Dorsal Fronto-Parietal Connections of the Human Brain: A Fiber Dissection Study of Their Composition and Anatomical Relationships

Authors

  • Igor L. Maldonado,

    Corresponding author
    1. Department of Neurosurgery, Montpellier University Hospital, Montpellier, France
    2. Investigation Group on “Central Nervous System Plasticity, Stem Cells and Glial Tumors,” INSERM-U1051 Laboratory, Institute for Neurosciences of Montpellier, University of Montpellier 1 and 2, Montpellier, France
    3. Department of Neuroradiology, Montpellier University Hospital, Montpellier, France
    • Département de Neuroradiologie, Hôpital Gui de Chauliac, CHU Montpellier, 80 Avenue Augustin Fliche, 34295 Montpellier, France
    Search for more papers by this author
    • Fax: 33-4-67-33-72-63.

  • Emmanuel Mandonnet,

    1. Department of Neurosurgery, Lariboisière Hospital, Paris, France
    Search for more papers by this author
  • Hugues Duffau

    1. Department of Neurosurgery, Montpellier University Hospital, Montpellier, France
    2. Investigation Group on “Central Nervous System Plasticity, Stem Cells and Glial Tumors,” INSERM-U1051 Laboratory, Institute for Neurosciences of Montpellier, University of Montpellier 1 and 2, Montpellier, France
    Search for more papers by this author

Abstract

The goal of this study was to detail the composition of dorsal fronto-parietal connections in the human brain, focusing on the dorsal component of superior longitudinal fasciculus (SLF), short association fibers, their three-dimensional organization, and relationships with adjacent projection and commissural fibers. Ten human cerebral hemispheres (five left and five right) were obtained from necropsy specimens. The technique for specimen preparation was adapted from that previously described by Ludwig and Klingler for spreading groups of white matter fibers, rendering tracts visible and dissectible. Near the superior border of the hemisphere, we observed an overall organization consisting of a succession of “U” fibers in both sides of a narrow and irregular intermediary layer of white matter. Dissection of the core fibers leads to the corona radiata (intermingled with the callosal radiations) on the lateral aspect and to the callosal radiations at the medial aspect of the hemisphere. Based on our findings, the fiber dissection technique does not provide evidence of the presence of long horizontal association fibers in such location, as suggested by brain imaging techniques. The results of this study lead us to hypothesize that dorsal regions of the frontal and parietal lobes superior to the level of the cingulate sulcus are connected by a succession of short association pathways. Dissectible long association fibers are only encountered in a zone restricted to a lower and deeper portion of the superior parietal lobule. These fibers are clearly integrated in the lower portions of the SLF/arcuate fasciculus complex. Anat Rec, 2012. © 2011 Wiley Periodicals, Inc.

Despite more than two centuries of research efforts, our understanding of the three-dimensional organization of fiber bundles in the white matter of the human brain is limited. The main reason is that fiber bundles are not easily seen using standard anatomical or imaging techniques.

The existence of the superior longitudinal fasciculus (SLF) was first demonstrated in the 19th century by the seminal works of Mayo, Burdach, Reil, and Dejerine and further documented in the 20th century by Ludwig and Klingler (1956a). The SLF is an important group of cerebral association fibers that connects distant sites of the frontal, parietal, and temporal lobes. Recently, Makris et al. (2005) identified and segmented the SLF, using magnetic resonance and diffusion tensor imaging (DTI) techniques in four normal male subjects. A subdivision is proposed based on the a priori knowledge from the animal (non-human primate) anatomy of four subcomponents (Petrides and Pandya,1984,1988). The dorsal component (SLF I) is depicted from the medial and dorsal regions of the parietal cortex to the dorsal and medial part of the frontal lobe. The major component (SLF II) connects the caudal inferior parietal lobe with the dorsolateral frontal region. The ventral component (SLF III) originates from the supramarginal gyrus and terminates in the ventral premotor and prefrontal areas. Finally, the fourth and final component is the arcuate fasciculus (AF), which connects mainly the temporal and frontal lobes.

In the Rhesus Monkey (Petrides and Pandya,1984; Schmahmann and Pandya,2006), the SLF I segment is a fascicle running from the medial posterior parietal region and the caudal superior parietal lobule (SPL) into the white matter of the superior frontal gyrus (F1), roughly parallel to the medial aspect of the brain hemisphere. Using autoradiographic techniques, this longitudinal pathway appears in the base as well as in the upper portions of F1 and SPL. In humans (Makris et al.,2005), antero-posterior diffusion was demonstrated using a six-direction DTI technique in a region comprised of the white matter of F1, SPL, paracentral lobule (PCL), and precuneus (PCN).

