Introduction

The recent revival of the white matter fiber dissection technique—first described by Klingler—and its incorporation not only to neuroanatomical education but also to neurocognitive research, in combination with the advent of sophisticated neuroimaging methods such as functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI) have led to a more profound understanding of the brain connectivity and anatomo-functional organization (Catani et al. 2002; Fernandez-Miranda et al. 2008a, b; Klingler 1935; Klingler and Ludwig 1956; Mamata et al. 2002; Mandonnet et al. 2006; Mori et al. 1999; Ture et al. 2000; Peltier et al. 2010a; Duffau et al. 2002, 2005; Arnts et al. 2014; Koutsarnakis et al. 2015). The coining of the modern concept of brain hodotopy (deriving from the Greek words hodos meaning road, pathway and topos meaning place) moves away from the classical localizationist view and introduces a new model of perceiving cerebral function not only as a cortical phenomenon but as an integral cortico-subcortical epiphenomenon (Duffau 2006, 2011, 2015, 2017; Duffau et al. 2014; Catani and de Schotten 2012; Catani et al. 2003, 2005; De Benedictis and Duffau 2011). In other words, apart from an intact cortical organization, the integrity of the white matter axonal connectivity is essential to cerebral physiology and function.

Since brain anatomy and function are tightly and reciprocally connected, a special interest has aroused over the past few years regarding the intricate morphology and architecture of white matter pathways previously described in core neuroanatomical texts, aiming not only to refine anatomical knowledge but mainly to improve our understanding of cerebral connectivity and function (Fernandez-Miranda et al. 2008b; Peltier et al. 2010a, b; Martino et al. 2010; Pascalau et al. 2018; Koutsarnakis et al. 2016). Recent literature has thus revisited the subcortical anatomy of various, eloquent and non-eloquent fiber tracts using white matter dissections and DTI techniques with the overarching goal to combine the extracted evidence with functional data, thus providing a refined concept of brain organization (Catani et al. 2002, 2005; Fernandez-Miranda et al. 2008a, b; Koutsarnakis et al. 2015, 2016, 2017b, 2018; Mamata et al. 2002; Mandonnet et al. 2006; Mori et al. 1999; Ture et al. 2000; Peltier et al. 2010a, b; Duffau et al. 2002, 2005, 2014; Duffau 2006, 2015; Catani and de Schotten 2012; Martino et al. 2010; Kier et al. 2004a, b).

In this regard, we opted to investigate, through cadaveric fiber microdissections and an vivo DTI study, the structural architecture of a recently identified fiber tract known as the “sledge runner fasciculus” (Gungor et al. 2017; Vergani et al. 2014; Baydin et al. 2017; Beyh et al. 2017). This white matter pathway has been allegedly implicated in the axonal connectivity of high level cortical areas, such as the parahippocampal place area (PPA) and posterior cingulate or retrosplenial cortex (PCC-RSC), which represent key hubs of the neural circuit underlying the cognitive ability of spatial navigation and visuospatial imagery (Whittingstall et al. 2014; Epstein et al. 2017, 2007a, b; Epstein 2008; Ino et al. 2002; Epstein and Kanwisher 1998; Spiers and Maguire 2006, 2007; Aguirre and D’Esposito 1999; O’Craven and Kanwisher 2000; Maguire et al. 1997; Beyh et al. 2017). Indeed, meticulous neuropsychological assessment of patients harbouring posterior circulation infarcts with posterior parahippocampal and/or retrosplenial area damage indicates increasing difficulty or even inability of these individuals to identify large scale scenes, such as landscapes or cityscapes, and further inability of proper self-navigation and way finding in familiar environments (Aguirre and D’Esposito 1999; Barrash et al. 2000; Pallis 1955; Hecaen et al. 1980; Takahashi et al. 1997; Ino et al. 2007; Katayama et al. 1999; Bottini et al. 1990; Osawa et al. 2008; Ploner et al. 2000).

In this study, we provide anatomic and imaging data on the topography, morphology and axonal connectivity of the sledge runner fasciculus (SRF) using the Klinger’s dissection technique and DTI tractography on the publicly available dataset from the Human Connectome Project. Our overarching goal is to clarify the intricate subcortical architecture of this recently identified white matter tract believed to subserve spatial navigation and visuospatial imagery.

Materials and methods

The study was divided in two parts, including cadaveric microanatomic white matter fiber dissections and a DTI imaging protocol in normal individuals.

White matter fiber micro-dissection

Twenty normal, adult, cadaveric cerebral hemispheres (11 right hemispheres—9 left hemispheres) obtained from 20 different cadavers previously fixed in a 10–15% formalin solution for a minimum period of 8 weeks, were investigated. Following careful removal of the arachnoid membrane and vessels, all specimens underwent the Klingler’s procedure (freeze–thaw process) and were subsequently investigated using the fiber dissection technique and the microscope (Carl Zeiss OPMIR Plus, Carl Zeiss AG, Oberkochen, Germany) (Klingler 1935; Klingler and Ludwig 1956; Ture et al. 2000; Koutsarnakis et al. 2015).

Given that the sledge runner fasciculus lies in the depth of the medial and medio-basal occipital lobe (Gungor et al. 2017; Vergani et al. 2014; Baydin et al. 2017), we opted to perform regional (confined to the occipital lobe), microanatomic dissections in a medial to lateral direction in all 20 hemispheres.

