The Mid-Fusiform Sulcus (sulcus sagittalis gyri fusiformis)
ABSTRACT
In the human brain, the mid-fusiform sulcus (MFS; sulcus sagittalis gyri fusiformis) divides the fusiform gyrus (FG) into lateral and medial partitions. Recent studies show that the MFS is identifiable in every hemisphere and is a landmark that identifies (a) cytoarchitectonic transitions among four areas of the FG, (b) functional transitions in many large-scale maps, and (c) the location of fine-scale functional regions. Thus, simply identifying the MFS in a person's brain provides researchers with knowledge regarding: (a) how cells are organized across layers within a particular cortical location, (b) how functional representations will be laid out in cortex, and (c) the precise location of functional regions from cortical folding alone. The predictive power of the MFS can guide future studies examining the anatomical-functional organization of the FG, as well as the development of translational applications for different patient populations. Nevertheless, progress has been slow in incorporating the MFS into the broader anatomical community and into neuroanatomical reference sources. For example, even though the MFS is a rare structural–functional landmark in human association cortex as just described, it is not recognized in the recently published Terminologia Neuroanatomica (TNA). In this review, I collate the anatomical and functional details of the MFS in one place for the first time. Together, this article serves as a comprehensive reference regarding the anatomical and functional details of the MFS, as well as provides a growing number of reasons to include the MFS as a recognized neuroanatomical structure in future revisions of the TNA. Anat Rec, 302:1491–1503, 2019. © 2018 American Association for Anatomy
INTRODUCTION
Located in ventral occipito-temporal cortex of the hominoid brain, the fusiform gyrus (FG) performs functionally specialized computations underlying face perception (Kanwisher et al., 1997; Puce et al., 1996, 1999; Nobre et al., 1998; Rossion et al., 2003; Grill-Spector et al., 2004; Rossion, 2008; Parvizi et al., 2012; Grill-Spector and Weiner, 2014; Rangarajan et al., 2014), object recognition (Malach et al., 1995; Gauthier et al., 1999; Gauthier et al., 2000; Grill-Spector 2003; Konen et al., 2011; Gauthier and Tarr, 2016), and reading (Cohen et al., 2000; Wandell et al., 2012; Glezer and Riesenhuber, 2013; Bouhali et al., 2014). A longitudinal sulcus known as the mid-fusiform sulcus (MFS) divides the FG into lateral and medial partitions (Puce et al., 1996; Nobre et al., 1998; Allison et al., 1999; Weiner and Grill-Spector 2010; Nasr et al., 2011; Weiner et al., 2014). Though the MFS was defined and labeled as the sulcus sagittalis gyri fusiformis in 1896 by Retzius (1896), it was mentioned only a handful of times in papers and atlases for the next 100 years until it re-appeared in the cognitive neuroscience literature in 1996 (Weiner and Zilles, 2016 for review). Within the last decade, a plethora of studies have shown that the MFS identifies (a) cytoarchitectonic transitions among four areas of the FG (Caspers et al., 2013; Weiner et al., 2014; Lorenz et al., 2017; Weiner et al., 2017a; Rosenke et al., 2018), (b) functional transitions in many large-scale functional maps (Weiner et al., 2014; Weiner et al., 2010; Nasr et al., 2011; Grill-Spector and Weiner, 2014; van den Hurk et al., 2015; Jacques et al., 2016; Kadipasaoglu et al., 2016), and (c) the location of fine-scale functional regions (Weiner and Grill-Spector 2010; McGugin et al., 2014; Weiner et al., 2014, 2017a; McGugin et al., 2015; Jacques et al., 2016; Kadipasaoglu et al., 2016) that are causally implicated in visual perception (Parvizi et al., 2012; Rangarajan et al., 2014). Together, the findings from these studies have established the MFS as an anatomical and functional landmark in the human brain.
