Anat Embryol (1992) 186:129-136
Anatomyand Efilbryolo 9 Springer-Verlag1992
Computer-assisted 3D-reconstruction and statistics of the limbic system 1. Computer-assisted 3D-reconstruction of the hippocampal formation, the fornix, and the mamillary bodies M. Gerke, T. Schiitz, and H.-J. Kretschmann Abteilung Neuroanatomie, MedizinischeHochschuleHannover, Konstanty-Gutschow-Strasse8, W-3000 Hannover 61, Federal Republic of Germany Accepted January 31, 1992
Summary. The hippocampal formation of eight perfusion-fixed human brains was examined using new methods according to stereotactic and morphometric principles (macrovibratome and computer-aided 3D reconstruction). The reconstructions form part of a neuroanatomical reference system (NeuRef). This reference system allows for 3D visualisation of the brain and its components on a computer graphic workstation, as well as for the presentation of the union set based on a neuroanatomical structure taken from this sample of brains. This retrievable knowledge of neurofunctional systems is important for the preoperative planning of neurosurgeons and the adjustment of radiotherapy.
lowing advantages for qualitative and quantitative analysis of the examined material:
Key words: Human - Hippocampus - Limbic system Computer graphics - Neuroimaging
This paper shows the application of our method to the 3D reconstructions of the hippocampal formation forming part of the limbic system. The limbic system, a term which was coined by MacLean (1952), includes structures of the mesencephalon, diencephalon and telencephalon (Nieuwenhuys et al. 1988). It is functionally associated with emotional reactions, visceral functions, vigilance, and memory (Hassler 1964). The hippocampus, the fornix and the mamillary bodies form the inner ring of the limbic system. With the periarchaeocortex as an outer ring, these structures constitute part of the circuit of Papez (1937). The limbic system may be understood as a functional concept (Nieuwenhuys et al. 1988) rather than a morphological system, and therefore cannot be defined by any specific nuclei and their associated fibre bundles (Starck 1982). Comparative examinations of primates presented extensively by Stephan (1975) and Stephan and Andy (1982), resulted in valuable morphological findings on the limbic system. The main demarcation criteria of the limbic system were taken from these works as well as from the monographs on the macroscopic anatomy of the hippocampus by Klingler (1948) and Duvernoy (1988). The periarchaeocortical areas and the subcortical nuclei of the limbic system (hypothalamus, corpus amygdaloideum and septum) will be dealt with in further papers.
Introduction Neuroimaging in medicine - computed tomography (CT), magnetic resonance imaging (MRI) and positron emission tomography (PET) - enable a detailed representation of in vivo brain conditions. Since neuroimaging capacities have increased dramatically over the last decade, these procedures are now recognized as indispensible tools for diagnosis and therapy. In order to take advantage of these techniques, precise knowledge of the topography and localization of morphological structures and neurofunctional systems of the human brain is of vital importance (Kretschmann and Weinrich 1991). This problem has been overcome using the neuroanatomical reference system (NeuRef). This system is made up of computer software for data acquisition and three-dimensional (3D) reconstruction, a neuroanatomical database, and application software. Computer-assisted 3D reconstruction of serial sections offers the folC o r r e s p o n d e n c e to .
H.-J. Kretschmann
1. The 3D spatial relations of neuroanatomical structures are clearly visible. 2. Quantitative evaluations of isolated structures are attainable. 3. Interindividual comparisons of localization and shape of neurofunctional systems are possible in any desired plane. 4. The ability to cut the structures in any desired plane offers a direct comparison with CT- and MR-images. 5. Hardcopy units can be used to document the images by photography and take over the task of draftsmen.
