Accepted Manuscript Title: Mapping the trajectory of the stria terminalis of the human limbic system using high spatial resolution diffusion tensor tractography Author: Arash Kamali David M. Yousem Doris D. Lin Haris I. Sair Siva P. Jasti Zafer Keser Roy F. Riascos Khader M. Hasan PII: DOI: Reference:
S0304-3940(15)30161-0 http://dx.doi.org/doi:10.1016/j.neulet.2015.09.035 NSL 31564
To appear in:
Received date: Revised date: Accepted date:
18-4-2015 16-9-2015 28-9-2015
Please cite this article as: Arash Kamali, David M.Yousem, Doris D.Lin, Haris I.Sair, Siva P.Jasti, Zafer Keser, Roy F.Riascos, Khader M.Hasan, Mapping the trajectory of the stria terminalis of the human limbic system using high spatial resolution diffusion tensor tractography, Neuroscience Letters http://dx.doi.org/10.1016/j.neulet.2015.09.035 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Mapping the trajectory of the stria terminalis of the human limbic system using high spatial resolution diffusion tensor tractography. Arash Kamali, MDa* [email protected]
, David M. Yousem, MDa, Doris D. Lin, MD, PhDa, Haris I. Sair, MDa, Siva P. Jasti, MDa, Zafer Keser, MDb, Roy F. Riascos, MDc, Khader M. Hasan, PhD.c a Department of Diagnostic Radiology, Division of Neuroradiology, Johns Hopkins University, Baltimore, MD, USA. b UTHealth Department of Physical Medicine Neuro Recovery Research Center c Department of Diagnostic Radiology. University of Texas at Houston, Texas, USA. * Corresponding Author: Department of Diagnostic Radiology, Division of Neuroradiology, Johns Hopkins University, 600 N. Wolfe St. Phipps B112B, Baltimore, MD 21287. Tel: 410 614 3146; Fax: 410 955 0962
• We present for the first time, detailed DTI based anatomic description of the stria terminalis connections. • Revealing the forniceal postcommissural fibers.
• Reviewing the advantages of our high resolution diffusion tensor tractography protocol.
Abstract The human limbic system is composed of gray and white matter structures which have been known to have a role in core processes such as motivation, memory, emotion, social behavior, self-awareness as well as certain primitive instincts. Multiple functional studies investigated some of these brain tasks in human brain limbic system. However, the underlying fine fiber pathways of the limbic system including the trajectory of the stria terminalis have not been delineated separately by prior diffusion weighted imaging. The ability to trace the underlying fiber anatomy noninvasively using diffusion tensor tractography (DTT) would be helpful to study the neurophysiology of these tracts in different functions in future functional studies. Few studies have focused on the stria terminalis using diffusion tensor tractography. Yet, the trajectory of the stria terminalis and some fine subtrajectories of the fornix have not been elucidated by prior DTT studies. We decided to further investigate these fine neuronal trajectory using tractography and high spatial resolution diffusion tensor imaging on 3T. Fifteen healthy right-handed men (age range 24–37 years) were studied. We delineated the detailed trajectories of the stria terminalis and fornix bilaterally in fifteen normal adult human brains. Using a high resolution DTT technique, we demonstrate for the first time, the trajectory of stria terminalis as well as detailed precommissural and postcommissural connectivity of the forniceal columns.
Abbreviations AC: anterior commissure DTI: diffusion tensor imaging
DTT: diffusion tensor tractography ROI: region of interest ST: stria terminalis.
Keywords: Amygdala; diffusion tensor imaging; fornix; high resolution; hypothalamus; hippocampus; limbic; septal nuclei; stria terminalis; tractography.