Nevertheless, progress has been slow in incorporating the MFS into the broader anatomical community and into neuroanatomical reference sources because most—if not all—of the findings summarized in the previous paragraph are published in neuroimaging and cognitive neuroscience journals. As such, even though the MFS is identifiable in every human brain to the point in which an algorithmic approach can be implemented to automatically identify the MFS on cortical surface reconstructions (Weiner et al., 2018), it is not recognized in the recently published Terminologia Neuroanatomica (TNA; FIPAT, 2017; Ten Donkelaar et al., 2017, 2018). Consequently, the main goal of the present work is to provide a comprehensive review of these empirical findings regarding the MFS, which can serve as a reference source for the extended anatomical and cognitive neuroscience fields moving forward.
To achieve this goal, this review can be divided into eight sections. First, I review recent findings revealing that the MFS displays morphological features that are identifiable from childhood to adulthood, as well as are identifiable in living and post-mortem brains. The second section reviews a series of recent studies that identified cytoarchitectonic transitions among four areas within the MFS using observer-independent methods: one in the posterior FG between areas FG1 and FG2 (Caspers et al., 2013; Weiner et al., 2014), as well as one in the middle portion of the FG between areas FG3 and FG4 (Lorenz et al., 2017). The third section shows that these cytoarchitectonic transitions correspond well with transitions in large-scale functional maps in which the cortical expanse lateral to the MFS is functionally distinct from the cortical expanse medial to the MFS. The fourth section further details that while the entire MFS identifies functional transitions in large-scale maps, particular features of the MFS identify the location of fine-scale functional regions. The fifth section discusses recent findings showing that the endpoints of white matter fascicles terminate in predictable locations relative to the MFS. The sixth section expands on how knowledge of the MFS opens new questions about the evolution of ventral occipito-temporal cortex in hominoids. The seventh section discusses how identifying and measuring different anatomical and functional features of the MFS has translational applications that have clinical and functional significance in different patient populations. Finally, the eighth section situates the MFS within nearby gyri and sulci presently accepted by the TNA (FIPAT, 2017; Ten Donkelaar et al., 2017, 2018). Together, this article serves as a comprehensive reference source regarding the anatomical and functional details of the MFS, as well as provides a growing number of reasons to include the MFS as a recognized neuroanatomical structure in future revisions of the TNA.
The MFS is Identifiable in Every Brain with Morphological Features that are Identifiable from Childhood to Adulthood, As Well As in Living and Post-Mortem Brains
“It is known, that the inferior surface of the temporal lobe is concave in anterior–posterior direction; this is particularly true for the Fusiform Gyrus. A sagittal sulcus along the midline can be very frequently found on its surface, which might be called Sulcus sagittalis gyri fusiformis. This sulcus can be uninterrupted and visible for a long distance, but is often subdivided into two or more parts, which may have branches and join neighboring sulci, which leads to a complicated pattern of the gyrus; if, however the Sulcus sagittalis is clearly visible and well developed, the surface of the gyrus can be subdivided into two parallel, sagittal convolutions, the Gyrus medius and lateralis, which are found separated in some cases or connected by bridges in other cases. These gyri can be followed far posterior, where they extend over the edge of the hemisphere and merge with the convolution of the occipital lobe, inferior temporal gyrus, and lingual gyrus in one or the other way”. P. 1421
After Retzius' observation, the MFS was mentioned (with different names) only a handful of times within the literature and in reference atlases for the next five decades (Mickle, 1897; Vogt 1904; Connolly, 1950; Bailey and von Bonin, 1951) until reappearing in the cognitive neuroscience literature 100 years after his original observation (Puce et al., 1996; Nobre et al., 1998; Allison et al., 1999; see Weiner and Zilles, 2016 for an extensive review).
Once the MFS re-surfaced in the literature in the late 1990s, it would almost be another 20 years before the stable and variable morphological features of the MFS would be assessed (Weiner et al., 2014). In terms of stability, the MFS was (a) identifiable in all 158 hemispheres included in that study, which included both living and post-mortem brains from individuals spanning in age from 7 to 85, and (b) consistently about half as deep compared to the OTS and CoS (Weiner et al., 2014). The difference in depth between the MFS and surrounding sulci generates a distinctive pattern on single coronal slices of the brain—whether within in vivo T1 images or post-mortem histological sections (Fig. 1).