The regional sulcal anatomy consisting of the parieto-occipital sulcus (demarcating the precuneus and isthmus of the cingulate gyrus from the cuneus), calcarine fissure (demarcating the cuneus from the lingual gyrus) and their common stem (demarcating the lingual and posterior parahippocampal gyri from the isthmus of the cingulate gyrus) the collateral sulcus (demarcating the parahippocampal from the fusiform gyrus) and sub-parietal sulcus (demarcating the isthmus of the cingulate gyrus from the precuneus) were readily identified in all specimens (Fig. 1). The fiber microdissections were initially focused on the area of the cuneus and lingual gyrus and were further extended to the adjacent cerebral region of the cingulate isthmus, posterior half of the precuneus and posterior part of the parahippocampal gyrus with the aim to study and record the morphology, topography and dimensions of the sledge runner fasciculus (Fig. 2). Special emphasis was also placed on the correlative anatomy of the sledge runner fasciculus with the sulci gyri, adjacent white matter pathways and ventricular compartments.

Fig. 1
figure 1

Sulcal and gyral anatomy of the medial and basal cerebral surface of a left (a) and right (b) hemisphere. The parieto-occipital sulcus (demarcating the precuneus, isthmus of the cingulate gyrus and cuneus), the calcarine fissure (demarcating the cuneus and the lingual gyrus), their common stem (demarcating the lingula posterior parahippocampal gyrus and the isthmus of the cingulate gyrus), the collateral sulcus (demarcating the parahippocampal and fusiform gyri) and sub-parietal sulcus (demarcating the isthmus of the cingulate gyrus and the precuneus) are illustrated. c The cerebral areas known as the Retrosplenial Cortex (RSC) and Parahippocampal Place Area are demarcated in the specimen shown in b to correlate the anatomical landmarks to their functional equivalent. d The arcuate or U fibers of the medial surface of a right hemisphere are depicted. The plane of the parieto-occipital sulcus, calcarine fissure, their common stem and sub-parietal sulcus is marked for better orientation. C cuneus, CiG cingulate gyrus, CF calcarine fissure, CoS collateral sulcus, IsCiG isthmus of the cingulate gyrus, Li lingual gyrus, PHG parahippocampal gurus, POS parieto-occipital sulcus, POS/CF parieto-occipital sulcus and calcarine fissure common stem, PPA parahippocampal place area, PrC precuneus, RSC retrosplenial cortex, SubPS sub-parietal sulcus

Full size image
Fig. 2
figure 2

Dissecting the U-fibers of a right (a) and a left (b) hemisphere reveals the fibers of the sledge runner fasciculus (SRF), the fibers of the cingulum arching over the splenium of the corpus callosum and a group of vertically oriented fibers located at the depth of the calcarine fissure, corresponding to the tract known as the “stratum calcarinum”. a Right hemisphere: the SRF is seen to course in an oblique direction- just under the U-fibers of the medial cerebral surface-connecting the cuneus, the anterior part of the lingual gyrus, the cingulate isthmus and the posterior part of the parahippocampal gyrus. Note the fibers of the stratum calcarinum connecting the superior and inferior banks of the calcarine fissure. The direction and trajectory of the SRF and stratum calcarinum (SC) are marked with curved arrows. This white matter morphology of both the SRF and SC was consistently encountered in all studied specimens. b Left hemisphere: The anatomical silhouette of the SRF and SC is delineated with curved arrows. Note how the most anterior fibers of the SRF cross the plane of the parieto-occipital sulcus as they travel ventrally towards the isthmus of the cingulum and posterior part of the parahippocampal gyrus. Note also the clear anatomical boundary between the fibers of the SRF and SC in this specimen (distance between the most anterior black curved arrows placed on th SC and the posteriorly placed white arrow on the SR). C cuneus, CiG cingulate gyrus, CF calcarine fissure, CoS collateral sulcus, IsCiG isthmus of the cingulate gyrus, Li lingual gyrus, PHG parahippocampal gurus, POS parieto-occipital sulcus, POS/CF parieto-occipital sulcus and calcarine fissure common stem, PrC precuneus, SC stratum calcarinum, SRF sledge runner fasciculus, SubPS Sub-parietal sulcus

Full size image

The dissection tools used consisted of fine metallic periosteal elevators, various sized anatomical forceps, and micro-scissors (Koutsarnakis et al. 2015, 2016). Numerous photographs were obtained during cadaveric dissections to illustrate the regional cortical and subcortical anatomy of interest. Of note is that the dissection photos included in this study have not been edited by picture enhancing software so that they closely resemble the anatomy encountered during standard fiber micro-dissections in the setting of a microneurosurgery laboratory (Koutsarnakis et al. 2015, 2016, 2017a).

DTI imaging

Data used in the preparation of this work were obtained from the Human Connectome Project (HCP) database (https://ida.loni.usc.edu/login.jsp). The HCP project (Principal Investigators: Bruce Rosen, M.D., Ph.D., Martinos Center at Massachusetts General Hospital; Arthur W. Toga, Ph.D., University of Southern California, Van J. Weeden, MD, Martinos Center at Massachusetts General Hospital) is supported by the National Institute of Dental and Craniofacial Research (NIDCR), the National Institute of Mental Health (NIMH) and the National Institute of Neurological Disorders and Stroke (NINDS). HCP is the result of efforts of co-investigators from the University of Southern California, Martinos Center for Biomedical Imaging at Massachusetts General Hospital (MGH), Washington University, and the University of Minnesota.

