related to ATL, difference scores computed on each of these tests consisted of the preoperative recognition discrimination score minus postoperative recognition discrimination score.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

A three-way repeated measures analyses of variance of corrected material-specific memory scores (Table 2) with Side of surgery (left vs. right) as the between-groups factor and side of injection (left vs. right) and Type of stimulus (words vs. designs) as within-groups factors showed a significant main effect of Side of injection [F(1, 68) = 12.86, p < 0.001]. Overall, patients obtained higher memory scores after injection of the right hemisphere. A significant interaction of Side of IAT injection × Side of surgery was also noted [F(1, 68) = 66.04, p < 0.001], such that patients undergoing LATL performed better on the IAT after injection of the epileptic hemisphere, when relying primarily on the nonepileptic hemisphere. Similarly, patients undergoing RATL performed better on the IAT after injection of the right, epileptic hemisphere, when relying primarily on the nonepileptic hemisphere. There was also a significant interaction of Side of injection × Type of stimulus [F(1, 68) = 24.09, p < 0.001]. IAT memory performance with words was significantly better when using the left hemisphere than the right hemisphere, whereas IAT memory for designs was slightly better when using the right hemisphere than the left hemisphere. The three-way interaction of Side of surgery × Side of injection × Type of stimulus was significant [F(1, 68) = 4.23, p < 0.05]. The previously noted Material × Side of injection interaction occurred in all IAT conditions except in the RATL group undergoing left hemisphere injection (Fig. 1). Specifically, LATL patients undergoing right hemisphere injection performed at higher levels in memory for words than designs [t(1, 27) = –2.37, p < 0.05]. In contrast, when undergoing left hemisphere injection, LATL patients performed significantly better with memory for designs than memory for words [t(1, 27) = 4.28, p < 0.001], when relying primarily on the right hemisphere. RATL patients also performed at significantly higher levels in memory for words than for designs when undergoing right hemisphere injection, relying on the left hemisphere [t(1, 41) = –2.85, p < 0.01]. Although RATL patients undergoing left hemisphere injection performed slightly better with memory for designs than with memory for words, this difference was not significant [t(1, 41) = 0.69, p < 0.49]. This latter finding is potentially because more RATL patients become obtunded after injection of the left hemisphere, consequently obtaining lower memory scores (36). However, the percentage of RATL patients that became obtunded after left injection was minimal, with percentages consistent with those in previously published research (36).

Table 2.  Mean scores on intracarotid amobarbital test memory stimuli by group (right anterior temporal lobectomy versus left anterior temporal lobectomy)
 Right ATL Mean (SD)Left ATL Mean (SD)
  • a

    Using the left hemisphere.

  • b

    Using the right hemisphere.

Right injectiona  
 Words2.06 (0.68)1.14 (1.0)
 Designs1.56 (1.04)0.52 (1.0)
 Objects2.86 (0.32)1.46 (1.78)
 Total6.52 (1.65)3.13 (2.08)
Left injectionb  
 Words0.40 (0.84)0.87 (0.86)
 Designs0.55 (0.99)1.82 (0.83)
 Objects1.33 (1.2)2.57 (0.65)
 Total2.35 (2.11)5.29 (1.53)
image

Figure 1. Means for the right anterior temporal lobectomy group (RATL; n = 42) and the left anterior temporal lobectomy group (LATL; n = 28) on the word memory recognition score (corrected for guessing) and the design memory recognition score (corrected for guessing) for the intracarotid amobarbital test (IAT) by hemisphere of injection (right, Rinj; left, Linj).

Download figure to PowerPoint

Two-way repeated-measures analyses of variance (ANOVAs) were performed to assess effects of RATL and LATL on verbal (CVLT) and visuospatial (FMT) recognition memory. Side of surgery (RATL vs. LATL) served as the between-groups factor, whereas Time of testing (preoperative vs. postoperative testing) served as the within-groups factor in each analysis.

An ANOVA of the CVLT corrected recognition scores did not show a significant main effect for Side of surgery or Time of testing. However, there was a significant interaction of Side of surgery × Time of testing [F(1, 68) = 11.58, p < 0.001], such that patients who underwent LATL showed a decrease in CVLT discriminability [t(26) = 19.89, p < 0.001] after surgery, whereas the patients who underwent RATL showed an increase in CVLT discriminability [t(40) = 38.08, p < 0.001] after surgery (Table 3).