Introduction Ample evidence points to the fundamental role of the stria terminalis (ST) and fornix connections in central limbic circuitry. These tracts serve as major connectivity pathways between components of the limbic system. The fornices are white matter tracts lying medially beneath the corpus callosum and are the major white matter connecting one hippocampus to the other and on to the hypothalamus. The ST runs parallel to the fornix and is one of the three major efferent pathways of the amygdala including the amygdalofugal pathway, stria terminalis and anterior commissure. The ST follows a Cshaped course lateral to the fornix and descend through the caudothalamic groove terminating in multiple forebrain limbic nuclei including the hypothalamic, septal and the bed nucleus of stria terminalis [Nobach, 2005; Nolte, 2002]. Diffusion tensor tractography (DTT) is a powerful technique based on diffusionweighted magnetic resonance imaging and allows the noninvasive delineation of the trajectory and course of the neuronal fiber tract. DTT may provide information about the course, anatomical connectivity, possible disruption, or integrity of neural pathways. For instance, this information may be helpful to explore the specific connections that are impaired or to visualize the surgical anatomy of the mesial temporal lobe to assist in surgical planning of partial lobectomy to avoid damaging the important limbic structures.
Multiple attempts have been made to delineate the stria terminalis and fornix by prior diffusion weighted tractography studies [Concha et al, 2005; Kwon et al, 2011; Catani et al, 2008; Catani et al, 2013]. However, the trajectories of the ST and precommissural and postcommissural connections of the fornix have not been elucidated
in part because of the complex connections and directionality as well as tight proximity with adjacent fiber tracts. The ability to trace the underlying fiber trajectory of the limbic circuitry separately using DTI tractography might be helpful for future functional studies to further unravel the specific functionality of these important fiber tracts in different limbic functions as well as specific role of these fine fiber tracts in variant pathologies involving the limbic system including the neuropsychiatric and neurodegenerative diseases such as in anxiety disorder, bipolar disorder, obsessive-compulsive disorder, schizophrenia, Parkinson disease or Alzheimer’s disease. This work aimed to present the detailed trajectory and descriptive anatomy of the stria terminalis and fornix of normal human limbic system using a tractography approach and high resolution DTI data on 3T.
Materials and Methods Study Subjects This work was approved by our institutional review board (IRB) and was health insurance portability and accountability act (HIPAA) compliant. Fifteen right-handed healthy men (age range 24-37 years) were included in this study and written informed consent was obtained from all the subjects. Conventional MRI Data Acquisition All MRI studies were performed on a 3T Philips Intera scanner with a dual quasar gradient system with a maximum gradient amplitude of 80 mT/m, maximum slew rate 200 mT/ms/m, and an eight channel SENSE-compatible head coil (Philips Medical Systems, Best, Netherlands). The conventional MRI (cMRI) protocol included axially
prescribed 3D spoiled gradient (repetition time, TR = 8 ms; echo time, TE = 4ms; and flip angle = TR/TE/flip angle = 8 ms / 4ms /6o ), 3-D proton density-weighted (TR/TE/flip angle = 10,000 ms /10 ms/90o and 3-D T2-weighted (TR/TE/flip angle = 10,000 ms/60 ms /90o), with a square field-of-view (FOV) = 256 mm x 256 mm and a matrix of 256x256 pixels. The slice thickness for the MRI sequences was 1.0 mm with 120 contiguous axial slices covering the entire brain (foramen magnum to vertex). DTI Data Acquisition Diffusion-weighted image (DWI) data were acquired axially from the same graphically prescribed cMRI volumes using a single-shot multi-slice 2D spin-echo diffusion sensitized and fat-suppressed echo planar imaging (EPI) sequence, with the balanced Icosa21 tensor encoding scheme [Hasan et al, 2009]. The b-factor = 500 sec mm-2, TR/TE = 14460/60 msec. The spatial coverage for DTI data matched the 3D cMRI spatial coverage (FOV = 256 mm x 256 mm and slice thickness / gap/ #slices = 1 mm / 0 mm / 120). The EPI phase encoding used a SENSE k-space undersampling factor of two, with an effective k-space matrix of 112x112 and an image matrix after zero-filling of 256x256. The acquisition spatial resolution for DTI data was ~ 2.29mm x 2.29mm x 1mm, and the nominal resolution after image construction was 1mm x 1mm x 1mm. The number of b-factor ~ 0 (b0) magnitude image averages was four. The total DTI acquisition time was ~ seven minutes for the diffusion-weighted acquisition. Three acquisitions were repeated and averaged after co-registeration to obtain one data set for DTI data [Hasan et al, 2009].
