Childs Nerv Syst (2003) 19:736–743 DOI 10.1007/s00381-003-0826-x
Eduardo Castro-Sierra Fernando Chico-Ponce de León Luis F. Gordillo-Domínguez Rocío Raya-Gutiérrez
Received: 20 June 2003 Published online: 27 August 2003 © Springer-Verlag 2003 E. Castro-Sierra (✉) Laboratory of Psychoacoustics and Auditory Physiology, Hospital Infantil de México Federico Gómez, Dr. Márquez No. 162, 06720 México D.F., Mexico e-mail:
[email protected] Tel.: +52-55-52289917 ext. 1302 F. Chico-Ponce de León L. F. Gordillo-Domínguez Department of Neurosurgery, Hospital Infantil de México Federico Gómez, Dr. Márquez No. 162, 06720 México D.F., Mexico R. Raya-Gutiérrez Faculty of Psychology, National Autonomous University of Mexico, 04100 México D.F., Mexico
O R I G I N A L PA P E R
Surgery of tumors of the cerebellum and prefrontal cortex, and sensory memory and motor planning disturbances in children
Abstract Object: This study addressed the integration of sensory short-term memory (SSTM) and motor planning (MP) in the lateral cerebellar region, where the dentate nucleus is localized, and in the prefrontal cortex (PF). Methods: Boucher and Lewis’s test of SSTM and MP was administered pre- and postsurgically to 8 patients of either sex, between 5 and 19 years of age, with tumors of the cerebellum, PF (area 9 medial) or the parieto-occipital region, and on one occasion to 8 corresponding controls. Results: Whereas lesions of the midline portions of the cerebellum and of the parietal-occipital region did not appear to cause any cognitive defects, patients with lesions affecting either the lateral cerebellar region or PF exhibited statistically significant deficits of SSTM and MP. Conclusions: The
Introduction The traditional point of view that the basal ganglia and the cerebellum are simply active in the control of movement has been questioned in recent years. One of the reasons for this re-evaluation has been new information referring to different connections of the basal ganglia and the cerebellum with the cerebral cortex [16]. Basically, recent anatomical studies have revealed that these connections may be organized in discrete circuits or loops. Rather than allowing access of diverse cortical areas to the motor system, these loops could reciprocally connect large and varied sets of cortical areas to either the basal ganglia or the cerebellum. The properties of neurons in
lateral cerebellum seems to act in concert with PF to integrate different cognitive activities related to holding objects in SSTM and planning motor strategies in reference to them. Keywords Dentate nucleus · Lateral cerebellar region · Cerebral cortex · Procedural learning · Memory integration · Boucher and Lewis’s test
the basal ganglia and the cerebellar components of these loops would resemble the properties of neurons in cortical areas subserved by the same corresponding loops. For instance, neuronal activity within the cerebellar loops connected to specific prefrontal cortex (PF) areas might have to do with aspects of cognitive function. More recently, studies with conventional retrograde markers as well as employing transneuronal transport of the McIntyre-B strain of H. simplex virus 1 have been carried out in Cebus apella monkeys [17]. These latter analyses have investigated the connections existing between the dentate nucleus (DN) of the lateral cerebellum and the PF of these primates, thus making it possible to determine the existence of a ventro-medial region in the
737
DN connecting with the ipsilateral ventro-lateral caudal nucleus of the thalamus. From this nucleus fibers depart towards areas 9 medial (9m), 9 lateral (9l), and 46 dorsal (46d) of the PF of both hemispheres of the brain. The contralateral thalamo-prefrontal connections are greater in number (85%) than the corresponding ipsilateral connections. These dentato-thalamo-prefrontal projections involved in cognitive activity have localizations, at both a thalamic and a prefrontal level, that are different from those of projections with a purely motor function stemming from the cerebellum. Encephalic activity in humans required for the learning of motor sequences seems to activate similar areas in PF [5, 8] and DN [9]. From January 1990 to December 1991, the Laboratory of Psychoacoustics and Auditory Physiology and the Department of Neurosurgery at the Hospital Infantil de México Federico Gómez carried out an investigation of sensory (visual and auditory) short-term memory (VSTM and ASTM) in 9 pediatric patients at this Hospital. The patients were of either sex, between 5 and 14 years of age, and suffered from tumor pathology of the cerebellum at either a vermal (midline) or a hemispheric (lateral) level. Results of the analyses undertaken indicated that these subjects had statistically significant deficits of ASTM [13]. In view of the data providing evidence of involvement of PF and DN in the learning of motor sequences [5, 8, 9], and of our own findings of the activity of different cerebellar areas in sensory shortterm memory [13], we decided to embark on research into new groups of pediatric patients with infratentorial (cerebellar, posterior fossa or fourth ventricle) or supratentorial (PF or parieto-occipital cortex) tumors in order to determine the degree of participation of the cerebellum and PF in the integration of sensory short-term memory (SSTM) into motor planning (MP) strategies during early phases of the learning of very simple motor sequences. On this occasion, we compared the data obtained in each patient directly with those obtained in a corresponding brother or sister, who served as an individual control. This statistical treatment was expected to provide a more detailed analysis of the presence or absence of cognitive disturbances in the patients than the one pursued in our first study [13], where results obtained in two groups of patients were compared with results obtained in the entire group of control subjects.