Table 3.  Means (standard deviations) on the California Verbal Learning Test Delayed Recognition Score and the Facial Memory Test Delayed Recognition Score by group (right vs. left anterior temporal lobectomy)
 Right ATL Mean (SD)Left ATL Mean (SD)
CVLT recognition discrimination,  preoperative0.77 (0.17)0.79 (0.14)
CVLT recognition discrimination,  postoperative0.85 (0.15)0.73 (0.20)
FMT delayed recognition,  preoperative43.71 (3.72)45.46 (4.82)
FMT delayed recognition,  postoperative43.17 (4.17)46.11 (4.05)

Repeated-measures ANOVA of the FMT delayed recognition memory scores showed a significant main effect of Side of surgery [F(1, 68) = 8.15, p < 0.01]. Overall, patients who underwent RATL performed at lower levels than did those who underwent LATL. The main effect of Time of testing was not significant. However, there was a significant interaction of Time of testing × Side of surgery [F(1, 68) = 4.25, p < 0.05], such that the RATL group showed a slight postoperative decline [t(40) = 10.69, p < 0.001], whereas the LATL group showed a substantial improvement [t(26) = 13.19, p < 0.001]. The aforementioned analyses were conducted by using the raw scores, but neuropsychological test means are presented as normalized Z-scores in Fig. 2 to illustrate the relationship of patients' memory performances to those of the general population. The Z-scores were computed by using means and standard deviations from published normative studies of the Graduate Hospital Facial Memory Test (35) and the California Verbal Learning Test (32).

image

Figure 2. Standardized scores for California Verbal Learning Test (CVLT) recognition performance and Facial Memory Test (FMT) recognition performance before to after anterior temporal lobectomy (ATL) by surgical group (left ATL vs. right ATL). Z scores are computed from normative studies on the CVLT (32) and the FMT (35).

Download figure to PowerPoint

Having established that the IAT yields material-specific measures related to the side of injection and having shown material-specific postoperative changes on the CVLT and FMT to be related to side of surgery, the relationship of IAT memory scores to memory outcome after ATL was examined next.

Multiple regression analyses were performed in which the independent, predictor, variable was the IAT memory asymmetry score (ipsilateral minus contralateral injection), and the dependent variable was the difference in memory performance before to after surgery. In one set of analyses, the CVLT preoperative to postoperative memory recognition difference score was the dependent measure, and in the other set of analyses, the FMT preoperative to postoperative memory recognition difference score was the dependent measure. Given the known relationship of age at first neurologic risk factor (16,17,37), and Full-Scale IQ (FSIQ) (38) to postoperative memory changes, as well as the potential confounding effect of group differences in education on postoperative cognitive outcome, the contribution of these three variables was considered before testing the effect of the IAT score on memory outcome. Therefore, the basic regression model forced FSIQ, education, and age at first neurologic risk factor for epilepsy into the equation first. The IAT memory recognition score was then entered into the equation as the final independent variable, with the exact score entered differing with each model (e.g., IAT total asymmetry score, IAT word asymmetry score).

With the IAT total memory asymmetry score as the predictor, the regression equation for predicting the CVLT recognition difference score was significant [F(4, 65) = 2.78, p < 0.05]. The model accounted for 14.63% of the variance in the CVLT recognition difference score (Table 4, Model 1). The IAT total asymmetry score significantly predicted the CVLT pre–post difference score [t(68) = 2.32, p < 0.05] after the variance associated with age at first risk, education, and WAIS-R FSIQ was removed from the equation. A greater asymmetry between IAT memory performance after ipsilateral injection and performance after contralateral injection was predictive of a smaller decline, or an improvement, on the postoperative CVLT. In this analysis as well as all subsequent analyses, WAIS-R FSIQ, education, and age at first risk were not significantly related to postoperative changes in verbal memory. With a similar regression analysis (Table 4, Model 2), the proposed model for predicting pre- to postoperative differences on the FMT was not significant [F(4, 65) = 0.18, p = 0.95]. Thus, unlike verbal memory outcome, visuospatial memory change after ATL was not significantly related to the total memory asymmetry score on the IAT.