As that region is part of the upper border of the brain, it is crossed by a number of transversal sulci of different depth along their course. This is the case for the precentral, central, and postcentral sulci on the lateral surface of the hemisphere, and the paracentral and marginal branches of the cingulate sulcus in the medial surface. As a consequence, this white matter is rich in short “U” association fibers that are obliquely or horizontally oriented and vary in length. This is potentially a confusing factor for diffusion-based imaging techniques. We consider that further investigation is justified; to our knowledge, this has not yet been studied in vitro in human anatomical specimens. The goal of this study is to detail the composition of dorsal fronto-parietal connections in human brains, focusing on the dorsal component of the SLF and its anatomical relationship with adjacent projection and commissural fibers.

MATERIAL AND METHODS

Ten human cerebral hemispheres (five left and five right) were obtained from necropsy specimens fixed in formalin. The technique used for specimen preparation and dissection was adapted from that previously described by Ludwig and Klingler (Klingler,1935; Ludwig and Klingler,1956b) and has previously been used for fiber dissection studies (Ture et al.,2000; De Castro et al.,2005; Fernandez-Miranda et al.,2008; Martino et al.,2009,2010). The pia mater, arachnoid membrane, and vascular structures were carefully removed, and the hemispheres frozen at −20°C for a minimum of 14 days. The crystallization of the water molecules and formalin inside the cerebral parenchyma disrupts the structure of gray matter, which enables the cortex to be removed from the brain. Freezing also slightly spreads groups of white matter fibers, rendering tracts visible and dissectible. After studying and photographing sulci and gyri, various sized hand-made wooden spatulas were used to peel away both the brain cortex and the short “U” association (intergyral) fibers. The dissection was performed gradually from the lateral to the medial surface and from the medial to the lateral surface of the hemisphere, with concomitant detailed photographic documentation. A surgical magnification loupe or operating microscope was used when necessary for dissection of delicate groups of fibers in the depth of the white matter.

To investigate the presence of long horizontal fronto-parietal fibers and their anatomical relations to adjacent groups, two different approaches were used. The first method involved removing the cortex and direct exploring the white matter of F1, upper portion of precentral and postcentral gyri, SPL, PCL, and PCN. The second method involved exploring the white matter from the frontal pole to the parieto-occipital sulcus in two 1-cm-thick cerebral parenchyma slices, obtained by performing two cuts parallel to the superior border of the hemisphere.

RESULTS

Detailed description of the gradual dissection of the white matter is provided in the figure legends. In eight of the 10 specimens, we dissected the medial and the lateral aspects of the cerebral hemisphere from F1 to the PCN (Fig. 1). After removing the gray matter, we studied and photographed the sulcal anatomy of both aspects of that region. Then, we proceeded with the delicate removal of the subcortical white matter, containing short “U” association fibers linking one gyrus to the other. Most groups of these fibers crossed transversally to the mean axis of sulci and were therefore horizontally or slightly obliquely oriented.

Figure 1.