130
Materials and methods
neighbouring points of a polygon and one point of a (semantically) neighbour polygon in the upper or lower plane. Automatic triangulation methods may not be correct in an anatomical sense. Consequently this work must be done interactively by an expert familiar with the subject to be reconstructed. The digitized outlines and the reconstructions of the structures serve to form a neuroanatomical database on which the application software of NeuRef can work. One application is the 3D visualisation of the reconstructions, which is done by means of common computergraphical methods on a computer graphic workstation IRIS 4D/35TG (Rogers 1985; Rogers and Adams 1990). The structures from different brains in an identical coordinate system can be displayed on one screen using the program 'disp'. This method can be used to present the union set of a structure from a number of brains to show the interindividual variability of neuroanatomical structures.
The Department of Anatomy of the Hannover Medical School provided eight human brains, all of which were intracranially fixed by perfusion (Table 1). The brains were derived from human corpses fixed in toto (fixed with a formalin-alcohol solution). The following requirements were laid down for inclusion into our material: 1. Maximal age of 70 years 2. Interval between death and fixation to be as short as possible 3. No evidence of diseases involving the central nervous system in past medical history and clinical findings 4. No macroscopic pathological changes 5. No pathological changes in the histological examinations After separation of the head from the body, the canthomeatal line (which joins the outer canthus of the eye to the centre of the external auditory meatus) was marked on both sides of the head. X-ray photographs of the head were taken in three projections: lateral-lateral, anterior-posterior, and superior-inferior. The cranium was sawed on a level 4 cm above the canthomeatal plane (as defined by the two canthomeatal lines). Subsequently the sawing plane was traced onto the brain. After withdrawal of the brain from the cranial cavity, the vessels and meninges were removed. The brain, resting on four adjustment screws, was aiigned with the canthomeatal plane and embedded in 4% agarose. The brainagarose block was cut into 2-mm-thick slices with a macrovibratome (Klekamp et al. 1985) parallel to the canthomeatal plane. The surface of the embedded brain-agarose block was photographed prior to each cut (Fig. 1 A). Copies in original scale were produced from the negatives. A transparency was placed upon each photograph and whilst under constant comparison with the original brain slice (inspection under a stereomicroscope illuminated by a cold-light lamp, Andres et al. 1981) the delimitation of the structures was drawn onto the transparency (Fig. 1 B). Verification of the macroscopic delimitation was controlled with histologically prepared human brains which had also been fixed by perfusion. Further work was done with the aid of computers and the NeuRef system (data aquisition and 3D reconstruction). The transparencies were scanned and a line following algorithm produced a polygon for each outline (Rosenfeld 1974). The permanent borderlines of the embedded block served as a frame for the reconstruction. Thus the reconstruction is independent of the contour lines of the object. The polygons of all slices were stacked onto each other. Afterwards, the structures were 3-dimensionally reconstructed using the triangulation method (Fuchs et al. 1977; Christiansen and Sederberg 1978). This method approximates the surface of the structure using triangles (Fig. 2), which are built from two
