Variation in the Anatomy of the Normal Human Optic Chiasm: An MRI Study

Original Contribution Section Editors: Clare Fraser, MD Susan Mollan, MD Variation in the Anatomy of the Normal Human Optic Chiasm: An MRI Study Nicholas S. I. Bosler, BBiomed, BSc, DipL, David Ashton, MBBS, FANZCR, Andrew J. Neely, BEng, MEngSc, PhD, FRAeS, Christian J. Lueck, MB, BChir, MA, PhD, FRACP FRCP(UK), FAAN Background: Compression of the optic chiasm typically leads to bitemporal hemianopia. This implies that decussating nasal fibers are selectively affected, but the precise mechanism is unclear. Stress on nasal fibers has been investigated using finite element modeling but requires accurate anatomical data to generate a meaningful output. The precise shape of the chiasm is unclear: A recent photomicrographic study suggested that nasal fibers decussate paracentrally and run parallel to each other in the central arm of an “H.” This study aimed to determine the population variation in chiasmal shape to inform future models. Methods: Sequential MRI scans of 68 healthy individuals were selected. 2D images of each chiasm were created and analyzed to determine the angle of elevation of the chiasm, the width of the chiasm, and the offset between the points of intersection of lines drawn down the centers of the optic nerves and contralateral optic tracts. Results: The mean width of the chiasm was 12.0 ± 1.5 mm (SD), and the mean offset was 4.7 ± 1.4 mm generating a mean offset:width ratio of 0.38 ± 0.09. No chiasm had an offset of zero. The mean incident angle of optic nerves was 56 ± 7°, and for optic tracts, it was 51 ± 7°. Conclusions: The human optic chiasm is “H” shaped, not “X” shaped. The findings are consistent with nasal fibers decussating an average of 2.4 mm lateral to the midline before travelling in parallel across the midline. This information will inform future models of chiasmal compression. Journal of Neuro-Ophthalmology 2021;41:194–199 doi: 10.1097/WNO.0000000000000907 © 2020 by North American Neuro-Ophthalmology Society Australian National University Medical School (NSIB, DA, CJL), Canberra, Australia; Departments of Neurology (NSIB, CJL) and Radiology (DA), The Canberra Hospital, Canberra, Australia; and School of Engineering and Information Technology (AJN), University of New South Wales, Canberra, Australia. The authors report no conflicts of interest. Address correspondence to Nicholas S. I. Bosler, BBiomed, BSc, DipL, Department of Neurology, the Canberra Hospital, PO Box 11, Woden 2606, Australia; E-mail: [e-mail] 194 T he optic chiasm is traditionally described as an X-shaped structure in the visual pathway in which there is partial decussation of the optic nerve fibers. The proportion of decussating fibers varies between species (1), but in humans, the fibers originating from the nasal half of the retina of each optic nerve decussate and enter the contralateral optic tract, whereas the temporal fibers remain ipsilateral and uncrossed. This is clinically important as lesions of the optic chiasm typically lead to bitemporal visual field defects (2), implying that the nasal fibers have been selectively damaged. There have been several theories as to why the nasal fibers are more affected than the nondecussated, temporal, fibers. These include compression of the arteries supplying the center of the chiasm leading to ischemia (3), mechanical stretching, particularly of the fibers related to the upper quadrants of vision (4), and increased pressure at the center of the chiasm in comparison to the periphery (5). One other theory suggests that it is simply the fact that crossing fibers cross each other that renders them selectively vulnerable. Crossing fibers have a decreased surface area of contact between adjacent neurons, and this potentially results in increased stress arising from compressive forces (6). Computer simulation using finite element modeling (FEM) is supportive of this theory (7–9). However, FEM requires accurate anatomical data to generate a meaningful output. In particular, it requires precise information about the anatomy of fiber decussation. To date, most anatomical studies have either focused on general descriptions of the chiasm such as its width and/or area (10) or its location relative to the pituitary gland (11–13), rather than on the internal arrangement of the decussating fibers. A recent cadaver study using