Gradual dissection of the white matter of the superior frontal gyrus, superior portion of pre and postcentral gyrus, SPL, PCL, and PCN of the left hemisphere of a human brain. A: Superior view. After fixation in formalin and freezing for a minimum of fourteen days, the arachnoid membrane and vessels were removed. The anatomy of gyri and sulci of the specimen was studied with serial photography at each process of the dissection. B: Oblique, supero-lateral view. The gray matter of sulci in the region being studied was progressively removed, and the short association fibers were exposed. At this step, a portion of the cortex was intentionally left in place at the most superficial portion of each gyrus. The preparation made the cortex friable and conserved the consistency of the white matter. The crystals of water and formalin inside the cerebral parenchyma slightly expanded and separated groups of fibers, facilitating the dissection. C: Oblique, supero-lateral view. Progression of the dissection. The cortical gray matter was completely removed from the area of interest. Gradual peeling of the white matter further exposed short “U” association fibers that interconnect adjacent gyri. Their orientation is perpendicular or slightly oblique to each sulcus. D: Medial aspect of the same brain hemisphere. As in the lateral aspect, the cortical gray matter in the region of interest has been removed and intergyral fibers exposed. Also, the cingulate fascicle, above the corpus callosum, was dissected and exposed in the anterior portion of the cingulate gyrus. The cingulate sulcus, which forms the inferior limit of the area of interest for this study, is widened with this procedure. In the frontal lobe, it separates the cingulate gyrus from the superior frontal gyrus. It presents an ascending (marginal) ramus along the posterior limit of the PCL, forming the anterior limit of the PCN. Note that in this specimen, the superior extremity of the central sulcus was not visible at this stage of the dissection, but the postcentral sulcus was very prominent. In the parietal lobe, the prolongation of the cingulate sulcus separates the cingulate gyrus from the PCN and is called subparietal sulcus. A deep and continuous sulcus, the parieto-occipital sulcus, separates the PCN from the cuneus of the occipital lobe. The disposition of sulci that vertically cross the medial aspect of the superior frontal gyrus, PCL, and PCN accounts for the high prevalence of horizontally oriented short association fibers in those regions (arrow). E: Supero-lateral view of the same specimen. Removal of intergyral fibers to expose the deep white matter began to show groups of fibers that are vertically disposed in both frontal and parietal lobes. The procedure failed to demonstrate horizontally oriented long association fibers, except for the lower portion of the SPL. F: The progression of the dissection to inferior portions of the hemispheric white matter exposed the AF and the external capsule. A line of white pins was placed to outline the location of the cingulate sulcus, which is the inferior limit of the superior frontal gyrus, PCL, and PCN on the medial aspect of the brain. Note that the horizontal association fibers exposed in the parietal lobe occupy only the lower portion of the region limited inferiorly by the cingulate sulcus and are not dissectible anterior to the upper central region. There is no space separating those fibers from those of the postero-superior portion of the AF. G: Detailed view on “F” in the parietal lobe. Note that the intergyral “U” fibers were preserved in the postcentral gyrus (arrow). Long association fibers are located deeper in the white matter, as are AF fibers. arc, arcuate fasciculus; calc, calcarine fissure; cc, corpus callosum; cg, cingulum (cingular fascicle); cgs, cingular sulcus; cng, cingular gyrus; cs, central sulcus; cs, cingular sulcus; ext, external capsule; F1, superior frontal gyrus; lv, lateral ventricle; mar, ascending (marginal) ramus of the cingulate sulcus; pcn, precuneus; pg, precentral gyrus; pcg, postcentral gyrus; pcs, postcentral sulcus; pos, parieto-occipital sulcus; spl, superior parietal lobule.

Following this procedure, the core and the upper portion of the white matter of the examined gyri were exposed. The core is comprised of a very thin irregular shaped and vertically oriented layer in which most fibers consisted of callosal radiations. After removal of the intergyral fibers, the dissection of both sides of F1, SPL, PCL, and PCN failed to demonstrate the presence of long horizontally oriented association fibers, except in the inferior portion of the SPL. As a consequence, the white matter running under the superior border of the hemisphere, from F1 to the caudal parietal lobe, was composed of a succession of short “U” association fibers located on either side of a central core of vertical fibers. The dissection of the core fibers lead to the corona radiata, intermingled with the callosal radiations, on the lateral aspect (Figs. 2, 3, 4, 2–4), and to the callosal radiations at the medial aspect of the hemisphere (Figs. 3, 4, 5, 3–5).

Figure 2.

Lateral view of the dissection of deep white matter of a human left cerebral hemisphere: a subsequent stage to Fig. 1. Fibers in the core of the region of study, near the superior border of the hemisphere in both frontal and parietal lobes were exposed. On the lateral aspect, they are vertically disposed and the progression of the dissection to caudal regions of the brain showed that those fibers originate from projection bundles of the corona radiata, which is situated deep to the AF. At this stage of the dissection, the procedure failed to demonstrate any horizontally oriented long association fibers in the deep white matter of the studied region. This region extended from the frontal pole anteriorly to the location of the parieto-occipital sulcus posteriorly, and from the superior border of the hemisphere superiorly to the location of the cingulate sulcus inferiorly. ac, anterior commissure; cer, cerebellum; cor, corona radiata; cp, cerebral peduncle; cs, central sulcus; F1, superior frontal gyrus; ic, internal capsule; med, medulla; oc, optic chiasma; ol, olivar body; ot, optic tract; pcn, precuneus; po, pons; ss, sagittal stratum.

Figure 3.