Anatomy of the retrocommissural hippocampus. The archaeocortical hippocampal formation is a large C-shaped structure, and forms the inner ring of the limbic lobe (Nieuwenhuys et al. 1988). The retrocommissural hippocampus represents the main portion of the hippocampal formation. Architectonically, it includes the fascia dentata, the cornu ammonis and the subiculum. The retrocommissural hippocampus ranges from the splenium corporis callosi to the corpus amygdaloideum. The temporal horn of the lateral ventricle represents the superior, lateral and posterior border of the hippocampus. Facing the ventricle, the alveus hippocampi, whose fibres merge into the fimbria hippocampi, lies on the hippocampus. The gyrus dentatus is macroscopically visible as a narrow edge between the fimbria hippocampi and the gyrus parahippocampalis. The subiculum merges without a macroscopically visible borderline into the cortex of the gyrus parahippocampalis. From anterior to posterior, the hippocampus can be subdivided into a head, a body, and a tail. The intraventricular hippocampal head features the digitationes hippocampi, one to five convolutions (in 90% two-three convolutions, Gertz et al. 1972), separated by small sulci. The hippocampal head and the amygdala are often joined together across the lateral ventricle (Duvernoy 1988). The hippocampus is connected with the uncus hippocampi in an inferior-anterior aspect. The uncus is subdivided from posterior to anterior aspect into the gyrus intralimbicus, the limbus Giacomini (anterior edge of gyrus dentatus) and the gyrus uncinatus. Because of its archaeocortical structure, the uncus has to be classed with the hippocampus (Stephan 1975). The gyrus uncinatus merges without a macroscopic borderline further anteriorly into the gyrus semilunaris (contains periamygdaloid cortex) and the gyrus ambiens (part of entorhinal cortex). The gyrus dentatus ends posteriorly, in the direction of the corpus callosum, in the fasciola cinerea, which is located next to
Table 1. Survey of the brains examined Brain
S S S S S S S
Age (Years)
Sex
Height (cm)
2/80 14/82 32/82 37/82 54/82 6/83 56/83
50 52 69 59 69 66 40
m f f m m f m
170 163 164 173 177 169 176
S 18/84
47
f
170
Cause of death
Suicide, hanged Suicide, E605 Heart failure Bronchial cancer Cardiac infarct Heart failure Primary hepato. cell. cancer Suicide, barbiturate
di
d2
Brainweight (g)
Skullvolume (ml)
Neuropathology: Macroscopical
Microscopical
Inconspicuous Inconspicuous Inconspicuous Inconspicuous Inconspicuous Light cerebral edema Light cerebral edema
Inconspicuous Inconspicuous Inconspicuous Inconspicuous Inconspicuous Inconspicuous Inconspicuous
Cerebral edema
Inconspicuous
5 5 3 5 5 4 2
34 23 21 22 12 15 7
1043 1119 1102 1118 1149 1334
1320 1300 1320 1580 1460 1340 1425
1
5
1353
1465
1191
dl, Days from death to conservation; d2, duration of conservation in months Average brain weight at the end of conservation time d2:1,185 g (1,043 g-1,353 g); average brain age: 56.3 years (40-69 years)
131
[ I I I I I I I [ I I [ I I I I I m
I
m
m
m
m
m
m
m
B I Fig. 1 A, B. Photograph of the embedded brain-agarose block in the course of sectioning with the macrovibratome (A) and the transparency with the measured delimitation of the hemispheres, the ventricles and the fornix (B). The absolute coordinates are obtained
I
[
I
I
I
I
I
I
I
I
I
I
I
I
I
from the visible frame of the macrovibratome, which serves as a constant reference to all structures in the embedded block. The cutting plane lies 13 mm above the midcommissural point
Macroscopic delimitation of the hippocampus, the fornix, and the mamillary bodies. The hippocampus (h, Fig. 3 A-D) can be reliably
Fig. 2. Triangulation. The outlines of two section levels were connected with triangles. Two points are located on one section level, the third point lies in the neighbouring section level the sulcus hippocampi. Further aspected anteriorly and separated by an allusive rim (sulcus dentato-fasciolaris), the cornu ammonis can be seen to extend as the gyrus fasciolaris (Nieuwenhuys et al. 1988). This structure can be differentiated from the fasciola cinerea by its darker colour (Stephan 1975). In the region of the corpus callosum, the fasciola cinerea and the gyrus fasciolaris join one another, and can be referred to by either name (Klingler 1948; Stephan 1975). Finally, the gyrus fasciolaris turns into the supracommissural hippocampus (indusium griseum). Convolutions located below the splenium corporis callosi, between the gyrus parahippocampalis (respectively the isthmus gyrus cinguli) and the gyrus dentatus (respectively the fasciola cinerea), are called Gyri Andreae Retzii (Klingler 1948). These structures of the hippocampus show great variability.