photomicroscopy (14) found that fibers actually decussated in the paracentral regions, not in the center of the chiasm where the crossing fibers actually ran parallel to each other in a mediolateral direction. This was at variance with the traditional view that Bosler et al: J Neuro-Ophthalmol 2021; 41: 194-199 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution fibers decussate in the center of the chiasm and suggested that the chiasm should be thought of as resembling an “H” rather than an “X.” This is particularly important for FEM as this would change the force distribution within the chiasm. Two recent studies investigating fiber orientation using diffusion tensor imaging (DTI) observed “H” shaped chiasms (15,16), whereas one study used 3D-polarized light imaging (3D-PLI) and found chiasms resembling a more traditional “X” shape (17). This raises the possibility of heterogeneity between individuals. Unfortunately, these studies were not designed to examine interindividual variation in chiasmal shape, something which is essential for accurate modeling. The aim of this study was to determine the variation in normal optic chiasm anatomy on MRI scans of individuals with otherwise normal visual pathways, with a specific focus on determining how many chiasms resembled an “H” and how many an “X.” METHODS MRI Scans The study was approved by the ACT Health Human Research Ethics Committee (ETHLR.17.122). Deidentified MRI scans were obtained from the Department of Radiology at the Canberra Hospital from sequential scans that had been performed as part of routine clinical practice. Scans were obtained using a 1.5 T Siemens Magnetom Avanto, with a slice thickness of 1 or 5 mm, depending on the clinical indication for the scan, and a spatial resolution of 1 mm2. The sequence with the highest resolution at the chiasm (T1, T2, or 3D FLAIR) was used for further analysis. Scans were excluded if they had been reported as having an abnormality, if the request form suggested that the patient had experienced any visual symptoms, or if the scan contained artifacts due to large slice thickness or unclear margins of greater than 1.5 mm (i.e., more than 10% of the expected average width of the chiasm (18)) covering 2 or more of the optic nerves and tracts. If only one optic nerve/tract was affected, the images were used for partial analysis (see below). Patients’ sex and age were provided anonymously. The scans were analyzed off-line using OsiriX (OsiriX 9.5; Pixmeo SARL, Bernex, Sweden) to obtain the best possible 2D planar image, defined as the plane which included the maximum area of nerves, chiasm and tracts and, ideally, had the clearest margins of these structures (Fig. 1). The angle of this chiasmal plane was determined with reference to a line joining the anterior and posterior commissures (the AC-PC line) in the sagittal plane (Fig. 2) using the Schaltenbrand method (19). This method was chosen as the AC-PC line is a standard reference plane used in MRI studies (19). Image Analysis Once the 2D images were extracted, chiasms were further analyzed using AutoCAD software (AutoCAD 2018; Bosler et al: J Neuro-Ophthalmol 2021; 41: 194-199 FIG. 1. MRI image of optic chiasm, nerves, and tracts. The plane of the optic chiasm was defined as the plane that included the maximum area of nerves, chiasm, and tracts. Autodesk, Mill Valley, CA). The steps were as follows (Figs. 3 and 4): 1. Determination of the “minimum width of the chiasm line” (MWCL). This was defined as the shortest distance between any 2 points on opposite sides of the chiasm and was found using an iterative process which involved generating circles centered on a point on one side of the chiasm with the edge of the circle only just touching the other side of the chiasm, thus making the radius of the circle equivalent to the width of the chiasm. This process was repeated at least 5 times until it was clear that the shortest possible line segment between the 2 sides of the chiasm had been found. 2. Construction of centerlines for both optic nerves and both optic tracts. This was achieved using the centerline function of AutoCAD. Because of the possibility of curvature of these structures with increasing distance from the chiasm, the relevant segments were arbitrarily limited to a distance of one chiasm width, as measured from the center of the MWCL segment (see above). The FIG. 2. Sagittal MRI showing the angle of the plane of the optic chiasm (BC) relative to the AC-PC line (AC). The plane of the optic chiasm was defined in Fig. 1. The AC-PC line was found using the Schaltenbrand method (19). The angle of inclination, being the angle between these 2 lines, is ^ represented by ACB. 195 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution 4. 