Lateral view of the dissection of deep white matter of a right cerebral hemisphere. The left hemisphere was also dissected and preserved in place. Vertical oriented fibers in the core of the studied region and along the superior border of the hemisphere were exposed. On the lateral aspect, the progression to caudal regions of the brain showed that those fibers originate from projection bundles of the corona radiata, which is situated deep to the AF. The superior portion of the projection fibers, next to the superior border of the frontal and parietal lobes was further dissected. This shows two partially intermingled layers, formed by corona radiata projections laterally (partially cut) and callosal radiations medially. At this stage of the dissection, the procedure failed to demonstrate any horizontally oriented long association fibers in the deep white matter of the studied region. This region extended from the frontal pole anteriorly to the location of the parieto-occipital sulcus posteriorly and from the superior border of the hemisphere superiorly to the location of the cingulate sulcus inferiorly. cal, callosal radiations; cor, corona radiata; put, putamen; ss, sagittal stratum.

Figure 4.

A: Superior view of the dissection of deep white matter of a human brain: same specimen as in Fig. 3. Vertical oriented fibers in the core of the region of study and along the superior border of both hemispheres were exposed. Ascending and descending fibers originate from two partially intermingled layers formed by projection bundles from the corona radiata (which is situated deeper to the AF) laterally and callosal radiations medially. The cingulum (cingulate fascicle), composed of antero-posteriorly oriented fibers, was also dissected and left in its natural place. This structure is situated inferiorly in relation to the region of study, above the corpus callosum and medial to the frontal and parietal callosal radiations. B: Detail on the exploration of the deep white matter in another brain specimen (a left hemisphere). The fibers from the corpus callosum cross under the cingulum before traveling to the opposite hemisphere. A and B: At this stage of the dissection, the procedure failed to demonstrate any horizontally oriented long association fibers in the deep white matter of the studied region. This region extended from the frontal pole anteriorly to the location of the parieto-occipital sulcus posteriorly and from the superior border of the hemisphere superiorly to the location of the cingulate sulcus inferiorly. A, anterior; P, posterior; M, medial; L, lateral; cal, callosal radiations; cc, corpus callosum; cg, cingulum (cingulate fascicle); cor, corona radiata; cs, central sulcus; F1, superior frontal gyrus; put, putamen.

Figure 5.

Further fiber dissection of the deep white matter of the region of study and along the superior border of both hemispheres (anterolateral view). The corona radiata was removed along as the majority of the lateral wall of both lateral ventricles. The remaining ascending/descending fibers near the medial aspect of the hemispheres are vertically oriented and arise from callosal radiations. The cingulum (cingulate fascicle) was dissected and left in place, above the corpus callosum and medial to callosal radiations. The fibers from the corpus callosum cross under the cingulum before traveling to the opposite hemisphere. At this stage of the dissection, the procedure failed to demonstrate any horizontally oriented long association fibers in the deep white matter of the studied region. This region extended from the frontal pole anteriorly to the location of the parieto-occipital sulcus posteriorly and from the superior border of the hemisphere superiorly to the location of the cingulate sulcus inferiorly. cal, callosal radiations; cg, cingulum (cingulate fascicle); ihf, inter-hemispheric fissure; tap, tapetum.

In each of the two remaining hemispheres, two 1-cm-thick slices were macroscopically studied from the frontal pole to the parieto-occipital sulcus. These curved slices were obtained parallel to the superior border of the hemisphere. As described by Ludwig and Klingler (1956a), successive freezing of the anatomical specimen may further spread groups of fibers and improve the quality of the preparation; thus, the slices were refrozen before completing each dissection. The study of those sections demonstrated an overall organization in vertically oriented lamellae, consisting of “U” fibers groups in both the medial and lateral aspects of the section and an intermediary white matter zone, which was also narrow and irregular (Fig. 6). In the intermediate layer, no horizontal long association fiber could be demonstrated using the fiber dissection technique.

Figure 6.