Thirty-five healthy participants [mean age 31.1 years, (20–59 years old); 16 females/19 males] from the MGH-USC Adult Diffusion Dataset (Human Connectome available dataset) underwent brain imaging in a 3-Tesla CONNECTOM magnetic resonance imaging (MRI) scanner (Setsompop et al. 2013) housed at the MGH/HST Athinoula A. Martinos Center for Biomedical Imaging with a custom-made 64-channel phased array head coil (Keil et al. 2013). Anatomic imaging was performed with a 1 mm isotropic Multi-echo Magnetization-Prepared Rapid Acquisition Gradient Echo three-dimensional T1-weighted sequence. DTI parameters included an axial spin-echo multiband echo-planar imaging sequence with 64 diffusion encoding directions; field of view: 210 mm; acquisition voxel size: 1.5 × 1.5 × 1.5 mm3; repetition time: 8800 milliseconds; echo time: 57 milliseconds; b factors with 0 s/mm2 (low b value) and 1000 s/mm2 (high b value). The acquisition consisted of 96 slices.

DTI images were processed by the Brainance™ DTI Suite (Advantis Medical imaging, Eindhoven, The Netherlands). We applied a multiple region-of-interest (ROI) approach (Fig. 3) based on anatomical landmarks derived from previous white matter fiber dissection studies and the high signal intensity of the expected SR on fractional anisotropy (FA) maps. Tracing of the ROIs was performed in the axial (ROI 1 and ROI 2) and the coronal (ROI 3, ROI 4, ROI 5) planes, utilizing five different slices and the ROIs were placed based on discrete anatomical boundaries (Petrides 2012). For the first ROI, the parieto-occipital sulcus was identified in a sagittal plane, where the trajectory of the sledge runner could be also identified as a continuous tract on FA map (Fig. 3). Then, an axial plane was selected at the level of the lateral ventricles where the genu and the splenium of the corpus callosum were clearly distinguishable (upper 1/3 of the parieto-occipital sulcus); ROI 1 included a circular ROI placed just posterior to the cingulate sulcus (Fig. 3). The second ROI was placed in the axial plane, below the first slice, at a level where the posterior part of the insula and only the caudate nucleus among the basal ganglia were clearly identified; ROI 2 included a circular ROI covering a similar region as ROI 1 posterior to the cingulate sulcus (Fig. 3). For the third ROI, the mid part of the parieto-occipital sulcus was identified in the sagittal plane and a circular ROI was placed in the coronal plane covering the cerebral area demarcated by the posterior part of the calcarine sulcus and the white matter included between the lingual sulcus and the collateral sulcus (Fig. 3). The fourth ROI was selected in the coronal plane at a level posterior to the splenium of the corpus callosum, where a part of the SRF medially to the lateral ventricle, below the calcarine sulcus and above the collateral sulcus/occipital ramus was identified (Fig. 3). The fifth and most anterior ROI was selected at the level of the middle part of the splenium of the corpus callosum, where a region of the SRF under the corpus callosum (medial to the lateral ventricles, below the calcarine sulcus and above the collateral sulcus/occipital ramus) was identified with a circular ROI (Fig. 3). All ROIs in individual space for each subject as well as subjects’ DWI are available as supplementary material (DWI_ROIs_all_subjects.zip folder contains 35 folders, one per subject. Each folder contains three .nii files (DWI and ROI data). ROIs files have been named based on the following convention. The first four digits correspond to the ID number of each subject of the MGH-USC Adult Diffusion Dataset (Human Connectome available dataset). The next two digits, i.e. “SR”, refer to the sledge runner fasciculus and the last digit corresponds to the left and right hemisphere). All tracts passing through these ROIs with a minimum FA of 0.15, maximum angle change of 90°, minimum length of 0 mm and maximum length of 200 mm were traced. The following DTI measures were then extracted for the right and left SRF: mean fractional anisotropy (FA); mean axial diffusivity (AD); mean radial diffusivity (RD); minimum, maximum and mean length; number of fibers (NoF). With regard to the topography of SRF in relation to adjacent white matter tracts, we further reconstructed the following tracts: forceps major, cingulum bundle, hippocampal—cingulum pathway, inferior longitudinal fasciculus. The reconstruction was based on a previous well-standardized protocol (Wakana et al. 2007) and Brainance™ DTI Suite default threshold values for FA, angle change and fiber length.

Fig. 3
figure 3

Representation of the multiple ROIs used for the reconstruction of the SRF and projection over an FA map. ROI region of interest, SRF sledge runner fasciculus, FA fractional anisotropy

Full size image

DTI data of ten participants were analyzed twice by a single experienced rater (E. K.) with an interval of 4 weeks, to assess intra-observer agreement. To determine inter-observer agreement for the SRF, the same dataset was additionally analyzed by a second experienced rater (F. C.), who was blinded to the results of the first rater. Intra- and inter-rater agreement (Table 1) was evaluated within the group of healthy participants with intraclass correlation coefficients (ICC) with an ICC greater than 0.75 being indicative of good agreement (Shrout and Fleiss 1979).