White Matter Fiber Tracking WM fiber tracking was performed using DTI Studio software (Johns Hopkins University, Baltimore, Maryland; http://cmrm.med.jhmi.edu/). Fiber tracking was based on the fiber assignment by continuous tracking (FACT) algorithm with a fractional anisotropy (FA) threshold of 0.22 and angle threshold of 60 degrees. Each subject was examined to assess for potential structural abnormalities, and all subject brains were found to be normal. Reproducibility of the fiber construction in both hemispheres was tested by two experienced raters on all subjects. Two ROIs were applied to obtain each fiber tract and an “AND” operation was performed to include the fibers passing through both of the ROIs (Figure 1). Results Complete trajectory of the stria terminalis (black) and fornix (yellow) are demonstrated in figure 2. A common anatomical course of the ST and fornix are observed among the fifteen subjects bilaterally (Fig. 3-5). These pathways are identified bilaterally in all subjects. The ST connects the amygdala to the anterior hypothalamic as well as septal nuclei (Fig. 5b). The ST fibers (black fibers in Fig. 2-5) have more lateral origin than the fornices (shown in yellow color in fig. 2 and in red color in fig. 3-5). Unlike the fornix which runs superior and medial to the thalami just above the third ventricle, the ST runs more laterally between the medial margin of the caudate and superior margin of the thalamus underneath the body of the lateral ventricles. The ST then course inferiorly along the anterior margin of the thalamus and posterior aspect of the caudate head within
the caudothalamic groove (CT groove in Fig. 3d, 5c) and terminates in the region of mammillary bodies and septal nuclei. Our results demonstrate that the precommissural fibers of the forniceal columns course toward the septal nuclei and postcommissural fibers of the forniceal columns approach toward the mammillary bodies (Fig. 2b, d, e, and 5 a, b). In addition, we further reveal the dorsal hippocampal commissure which join together and form the body of the fornix (Fig. 2f) and finally the horizontal fibers (HF) arising from the distal precommissural fibers within the septal nuclei which course laterally toward the mesial temporal lobes (Fig. 2d, e). We further illustrate the sagittal, coronal and axial mappings of the fornix and stria terminalis trajectories (Fig. 3, 4 and 5). Stria terminalis: The stria terminalis (ST) (black fibers in Fig. 2-5) arises from the amygdala and follows a C-shaped course along the anterior surface of the temporal horns and atrium of the lateral ventricles. The ST runs lateral to the fimbria and crura of the fornix (black fibers in Fig. 3a-c, 4, 5). The ST extends cranially from the amygdala coursing laterally and more cranially in relation with the fornix along the superior margin of the temporal horn of the lateral ventricle and lateral margin of the thalamus (Fig. 3c). The ST ascends to the level of atrium of the lateral ventricle and turns frontally at the atrium along the superior margin of the thalamus and undersurface of the body of the lateral ventricle (Fig. 4c-d). At this point, the ST courses just medial and side by side with the body of the superior fronto-occipital fasciculus as they both run anteriorly (red fibers in Fig. 2f). The ST then follows the caudothalamic groove between the medial edge of the caudate and superior margin of the thalamus (Fig. 3d, 5c) marking a line of separation between the
thalamus and the caudate nucleus. The ST then reaches the region of the anterior commissure where it branches out to multiple septal and hypothalamic nuclei (Fig. 5b). Fornix The fornix originates from the hippocampal formation on each hemisphere curving through the fimbria. The course of the fornix is shown traversing through the amygdala and ascending cranially (Fig. 2, 3-5) along the anterior and superior margin of the temporal horns of the lateral ventricles (Fig. 3a-c, 4, 5). Bilateral crura arch anteriorly at the ventricular atrium under the splenium of the corpus collosum (Fig. 4c-d) and join to form the body of the fornix. Just before the body of fornix forms, some crura fibers project medially, crossing the midline, and form the hippocampal commissure (white solid arrow in Fig. 2f). The body of the fornix continues this curve course, arching anteriorly along the undersurface of the corpus collosum and medial aspect of the body of lateral ventricle in the midline (Fig. 4 a). At the interventricular foramen, the fornix turns inferiorly and posteriorly toward the hypothalamus and divides into the right and left columns of the fornix (Fig. 5b). The forniceal columns descend encasing the foramen of Monro and end in the region of mammillary bodies (Fig. 3d-e). Forniceal columns are also divided into the anterior or precommissural fibers and posterior or postcommissural branches around the anterior commissure (AC, red fibers in Fig. 2d, e) (Fig. 5b). The precommissural fibers course ventral to the anterior commissure descending anterior to the lamina terminalis and end in the region of septal nuclei (Fig. 2d, e). The postcommissural fibers of the fornix descend dorsal to the anterior commissure entering the hypothalamus (Fig. 2d, e). The postcommissural branch courses inferiorly through the hypothalamic nuclei