Table 4.  Summary of hierarchical regression analyses for variables predicting California Verbal Learning Test difference score and Facial Memory Test difference score before to after surgery in 70 patients undergoing anterior temporal lobectomy
ModelaPredictor variableDependent variableTotal R2Adjusted R2
  • IAT, intracarotid amobarbital test; MAS, memory asymmetry score (IAT total memory score after ipsilateral injection minus IAT total memory score after contralateral injection).

  • a

    Model: All regression models are as follows: dependent variable: CVLT or FMT (specified within each model); predictor variables: Block 1, Education, WAIS-R Full Scale IQ; Block 2, Age at which the first neurologic risk factor for epilepsy was sustained; Block 3, IAT memory score (specified within each model).

  • b

    p < 0.05

  • c

    p = 0.056.

Model 1IAT total MASCVLT0.1460.094b
Model 2IAT total MASFMT0.011−0.050
Model 3IAT word MASCVLT0.1270.073c
Model 4IAT word MASFMT0.017−0.043
Model 5IAT design MASFMT0.010−0.051
Model 6IAT design MASCVLT0.0820.026
Model 7IAT total memory score after ipsilateral injectionCVLT0.1530.101b
Model 8IAT word memory score after ipsilateral injectionCVLT0.1640.113b
Model 9IAT total memory score after contralateral injectionCVLT0.0860.029
Model 10IAT word memory score after contralateral injectionCVLT0.0770.020
Model 11IAT total memory score after ipsilateral injectionFMT0.011−0.050
Model 12IAT design memory score after ipsilateral injectionFMT0.017−0.044
Model 13IAT total memory score after contralateral injectionFMT0.010−0.051
Model 14IAT design memory score after contralateral injectionFMT0.021−0.039

The specific contribution of IAT verbal memory measures to predicting pre- to postoperative memory difference scores was examined in a series of multiple regression analyses. The first analysis used the IAT word memory asymmetry score as a predictor of CVLT pre–post memory difference, controlling for age at first risk, education, and WAIS-R FSIQ. The model approached significance [F(4, 65) = 2.36, p = 0.06], and accounted for 12.67% of the variance in the CVLT pre–post difference score (Table 4, Model 3). A greater asymmetry in IAT memory performance was related to a smaller decline, or an improvement, on the postoperative CVLT recognition score. The IAT verbal memory asymmetry score did not predict changes on the FMT (Table 4, Model 4).

To examine the specific contribution of IAT design memory recognition scores to predicting postoperative memory change, a multiple regression was performed by using the IAT design memory asymmetry as a predictor of FMT pre–post operative memory change, controlling for age at first risk, education, and WAIS-R FSIQ. The overall model was not significant, nor were any of the predictor variables significantly related to outcome. In contrast to the relationship between IAT word memory and verbal memory outcome, no significant relationship was noted between IAT design memory and visuospatial memory outcome (Table 4, Model 5). Additionally, IAT design memory asymmetry scores did not predict changes in verbal memory measures after ATL (Table 4, Model 6).

Finally, a series of analyses was conducted to determine if the observed relationship between IAT memory asymmetry scores and verbal memory outcome was related to performance after injection of the hemisphere ipsilateral to the seizure focus, when the patient is presumed to be relying primarily on the nonepileptic hemisphere, or performance after injection contralateral to the seizure focus, when relying primarily on the epileptic hemisphere. Multiple regressions were performed with IAT total memory recognition scores for injection of each hemisphere separately as predictor variables, and the CVLT pre- to postoperative difference score as the dependent variable, controlling for education, FSIQ, and age at first neurologic risk. The IAT total memory score after ipsilateral injection was significantly related to CVLT recognition change score [F(4, 65) = –2.43, p < 0.05]. This model accounted for 15.28% of the variance in the CVLT pre–post difference score (Table 4, Model 7). A similar multiple regression analysis performed by using the IAT word memory score after injection of the ipsilateral hemisphere was also significant [F(4, 65) = 3.20, p < 0.05], and the model accounted for 16.4% of the variance associated with the CVLT pre–postoperative difference score (Table 4, Model 8).