In contrast to the stable shallowness of the MFS, the MFS can vary significantly in terms of its length—ranging from 2.0 mm to 56.3 mm (Nasr et al., 2011; Weiner et al., 2014). The difference in length affects the location of the posterior, not the anterior, tip of the MFS. Specifically, the anterior tip of the MFS aligns with the posterior tip of the hippocampus (Grill-Spector and Weiner, 2014). The MFS also varies in fractionation and intersection with the OTS and CoS. From previous analyses, in nearly half (48.55%) of the hemispheres examined, the MFS appears as a single longitudinal sulcus independent of the OTS and CoS in both children and adults (Weiner et al., 2014; Fig. 1). In the rest of the hemispheres examined (51.45%), the MFS varies in terms of its fractionation, as well as its intersection with the OTS and CoS. These stable and variable features are equally as likely to appear in (a) a child's brain or an adult's brain and (b) in a living brain or a post-mortem brain. Importantly, it does not take the expert eyes of trained anatomists to identify the MFS. Algorithmic approaches that leverage cortex-based alignment among individuals (Dale et al., 1999; Fischl et al., 1999) automatically identify the MFS and accurately discriminate it from surrounding sulci (Fig. 1).
Taken together, the macroscopic description of the MFS has not changed between Retzius' original observations and present quantifications. In particular, Retzius described the MFS as a sulcus dividing the FG into medial and lateral partitions, which is how present neuroanatomists and cognitive neuroscientists also define the MFS. Additionally, the first morphological analyses of the MFS revealed that (a) the MFS is identifiable in every brain, (b) its most stable morphological feature is its shallowness, (c) its most variable feature is its length, and (d) despite the variable features, the stable features among individual hemispheres enable the MFS to be defined automatically using tools that leverage cortex-based alignment.
The MFS is a Cytoarchitectonic and Receptor Architectonic Landmark
Classic cytoarchitectonic studies of the human brain commonly concluded that sulci seldom served as landmarks of cytoarchitectonic transitions outside of primary sensory cortices (Smith, 1907; Brodmann, 1909; Economo and Koskinas, 1925; Bailey and von Bonin, 1951). However, a main limitation of those classic studies is that neuroanatomists were manually examining histological tissue under a microscope and deciding, in an observer-dependent fashion, when one part of the tissue was cytoarchitectonically dissociable from an adjacent piece of tissue. Over the last few decades, a major methodological advancement occurred when researchers treated cytoarchitectonic analyses as an image processing problem. That is, instead of manually examining the tissue under a microscope, automated algorithms were devised to traverse the cortical ribbon and to determine if and where adjacent pieces of tissue were cyotarchitectonically different from one another (Zilles, 1978; Zilles et al., 1978; Zilles et al., 1980; Wree et al., 1982; Istomin and Shkliarov, 1984; Schleicher et al., 1986; Serra, 1986; Zilles et al., 1986; Rauch et al., 1989; Schleicher and Zilles, 1990; Schleicher and Zilles, 1990; Ahrens et al., 1990; Istomin and Amunts, 1992; Amunts et al., 1995; Schlaug et al., 1995; Schleicher et al., 1999; Schleicher et al., 2005; Amunts and Zilles, 2015). Using these methods, four areas have been parcellated in the human FG (Figs. 2 and 3). In the posterior FG, FG1 is located on the medial FG and extends into the CoS, while FG2 is located on the lateral FG and extends into the OTS (Caspers et al., 2013). Moving more anteriorly to the middle portion of the FG, FG3 is located on the medial FG and extends into the CoS, while FG4 is located on the lateral FG and extends into the OTS (Lorenz et al., 2017).