5. FIG. 3. MR image of optic chiasm. centerlines represented the average distance between 2 lines plotted at the edges of each nerve/tract. Averaging in this way reduced the potential error which might have arisen from partial voluming artifacts at the edges of the nerves and tracts. 3. Determination of intersection offset. The points at which ipsilateral optic nerve and optic tract centerlines intersected were found, and a line was drawn between the points of intersection. This line was labeled the “offset,” and its length was measured. The offset was considered the determining measure distinguishing “H” from “X” shaped chiasms. An “X” would have an offset 6. 7. 8. of zero, and an increase in offset would increase the width of the horizontal bar of the “H.” Determination of the “offset projection.” On occasion, the points of intersection did not lie on the MCWL segment. To explore the effect of this and to cater for possible error generated by artifacts such as partial voluming, the offset was projected back onto the MWCL segment and the resultant line segment was labeled the “offset projection;” its length was measured. Determination of proportional length of the “H” crossbar. The offset projection was divided by the MWCL. This provided an estimate of the length that crossing fibers had to travel in parallel at the center of the chiasm, relative to the width of the chiasm. Determination of “single optic nerve and tract offset” (SONTO). To accommodate those scans in which an optic nerve/tract was not clearly visible, the offset was calculated using an alternative measurement. This involved determining the intersections between the MWCL and an optic nerve, and the MWCL and the contralateral optic tract. The SONTO was defined as the distance between these points. This methodology was performed for all scans to allow for comparison with the “offset projection.” Determination of relevant angles. The shape of the chiasm was further defined by measuring the angles subtended by the optic nerves and tracts and the MWCL segment. All angle measurements used the Measuregeom function in AutoCAD. All measurements were tested for normal distribution using the Shapiro–Wilk test and were also regressed against patient age and sex. RESULTS FIG. 4. Associated analysis of MRI of optic chiasm (Fig. 3). The width of the chiasm is represented by the AB line segment, with the I circle (dashed) representing the circle(s) used to find points on opposite side of chiasm. A second circle (J) centered at the centerpoint of the width line segment with radius equivalent to the chiasmal width was created. All line segments (dashed white) corresponding to the optic nerve and tract edges were limited by this circle. Centerlines of the optic nerves (EC and FD) and tracts (GC and HD) were derived from these edges, and the intersections between ipsilateral optic nerve and tract pairs were found (C and D). The offset (CD) was found using these intersections. See methods section for full description of methodology. 196 A total of 68 MRI scans were obtained. Seven scans had to be excluded because the chiasm could not be clearly identified on any sequence (one scan had a visible abnormality at the chiasm, 3 had inadequate resolution to allow accurate measurement, and 3 had artifacts due to partial voluming). Of the 61 remaining scans, 4 had artifacts involving a single nerve or tract, and another 7 had unclear margins of a single nerve or tract. These scans were only used for SONTO analysis. This meant that of the 61 scans, 50 were suitable for full analysis and 11 could only be used for SONTO analysis. In addition, 2 scans had insufficient resolution in the sagittal plane to allow determination of the inclination angle, and information about sex was not available for 2 scans. In total, the sequence offering the best definition was T1 in 39 scans, T2 in 18, and 3D FLAIR in 4. The mean age of the 61 participants was 50.3 ± 15.8 (range 20–78 years), and the sex distribution was equal. Details of the various measurements are