By contrast, multiple regression analyses using the IAT total memory score and IAT word memory score after contralateral injection did not show a significant relationship to the pre- to postoperative CVLT difference scores (Table 4, Models 9, 10).

The same set of multiple regression analyses using unilateral IAT memory scores, as described earlier, were performed by using FMT change scores as dependent measures. When either ipsilateral injection or contralateral injection was considered separately, there was no relationship between IAT total memory or IAT design memory scores and changes on the FMT pre- to postoperatively (Table 4, Models 11–14).

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

There are five major results of this study:

  • 1
    This study rEPIicates well-recognized findings indicating reliable changes in verbal memory and visuospatial memory after unilateral ATL. LATL patients showed a postoperative decline on a verbal learning test, whereas the RATL patients showed a slight improvement. Conversely, with visuospatial materials, RATL patients showed a slight decline in recognition memory for unfamiliar faces postoperatively, whereas LATL patients showed an improvement.
  • 2
    Before surgery, material-specific hemispheric differences in recognition memory were also apparent on the IAT. After right hemisphere injection, when patients were relying primarily on the language-dominant left hemisphere, they showed better memory for words than for designs, whereas after left hemisphere injection, memory for designs tended to be better than memory for words.
  • 3
    When examining the relationship between these two sets of measures, we found that a greater difference or asymmetry in memory scores between ipsilateral and contralateral injection on the IAT significantly predicted changes in verbal memory after ATL, even after controlling for the variance associated with WAIS-R FSIQ, education, and age at first neurologic risk. The total memory hemispheric asymmetry score (combined for words, objects, and designs) and the word memory hemispheric asymmetry score predicted verbal memory outcome after ATL, but the design memory hemispheric asymmetry score did not. IAT performance was not significantly related to changes in visuospatial facial recognition memory after ATL.
  • 4
    Results indicated that IAT memory performance after the ipsilateral injection, when patients are relying primarily on the nonepileptic hemisphere for memory encoding, predicts changes in verbal memory after ATL, providing evidence for the functional reserve theory of postoperative memory change (28).
  • 5
    IAT memory performance after contralateral injection was not related to changes in postoperative material-specific memory functioning, although this result must be interpreted cautiously, as patients were specifically selected to undergo IAT if there was substantial evidence of a unilateral medial temporal lobe abnormality on the side of the seizure focus.

Results of this study support previous research documenting a significant, but modest, relationship between IAT memory scores and postsurgical changes in verbal memory (4,7). Previous studies noted greater verbal memory decline after ATL to be associated with less IAT memory asymmetry between performances after injection of the ipsilateral and contralateral hemispheres, a finding rEPIicated with our patient series. There is greater verbal memory decline after ATL in patients who presurgically demonstrate less of a difference in total recognition memory between ipsilateral and contralateral IAT injection. In contrast to previous studies that examined the relationship between IAT memory for dually encodable stimuli and postoperative material-specific memory changes, we also assessed the relationship between material-specific IAT memory performance and verbal memory outcome. Greater decline or less improvement in verbal memory after ATL was predicted by less IAT memory asymmetry for low-imagery words. IAT design memory scores did not predict verbal memory outcome, and IAT verbal memory did not predict visuospatial memory outcome.

The relationship noted between the IAT memory asymmetry scores and the verbal memory change score remains significant when controlling for age at first risk and WAIS-R FSIQ, both factors known to be related to postsurgical cognitive outcome. Although WAIS-R FSIQ has been shown to influence postsurgical cognitive outcome (38), in the present study, overall intellectual capacities were not a significant predictor of postoperative changes in verbal memory abilities. Also in contrast to previous research showing that that age of first brain insult is related to postoperative cognitive changes (16,17), this factor was not a significant predictor of postsurgical memory outcome in our analyses. However, the IAT total recognition memory asymmetry score and the IAT word recognition memory asymmetry score were significant predictors of verbal memory change in this patient series, after controlling for the variance associated with these other variables.