Table 1 Indices of intra- and inter-rater agreement for SRF DTI measures for the group of healthy participants
Full size table

Results

Microanatomic dissection

A medial to lateral microanatomic fiber dissection focused on the area of the cuneus, lingual gyrus, posterior half of precuneus, isthmus of the cingulate gyrus and posterior parahippocampal gyrus was employed to study and record the morphology, connectivity and correlative topographic anatomy of the sledge runner fasciculus (Fig. 1). Initially, peeling away the cortex of the aforementioned gyri exposes the arcuate or U fibers (Fig. 1). Progressive dissection of the regional U fibers reveals the Sledge Runner fasciculus along with fibers arching from the cingulum over the splenium of the corpus callosum (distal part of the superior arm and proximal part of the inferior arm of the cingulum). A distinct group of fibers exhibiting a vertical trajectory and encountered at the depth of the calcarine fissure correspond to the tract previously described by H.Sachs and known as “stratum calcarinum” (Fig. 2).

Connectivity and morphology of the sledge runner fasciculus (SRF)

The sledge runner fasciculus (SRF) was consistently identified as an oblique band of fibers travelling under the U-fibers of the medial surface of the occipital lobe, connecting the areas of the anterior cuneus, anterior half of the lingual gyrus, cingulate isthmus and posterior parahippocampal gyrus (Fig. 2). With respect to its trajectory and configuration, the SRF demonstrated a dorsomedial–ventrolateral direction with two medial curves—one at the level of the major forceps and the other at the inferior margin of the medial wall of the ventricular atrium- thus, creating its peculiar sled-like shape (Fig. 4). It was also consistently found to exhibit a superior narrow part, corresponding to the antero-superior part of the cuneus, which was seen to widen progressively as the tract descended towards the lingual gyrus (Fig. 4). The average length of the SRF was 4.9 cm (range 4.7–5.2 cm), while its average width was 1.7 cm (range 1.3–2) at its narrowest and 2,1 cm (range 1.8–2.5) at its widest part.

Fig. 4
figure 4

a, b The morphology and configuration of the SRF is illustrated in two left hemispheres. The standard and consistent dorsomedial–ventrolateral direction of the SRF is depicted in both specimens. Note how the tract progressively widens from its dorsal narrow part, corresponding to the area of the cuneus (marked with the grey arrow), to its wider ventral part at the area of the cingulum and posterior parahippocampal gyrus (marked with the white arrow). During its course the SRF usually exhibits two medially directed curves, as seen and marked with black and white stars in these specimens. In the specimen included in b, we preserved the fibers of the SRF (marked with the white strap) while deepening the dissection plane posteriorly and exposing the more medially located fibers of the forceps major, thus illustrating the upper curve to lie at the level of the major forceps (marked in both specimens with the black star). The lower curve, marked with the white star, roughly corresponds to the level of the infero-medial wall of the ventricular atrium. C cuneus, CiG cingulate gyrus, FrM forceps major, IsCiG isthmus of the cingulate gyrus, Li lingual gyrus, PHG parahippocampal gurus, PrC precuneus, SC stratum calcarinum, SRF sledge runner fasciculus

Full size image

SRF correlative anatomy with respect to the medial surface sulci and gyri

The characteristic aforementioned silhouette of the sledge runner fasciculus places its superior fibers in a deep plane just posterior to the superior two-thirds of the parieto-occipital sulcus. As the bundle travels anteriorly, to connect the area of the cuneus with the isthmus of the cingulate gyrus, its respective fibers are found deep at the level of the inferior one-third of the parieto-occipital sulcus. There were no SRF fibers crossing the plane of the sub-parietal sulcus or entering the white matter of the precuneal lobule.

Deep to the calcarine fissure, the SRF fibers were seen to cross the distal part of the fissure in an almost vertical fashion, connecting the cuneus with the anterior half of the lingual gyrus. In relation to the common trunk formed by the proximal parts of the calcarine fissure and the parieto-occipital sulcus, the SRF fibers course in an almost parallel fashion deep to the trajectory of the sulcus, extending both superiorly and inferiorly to reach the level of the cingulate isthmus and posterior part of the parahippocampal gyrus, respectively (Fig. 2). With respect to the regional gyral anatomy, the SRF lies deep at the level of anterior half of cuneus, anterior half of the lingual gyrus, isthmus of cingulate gyrus and posterior part of parahippocampal gyrus.

SRF Subcortical correlative anatomy

We observed the fibers of the cingulum lying adjacent to the fibers of the SRF. This relation was particularly tight with regard to the most anteroinferior part of the SRF, which runs deep to the isthmus of the cingulate gyrus and posterior part of the parahippocampal gyrus. To clarify and illustrate more vividly the regional anatomy, we dissected this part of the sledge runner fasciculus and we further observed that it lies medial and posterior to the fibers of the cingulum, displaying a postero-supero-medial to antero-infero-lateral direction while the cingulum fibers travel antero-supero-medially to postero-infero-laterally. (Fig. 5).

Fig. 5
figure 5

Photos of a right (a) and left (b) cerebral hemisphere zooming in the correlative topography of the SRF with regard to the fibers of the cingulum. In both figures, the anatomical vicinity between the most anteriorly descending fibers of the SRF and the fibers of the inferior arm of the cingulum is illustrated. The different trajectories of the aforementioned fibers are marked with arrows. Note that the SRF travels in a postero-supero-medial to antero-infero-lateral direction while the fibers of the cingulum take an antero-supero-medial to postero-infero-lateral course. CC corpus callosum, Ci cingulum, SR Sledge runner fasciculus

Full size image

Deep to the plane of the calcarine fissure, we consistently observed a distinct group of vertically oriented fibers known as the stratum calcarinum, running between the upper and lower banks of the fissure and connecting the upper and lower lips of the calcarine cortex. This elegant tract lies posterior to the SRF and in 60% of the studied specimens (12/20) it exhibits a clear demarcation from the SRF (Fig. 4). In the remaining 40% of hemispheres (8/20) this tract was seen to intermingle with the fibers of the SRF anteriorly, displaying however a different orientation and trajectory (vertical for the SC vs. oblique for the SRF) (Fig. 6).