Dissection of the white matter of the upper portion of the cerebral hemisphere in the region of study in the frontal and parietal lobes. The sulci and gyri were studied and photographed. Then, the white matter of the superior frontal gyrus, PCL, SPL, and PCN was explored, by dissecting two 1-cm-thick cerebral parenchyma slices obtained by performing two cuts parallel to the superior border of the hemisphere, after fixation and freezing as recommended for Klingler's technique. A: Superior aspect of the second and inferior slice, passing through the base of the superior frontal gyrus and near the location of the cingulate sulcus of a right hemisphere before removal of the gray matter. The core of the white matter in the posterior half of the superior frontal gyrus is somewhat straighter and relatively thicker. In most parts of the slice, there is a high prevalence of short association fibers on both lateral and medial sides (arrow heads), delineating a very thin and irregular intermediate layer of deep white matter. B: After removal of the cortical gray matter, a left hemisphere. Superior view of the first (left image) and second (right image) slices passing through F1, SPL, and PCN at their upper and lower portions, respectively. The significant number of vertically oriented sulci, in both medial and lateral aspects, correlates to the particularly high concentration of short “U” association fibers. Note the narrowness and the irregularity of the white matter; even if intergyral fibers are left. cs, central sulcus; F1, superior frontal gyrus; mar, marginal ramus of the cingulate sulcus; occ, occipital lobe; pcr, paracentral ramus of the cingulate sulcus; pcn, precuneus; pcs, postcentral sulcus; pos, parieto-occipital sulcus; ps, precentral sulcus.

Similar observations could be made from one specimen to the other and are worth being detailed as follows: the white matter of the core of the superior frontal gyrus is slightly straighter and thicker than many other regions of the same slice, especially for its posterior half. This region is significantly less affected by transverse cerebral sulci than the rest of the slice. Some horizontal long association fibers were identified while dissecting the lateral aspect of the hemisphere in the deep base, lower portion, of the SPL; however, they were only dissectible in the parietal lobe. Although part of these fibers was found superior to the level of the cingulate sulcus, which was the inferior limit of the region of study, they formed a single contingent with the upper parietal portion of the AF (Fig. 1F,G).

DISCUSSION

When entering the white matter of the lateral aspect of the brain, the first long association fascicle horizontally oriented is the SLF, which runs deeper to the subcortical short association U-fibers (Dejerine,1895; Ludwig and Klingler,1956a; Mori,2005; Catani and Thiebaut de Schotten,2008; Fernandez-Miranda et al.,2008). It has been assumed that this fasciculus associates distant areas of the frontal, parietal, and temporal lobes, dealing with high order functions, particularly language in the dominant hemisphere (Anderson et al.,1999; Catani et al.,2005; Bernal and Ardila,2009).

According to DTI results (Makris et al.,2005), the dorsal component of the superior longitudinal fasciculus (SLF I) occupies the dorsomedial portion of the white matter of F1. Posteriorly, it remains in a dorsomedial location within the white matter of the same gyrus and the supplementary motor area. Caudal to the splenium of the corpus callosum, the SLF I was identified in the white matter of the SPL and PCN. In a sagittal view, the overall orientation of fibers was horizontal, and they were situated around the cingulate sulcus, continuing further anteriorly in the medial white matter of F1. Despite limitations of DTI techniques in determining origins and terminations of fiber bundles, it was inferred from the anatomy of the monkey that SLF I originates from the superior and medial parietal cortex. The SLF I may run through the medial superior white matter of the parietal and frontal lobes, terminate in the dorsal part of areas 6, 8, and 9, and be bidirectional (Petrides and Pandya,2002; Makris et al.,2005). From a functional point of view, it is suggested that this bundle could participate in the regulation of higher aspects of motor behavior, while taking into consideration the interconnections it may provide among SPL, SMA, and dorsal premotor cortex.

In this study, we were particularly interested in the three-dimensional organization of fiber bundles inside the white matter of F1, upper pre- and postcentral gyri, PCL, SPL, and PCN. These regions form, from anterior to posterior, the superior border of the brain hemisphere. We used the Klingler's technique, which has been recognized as the “gold standard” for gross dissection of the white matter. A specific protocol comprising of the fixation and freezing of brain specimens, enabled the spreading of groups of fibers, rendering them visible and dissectible.

Peeling the gray matter and the superficial short association fibers exposed the deep white matter and enabled its three-dimensional organization to be examined. This procedure demonstrated that the core of the F1-SPL-PCL-PCN region is composed of a very thin (∼2 mm) and irregular intermediate layer of white matter. Inside this layer, no long horizontally orientated association fibers were found, except close to the most superior fibers of the AF in its parietal, postero-superior portion. However, this procedure exposed the core of F1-SPL-PCL-PCN, rich in vertically orientated fiber groups continuous to those originating from the corona radiata and callosal radiations. Here, the callosal radiation fibers predominate medially and the projection fibers laterally.

The fact that the central layer of white matter in this region is thin and extremely irregular makes it unlikely that a long horizontal association tract runs from the frontal to the parietal lobe. The presence of deep, vertically orientated sulci is responsible for its intermittent narrowing, except at the level of its lower portion and in proximity to the cingulate sulcus. This is seen particularly in the precentral and central sulcus of the lateral aspect and the paracentral and marginal branches of the cingulate sulcus on the medial surface.