brother or sister, without lesions or deficits at an encephalic level, who served as a control for his/her affected sibling (CONT Group).
INFRAT Group 1. V.C.O., a girl, aged 5 years and 11 months, with a midline cerebellar astrocytoma. Postsurgical CT showed two sequelae of the initial operation (1990). The first was a relapse (1992) with a nodule (1 cm in diameter), located behind the fourth ventricle and in intimate contact with both middle cerebellar peduncles, at the union of the pons and the midbrain. In the second relapse (2000), only a fragment of the inferior medullary velum could be distinguished; the remainder of the fourth ventricle had disappeared. Both cerebellar hemispheres were slightly affected in their medial aspects (x=2.5 cm, y=4 cm, z=3.2 cm) 2. A.V.A., a boy, aged 6 years and 2 months, with a posterior fossa juvenile astrocytoma affecting the right cerebellar hemisphere. Postsurgical CT showed that the lesion extended from the foramen magnum all the way to the culmen. It was of heterogeneous density, its postero-inferior part being cystic and its antero-superior part nodular. This last segment did not allow visualization of the upper part of the fourth ventricle (x=5 cm, y=5 cm, z=7 cm) 3. E.L.R., a boy, aged 11 years and 2 months, with a midline cerebellar astrocytoma. Presurgical CT showed a central cystic lesion of the cerebellum extending from above the occipital foramen down to the culmen. A thin mural nodule inside the lesion slightly affected the left middle cerebellar peduncle (x=4.9 cm, y=5 cm, z=4.5 cm) 4. C.R.M., a boy, aged 14 years and 5 months, with an astrocytoma adhering to the floor of the fourth ventricle. Presurgical CT showed that the lesion extended from the foramen magnum, in the central part of the cerebellum, towards the culmen. Its density was heterogeneous, with a cystic postero-inferior segment, and a solid antero-superior segment. This last segment did not allow visualization of the upper part of the fourth ventricle. The lesion affected the right middle cerebellar peduncle (x=5 cm, y=5 cm, z=7 cm) 5. J.S.A., a boy, aged 15 years and 6 months, with a midline fourth ventricle ependymoma. Presurgical CT showed a cerebellar lesion consisting of two segments. The first was a cystic lesion pushing the fourth ventricle to one side and extending from the bulbar area to the highest part of the ventricle; the second was a nodular lesion especially apparent at the bottom of the ventricle, as well as in the middle and lower bulbar sections. These lesions affected the cephalic cerebellar area (cystic segment: x=3 cm, y=3 cm, z=3 cm; nodular segment: x=2 cm, y=1.5 cm, z=1.5 cm) 6. J.J.S., a girl, aged 15 years and 7 months, with a right cerebellar hemisphere astrocytoma. Presurgical CT showed that the tumor mass displaced the fourth ventricle and the medial part of the left hemisphere leftwards. There was a nodular posteromedial segment and a cystic antero-lateral segment (x=4 cm, y=4 cm, z=3.5 cm)
Patients and methods Patient and control populations
SUPRAT Group
Six of the subjects studied were patients with infratentorial tumors at a cerebellar, posterior fossa or fourth ventricle level (INFRAT Group). Besides them, there were 2 more subjects with supratentorial tumors at either a PF (area 9m) or a parieto-occipital level (SUPRAT Group). We included in this study subjects with astrocytomas, ependymomas or meningiomas, and excluded from it those with medulloblastomas. Each patient was accompanied by a
1. V.M.D., a boy, aged 11 years and 6 months, with a right frontal meningioma of the internal frontal gyrus at the level of the coronal suture. Presurgical CT showed that the lesion adhered to the falx cerebri at the level of the right bregma. It pushed aside and compressed the medial side of the hemisphere at this level, especially the right internal frontal gyrus (area 9m: x=1 cm, y=1.5 cm, z=3.2 cm)
738
2. I.G.Y., a man, aged 19 years and 1 month, with a right parietooccipital cystic astrocytoma. On postsurgical CT parieto-occipital sequelae with a cystic zone could be observed. The cyst appeared to contact the body of the right lateral ventricle, and was triangular-shaped, with its base at the convexity of the brain and its apex toward the medial part of the brain (x=2.5 cm, y=3.5 cm, z=4 cm; x: left-right diameter, y: anterior-posterior diameter, z: cephalic-podalic diameter) CONT Group 1. 2. 3. 4. 5. 6. 7. 8.