Fig. 6
figure 6

High magnification of the regional anatomy of a right hemisphere at the level of the SRF and SC fasciculi i.e. just under the U fibers of the medial surface of the occipital lobe. Note the tight anatomical relationship between the fibers of the SR and SC illustrated with the white circle. Continuous and dashed arrows delineate the trajectory of the SR and SC respectively. SC stratum calcarinum, SRF sledge runner fasciculus

Full size image

Gradual dissection of the SRF reveals the fibers of the forceps major, running from the splenium of the corpus callosum towards the medial wall of the atrium and occipital horn (Fig. 7). By removing the forceps major and the subsequently encountered ventricular ependyma we finally enter the atrium and occipital horn. During this process we observed the superior part of the SRF lying medially and coursing almost vertically in relation to the location and trajectory of the forceps major while the inferior SRF part was always found to correspond to the medial wall of the atrium.

Fig. 7
figure 7

Oblique view of a right hemisphere. Gradual white matter dissection exposes the fibers of the major forceps and the ventricular ependyma of the medial wall of the atrium. Ci cingulum, Ep ependyma, FrM forceps major, SpR splenial radiations

Full size image

DTI imaging

The applied multi-ROI tractography protocol was successful for the reconstruction of the SRF and the extracted results were in line with the trajectory, direction and connectivity of the SRF fibers found during the white matter fiber dissection part of the study. The reconstructed SRF for a single subject is superimposed on a sagittal 3D T1 weighted image (Fig. 8), where the trajectory of the SRT along the anterior cuneus, anterior part of the lingual gyrus and posterior parahippocampal gyrus is delineated. We failed to reconstruct the left SRF in one subject with the predefined thresholds for FA, angle threshold and step size.

Fig. 8
figure 8

Representation of the trajectory of the right SRF for a healthy participant and projection over an anatomical 3D high-resolution T1 weighted image (sagittal plane). SRF sledge runner fasciculus, POS parieto-occipital sulcus, C cuneus, IsCIG cingulate isthmus, PHG parahippocampal gyrus, Li lingual gyrus

Full size image

DTI indices and CV% are presented in Table 2. The variability of the reconstructed tracts for the left and right hemispheres for all participants is depicted by CV%.

Table 2 DTI indices of bilateral sledge runner fasciculus reconstructions
Full size table

To further delineate the SRF white matter topography, we also reconstructed the forceps major, the cingulum bundle, the hippocampal–cingulum pathway and the inferior longitudinal fasciculus. Figure 9 depicts the anatomy of SRF in association with the adjacent white matter tracts in one subject.

Fig. 9
figure 9

Representation of the topography of SRF (colored in yellow) in association with adjacent white matter tracts [cingulum bundle (colored in blue); hippocampal-cingulum pathway (colored in green); forceps major (colored in purple); inferior longitudinal fasciculus (colored in red)] and projection over a FA image (sagittal/axial plane). SRF sledge runner fasciculus, FA fractional anisotropy

Full size image

Discussion

Although many texts and atlases have previously investigated the anatomy and fiber tract morphology of the occipital lobe (Catani et al. 2002; Fernandez-Miranda et al. 2008a; Mamata et al. 2002; Mori et al. 1999, 2005; Ture et al. 2000; Koutsarnakis et al. 2015; Catani and de Schotten 2012; Alves et al. 2012; Forkel et al. 2014; Thiebaut de Schotten et al. 2014; Schmahmann and Pandya 2009; Sachs 1892; Catani and Thiebaut de Schotten 2008), it was not until the publication of the laboratory anatomical report by Vergani and colleagues entitled “Intralobar fibres of the occipital lobe: A post mortem dissection study” in which the authors identified a novel fiber bundle that was not previously described and named it “sledge runner fasciculus” after its peculiar shape (Vergani et al. 2014). In this study the authors explored the white matter architecture of the occipital lobe in three cadaveric hemispheres through the fiber dissection (Klingler’s) technique and compared their results to the anatomic atlas written by H.Sachs “The hemispheric white matter of the human brain. Part I: The occipital lobe” (Sachs 1892). During their dissections they found a very close concordance between the histological preparations of Sachs and the Klingler’s technique that they used, with the only difference being the identification of an unrecognized fiber pathway located in a plane deep to the calcarine fissure. However, Vergani and colleagues did not proceed to a more detailed description of this tract in terms of morphology, topography and connectivity since it was out of the scope of their manuscript. Since then, there have been only two published anatomo-tractographic studies entitled “The white matter tracts of the cerebrum in ventricular surgery and hydrocephalus” by Gungor et al. (2017) and “Fiber Tracts of the Medial and Inferior Surfaces of the Cerebrum” by Baydin et al. (2017) that refer to this fiber tract, but again the anatomic descriptions provided are not detailed due to the more generic objective of these laboratory investigations.