toward the mammillary bodies (Fig. 3d-e). Some of the fibers of the precommissural subdivisions continue horizontally from the septal nuclei parallel and anterior to the anterior commissure (horizontal fibers, HF in Fig. 2d, e) and course laterally toward the uncus of the medial temporal lobes which give the appearance of mustache to the forniceal columns (Fig. 2d, e). Discussion To the best of our knowledge, the current study is the first to reconstruct and present the accurate DTT reconstruction of the ST using high spatial resolution DTI data. The detailed anatomy of the fornix, for example, the DTT reconstruction of the precommissural and postcommissural trajectories of the forniceal columns has not been demonstrated before. To our knowledge, only one prior study claimed to demonstrate the tractography of the ST [Kwon et al, 2011]. However by observing the course of the fibers, it is obvious that their 3D reconstruction of the fornix and ST is anatomically incorrect since the ST is supposed to be located lateral to the fornix; however, their 3D reconstruction represents medial position of the ST in relation to the fornix. Our tractography results are in line with the known anatomy of the stria terminalis, fornix and their subtrajectories from prior animal studies [De Olmos et al, 1972; Conteras, et al, 1992; Lanuzaet al, 1997] and atlas of human neuroanatomy [Nobach, 2005; Nolte, 2002]. Our results confirm that the ST is a long white matter structure in the brain consisting of a band of fibers running along and parallel to the fornix serving as the largest major output pathway of the amygdala. We present for the first time the exact regions of interest to delineate the trajectory of the ST from adjacent fiber tracts. The ST has been shown by animal studies to be involved in social behaviors, memory, sex, motivational circuits,
reaction to stressful challenges, pain and affective defense [Brutus et al, 1986; Del Abril et al, 1987; Cai 1990; Wood et al, 2005; Ulrich et al, 2006].
We also demonstrate for the first time 3D DTT reconstruction of precommissural and postcommissural fibers of the forniceal column. Our results show that the precommissural fibers reach the septal area and postcommissural fibers coursing toward the mammillary bodies. Septal nuclei are a set of gray matter structure that lies below the rostrum of the corpus callosum, anterior to the lamina terminalis. Septal nuclei have extensive reciprocal connections with the hippocampus via the fornix, lamina terminalis and amygdalofugal tracts [Nobach, 2005; Nolte, 2002]. We also demonstrate the hippocampal/fornix commissure which is cross connection of the crura fibers with the other side as they turn and form the body of the fornix. The DTI trajectory of the horizontal extension of the precommissural fibers is also demonstrated for the first time which runs parallel to the anterior commissure toward the uncus of the medial temporal lobe. These horizontal connections may have a role in transmitting the olfactory information to the septal nuclei and from there to the other parts of the limbic system.
Diffusion tensor tractography of white matter connections in highly crossing area for example the corona radiata at the level of the lateral ventricles is challenged by the signal-to-noise ratio (SNR) due to overestimation of anisotropy at low SNR and partial volume averaging upon using large voxel volumes. These delicate limbic structures have not been elucidated on prior DTI studies primarily because of inadequate spatial resolution and excessive partial volume averaging from close neighboring fiber tracts. This is attributable
to mixing of fibers in different orientations within the voxel leading to fiber kissing and switching (false positive result) or fiber termination (false negative result) by the tractography algorithm. This issue was remedied to some extent by using thinner slices and smaller voxel volume. In our experiment the high resolution DTI method increased the detectable anisotropy within the gray matter structures and helped to trace the thin white matter fibers passing through the hippocampus and hypothalamus nuclei [Kamali et al 2010]. For instance, we demonstrated, for the first time, the prefronto-caudate projections of the frontostriatal pathways coursing through the caudate nuclei [Kamali et al 2010] using the same high resolution tractography technique [Hasan et al 2009]. A given DTI-based trajectory may not correspond exactly to the fine distribution of actual axonal bundles due to crossing and kissing fibers that may cause false positive or false negative results [Jones et al, 2008; Kamali and Hasan, 2010]. However, our anatomical observations were consistent across all subjects. In our experience, the main challenge on the way of tracking the ST was marked crossing with the major anteriorposteriorly oriented fibers running adjacent to the body of the lateral ventricles such as the superior fronto-occipital fasciculus (red fibers in Fig. 2f). This crossing results in fiber switch due to kissing and crossing fiber phenomenon and false positive results [Kamali and Hasan, 2014]. To avoid this confusion we also traced the superior frontooccipital fasciculus on each subject to avoid contamination with the ST and fornix as we illustrated on one of the subjects to better orient the readers to the close proximity of these white matter bundles (red fiber in Fig. 2f) which may result in fiber switch (false positive result) or abrupt termination of the tracking algorithm (false negative result).