provided in Table 1 and 2. Bosler et al: J Neuro-Ophthalmol 2021; 41: 194-199 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution TABLE 1. Chiasm measurements Parameter Sex Age (years) MWCL (mm) Offset (mm) Offset/MWCL ratio Angle between offset and MWCL (degrees) Offset projection (mm) Offset projection/width ratio SONTO offset (mm) Sample Number Mean Value (95% Confidence Interval) SD Coefficient of Minimum Maximum Normality Variation Value Value (Shapiro–Wilk Sig.) 59 61 61 50 50 50 M = 29, F = 30 50.2 (46.3–54.2) 12.0 (11.6–12.4) 4.7 (4.3–5.1) 0.38 (0.36–0.41) 7.9° (6.1–9.7°) N/A 15.8 1.5 1.4 0.87 6.4° N/A 0.31 0.13 0.30 2.29 0.81 N/A 20 8.9 1.6 0.14 0.0° N/A 78 14.8 8.6 0.59 29.0° N/A 0.030 0.290 0.697 0.521 ,0.001 50 50 61 4.6 (4.2–5.0) 0.38 (0.35–0.40) 4.3 (4.0–4.7) 1.5 0.09 1.5 0.33 0.24 0.35 1.6 0.14 1.5 8.6 0.59 8.5 0.679 0.563 0.533 P . 0.05 was used to determine normality in the Shapiro–Wilk test. MWCL, minimum width of chiasm line; SONTO, single optic nerve or tract offset. The mean width of the optic chiasm was 12.0 mm (95% confidence interval: 11.6–12.4 mm) ± 1.5 mm (SD), range 8.9–14.8 mm. The data were normally distributed (Shapiro–Wilk test for normality, P = 0.29). Measurements were slightly larger on T2 than T1 (13.2 mm vs 11.5 mm, respectively), and this was statistically significant (P , 0.001, t test). The mean offset was 4.7 mm (95% CI: 4.3–5.1 mm) ± 1.4 mm, range 1.6–8.6 mm, and was normally distributed (P = 0.70). No chiasm had an offset of 0, that is, a pure “X” shape. Offset was linearly correlated with chiasmal width (Pearson correlation, r = 0.736, P , 0.01). The mean offset-to-width ratio was 0.38 (95% CI: 0.36–0.41). The mean value of the offset projection was 4.6 mm, slightly smaller than the mean offset. The mean SONTO was 4.3 mm, slightly, but significantly, smaller than both the mean offset and the mean offset projection (P , 0.001, paired t test). The mean inclination of the chiasm relative to the ACPC line was 25.1° (95% CI: 23.3°–26.9°) ± 7°. The mean incident angle of the optic nerves was slightly larger than the optic tracts (55°/56° and 51°, respectively). There was no significant correlation between age and any of the chiasm metrics (Pearson correlation, jrj,0.3, and P . 0.05 for all tests). CONCLUSIONS Using MRI scans to study optic chiasms has allowed assessment of multiple individuals to investigate the normal variation in human anatomy. We found that the mean width of the optic chiasm was 12.0 mm with a mean inclination angle of 25.1° relative to the AC-PC line. Interestingly, T2 returned slightly, but significantly, wider chiasms than T1. Importantly, no chiasm was “X”-shaped. The mean length of the crossbar of the “H” was 4.7 mm, which would be consistent with paracentral crossings of nasal fibers at a mean distance of approximately 2.4 mm lateral to the center of the chiasm. This distance was positively correlated with increased chiasmal width, the mean offset-to-width ratio being 0.38. There was no correlation between sex or age and any of the chiasm metrics. To the best of the authors’ knowledge, this is the first study that has looked at inclination angle using the AC-PC TABLE 2. Chiasm angle measurements Parameter Inclination angle (°) Incoming left optic nerve angle (°) Incoming right optic nerve angle (°) Outgoing left optic tract angle (°) Outgoing right optic tract angle (°) Sample Number Mean Value (95% Confidence Interval) 59 50 25.1° (23.3–27.0°) 56° (53–58°) SD Coefficient of Variation Minimum Value Maximum Value Normality (Shapiro– Wilk Sig.) 7.0° 8° 0.28 0.14 12.1° 39° 41.1° 73° 0.068 0.922 50 55° (53–57°) 7° 0.13 41° 70° 0.116 50 51° (49–53°) 7° 0.14 38° 74° 0.046 50 51° (49–53°) 6° 0.12 34° 66° 0.822 P . 0.05 was used to determine normality in the Shapiro–Wilk test. Bosler et al: J Neuro-Ophthalmol 2021; 41: 194-199 197 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution line as a reference. Other studies have typically compared chiasmal inclination to the pituitary gland itself (12). The chiasmal widths measured in this study were slightly smaller than those of other MRI studies which ranged from 12.9 to 15.0 mm (10,18,20). It is possible that this relates to postmortem changes or fixation artifacts, as well as the choice of MRI sequence. One study found that chiasmal width decreased in older patients (10), but this was not observed in this study. Several studies