In this context, the present study attempts to address this gap and enhance anatomical knowledge regarding the sledge runner fasciculus (SRF) through cadaveric white matter dissections combined with a DTI protocol carried out in a public available dataset from the HCP. Additional available ROIs in individual space, descriptive measures and degree of variability for the most commonly used DTI metrics are also provided. DTI based tractography is a sophisticated neuroimaging technique widely used to reconstruct and noninvasively map the cerebral white matter architecture. It is however prone to anatomic inaccuracies and artifacts, particularly when applied to complex subcortical areas, related not only to the MR noise but also to the crossing, kissing and bending effect of the fiber tracts and thus it is deemed unfit to discover and describe new white matter pathways (Oouchi et al. 2007; Vos et al. 2011; Johansen-Berg and Rushworth 2009; Le Bihan et al. 2006; Fernandez-Miranda et al. 2012; Tournier et al. 2012.) We therefore opted to combine DTI tractography with the fiber microdissection technique, which is considered the gold standard method for investigating novel fiber tracts and is used to validate evidence coming from DTI studies (Zemmoura et al. 2016; Silva and Andrade 2016). More specifically, 20 normal adult cadaveric hemispheres were dissected using the Klingler’s technique and the microscope and 35 normal individuals with available DTI data from the HCP were analyzed using ROI-based DTI tractography, with the aim to describe in detail the topography, morphology and connectivity of this elegant fiber tract, establish the degree of correspondence between microanatomic dissection and in vivo tractography findings and provide descriptive DTI measures for the SRF. We therefore consistently identified the SRF as a distinct white matter pathway lying under the U fibers of the medial occipital lobe, exhibiting an oblique dorsomedial–ventrolateral direction and connecting the areas of the anterior cuneus, anterior lingula, isthmus of the cingulum and posterior parahippocampal gyrus. It originates as a relatively narrow tract, at the area of the cuneus, and during its course, it fans out and progressively widens as it descends towards the parahippocampal region, displaying two medial curves at the level of the forceps major and inferior part of the ventricular atrium respectively. With regard to its topography, the SRF was seen to cross the plane of the calcarine fissure and that of the common trunk formed by the calcarine fissure and parieto-occipital sulcus, thus connecting the cuneus to the lingula and posterior parahippocampal region, while it was observed to course deep to the proximal 1/3 of the parieto-occipital sulcus, connecting the cuneus to the isthmus of the cingulum. It was consistently found not to cross the plane of the sub-parietal sulcus or to enter the white matter of the precuneal lobule. The fibers of the SRF lie in the same plane and just anterior to the fibers of the stratum calcarinum, a vertically oriented white matter tract that connects the upper and lower lips of the calcarine cortex that has been previously described in the anatomic atlas of H. Sachs (Sachs 1892). The two tracts display a clear boundary in 60% of studied specimens, while in the remaining 40% they were observed to progressively intermingle in their ventral projection, exhibiting however a different trajectory, which aids in their proper identification and differentiation. Due to this elegant morphology, these fiber pathways should be carefully identified and dissected since anatomic misinterpretation can very easily occur. Topographically and in deeper dissection plane, the dorsal part of the SRF was found to correspond to the forceps major while its ventral part to the medial wall of the ventricular atrium.

A particularly tight anatomic relationship between the SRF and the cingulum bundle at the area of the cingulate isthmus, just posterior to the splenium of the corpus callosum was revealed in our dissections. The cingulum bundle is a white matter tract which encircles the corpus callosum and through its radiations supports the connection of the frontal, temporal and parietal lobes (Jones et al. 2013a; Wu et al. 2016). In particular, the posterior projection of this bundle known as posterior cingulum (Ito et al. 2015) or parahippocampal cingulum (Jones et al. 2013a) or Cingulum Bundle V (Wu et al. 2016) connects the parahippocampal gyrus with the precuneus (Baydin et al. 2017). As such, to reach the precuneus, the cingulum travels in the depth of the parahippocampal gyrus and cingulate isthmus sharing a similar trajectory with the sledge runner at this level. In this context, we aimed to elucidate this relationship by heavily focusing our dissections on the specific areas. We observed that as the SRF courses deep to the cingulate isthmus it lies superficially (medially) to the cingulum fibers with a postero-supero-medial to antero-infero-lateral direction while the cingulum fibers course from antero-supero-medially to postero-infero-laterally before they arch over the splenium of the corpus callosum (Fig. 5). Moreover, the termination fibers of the SRF were encountered at the posterior part of the parahippocampal gyrus while the cingulum was observed to continue its course towards the anterior temporal lobe (Fig. 5).