delimited by the ventricles at the cornu inferior of the lateral ventricles (according to the cutting plane to a more lateral, anterior or posterior aspect) (lv, Fig. 3 B-D). The whitish alveus facilitates orientation into the inferior part of the uncus hippocampi in the direction of the corpus amygdaloideum (u, Fig. 3 D). The elusive boundary of cortex and white substance in the presubiculum (DeArmond et al. 1976; Braak 1980) and the sometimes distinctly visible fibre bundles (Papez 1937; Stephan 1975), which perforate and extend into the cortex, serve as demarcations of the hippocampus in the region of the periarchaeocortex (p, Fig. 3 A D ) . The delimitation of this region could be verified by examination of the histologically prepared brains. Also included in the hippocampus are the alveus, the fimbria hippocampi as fibrous parts, the fasciola cinerea and the gyrus fasciolaris (h, Fig. 3 A). The fornix (f, Fig. 3A-C) is marked as a separate part following its detachment from the hippocampal formation (closely above Fig. 3 A). The mamitlary bodies (rob, Fig. 3 D) are the only nuclear subregions of the hypothalamus that can be delimited macroscopically (Schaltenbrand and Wahren 1977).
Results T h e results o f c o m p u t e r - a s s i s t e d 3D r e c o n s t r u c t i o n s o f b r a i n S 32/82 are p r e s e n t e d in Figs. 4 a n d 5. T h e surface o f the structures are s m o o t h e d with G o u r a u d shading. A n y desired perspective v i s u a l i z a t i o n o f the r e c o n s t r u c t ed structures c a n be o b t a i n e d . As viewed f r o m a b o v e (Fig. 4 A ) the h i p p o c a m p u s (h) forms a n arch posteriorly, w i d e n i n g a n t e r i o r l y to a
I
-~
I
anterior
anterior
-~
I
o J
I
20 I
I
I
60
-~,-~
so ]
,
~
,
2,o,
,~
,
60
.60
6O
6O
~
I:
40
20
-0
-20
-20
-40
-40
-60
"0
.40
-I
-60
-80
-80
-so 1 A -100
I
-4o
-2o
'
! 6. '
postenor
~o
iOm'~ 6o
-100
80
-=1
B -eb
anterior 80
-~
!
-40
, -40 '
-20
' -2,0
6
ZO
posterior anterior
'
0I
'
20
40 ~m do
, 40,
,
80
400
80
'60
40-
40
20-
2O
0-
-0
-0
-20
-20
-20
-40
-40
-40
-60-
-60
-80
-8r
D -6()
-413
-21)
;.
postenor
~
dO m'm 60
oo
-lO0 l
!
-~
' -,/o
-2o
6
posterior
2o
.0 .,;. do
-lOO
Fig. 3 A-D. Delimitation of the limbic system upon the brain slices of plane z = 8 mm (A), z = - 2 m m (B), z = - 8 mm (C), z = 12 mm (D) (z= distance to the bicommissural plane) of the brain S 32/82. The projection of the midcommissural point on each plane
serves as the origin of the coordinate system.ffornix; h, hippocampus; ls, lateral sulcus; lv, lateral ventricle; rnb mamillary body; p, periarchaeocortical areas; tv, third ventricle; u, uncus hippocampi
Fig. 4A-C. 3D reconstructions of the hippocampus, the fornix and the mamillary body (red) with the ventricular system (white, vitreous) and the left hemisphere (yellow) in orthogonal projections of the brain S 32/82. On the right side are drawings with the out-
lines of these structures. A Superior view onto the bicommissural plane. B Anterior view in the bicommissural oriented coordinate system. C Lateral view from the right side. fv, fourth ventricle; gf, gyrus fasciolaris; H, left hemisphere
...~:.~ gf
""
..-
Iv
"'fv'"~
::fv
Fig. 4A-C
.
....
........ "..~?