Delineation of the limbic forebrain nuclei such as the septal and hypothalamic nuclei and bed nucleus of stria terminalis is beyond the resolution of the current study. The complexity of the fibers within a voxel remains to be a limitation to diffusion tensor tractography model. Future studies using high magnetic field MRI and higher angular diffusion technique such as diffusion spectrum imaging [Wedeen et al, 2008] with focusing on the limbic system may further validate our results and overcome some of the challenges due to crossing fibers and multidirectional fiber orientation within a voxel. Conclusions In conclusion, our results complement prior diffusion weighted tractographic studies of the limbic system. In this report we demonstrate the detailed DTI trajectory of the stria terminalis and fornix using a high spatial resolution diffusion tensor tractography technique on 3T. Acknowledgements Acknowledgments—This work is funded by the American National Institutes of Health (NIH)-Institute for Neurological Diseases and Stroke (NIH-NINDS: R01-NS052505-04) and the Dunn Research Foundation.
References 1. Brutus M, Shaikh MB, Edinger H, Siegel A (1986) Effects of experimental temporal lobe seizures upon hypothalamically elicited aggressive behavior in the cat. Brain Res. 366:53-63. 2. Cai ZJ (1990) The neural mechanism of declarative memory consolidation and retrieval: a hypothesis. Neurosci Biobehav Rev. 14:295-304. 3. Canteras NS, Swanson LW (1992) Projections of the ventral subiculum to the amygdala, septum, and hypothalamus: a PHAL anterograde tract-tracing study in the rat. J Comp Neurol. 324:180-94. 4. Catani M, Thiebaut de Schotten M (2008) A diffusion tensor imaging tractography atlas for virtual in vivo dissections. Cortex. 44:1105-32. 5. Catani M, Dell'acqua F, Thiebaut de Schotten M (2013) A revised limbic system model for memory, emotion and behaviour. Neurosci Biobehav Rev. 37:1724-37. 6. Concha L, Gross DW, Beaulieu C (2005) Diffusion tensor tractography of the limbic system. AJNR Am J Neuroradiol 26:2267–2274. 7. Del Abril A, Segovia S, Guillamón A (1987) The bed nucleus of the stria terminalis in the rat: regional sex differences controlled by gonadal steroids early after birth. Brain Res. 429:295-300. 8. De Olmos JS, Ingram WR (1972) The projection field of the stria terminalis in the rat brain. An experimental study. J Comp Neurol. 1972 146:303-34. 9. Hasan KM, Kamali A, Kramer LA (2009) Mapping the human brain white matter tracts relative to cortical and deep gray matter using diffusion tensor imaging at high spatial resolution. Magn Reson Imaging. 27:631-636.
10. Jones DK (2008) Studying connections in the living human brain with diffusion MRI. Cortex. 44:936-952. 11. Kamali A, Kramer LA, Hasan KM (2010) Feasibility of prefronto-caudate pathway tractography using high resolution diffusion tensor tractography data at 3T. J Neurosci Methods. 191:249-254.
12. Kwon HG, Byun WM, Ahn SH, Son SM, Jang SH (2011) The anatomical characteristics of the stria terminalis in the human brain: a diffusion tensor tractography study. Neurosci Lett. 500:99-102.