J.C.O. (sister of V.C.O., 7 years and 10 months) E.V.A. (brother of A.V.A., 11 years and 3 months) M.L.R. (brother of E.L.R., 10 years and 1 month) N.R.M. (sister of C.R.M., 12 years and 3 months) E.S.A. (sister of J.S.A., 14 years and 0 months) Jo.J.S. (sister of J.J.S., 20 years and 2 months) R.M.D. (brother of V.M.D., 8 years and 5 months) Iv.G.Y. (brother of I.G.Y., 13 years and 9 months)
Sensory memory and motor planning procedure Boucher and Lewis’s test (BLT) has been utilized to study the VSTM and ASTM of autistic children [1]. After instructions are given either visually or auditorily, subjects must plan motor strategies directed at placing small plastic objects of different colors and shapes in different positions in relation to diverse containers or receptacles. Although there are as yet no international normative data, the informative results previously obtained by us with this test in children suffering from tumor pathology of the cerebellum, at both a midline and a lateral level [13], encouraged us to further employ it in the present study of sensory short-term memory and motor planning. Moreover, the application of the BLT to the study of SSTM and MP of subjects with cerebellar lesions was decided upon because of the suggested relationship of attentional disturbances of the autistic syndrome to similar disturbances accompanying tumoral, vascular and degenerative pathology of the cerebellar vermis [3]. The test was carried out in the majority of the patients immediately before and within 1 month of surgical treatment for their respective tumors. In 1 patient (E.L.R), the postsurgical administration of the BLT occurred 1 year after surgery. In another patient (V.C.O.), the test was administered immediately and 1 year after surgery, on the occasion of a relapse. Finally, in 2 other patients, the test was also used 6 months (A.V.A. and V.M.D.) and 1 year (A.V.A.) after surgery. The test was used only once in each of the controls. BLT involves the presentation to the subject by the investigator of visual or auditory instructions relating to small plastic toys of diverse shapes (hooks, soldiers, dogs, or wheels), colors (red, green, blue, or yellow), and numbers (1, 2, or 3). Subjects must place them in specific positions (on top of, under, in front of, or behind) in relation to different receptacles (wooden box, wicker basket, cardboard plate, or plastic vase). The toys lie on a table, and the investigator and the subject sit facing each other at opposite sides of the table. In the visual mode of presentation, the investigator fixes the visual attention of the subject. He then holds the object(s) and places it (them) in the prescribed position; after a pause, he returns it (them) to the original position. The subject must then place the object(s) according to what he/she has seen the investigator do. In the auditory mode of presentation, the subject stands facing away from the table. The investigator then reads out each instruction. Subsequently, the subject turns around and faces the table proceeding to place the object(s) according to the instruction given by the investigator. The test is given first in one (visual/auditory) and then in the other (visual/auditory) mode of presentation to each subject. The order of presentation is random, and there are two different series
of 11 instructions. BLT, thus, consists of 22 instructions with a total of 76 content words. For example, “Put three red soldiers under the box” has five content words (in italics). There are five different conditions: shape of object, color, number, position, and receptacle. An error in any of the five conditions in each instruction is penalized and counted separately. In the present study, errors in the conditions “position” and “receptacle” were considered to reflect errors of MP relating to the placement of the toys in specific positions and receptacles and those in the conditions “object shape,” “color”, and “number” were considered to reflect errors of SSTM relating to shape, color, and number of toys employed. Statistical analyses 1. Errors relating to each of the five conditions of every instruction of the BLT committed either presurgically or postsurgically by each patient of INFRAT Group or SUPRAT Group were compared with those committed by his/her own control (CONT Group) using the Mann-Whitney-Wilcoxon z test. The data obtained from presentation in the visual mode were analyzed separately from those obtained from presentation in the auditory mode 2. Errors relating to each of the five conditions of every instruction of the BLT committed in the presurgical application of the test by each patient of INFRAT Group and SUPRAT Group were compared with those