Fig. 5A-C
135 club-like shape. The uncus hippocampi (u) lies at the tip of the temporal horn of the lateral ventricle. The gyrus fasciolaris (gf) is positioned in a posterior aspect to the fornix (1) which lies close to the median line. The same structures are visible from an anterior view (Fig. 4B). Viewed from this angle, the gyrus fasciolaris (gJ) is visible as a little tip under the arch of the fornix. The mamillary body (mb) is located at the anterior and inferior end of the fornix (/). Viewed laterally (Fig. 4 C), the hippocampus (h) resembles a bow curving around the midcommissural point (+), extending anteriorly by the arch of the fornix (D. The mamillary body is covered by the hippocampus. Figure 5 shows the overlapping surfaces of the left hippocampus derived from the eight brains. Each hippocampus is represented by a different colour (brain S 32/ 82 red). This is an appropriate method to achieve a union set of structures. The right hippocampus from brain S 32/82 is shown for comparison. Figure 5 C is an oblique view from a right anterior aspect. Figure 5 also shows some details of the right hippocampal formation. Uncus hippocampi (u), fimbria hippocampi (fh), subiculum (s) and gyrus fasciolaris (gJ) are clearly visible. The fornix (D is cut off at its detachment from the hippocampal formation. The amygdaloid body lies in front of the anterior end of the hippocampal formation. For this reason an inlet appears at this location (ab). Discussion
The mean age of the eight brains examined was 57 years (extremes 40-69 years). The mean brain weight after fixation in situ was 1,176 g (1,043 g-1,353 g). Selective use of intracranially fixed brains was necessary in order to prevent the typical deformation of the angle between the brainstem and the forebrain in extracranially fixed brains. Minimal volume loss (1-5%, Klekamp et al. 1985) was obtained by embedding the brain in agarose. By cutting the brain carefully with the macrovibratome, a regular and thin series of cuts was produced in a coordinate system without any loss of substance. The brain slices were stored in welded plastic bags filled with formalin, resulting in a shrinkage of 3-5 mm (frontal to occipital) over a period of 2 years. For this reason copies of the photographs in original scale serve as the basis for quantitative evaluation. During data input errors may arise due to non-linear distortions of the enlarged
(
Fig. 5A-C. 3D reconstructions of eight hippocampi, each in a different colour from the left hemisphere, and one hippocampus from the right hemisphere (red, S 32/82) in orthogonal projections. The overlapping surfaces of the eight hippocampi form the union set of these structures. A Superior view onto the bicommissuralplane. B Anterior view in the bicommissural oriented coordinate system. C Oblique view from right-anterior, ab, inlet in front of which the amygdaloidbody lies,f, fornix (cut off);f h, fimbriahippocampi; s, subiculum
photographs, by inaccuracy when placing the transparencies upon the photographs and by the line thickness of the pencil used. The sum of these errors is less than 1 mm. Reconstructions from serial sections impose one basic problem. Structures which are located inside a slice without crossing the cutting plane are not taken into account in the following delimitation. This problem affects the small structures running parallel to the cutting plane as well as the superior and inferior border of structures. These surfaces are covered by caps in 3D reconstruction. The height of the cap can be adapted to the thickness of the plane. An increase in resolution can only be gained by diminishing the cutting thickness, resulting in an increase of measured volume points. A macroscopic-microscopic delimitation of the limbic system is difficult in those areas which have no furrow or prominent structures. These particular areas in the field of the limbic system form the transition zone between the hippocampus and the periarchaeocortex and the border between