13. Lanuza E, Font C, Martínez-Marcos A, Martínez-García F (1997) Amygdalohypothalamic projections in the lizard Podarcis hispanica: a combined anterograde and retrograde tracing study. J Comp Neurol. 384:537-55. 14. Noback CR, Strominger NL, Demarest RJ, Ruggiero DA (2005) The human nervous system. Structure and function, 6th edn. New Jersey: Humana Press. 15. Nolte J (2002) The human brain. An introduction to its functional anatomy, 5th edn. Mosby, Missouri. 16. Ulrich-Lai YM, Xie W, Meij JT, Dolgas CM, Yu L, Herman JP (2006) Limbic and HPA axis function in an animal model of chronic neuropathic pain. Physiol Behav. 88:67-76.
17. Wedeen VJ, Wang RP, Schmahmann JD, Benner T, Tseng WY, Dai G, Pandya DN, Hagmann P, D’Arceuil H, de Crespigny AJ (2008) Diffusion spectrum magnetic resonance imaging (DSI) tractography of crossing fibers. Neuroimage 41:1267–1277. 18. Wood RI, Swann JM (2005) The bed nucleus of the stria terminalis in the Syrian hamster: subnuclei and connections of the posterior division. Neuroscience. 135:155-79.
Figure Captions Figure 1. The location of the regions of interest for the fornix (A and B) and the stria terminalis (ST) (C and D) on the DTI color-coded maps. The first ROI for the fornix is drawn on the axial plane at the middle of the fornix on the midline sagittal cut. The ROI2 (B) is placed on the fiber generated on the coronal plane through the mid fornix on the midline sagittal plane. The ROI1 (C) for the ST is drawn over the caudothalamic groove on the coronal plane through the crossing the superior cerebellar peduncles or the coronal plane passing through the posterior most of the fornix on the mid sagittal plane. The ROI2 (D) is placed over the fibers generated inferolateral to the thalamus on the same coronal plane.
Figure 2. Three-dimensional reconstructions of the fornix (yellow in 1a, b, d, e and f) and ST (black). The ST (black fibers) is tracking cranially from the amygdala coursing laterally in relation with the fornix (yellow). Also illustrated are the precommissural (preC) and post commissural (posC) subdivisions of the forniceal columns (d, e). The horizontal fiber extension (HF in d, e) of the preC fibers are also shown running parallel and anterior to the AC (AC in red in d and e) and course laterally toward the uncus of the medial temporal lobes. Posterior-superior view of the fornix is showing the forniceal commissure (arrow in 1f) and close relationship of the superior fronto-occipital fasciculus (SFO in red) with the fornix and ST.
Figure 3. Six successive T1 weighted sagittal planes demonstrating the relationship of the fornix (red), and ST (black) as they course parallel to each other. The ST (black
fibers) is coursing cranially from the amygdala and laterally in relation with the fornix (red in a-c). As it’s visible in Fig. 3f, the descending portion of the ST courses dorsal to the caudate head and ventral to the thalamus and makes a line of separation between caudate head and the thalamus along the caudothalamic groove (CT grove in d) as it continues toward the hypothalamic nuclei (d, e). Septal nuclei (SN) and the mammillary bodies (MB) are shown by yellow arrows in Fig. 3e.
Figure 4. The courses of the fornix (red) and ST (black) are shown on four consecutive coronal sections from the anterior (4a) to most posterior (4d). The ST is seen to ascend alongside the fornix (4a). The fornix (red) is seen coursing medially toward the mammillary bodies (MB in 4b). The course of the fornix (red) and ST along the anterior and superior margin of the temporal horns of the lateral ventricles (4c and d), turning anteriorly at the ventricular atria (4d) continuing anteriorly along the undersurface of the corpus collosum.
Figure 5. Four consecutive axial T1 weighted images are shown from inferior (5a) to the superior (5d) for illustration of the trajectory of the fornix and ST tracts (black). Precommissural connections of the forniceal columns to the septal nuclei are demonstrated by yellow arrow in Fig. 5a. Postcommissural fibers of the forniceal columns are shown coursing toward the mammillary bodies in Fig. 5b. The ST (black) courses parallel and lateral to the fornix fibers (red).
D 2 Figure 1
AC HF PosC
b ST Fx MB
d ST Fx
c CT groove