In both the posterior and middle FG, the algorithmic approach identified cytoarchitectonic transitions within the MFS (Weiner et al., 2014; Lorenz et al., 2017). This is worth emphasizing because the algorithmic cytoarchitectonic approach identifies cytoarchitectonic boundaries independent of cortical folding. Thus, if the algorithm identifies the boundary at a particular cortical location that is reproducible between hemispheres in the same person and brains from multiple participants, then it is meaningful and serves as a landmark. In the posterior FG, the MFS identifies a cytoarchitectonic transition between FG1 and FG2 (Weiner et al., 2014). FG1 displays a columnar arrangement of small pyramidal cells and a thin and cell sparse layer IV (Caspers et al., 2013). FG2 shows large pyramidal cells in layer III, a prominent layer IV, but a less pronounced columnar organization. Additionally, FG2 is characterized by a higher cell density compared to FG1 (Caspers et al., 2013; Figs. 3 and 4). In the middle portion of the FG, the MFS also identifies a cytoarchitectonic transition between FG3 and FG4 (Lorenz et al., 2017). FG3 shows a compact and dense layer II, a prominent sub-layer IIIc with medium-sized pyramidal cells, and little clusters of granular cells in layer IV (Lorenz et al., 2017). FG4 has a less densely packed layer II, broad layer III, a thin, moderately dense layer IV, and a cell dense layer VI (Lorenz et al., 2017). Concomitantly, the MFS is a landmark identifying cytoarchitectonic transitions among four areas within the posterior and middle portions of the FG.
Since these FG areas have only been parcellated within the last 5 years, a useful exercise is to compare their cortical location to the classic cytoarchitectonic parcellations within the FG. This is possible because classic cytoarchitectonic areas of the human brain have been aligned to the FreeSurfer average cortical surface. For instance, Brodmann's parcellation (defined in the PALS-B12 atlas by Van Essen, 2005) and the Economo and Koskinas (1925) parcellation (manually delineated by Scholtens et al., 2018) have both been aligned to the FreeSurfer average surface. Consequently, qualitative comparisons can be made among the areas of the classic, observer-dependent parcellations and those of the modern, observer-independent parcellations within the FG. In comparison to Brodmann's scheme, FG areas 1–4 from the observer-independent scheme overlap with Area 37. Additionally, FG1 and FG2 overlap with Area 19, while FG3 and FG4 overlap with Area 20. In comparison to the scheme of Economo and Koskinas (1925), FG1 and FG2 are largely contained within Area PH, while FG3 and FG4 are largely contained within area TF. Interestingly, if we consider (a) FG1 and FG2 as a posterior cluster and (b) FG3 and FG4 as an anterior cluster, the observer-dependent TF/PH boundary resembles the observer-independent boundary differentiating the FG1/FG2 posterior cluster from the FG3/FG4 anterior cluster (Fig. 4A), especially in the left hemisphere at the group level.
As a final point in this section, it is worth emphasizing that the MFS not only identifies cytoarchitectonic transitions, but also differences in receptor density across cortical layers (known as receptor architecture) in the posterior FG (Caspers et al., 2015a, b; Fig. 4B). As transmitter receptors are key molecules of neurotransmission, differences in receptor architecture reflect differences in functional architecture. Figure 4B illustrates the relationship among the algorithmic cytoarchitectonic delineation of FG1 and FG2, the MFS, and the laminar distribution of 5-HT1A receptors along the cortical ribbon (Caspers, 2013; Caspers et al., 2015a, b). Quantifying the densities of different receptor binding sites across cortical layers within areas FG1 and FG2 empirically showed that these areas differ in the densities of 5-HT1A, NMDA, GABAA, GABAB, M3, and nicotinic α4/β2 receptors. Thus, knowing the location of the MFS also predicts differences in receptor density across cortical layers in the posterior FG, which has functional implications for interpreting cellular architecture (Fig. 4B, right).
Together, these findings across studies indicate that the MFS is a rare landmark identifying microarchitectonic transitions in human association cortex. These findings boast an impressive amount of predictive power: cellular insight can be gleaned from a macroanatomical location on the cortical surface.