the archaeocortex and the palaeocortex respectively the isocortex. The delimitation of these areas was checked by comparing the neurohistological sections of different brains. The precommissural and supracallosal parts of the hippocampus are not available with the macroscopical technique presented in this paper. Structures derived from neuroanatomical sections or clinical CT- and MRI-data are referred to as contour lines. In order to achieve a good spatial impression the surface of the structures must be reconstructed from these contour lines. There are two possible methods: 1. The 'Cuberille Model' (Herman and Liu 1979). This volume-based model is a dissection of space into equal cubes (called voxels) by three orthogonal sets of parallel planes, A structure is represented by a subset of the set of all voxels. This representation can be performed automatically from the contours of the structure. 2. The 'Triangulation' (Keppel 1975) of the contour lines to get a surface-based model. This method has the advantage of a higher degree of anatomical exactness; however, the reconstruction must be done interactively. We chose the triangulation method for 3D visualisation and a voxel based model for 3D statistics. Compared to past methods (waxplate models, Born 1883; plastic plate models, Blechschmidt 1954) which produce particular models of a definite scale, our reconstruction technique has the advantage in that the database could be duplicated and transferred to other computer graphic workstations, and visualised in any desired modality. But the major advantage is the possibility of matching the neuroanatomical structures with CT-, MR-, and PET scans using the same coordinate system. In clinical practice, neurosurgery and radiotherapy have to observe neuroanatomical risk structures to avoid clinical deficits of neurofunctional systems. For this reason, computer-based knowledge of interindividual variability of neuroanatomical structures can be of use. Interindividual variability can be estimated by the union
136 o f structures f r o m m o d e l brains in the same b i c o m m issural c o o r d i n a t e system. Sections and u n i o n sets can be c o m p u t e d f r o m 3D reconstructions a n d visualized by ' d i s p ' ( S c h a u m a n n 1989). This c o r r e s p o n d s to a representation o f the extreme values. A v o l u m e - b a s e d m o d e l o f the brain is required to obtain the probabilities o f residence between these two extreme values. The m a i n parts o f the h i p p o c a m p a l f o r m a t i o n , the fornix and the mamillary bodies can be displayed in M R I (Naidich et al. 1988; Baulac et al. 1988; Renella 1989), b u t histological boundaries c a n n o t be discerned by using M R I alone. A t this location o u r reference system can be cited for direct c o m p a r i s o n with CT- a n d M R - i m a g e s . The 3D reconstructions were o b t a i n e d f r o m different brains a n d scaled to reduce interindividual variability (Talairach et al. 1967; Talairach and T o u r n o u x 1988). This m o d e l is scaled to the dimensions o f the patient's brain. It can be cut in any desired plane and projected o n t o CT- a n d M R - i m a g e s . A n a u t o m a t e d identification process o f n e u r o a n a t o m i c a l structures n o t visible in CT- and M R - i m a g e s is also possible (Schfitz et al. 1989).
Acknowledgements. We would like to thank Mr. Rust and Mrs. Fahlbusch for the anatomical and histological preparation and Mrs. Heike for her excellent drawings. This work was supported by the Deutsche Forschungsgemeinschaft (Kr 289/14-1, Kr 289/142).
