n’purop,,choioyru, Vol 2X. No. 2. pp I I l-120. Prmted I” Great Bntsm.

002X-3932 90 S3.00+0 00 c 1990 Pergamon Press plc

1990

SUPRASPAN LEARNING IN PATIENTS WITH UNILATERAL ANTERIOR TEMPORAL LOBE RESECTIONS REBECCA RAUSCH*~: and CATHERINE

M. ARY*

*Reed Neurological Research Center, tDepartments of Psychiatry and Biobehavioral Sciences and $Neurology, University of California, Los Angeles School of Medicine, Los Angeles, CA 90024. U.S.A. (Received 21 Ocroher

1988: accepted 21 August

1989)

Abstract-The present data address the role of the anterior temporal lobe in verbal and visual supraspan learning. Patient subjects had either a left or right standardized temporal lobe resection, which involved extensive removal of mesial temporal lobe structures. Neither patient group showed marked deficits on either the verbal or visual supraspan learning tasks, although an unexpected performance variation occurred. A tendency to mild depression of performance on the verbal supraspan learning task was observed among right temporal-lobectomy patients. Nevertheless, the overall results indicate that anterior temporal lobe structures are not independently critical for supraspan learning. These findings are consistent with a previous report that posterior brain regions are involved in mediating such tasks.

INTRODUCTION USINGa variety of techniques, researchers have demonstrated that supraspan learning, that is, learning a sequence beyond the immediate memory span, is sensitive to brain dysfunction (cf. [2], also discussed in [9, 13, 271). Patients with brain damage have shown deficits in supraspan learning relative to the immediate span of normal controls (i.e. 8 or 9 items) or to their own immediate span (e.g. one digit over the individual’s immediate span). Deficiencies have been found in both verbal and visual modalities. In visual supraspan learning there is evidence that the posterior brain areas may predominate. DE RENZI et al. [S] reported that patients with lesions of either brain hemisphere and known visual field defects were more likely to be impaired on a visual supraspan learning task than corresponding patients without visual field deficits. The authors interpreted the association of the supraspan deficit with visual field defects as reflecting the participation of retro-Rolandic areas in visual supraspan learning. DE RENZIet al. [S] also emphasized that the right posterior hemishere plays a predominant role in such learning since patients with lesions of the right hemisphere (and with visual field defects) showed more marked deficits than all other patient groups. Based on his work with temporal lobectomy patients from the Montreal Neurological Institute (MN]), CORSI [2] also found the right hemisphere to be more involved than the left hemisphere in visual supraspan learning. He reported that epileptic patients with resections of the right anterior temporal lobe were impaired as compared to normal controls in visual supraspan learning, if the excision invaded a large portion of the mesial temporal lobe area, which included the hippocampus and hippocampal gyrus. Patients with right anterior temporal lobe resections that did not invade the hippocampal region or resected only the anterior hippocampal pes were not impaired on this task. CORSI [2] III

REBECCA RAUSCH and CATHEKINE M. ARY

112

reported a similar effect with left anterior temporal lobe resections and verbal supraspan learning. That is, patients with temporal lobe resections that invaded the mesial temporal lobe structures showed impaired verbal supraspan learning, while temporal lobectomy patients without invasion of the left mesial structures showed no corresponding verbal deficit. Furthermore, his data indicated a relationship between extent of mesial resection and the degree of deficit observed. These findings were interpreted as evidence that the hippocampal region was critically involved in these behaviors and that the extent of the damage to the hippocampal region was related to the severity of the behavior deficit (cf. [2.