The MFS is a Landmark Identifying Transitions in Many, Large-Scale Functional Maps
Over the last several decades, neuroimaging studies have identified many large-scale functional gradients, or maps, in human ventral occipito-temporal cortex. Interestingly, each of these maps contains a similar functional topology: neural responses on the lateral FG extending into the OTS are functionally distinct from neural responses on the medial FG extending into the CoS. For example, preferential neural responses for processing visual stimuli presented in the center of the visual field (known as a foveal bias) are located on the lateral FG and OTS, while preferential neural responses for processing visual stimuli presented in the peripheral portion of the visual field (known as a peripheral bias) are located on the medial FG and CoS. This orderly representation of functional responses on the cortical sheet is known as an eccentricity bias map (Fig. 5A; Malach et al., 2002). Intriguingly, the functional transition in this map is predicted by the MFS. Specifically, in children, adolescents, and adults (ranging in age from 7 to 40), the functional transition in this map occurs 4.1–4.6 mm from the fundus of the MFS. Though that was the only study to explicitly quantify the functional transition in a large-scale map relative to the MFS, a recent review article (Grill-Spector and Weiner, 2014) showed that the MFS qualitatively identified the functional boundary in additional functional maps containing representations of animacy (Haxby et al., 2011), real world object size (Konkle and Oliva, 2012), semantics (Huth et al., 2012), domain selectivity (Nasr et al., 2011), and conceptual knowledge (Martin, 2007). Importantly, this organization generalizes across multiple types of neuroimaging techniques. For example, while the measurements just described were conducted with functional magnetic resonance imaging (fMRI), these functional transitions have also been identified with invasive measurements in patient populations using electrocorticography (Jacques et al., 2016; Kadipasaoglu et al., 2016) or intracranial depth electrodes (Jonas et al., 2016; Rossion et al., 2018).
It should also be stated that though the MFS is a crucial landmark in visual cortex, being able to see is not a pre-requisite for the MFS to be a functional landmark. Indeed, the MFS also identifies functional representations in blind individuals (van den Hurk et al., 2015). Similar to the ending of the previous section, these findings also boast an impressive amount of predictive power: simply identifying the MFS in a person's brain predicts how functional representations will be laid out in cortex and additionally, where distinctions, or boundaries, of functional representations will occur.
The Anterior Tip of the MFS is a Landmark Identifying the Location of a Fine-Scale Functional Region Selective for Images of Faces
The FG has long been associated with visual perception. For example, neuropsychological case studies reveal that damage to the FG results in different types of perceptual disorders such as object agnosia and prosopagnosia (Damasio et al., 1982; Farah, 1990; Rossion, 2008; Konen et al., 2011). Additionally, since the early 1990s, neuroimaging studies have identified regions that are face-selective (e.g., selective in the sense that neural responses are higher to images of faces compared to neural responses to images of non-face categories) on the FG (Sergent et al., 1992; Haxby et al., 1994; Puce et al., 1995; Kanwisher et al., 1997). Within the last decade, improved neuroimaging methods and data analyses within individual subjects noted that face-selective regions have a tight correspondence with cortical folding (Weiner and Grill-Spector, 2010; Weiner et al., 2010, 2014; Nasr et al., 2011). Specifically, the MFS serves as a landmark identifying face-selective regions on both a lateral–medial axis, as well as an anterior–posterior axis. On a lateral–medial axis, the MFS reliably discriminates face-selective regions on the lateral FG from place-selective regions within the CoS (Weiner and Grill-Spector, 2010; Weiner et al., 2010, 2014; Nasr et al., 2011). On an anterior–posterior axis, the MFS accurately discriminates the location of face-selective regions from one another. In fact, building a 1 cm disk at the anterior tip of the MFS defines 83 ± 7% of mFus-faces/FFA-2 within the right hemisphere across individuals (Weiner et al., 2014; Fig. 5B). On the contrary, a 1 cm disk at the posterior tip of the MFS only defines 48 ± 9% of pFus-faces/FFA-1 within the right hemisphere across individuals (Fig. 5B). Consequently, the morphological stability of the anterior tip of the MFS compared to the morphological variability of the posterior tip of the MFS (which was described in the first section) has functional implications in which the anterior tip of the MFS is a landmark predicting the cortical location of a functional region selective for faces. It is important to emphasize that this structural–functional correspondence is not epiphenomenal as electrical stimulation to face-selective regions located lateral to the MFS induces perceptual distortions of faces (Parvizi et al., 2012; Rangarajan et al., 2014). Thus, simply identifying the anterior tip of the MFS not only identifies the location of a functional region implicated in high-level visual processing, but also offers causal insight into face perception.