References Andres KH, von Dfihring M (1981) General methods of characterization of brain regions. In: Heym C, Forssmann WG (eds) Techniques in neuroanatomical research. Springer, Berlin Heidelberg New York, pp 10(~108 Baulac M, Vitte E, Dormont D, Hasboun D, Chiras J, Sarcy JJ, Bories J, Signoret JL (1988) The limbic system: identification of its structures on brain slices. In: Gouaze A, Salamon G (eds) Brain anatomy and magnetic resonance imaging. Springer, Berlin Heidelberg New York, pp 140-149 Blechschmidt E (1954) Rekonstruktionsverfahren mit Verwendung von Kunststoffen. Z Anat Entwicklungsgesch 118 : 170-174 Born G (1883) Die Plattenmodelliermethode. Arch Mikrosk Anat 22:584-599 Braak H (1980) Architectonics of the human telencephalic cortex. Springer, Berlin Heidelberg New York Christiansen HN, Sederberg TW (1978) Conversion of complex contour line definitions into polygonal element mosaics. Cornput Graph 12:187-192 DeArmond SJ, Fusco MM, Dewey MM (1976) Structure of the human brain. 2nd edn. Oxford University Press, New York Duvernoy HM (1988) The human hippocampus. Bergmann, Mfinchen Fuchs H, Kedem ZM, Uselton SP (1977) Optimal surface reconstruction from planar contours. Comm. ACM 20 : 693-702
Gertz DS, Lindenberg R, Piavis GW (1972) Structural variations in the rostral human hippocampus. Hopkins Med J 130:367 376 Hassler R (1964) Zur funktionellen Anatomie des limbischen Systems. Nervenarzt 35:386-396 Herman GT, Liu HK 0979) Three-dimensional display of human organs from computed tomograms. Proc Comput Graph Imaging 9 : 1-21 Keppel E (1975) Approximating complex surfaces by triangulation of contour lines. IBM J Research Dev 19 : 2-11 Klekamp J, Riedel A, Herrmann A, Kretschmann HJ (1985) A new embedding and sectioning technique (macrovibratome) for macroscopic and morphometric examinations especially of the human brain. J Hirnforsch 26:33-40 Klingler J 0948) Die makroskopische Anatomic der Ammonsformation. Denkschr Schweiz Naturforsch Ges 78 : 1-82 Kretschmann H J, Weinrich W 0991) Klinische Neuroanatomie und kranielle Bilddiagnostik. Thieme, Stuttgart MacLean PD (1952) Some psychiatric implications of physiological studies on frontotemporal portion of limbic system (visceral brain). Electroencephalogr Clin Neurophysiol 4: 407-418 Naidich TP, Daniels DL, Haughton VM, Williams A, Pech P, Pojunas K, Palacios E (1988) The hippocampal formation and related structures of the limbic lobe : Anatomic - magnetic resonance correlation. In: Gouaze A, Salamon G (eds) Brain anatomy and magnetic resonance imaging. Springer, Berlin Heidelberg New York, pp 32-64 Nieuwenhuys R, Voogd J, van Huijzen C (1988) The human central nervous system. 3rd edn. Springer, Berlin Heidelberg New York Papez JW (1937) A proposed mechanism of emotion. Arch Neurol Psychiatr 38:725-743 Renella RR (1989) Microsurgery of the temporo-medial region. Springer, Wien Rogers DF (1985) Procedural elements for computer graphics. McGraw-Hill, New York Rogers DF, Adams JA (1990) Mathematical elements for computer graphics. McGraw-Hill, New York Rosenfeld A (1974) Digital straight line segments. IEEE Trans Comput C-23 : 1264-1269 Schaltenbrand G, Wahren W (1977) Atlas for stereotaxy of the human brain. 2nd edn. Thieme, Stuttgart Schaumann A (1989) Computergestfitzte Berechnung der Vereinigungs-, Schnitt- und Differenzmenge triangulierter Rekonstruktionen neuroanatomischer Strukturen. Diplomarbeit Fachbereich Mathematik, Universit/it Hannover Schfitz T, Kretschmann H J, M/.iller D (1989) Automated identification of neuroanatomical structures in CT- and MR-images. In: Lemke HU, Rhodes ML, Jaffe CC, Felix R (eds) Computer assisted radiology CAR'89. Springer, Berlin Heidelberg New York, pp 287-291 Starck D (1982) Vergleichende Anatomic der Wirbeltiere, vol 3. Springer, Berlin Heidelberg New York Stephan H (1975) Allocortex. In: Bargmann W (ed) Handbuch der mikroskopischen Anatomie des Menschen, vol 4, part 9. Springer, Berlin Heidelberg New York Stephan H, Andy OJ (1982) Anatomy of the limbic system. In: Schaltenbrand G, Walker AE (eds) Stereotaxy of the human brain. 2rid edn. Thieme, Stuttgart, pp 269 292 Talairach J, Szikla G, Tournoux P, Prossalentis A, Bordas-Ferrer M, Covello L, Iacob M, Mempel E (1967) Atlas d'anatomie st~r~otaxique du t~lenc~phale. Masson, Paris Talairach J, Tournoux P (1988) Co-planar stereotaxic atlas of the human brain. Thieme, Stuttgart