have observed paracentral nasal fibers decussation (14–16). However, these studies did not quantify how far away from the midline the decussation occurred. Historically, nasal fibers have been considered to cross the chiasm at its center, thereby forming a traditional “X” shape (21). The fact that all chiasms found in this study were “H” shaped runs contrary to this notion. These findings have important implications for computer modeling of the chiasm, as previous FEM studies have assumed central nasal fiber crossings (7,8). The findings of this study will inform further investigation into the cause of bitemporal hemianopia, including the “crossing-theory” (6). It is also helpful to have information regarding the incident angles of optic nerves and tracts, as well as the angle of decussation (FEM studies have assumed perpendicular fiber crossing (7,8), but the true angle is likely nearer to 106°). There were several limitations to this study. First, the number of subjects was relatively small. Second, the scan thickness of 1–5 mm meant that considerable partial voluming occurred, generating difficulty in precise identification of some of the edges of the nerves and tracts. The resultant artifacts led to the exclusion of roughly 25% of scans from full analysis. Third, this study only examined the chiasm in a single plane, thereby limiting full assessment of chiasms which might have been curved. Importantly, this study only investigated the macroscopic envelope of the optic chiasm. Several studies have suggested variation in nerve fiber size and orientation at a microscopic level (22,23). The DTI study by Sarlls and Pierpaoli (15) suggested that some nerve fibers travelled in a superior-inferior direction, consistent with previous ferret studies (22). The latter study also observed that different sizes of axons crossed the chiasm at different locations: large diameter fibers more rostrally, small fibers at the midpoint, and medium-sized fibers more caudally (22). Our study did not address these variations in nerve fiber anatomy which would add further complexities to computer modeling. Finally, our study did not look at the precise location of the pituitary gland and the diaphragma sellae in relation to the chiasm. Most pituitary adenomas compress the chiasm after expanding upward through the foramen in the diaphragma sellae, and the location of this foramen in relation to the chiasm can be variable (24). In more than 80% of cases, it is located below the center of the chiasm (24). Variation in the location of the foramen might impact on the anatomy of chiasmal compression and hence which nerve fibers are affected. 198 A future, dedicated, study looking at increased numbers of subjects is warranted, incorporating high-resolution, thin-sliced MRI images, and investigating other relationships such as the anteroposterior relation of the chiasm to the pituitary. It would also be of interest to study how medical conditions affect the anatomy of the chiasm. For example, albinism can lead to decreased chiasmal width (18). In summary, this study found that the human optic chiasm has a mean width of 12.0 ± 1.5 mm and an angle of inclination of 25.1° ± 7° relative to the AC-PC line. Importantly, the human optic chiasm seems to be “H”-shaped and not “X”-shaped, consistent with previous findings that nasal retinal fibers decussate paracentrally rather than centrally in the chiasm. Our study is consistent with fiber-crossings being located approximately 2.4 mm lateral to the midline and occurring at an angle of approximately 110°. These findings will be used to inform future models of chiasmal compression and so add to our understanding of the underlying mechanism by which bitemporal hemianopia is generated. STATEMENT OF AUTHORSHIP Category 1: a. Conception and design: Nicholas S. I. Bosler, David Ashton, Andrew J. Neely, and Christian J. Lueck; b. Acquisition of data: Nicholas S. I. Bosler and David Ashton; c. Analysis and interpretation of data: Nicholas S. I. Bosler and Christian J. Lueck. Category 2: a. Drafting the manuscript: Nicholas S. I. Bosler and Christian J. Lueck; b. Revising it for intellectual content: Nicholas S. I. Bosler, David Ashton, Andrew J. Neely, and Christian J. Lueck. Category 3: a. Final approval of the completed manuscript: Nicholas S.I. Bosler, David Ashton, Andrew J. Neely, and Christian J. Lueck. ACKNOWLEDGMENTS The authors thank Dr. Jouke de Baar and Alexander D’Souza for helpful comments on the methodology. REFERENCES 1. Glaser JS. Romancing the chiasm: vision, vocalization, and virtuosity. J Neuroophthalmol. 