In contrast to positive findings for verbal memory outcome, no significant relationship was found between IAT memory measures and changes in visuospatial, facial memory scores after surgery. Facial memory scores were not related to IAT performance, although there was a reliable decline in facial memory after RATL and despite the finding of hemispheric differences in material-specific visuospatial memory on the IAT. Our failure to find a relationship between IAT memory performance and postoperative changes in visuospatial memory is consistent with other studies that have yielded mixed results regarding deficits in visuospatial memory after ATL. Several factors may underlie this inconsistency. One possibility is that visuospatial memory is organized in a more diffuse manner in the temporal lobe than is verbal memory, an explanation offered by other researchers (16,39,40). So there would be a less robust relationship between degree of pathology in the medial temporal lobe and degree of memory impairment. Alternatively, different tasks may tap into aspects of visuospatial memory mediated by different brain regions. The IAT memory test using recognition of abstract designs may reflect functions of different brain regions than the facial recognition memory test used to measure pre- to postoperative visuospatial memory change. Studies of visual processing in animals and humans have indicated distinct memory subsystems for the processing of complex pattern information, such as unfamiliar faces, and spatial-location information, parallel to the different dorsal and ventral visual systems (12,13,41–44). Considering these different visuospatial memory systems, future studies may need to use the same types of stimuli on the IAT as in pre- and postsurgical neuropsychological evaluations to attain a predictive relationship between the two tests.

In attempting to explain memory decline after ATL, reference is frequently made to the theories of functional adequacy and functional reserve (28). The distinguishing feature between the two theories is the relative importance of each hemisphere's presurgical functional capacities in predicting postoperative memory changes. The validity of these models can be tested by assessing whether ipsilateral or contralateral IAT injection is more predictive of postoperative memory change. Results of this study showed that the IAT memory recognition scores, when relying on the hemisphere contralateral to the seizure focus, were sufficient in predicting pre–post verbal memory difference scores. There was no significant relationship between IAT memory using the hemisphere ipsilateral to the seizure focus and verbal memory outcome. These findings indicate that the functional integrity or reserve of the mesial temporal lobe tissue contralateral to the seizure focus may be more important in determining memory capacities postoperatively than the functional adequacy of the to-be-resected brain region. However, a role for functional adequacy of the ipsilateral medial temporal lobe cannot be discounted because of the preselection of patients undergoing IAT. We note that similar results showing the importance of functional reserve in the contralateral mesial temporal lobe were obtained in an functional magnetic resonance imaging (fMRI) memory activation study that examined memory changes after ATL (45).

Although this study provides compelling evidence in support of the functional reserve theory of postoperative memory change (28), the current methodology did have some limitations. Specifically, because of the fast-paced nature of the IAT and the speed with which patients recover from the effects of amobarbital, the number of stimuli presented during the IAT was limited. This limitation restricts the range of IAT scores that can be obtained for individual patients, thus limiting the variance and consequent predictive power. Additionally, there is an obvious difference in the type of visuospatial memory stimuli used in the IAT and the type of visuospatial memory measure used for pre- and postoperative neuropsychological evaluation. Whereas it is difficult to change the memory stimuli used in a procedure such as the IAT, future studies should attempt to engage the same type of visuospatial memory in both neuropsychological assessment and the IAT (i.e., using faces in both protocols). Despite these limitations, the current study indicates a specific relationship between postoperative verbal memory change and IAT verbal memory performance after injection ipsilateral to the seizure focus, when patients are relying primarily on the contralateral hemisphere. This finding is consistent with the functional reserve theory of postoperative memory change.