committed in the postsurgical application of the test by the same subject using Wilcoxon’s matched-pairs signed-ranks test (WMRT). The data obtained from presentation in the visual mode were analyzed separately from those obtained from presentation in the auditory mode 3. WMRT was also applied to a comparison of errors in each of the five conditions of every instruction of the BLT committed by subjects A.V.A. and V.M.D. on the application of the test before surgery with those committed by these subjects in the test a few days (A.V.A. and V.M.D.), 6 months (A.V.A. and V.M.D.), and 1 year (A.V.A.) after surgery. The data obtained from presentation in the visual mode were analyzed separately from those obtained from presentation in the auditory mode 4. Errors relating to the position-receptacle conditions of every instruction of BLT committed by each of the subjects of all three groups were compared with those relating to the shapeof-object-color-number conditions that had been committed by the same subject using the Kruskall-Wallis one-way analysis of variance. The data obtained from presentation in the visual mode were analyzed separately from those obtained from presentation in the auditory mode
Results Figures 1 and 2 show errors committed pre- or postsurgically in each of the five conditions of the BLT, after either visual or auditory presentation of the test, by 1. Three patients (J.S.A., C.R.M., I.G.Y.) with either fourth ventricle or parieto-occipital tumors (Fig. 1a–c) 2. Three patients (E.L.R., V.C.O., J.J.S.) with midline or lateral cerebellar tumors (Fig. 2a–c) in comparison to those committed by the corresponding controls. These graphs are accompanied by contrast-enhanced CT of the subjects’ lesions before or after surgical intervention (Figs. 1d–f, 2d–f). The patients with midline lesions affecting the cerebellar hemispheres or
739
Fig. 1 Errors committed pre- or postsurgically in Boucher and Lewis’s test by 3 patients with a, b fourth ventricle or c parieto-occipital tumors. Contrast-enhanced CT of the respective lesion performed d, e before or f after surgical treatment
Fig. 2 Errors committed pre- or postsurgically in Boucher and Lewis’s test by 3 patients with a, b midline or c lateral cerebellar tumors. Contrast-enhanced CT of the respective lesion performed d, f before or e after surgical treatment
740
Fig. 3 Errors committed pre- or postsurgically in Boucher and Lewis’s test by 2 patients with a posterior fossa or b prefrontal area 9m tumors. Contrast-enhanced CT of the respective lesion performed d before or c after surgical treatment. Evolution of errors committed in BLT e, f postsurgically, e, f 6 months, and e 1 year after treatment
with lateral cerebellar lesions (C.R.M., V.C.O., J.J.S.) made significantly more errors after both the visual (V.C.O., post, z=-1.9152 [p=0.0277], 1 year, z=-2.5220 [p=0.0058]; J.J.S., post, z=-1.9007 [p=0.0287]) and the auditory (C.R.M., post, z=-2.5220 [p=0.0058]; J.J.S., pre, z=-2.2778 [p=0.0114], and post, z=-1.71132 [p=0.0433]) BLT. In contrast, the errors of patients with fundamentally midline lesions (J.S.A., E.L.R.) were not significantly different. There was a statistically significant difference between the errors committed by I.G.Y., with a lesion distant from both the lateral cerebellum and PF, and those committed by his control in the auditory mode of presentation (pre, z=2.7491 [p=0.0030]; post, z=2.0153 [p=0.0219]); however, the errors committed by the control were greater than those committed by the patient. The left side of Fig. 3 shows errors committed pre- or postsurgically in each of the five conditions of the BLT, after both the visual and the auditory presentation of the test, by 2 patients (A.V.A. and V.M.D.), with either a posterior fossa tumor affecting the lateral cerebellum or a PF tumor (Fig. 3a, b), in comparison to those committed by the corresponding controls. These graphs are ac-
companied by CT of these subjects’ lesions performed before or after surgery (Fig. 3c, d). Both patients made significantly more errors after both visual (A.V.A., pre, z=-2.5456 [p=0.0055], and post, z=-2.2274 [p=0.0130]) and auditory (A.V.A., pre, z=-2.4096 [p=0.0080], and post, z=-2.5143 [p=0.0060]; V.M.D., pre, z=-1.8320 [p=0.0335], and post, z=-1.8320 [p=0.0335]) presentations of the BLT. The right side of Fig. 3 shows the evolution of errors committed in the BLT a few days (A.V.A. and V.M.D.), 6 months (A.V.A. and V.M.D.) and 1 year (A.V.A.) after surgical intervention, in comparison to those committed presurgically by these 2 patients (Fig. 3e, f). Although the number of errors committed by A.V.A. after intervention significantly decreased in both the visual (postsurgical, T=15 [p