Images obtained with DTI techniques show antero-posterior diffusion within the lower and upper portions of the F1 and SPL. This can be found not only in the deep portions but also for relatively superficial white matter (Makris et al.,2005). An explanation for this apparent disparity between DTI and anatomical findings involves the three-dimensional organization of the fiber groups in these areas. Near the border of the hemisphere, where the core white matter is marked by its narrowness and the presence of deep vertical sulci, fiber dissection does not provide arguments for the presence of a horizontal association bundle, except for intergyral fibers. At that level, frontal and parietal lobes seem to be connected by a succession of short association fibers. As DTI detects the preferential direction of the diffusion of the water molecules, sites of antero-posterior diffusion may be depicted even in the most superior and superficial portions of F1, SPL, PCL, and PCN. However, this does not necessarily imply that a long fronto-parietal tract is being detected. The results of this study suggest an organization in successive short distance association pathways. While progressing to the base of the gyri, closer to the AF, longer fibers appear.

Previous studies have also identified discrepancies between the DTI and fiber dissection findings. This was the case for a small region of white matter situated laterally to the angle between the inferior aspect of the corpus callosum and the supero-medial aspect of the caudate nucleus. It has been suggested that this region harbors the superior occipitofrontal fasciculus (Dejerine,1895). Examination of the three-dimensional anatomy using the Klingler's technique does not demonstrate any superior occipitofrontal fasciculus (Ture et al.,1997). Progressive dissection from the structure that was previously thought to be the superior occipitofrontal fasciculus leads to the observation that its fibers form an angle extending inferiorly to the thalamus and therefore belonging to the superior thalamic peduncle. However, a zone of predominant antero-posterior diffusion can be depicted using DTI (Mori,2005; Makris et al.,2007).

Both gross fiber dissection and diffusion-tensor imaging techniques present limitations. Very sparse and delicate groups of fibers may be difficult to dissect and partially destroyed when the neighboring white matter is removed. Imaging techniques are not limited by this problem; however, they are susceptible to noise, resolution problems, contamination from adjacent bundles, and abrupt changes in fiber direction. In the F1-SPL-PCL-PCN region, the abundance of short associate fibers in both the medial and lateral aspects could be a confounding factor for DTI.

It is possible that more ventral parts of the horizontal portion of the SLF provide dorsal fronto-parietal connections, which might have been underestimated by the current classification and by this study. As some intergyral fibers also cross the cingulate sulcus in a relatively superficial level, there are two other potential routes for dorsal fronto-parietal connections, indirectly, through the cingulate gyrus, or directly, through the cingulate fasciculus. Currently, this cannot be confirmed. However, more information may arise from further studies and from other techniques such as chronometric experiments, direct subcortical stimulation, and from the advancement of imaging and laboratory methods. We consider that these techniques are highly complementary and together have the potential to help neuroscientists perceive the hodological organization of cerebral white matter.

CONCLUSION

On the basis of our findings, fiber dissection technique does not provide any evidence of the presence of a long horizontal fronto-parietal association bundle in the white matter of upper and middle portions of the superior frontal gyrus, SPL, PCL, and PCN. At this level, we observed an extreme tortuosity and narrowness of the white matter after removal of “U” fibers, which are highly present in both aspects of an intermediary core, containing projection and commissural fibers. Here, callosal radiations predominate medially and projection fibers laterally.

Although DTI techniques indicated the presence of a dorsal-most component of the SLF running parallel to the superior border of the brain hemisphere, connecting superior and medial portions of the frontal and parietal lobes, this was not confirmed by the results of this study. This leads us to hypothesize that, in this particular region, frontal and parietal lobes are connected by a succession of short association pathways. In the deep and lower portion of the SPL, there are dissectible long association fibers. These are clearly integrated in lower portions of the SLF/AF complex.

Acknowledgements

The authors thank the cadaver donors, who gave their own bodies to the University of Montpellier School of Medicine for medical research. They acknowledge Professor François Bonnel, M.D., Ph.D. for his constant and enthusiastic incentive and Kyriakos Lobotesis, M.D. for his useful comments on the article. They also thank Frank Meyer and Jean Marc Gory for helping in preparing the anatomic specimens and Professor François Canovas, M.D., Ph.D. for his administrative assistance.

Ancillary