Acknowledgment: Our thanks to Drs. Gordon Baltuch, Jacqueline French, Joyce Liporace, Michael O'Connor, Joseph Sirven, and Michael Sperling for the opportunity to examine their patients.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    Milner B. Intellectual function of the temporal lobes. Psychol Bull 1954;51:4262.
  • 2
    Chelune G, Awad I, Luders H. Verbal memory deficits after temporal lobectomy: independent or confounded by language. Epilepsia 1989;30:712.
  • 3
    Chelune G, Naugle R, Luders H, et al. Individual change after epilepsy surgery: practice effects and base-rate information. Neuropsychology 1993;7:4152.DOI: 10.1037//0894-4105.7.1.41
  • 4
    Kneebone AC, Chelune GJ, Dinner DS, et al. Intracarotid amobarbital procedure as a predictor of material-specific memory change after anterior temporal lobectomy. Epilepsia 1995;36:85765.
  • 5
    Ivnik R, Malac J, Sharbrough F, et. al Traditional and computerized assessment procedures applied to the evaluation of memory change after temporal lobectomy. Arch Clin Neuropsychol 1993;8:6981.
  • 6
    Seidenberg M, Hermann B, Wyler A, et al. Neuropsychological outcome following anterior temporal lobectomy in patients with and without the syndrome of mesial temporal lobe epilepsy. Neuropsychology 1998;12:30316.DOI: 10.1037//0894-4105.12.2.303
  • 7
    Loring DW, Meador KJ, Lee GP, et al. Wada memory asymmetries predict verbal memory decline after anterior temporal lobectomy. Neurology 1995;45:132933.
  • 8
    Rausch R. Role of neuropsychological evaluation and the intracarotid sodium amobarbital procedure in the surgical treatment for epilepsy. In: TheodoreWH, ed. Surgical treatment of epilepsy. New York: Elsevier Science Publishers BV, 1992:7785.
  • 9
    McClone J. Memory complaints before and after temporal lobectomy: do they predict memory performance or lesion laterality? Epilepsia 1994;35:52939.
  • 10
    Wyllie E, Naugle R, Awad I, et al. Intracarotid amobarbital procedure, I: prediction of decreased modality-specific memory scores after temporal lobectomy. Epilepsia 1991;32:85764.
  • 11
    Milner B. Visual recognition and recall after right temporal-lobe excision in man. Neuropsychologia 1968;6:191209.
  • 12
    Barr WB. Examining the right temporal lobe's role in nonverbal memory. Brain Cogn 1997;35:2641.DOI: 10.1006/brcg.1997.0925
  • 13
    Rapcsak SZ, Polster MR, Comer JF, et al. False recognition and misidentification of faces following right hemisphere damage. Cortex 1994;30:56583.
  • 14
    Zatorre R. Discrimination and recognition of tonal melodies after unilateral cerebral excisions. Neuropsychologia. 1985;23:3141.
  • 15
    Jones-Gotman M. Memory for designs: the hippocampal contribution. Neuropsychologia 1986;24:193203.
  • 16
    Saykin AJ, Gur RC, Sussman NM, et al. Memory deficits before and after temporal lobectomy: effect of laterality and age of onset. Brain Cogn 1989;9:191200.
  • 17
    Stafiniak BA, Saykin AJ, Sperling MR, et al. Acute naming deficits following dominant temporal lobectomy: prediction by age at first risk for seizures. Neurology 1990;40:150912.
  • 18
    Corkin S. Tactually guided maze learning in man: effects of unilateral cortical excisions and bilateral hippocampal lesions. Neuropsychologia 1965;3:33952.
  • 19
    Hermann BP, Wyler AR, Somes G. Memory loss following left anterior temporal lobectomy is associated with hippocampal pathology and not extent of hippocampal resection. J Clin Exp Neuropsychol 1993;15:24.
  • 20
    Loring DW, Lee GP, Meador KJ, et al. Hippocampal contribution to verbal recent memory following dominant hemisphere temporal lobectomy. J Clin Exp Neuropsychol 1991;13:57586.
  • 21
    Wolfe RL, Ivnik RJ, Herschorn KA, et al. Neurocognitive efficiency following left temporal lobectomy: standard versus limited resection. J Neurosurg 1993;79:7683.
  • 22
    Rausch R & Crandall PH. Psychological status related to surgical control or temporal lobe seizures. Epilepsia 1982;23:191202.