Regarding the in vivo reconstruction of the SRF, we applied a multi-ROI protocol in publicly available data from the HCP, which is fully described for replication and yielded high intra- and inter-rater reliability for the examined DTI indices. We also used a robust DTI algorithm, which has been previously shown to be reliable for the reconstruction of white matter tracts including fibers (e.g. lateral cortical projections and crossing pontine fibers of the corticospinal tract) that cannot be reconstructed with previous deterministic algorithms (Christidi et al. 2016). By qualitatively comparing dissection and tractography findings, our study provides evidence of similar findings concerning the anatomical trajectory of the SRF (i.e. connections between the anterior cuneus, anterior lingual, isthmus of the cingulate gyrus and posterior parahippocampal gyrus) and its topography in relation to adjacent white matter tracts (i.e. cingulum bundle, hippocampal–cingulate pathway, forceps major). However, using predefined thresholds, we failed to reconstruct the left SRF in one subject; by applying different thresholds by means of FA and angle degree, the resulted tract was considered erroneous and unreliable (e.g. false-positive). The tracking of the SRF is challenging because of the turning angle and the curving features along its course as well as fiber intercrossing. Elevated CV% of DTI metrics might indicate the complexity of the SRF reconstruction. System—(e.g. B0 inhomogeneity or gradient non-linearity) or radiographer—(e.g. subject positioning, slice tilt) related factors can contribute to the increased intersubject variation. However, in a well-designed study where system—and user-related errors had been minimized, there was increased variability in DTI indices using ROIs measures in different brain anatomical structures (Veenith et al. 2013) which is in line with previous studies (Bisdas et al. 2008; Danielian et al. 2010; Heiervang et al. 2006; Takao et al. 2012). In our study, the calculated CV% values of FA, AD and RD metrics are in accordance with previous studies. The above-mentioned findings as well as the high complexity of SRF and the high intra- and inter-observer agreement strengthen the reliability of our findings. A multimodal task-related fMRI-DTI study is definitely warranted to fully describe the structural trajectory and the functional role of SRF.

From an anatomical–functional point of view, we reconstructed a white matter tract which was so far under-recognized as a distinct white matter tract, has a significant functional role and its anatomical trajectory includes regions that are often involved in neurosurgical procedures. The reconstruction of white matter tracts in everyday radiological practice is often hampered by low-resolution DWI sequences which were initially included in the clinical protocols to exclude and/or identify recently onset vascular lesions and not to reconstruct specific white matter tracts and especially those that are highly relied on non-low resolution DTI data. In our study, we applied a DTI reconstruction algorithm which has previously led us to reliably reconstruct not only common associative white matter tracts such as the uncinate fasciculus but also the hippocampal perforant pathway (Christidi et al. 2017) and the ipsilateral/contralateral cortico-cerebellar connections (Karavasilis et al. 2018). Tractography protocols are applied using manually placed or atlas-based ROIs. Even though using atlas-based ROIs created in a standard space such as MNI can be easily used to compare between subjects or studies (Froeling et al. 2016), this approach cannot be applied in misaligned data or in white matter tracts with significant anatomical variations (such as the one examined in our study). The main purpose of our study was to identify the SRF using a reconstruction protocol that can be further used by experienced or non-experienced DTI users in clinical and research practice using available DTI platforms provided either by the MRI scanner manufacturers or other companies. Thus, by providing a detailed description of the anatomical trajectory of SRF on DTI data with a public access (i.e. Human Connectome Project) and the identification of multiple ROIs which are publicly available in individual space, our protocol might be helpful for both experienced and non-experienced raters on the identification and reconstruction of SRF.

Given its subcortical architecture, the SRF was consistently recorded to connect the cortices of the anterior cuneus, anterior lingula, isthmus of the cingulate gyrus and posterior parahippocampal gyrus in all studied hemispheres. Growing body of evidence suggests that these cerebral areas are primarily involved in the perception of complex visual scenes and in the recovery of familiar spatial knowledge, thus subserving spatial memory, wayfinding and visuospatial imagery (Whittingstall et al. 2014; Epstein et al. 2001, 2007a, b, 2017; Epstein 2008; Ino et al. 2002; Epstein and Kanwisher 1998; Spiers and Maguire 2006, 2007; Aguirre and D’Esposito 1999; O’Craven and Kanwisher 2000; Barrash et al. 2000; Pallis 1955; Maguire et al. 1997; Henderson et al. 2008; Sugiura et al. 2005; Wolbers and Buchel 2005). More specifically, the posterior part of the parahippocampal gyrus, widely known in neuropsychology as the parahippocampal place area (PPA), has been documented through functional imaging and stroke studies to activate strongly to complex landscapes and cityscapes, mainly by encoding the main spatial outline of the scene (Epstein et al. 2001, 2007b, 2017; Epstein 2008; Knauff et al. 2000). To this end, the PPA seems to respond to a large topographical entity by treating it in a unified manner i.e. as a single discrete object, in contrast to hippocampal activity, which primarily encodes information about specific objects and their respective spatial location within a certain landscape, thus supporting a concept of double dissociation between scene and object recognition (Janzen and van Turennout 2004; Lee et al. 2005; King et al. 2002; Goh et al. 2004; Buffalo et al. 2006). Interestingly, PPA activity is not exclusively dependent on visual stimuli but can also be triggered by scenes and landscapes brought in mind during mental navigation tasks (O’Craven and Kanwisher 2000; Maguire et al. 1997; Bridge et al. 2012; Rosenbaum et al. 2004).

The cerebral territory, in turn, demarcated by the anterior lingula, anterior cuneus and cingulate isthmus, collectively referred as the retrosplenial cortex (RSC) in the field of neurocognitive research, has been documented to respond mainly to familiar topographical entities (Ino et al. 2002; O’Craven and Kanwisher 2000; Epstein et al. 2007b). Indeed, RSC is involved in the recovery of familiar spatial knowledge thus participating in the neural circuit of long term spatial memory (Epstein et al. 2007a; Sugiura et al. 2005; Vann et al. 2009). In addition, evidence from fMRI studies indicates that the retrosplenial cortex also activates when subjects imagine orienting and navigating themselves in familiar places, hence playing a significant role in integrating visuospatial imagery (Whittingstall et al. 2014; Epstein et al. 2017; Epstein 2008; Ino et al. 2002; O’Craven and Kanwisher 2000; Knauff et al. 2000; Rosenbaum et al. 2004). As such, damage to this brain area results in wayfinding and orientation issues, whereas lesions confined to the PPA manifest with difficulties in encoding the topographical structure and layout of newly introduced landscapes (Aguirre and D’Esposito 1999; Barrash et al. 2000; Pallis 1955; Hecaen et al. 1980; Takahashi et al. 1997; Ino et al. 2007; Katayama et al. 1999; Bottini et al. 1990; Osawa et al. 2008; Ploner et al. 2000).