131). Recently, studies from several different surgical centers have addressed the selective learning and memory deficits associated with temporal lobe resections (e.g. [4, 14, 16.2 I ] 1. Efforts have been made to relate functional significance to subregions within the temporal lobe (cf. [ 183). It is imperative to our understanding of temporal lobe functions to delineate as accurately as possible the temporal lobe regions associated with selective behaviors and to identify variations in surgical effects from different centers. The purpose of this study is to re-evaluate the role of mesial temporal lobe structures in supraspan learning. We have attempted to replicate the CORSI [2] study with a different surgical patient population. We present findings on verbal and visual supraspan learning in groups of epileptic patients with standardized left and right anterior temporal lobe resections (LTL and RTL, repectively). The amount of mesial temporal lobe removal in the patients in this series was at least as large as those described by CORSI [2] as “radical” mesial temporal lobe resections. Hence, if supraspan learning is dependent upon the mesial temporal lobe structures, we would predict similarly marked deficits for patients in this study. METHOD Subjects Patient groups consisted of subjects with intractable complex partial seizures who had undergone a standard anterior temporal lobe resection [3] of either the left (n = 27) or right (n = 36) hemisphere. The patients had focal epilepsy originating from anterior mesial temporal lobe areas. Documentation of the seizure focus was by surface telemetry (n = 11). depth electrode recordings (n = 51), or subdural grid recording ()I = 1). Electroencephalographic seizure data were confirmed by tests offocal dysfunction (i.e. neuropsychological data. intracarotid Sodium Amytal Procedure (IAP) or Positron Emission Tomography: cf. [7]). Patients with evidence of bilateral or diffuse damage were excluded. The surgery involved resection of the lateral temporal neocortex, 3 of the hippocampus (i.e. 3 3i cm of the anterior hippocampus), the hippocampal gyrus. uncus, and the lateral $ of the amygdala (see Fig. I ). While the amount of hippocampus removed from each hemisphere was the same, the lateral neocortical resection was shorter in the left than in the right hemisphere (e.g. mostly 5.0 cm as compared to 6.0 cm as measured along the middle temporal gyrus from the anterior tip). Patients were studied at least 6 months postoperatlvely. All patients had left hemisphere language dominance as determined by the IAP in 54 patients and by postoperative clintcal evaluation in 9 patients. Normal control subjects (NC) included 43 paid volunteers who were matched in age, gender and educatmn to the patlent groups. Control subjects had no history of neurological disorder or current use of any mcdrcation. Clinical parameters of the subject groups are shown in Table I. No significant difference among the groups was found. Postoperative neuropsychological performance of the patient groups are shown In Table 2. Stgniticant differences between patient groups were found on five memory measures. Patients with LTL resections performed worse on verbal memory tasks, while patients with RTL resections performed worse on a nonverbal visual memory task. Similar selective learning deficits associated with unilateral temporal lobe resectlons have been reported from several centers [I. 12. 173. Materials

and proc’udures

b’~rhal .sup-aspan Ieamrny. The general protocol was based on the original procedure derived from HEHH [IO] and modified by C~RSI [2]. Verbal span was determined by the method of WKHSLEK 1231 for dIgIts forward. Each

SUPRASPAN

LEARNING

IN TEMPORAL

TEMPORAL

POLE

LOBECTOMY

PES

113

PATIENTS

HIPPOCAMPI

,UNCUS

,GYRUS

LTL -

h-

HIPPOCAMPI

FIMBRIA

HIPPOCAMPI

RTL -

‘CALCARINE

SULCUS

Fig. 1. Schematic of the en bloc anterior temporal lobectomy. Resection line of the mesial temporal lobe structures is shown in bold and the extent oflateral neocortical excision is indicated for the LTL and RTL surgeries (approximately 5.0 cm and 6.0 cm posterior to the anterior pole, respectively) by dashed lines. Drawing modified from SOBOTTA [22].

SUPRASPAN

LEARNIKG

Table

Number of subjects: Gender (n) Handedness

I

Years of education Age at testing (years)

Table 2. Comparison

of postoperative

(years)

Wechsler Adult Intelligence Scale [24] Full scale IQ Verbal IQ Performance IQ Wechsler Memory Scale [25] Logical memory Immediate Delayed Hard-paired-associates Immediate Delayed Visual reproduction Immediate Delayed Auditory Attention Tests Seashore rhythms test [ 191 Speech sounds perception Test (errors) [S] WAIS arithmetic subtest (age-corrected) [24]

115

PATIENTS

of subjects

Left temporal lobectomy

Right temporal lobectomy

Normal controls

27 15 12 2 25 13.5 (2.4) 30.8 (7.3)

36 17 19 0 36 14.0 (2.9) 31.0 (8.2)

43 20 23 0 36 14.5 (2.0) 28.9 (7.4)

neuropsychological

Lobectomy Left temporal (n=27)

Measure

LOBECTOMY

1. Clinical parameters

Male: Female: Left : Right: [R(&SEM)]: [X(&SEM)]:

In)

Time postoperative

IN TEMPORAL

1.6 (1.4)

group

performance [R ( f SD)] Right temporal (n=36)

of patient

groups

(P-value)

1.6 (1.1)

ns

103.0 (12.0) 101.1 (13.0) 105.1 (12.9)

106.9 (12.1) 106.0 (13.1) 107.0 (12.8)

ns ns ns

6.7 (2.3) 4.2 (2.2)

10.2 (3.3) 8.0 (3.5)

P