  • 23
    Penfield W & Mathieson G. An autopsy and a discussion of the role of the hippocampus in experiential recall. Arch Neurol 1974;31:14554.
  • 24
    Loring DW, Hermann BP, Meador KJ, et al. Amnesia after unilateral temporal lobectomy: a case report. Epilepsia 1994;35:75763.
  • 25
    Warrington EK & Duchen LW. A re-appraisal of a case of persistent global amnesia following right temporal lobectomy: a clinico-pathological study. Neuropsychologia 1992;30:43750.
  • 26
    Wada J & Rasmussen T. Intracarotid injection of sodium amytal for the lateralization of cerebral speech dominance: experimental and clinical observations. J Neurosurg 1960;17:26682.
  • 27
    Milner B, Branch C, Rasmussen T. Evidence for bilateral speech representation in some nonrighthanders. Trans Am Neurol Assoc 1966;91:3068.
  • 28
    Chelune G. Hippocampal adequacy versus functional reserve: predicting memory functions following temporal lobectomy. Arch Clin Neuropsychol 1995;10:41332.DOI: 10.1016/0887-6177(95)00015-v
  • 29
    Sperling MR, O'Connor MJ, Saykin AJ, et al. A non-invasive protocol for anterior temporal lobectomy. Neurology 1992;42:41622.
  • 30
    Glosser G, Saykin AJ, Deutsch GK, et al. Neural organization of material-specific memory functions in temporal lobe epilepsy patients as assessed by the intracarotid amobarbital test. Neuropsychology 1995;9:44956.DOI: 10.1037//0894-4105.9.4.449
  • 31
    Glosser G, Deutsch GK, Cole LC, Corwin J, Saykin AJ. Differential lateralization of memory discrimination and response bias in temporal lobe epilepsy patients. J Int Neuropsychol Soc 1998;4:50211.DOI: 10.1017/s1355617798455097
  • 32
    Delis D, Kramer J, Kaplan E, et al. The California Verbal Learning Test. New York: The Psychological Corporation, 1987.
  • 33
    Saykin AJ, Robinson LJ, Stafiniak P, et al. Neuropsychological changes after anterior temporal lobectomy: acute effects on memory, language, and music. In: BennettTL, ed. The neuropsychology of epilepsy. New York: Plenum Press, 1992:26390.
  • 34
    Snodgrass JG & Corwin J. Pragmatics of measuring recognition memory: applications to dementia and amnesia. J Exp Psychol 1988;117:3450.
  • 35
    Saykin AJ, Gur RC, Gur RE, et al. Normative neuropsychological test performance: effects of age, education, gender, and ethnicity. Appl Neuropsychol 1995;2:7988.
  • 36
    Glosser G, Cole LC, Deutsch GK, et al. Hemispheric asymmetries in arousal affect outcome of the intracarotid amobarbital test. Neurology 1999;52:158390.
  • 37
    Seidenberg M, Hermann B, Schoenfeld J, et al. Reorganization of verbal memory function in early onset left temporal lobe epilepsy. Brain Cogn 1997;35:13248.DOI: 10.1006/brcg.1997.0931
  • 38
    Powell GE, Polkey CE, McMillan T. The new Maudsley series of temporal lobectomy, I: short term cognitive effects. Br J Clin Psychol 1985;24:10924.
  • 39
    Semmes J. Hemispheric specialization: a possible clue to mechanisms. Neuropsychologia 1968;6:1126.
  • 40
    Rausch R & Ary C. Supraspan learning in patients with unilateral anterior temporal lobe resections. Neuropychologia 1990;28:11120.
  • 41
    Pohl W. Dissociation of spatial discrimination deficits following frontal and parietal lesions in monkeys. J Comp Physiol Psychol 1973;82:22739.
  • 42
    Hermann BP, Seidenberg M, Wyler A, et al. Dissociation of object recognition and spatial localization abilities following temporal lobe lesions in humans. Neuropsychology 1993;7:34350.DOI: 10.1037//0894-4105.7.3.343
  • 43
    Mishkin M, Ugerleider L, Macko K. Object vision and spatial vision: two cortical pathways. Trends Neurosci 1983;6:4147.
  • 44
    Kapur N, Friston KJ, Young A, et al. Activation of human hippocampal formation during memory for faces: a PET study. Cortex 1995;31:99108.
  • 45
    Casasanto D, Glosser G, Killgore WDS, et al. Presurgical fMRI predicts memory outcome following anterior temporal lobectomy. J Int Neuropsychol Soc 2001;7:183.