Whether these two distinct but functionally complementary cortical regions are directly connected through a subcortical network or act alternatively as key hubs that promote an indirect connectivity between distant cerebral areas, has been a topic of recent research (Whittingstall et al. 2014; Epstein et al. 2017; Beyh et al. 2017; Powell et al. 2004; Malykhin et al. 2008; Catani et al. 2003). To this end, the implementation of methodological advances in diffusion MRI (dMRI), such as the high angular resolution diffusion imaging (HARDI) protocol (Tournier et al. 2012; Descoteaux and Poupon 2012; Seunarine and Alexander 2014), along with the use of fMRI–dMRI techniques (Gong et al. 2009; Hagmann et al. 2007; Honey et al. 2007; Johansen-Berg and Behrens 2006) have offered valuable insights into the structural connectivity of the neural correlates of spatial navigation and visuospatial imagery. Indeed, growing body of evidence points towards a common white matter pathway that directly links the cortical areas of interest i.e. the PPA and RSC (Whittingstall et al. 2014; Beyh et al. 2017; Kravitz et al. 2011). This fiber tract, allegedly conveying the core cognitive ability of spatial navigation and visuospatial imagery, has been described in a recent tractographic study and has been alternatively named as the medial occipital longitudinal tract (MOLT) (Beyh et al. 2017). Here, we provide anatomic and imaging data on the morphology and axonal connectivity of the SRF through microanatomic dissections and DTI tractography thus raising awareness on the structural architecture of this recently identified WM tract believed to participate in the neural circuit of spatial navigation and visuospatial imagery.

Study limitations

Magnetic resonance (MR) diffusion tractography is a sophisticated, in vivo, fast, non-invasive but indirect method to map the cerebral white matter architecture, based on the diffusion properties of water molecules along the direction of nerve axons. Anatomic data deriving from this technique can significantly vary due to various differences in acquisition and post processing parameters used (Mori and Zhang 2006; Jones and Cercignani 2010; Jones et al. 2013b; Nimsky et al. 2016; Thomas et al. 2014). Additionally, as already highlighted in the “Discussion” section, DTI tractography is prone to inaccurate results and artifacts, having the tendency to create erroneous white matter tracts, particularly when applied to narrow and anatomically complex cerebral areas, due to the crossing, kissing and bending effects as well as to the partial volume effect generated near CSF filled cavities (Oouchi et al. 2007; Vos et al. 2011; Johansen-Berg and Rushworth 2009; Le Bihan et al. 2006; Fernandez-Miranda et al. 2012; Tournier et al. 2012). For these reasons, DTI results and especially those regarding novel fiber tracts have to be fully validated through other more robust methods.

The Klingler’s technique in turn, is used for the blunt dissection of white matter tracts and entails the fixation of cadaveric specimens in a formalin solution followed by a freeze-thaw process, during which the formation of ice crystals separates the white matter fibers thus rendering their dissection feasible. This method -first described by Klingler (1935) has been recently revived after a long period of neglect, mainly due to the advent of MR tractography. The question of whether this technique of formalin fixation and freezing- defrosting procedure alters the integrity of myelin sheaths and consequently interferes with the axonal ultrastructure has been thoroughly answered by Zemmoura et al. in their recently published study (Zemmoura et al. 2016). The authors have proved that the Klingler’s preparation method maintains the white matter axonal integrity and structure, interfering only with the extracellular matrix, and therefore the results obtained are accurate. Hence, Klingler’s dissection is considered the gold standard technique for testing the validity of DTI studies.

Apart from being a very delicate, operator dependent, time consuming procedure the Klingler’s technique offers lower spatial resolutions in comparison to histology, polarized light imaging and optical coherence tomography, maintaining on the contrary the three-dimensional coherence of the white matter tracts (Dammers et al. 2010; Goergen et al. 2012; Magnain et al. 2014; Palm et al. 2010; Wang et al. 2011). Most significant however is the fact that by the Klingler’s method one is unable to simultaneously investigate tracts that exhibit fibers with crossing directions since the proper dissection of the one means the destruction of the other. Lastly, it remains ambiguous whether the cortical terminations of the white matter pathways can be adequately demonstrated through this technique, even though recent studies have reported so (Martino et al. 2013; Sarubbo et al. 2013).

Conclusion

Laboratory white matter dissections and DTI studies were employed with the aim to investigate and record the structural architecture of a fiber tract residing in the medial part of the occipital lobe known as the SRF. By combining the Klingler’s dissection method and a reliable DTI protocol with public available ROIs on dataset from the HCP yielding high intra- and inter-rater agreement values and providing descriptive and variability measures for DTI indices, our results support the hypothesis that the SRF is consistently involved in the axonal connectivity of cerebral areas that are believed to be strongly implicated in the cognitive ability of spatial navigation and visuospatial imagery.