2008;28:131–143. 2. Lee IH, Miller NR, Zan E, Tavares F, Blitz AM, Sung H, Yousem DM, Boland MV. Visual defects in patients with pituitary adenomas: the myth of bitemporal hemianopsia. AJR Am J Roentgenol. 2015;205:W512–W518. 3. Bergland R. The arterial supply of the human optic chiasm. J Neurosurg. 1969;31:327–334. 4. Hedges TR. Preservation of the upper nasal field in the chiasmal syndrome: an anatomic explanation. Trans Am Ophthalmol Soc. 1969;67:131–141. 5. Kosmorsky GS, Dupps WJ Jr, Drake RL. Nonuniform pressure generation in the optic chiasm may explain bitemporal hemianopsia. Ophthalmology. 2008;115:560–565. 6. McIlwaine GG, Carrim ZI, Lueck CJ, Chrisp TM. A mechanical theory to account for bitemporal hemianopia from chiasmal compression. J Neuroophthalmol. 2005;25:40–43. 7. Wang X, Neely AJ, McIlwaine GG, Lueck CJ. Multi-scale analysis of optic chiasmal compression by finite element modelling. J Biomech. 2014;47:2292–2299. Bosler et al: J Neuro-Ophthalmol 2021; 41: 194-199 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution 8. Wang X, Neely AJ, McIlwaine GG, Tahtali M, Lillicrap TP, Lueck CJ. Finite element modeling of optic chiasmal compression. J Neuroophthalmol. 2014;34:324–330. 9. Sigal IA, Flanagan JG, Tertinegg I, Ethier CR. Finite element modeling of optic nerve head biomechanics. Invest Ophthalmol Vis Sci. 2004;45:4378–4387. 10. Wagner AL, Murtagh FR, Hazlett KS, Arrington JA. Measurement of the normal optic chiasm on coronal MR images. AJNR Am J Neuroradiol. 1997;18:723–726. 11. Schaeffer JP. Some points in the regional anatomy of the optic pathway, with especial reference to tumours of the hypophysis cerebri and resulting ocular changes. Anatomical Rec. 1924;28:243–270. 12. Long H, Qi St, Song Y, Pan J, Zhang XA, Yang Kj. Topographic variations of the optic chiasm and the pituitary stalk: a morphometric study based on midsagittal T2-weighted MR images. Surg Radiol Anat. 2014;36:775–781. 13. Noble JH, Dawant BM. An atlas-navigated optimal medial axis and deformable model algorithm (NOMAD) for the segmentation of the optic nerves and chiasm in MR and CT images. Med Image Anal. 2011;15:877–884. 14. Jain NS, Jain SV, Wang X, Neely AJ, Tahtali M, Jain S, Lueck CJ. Visualization of nerve fiber orientations in the human optic chiasm using photomicrographic image analysis. Invest Ophthalmol Vis Sci. 2015;56:6734–6739. 15. Sarlls JE, Pierpaoli C. In vivo diffusion tensor imaging of the human optic chiasm at sub-millimeter resolution. Neuroimage. 2009;47:1244–1251. 16. Roebroeck A, Galuske R, Formisano E, Chiry O, Bratzke H, Ronen I, Kim DS, Goebel R. High-resolution diffusion tensor Bosler et al: J Neuro-Ophthalmol 2021; 41: 194-199 17. 18. 19. 20. 21. 22. 23. 24. imaging and tractography of the human optic chiasm at 9.4 T. Neuroimage. 2008;39:157–168. Axer M, Strohmer S, Grassel D, Bucker O, Dohmen M, Reckfort J, Zilles K, Amunts K. Estimating fiber orientation distribution functions in 3D-Polarized Light Imaging. Front Neuroanat. 2016;10:40. Schmitz B, Schaefer T, Krick CM, Reith W, Backens M, Kasmann-Kellner B. Configuration of the optic chiasm in humans with albinism as revealed by magnetic resonance imaging. Invest Ophthalmol Vis Sci. 2003;44:16–21. Otake S, Taoka T, Maeda M, Yuh WT. A guide to identification and selection of axial planes in magnetic resonance imaging of the brain. Neuroradiol J. 2018:31:336–344. Parravano JG, Toledo A, Kucharczyk W. Dimensions of the optic nerves, chiasm, and tracts: MR quantitative comparison between patients with optic atrophy and normals. J Comput Assist Tomogr. 1993;17:688–690. Rucker CW. The concept of a semi-decussation of the optic nerves. Trans Am Ophthalmol Soc. 1957;55:87–107. Reese BE, Baker GE. The re-establishment of the representation of the dorso-ventral retinal axis in the chiasmatic region of the ferret. Vis Neurosci. 1993;10:957– 968. Reese BE, Baker GE. Changes in fiber organization within the chiasmatic region of mammals. Vis Neurosci. 1992;9:527– 533. Won HS, Han SH, Oh CS, Lee JI, Chung IH, Kim SH. Topographic variations of the optic chiasm and the foramen diaphragma sellae. Surg Radiol Anat. 2010;32:653–657. 199 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.