The Relationship between Anatomical Connectivity and the MFS
Diffusion MRI (dMRI) and tractography algorithms enable the examination of white matter tracts in the living human brain (Mori and van Zijl, 2002; Catani et al., 2003; Catani and Thiebaut de Schotten, 2008; Tournier et al., 2012; Sotiropoulos et al., 2013; Pestilli et al., 2014; Takemura et al., 2016; Wandell, 2016; Maier-Hein et al., 2017; Yeatman et al., 2018; and many others). Using these methods, findings from recent studies have revealed an elegant correspondence between white matter association fibers and the MFS.
Specifically, recent dMRI studies have identified a consistent topological relationship between the MFS and both vertical and longitudinal white matter tracts. For example, examining the cortical endpoint terminations of the vertical occipital fasciculus (VOF) and the posterior arcuate fasciculus (pAF) relative to cortical folding revealed that the anterior boundary of the VOF and the posterior boundary of the pAF are located near the midpoint of the MFS (Fig. 6A; Yeatman et al., 2014; Weiner et al., 2017b). To link to previous sections of the present article that discussed cytoarchitectonic areas within the FG, the VOF largely terminates posterior to FG3 and FG4 within FG1 and FG2 and surrounding areas (at least at the group level as illustrated in Fig. 6A). In terms of longitudinal tracts, Gomez et al. (2015) identified white matter fascicles that intersected with functional regions selective for either faces or places in ventral occipito-temporal cortex. These tracts were (a) oriented longitudinally, (b) located below the inferior longitudinal fasciculus, and (c) surrounded the MFS in which face-selective fascicles were positioned lateral to the MFS and place-selective fascicles medial to the MFS (Fig. 6B). These vertical and longitudinal white matter tracts likely contributed to a recent data-driven parcellation of the FG based on connectivity, which delineated three distinct areas that also had a consistent topological relationship relative to the MFS (Zhang et al., 2016). Future seed-based analyses may provide additional clarity regarding the similarity and differences in whole brain connectomes when positioning different seeds on either side of the MFS.
It should also be stated that in addition to non-invasive dMRI and tractography, many invasive methods are used to examine white matter and connectivity in post-mortem human brains such as the Klingler technique (Klingler, 1935), fiber dissections (Curran, 1909), and the Nauta method (Clarke and Miklossey, 1990; Clarke, 1994), among others. To my knowledge, no study has yet examined connectivity relative to the MFS in post-mortem human brains. To fill this gap in knowledge, it would be ideal if one could draw insights from the long history of anatomical tracer studies in macaques. However, as stated in a previous section of the present manuscript, the FG is a hominoid-specific structure and is not present in macaques (Zeki and Marini, 1998; Nasr et al., 2011; Weiner and Zilles, 2016). Thus, relating findings from the plethora of anatomical connectivity studies in macaques to glean insights specifically regarding connectivity of the MFS in humans is impossible since macaques also do not have an MFS. Consequently, future studies, perhaps using new methods to measure fine-scale connections in post-mortem human brains such as polarized light imaging (Caspers et al., 2015b), could further examine the similarity and differences in the connectivity on either side of the MFS. Taken together, the MFS is a landmark linking cytoarchitectonic transitions, functional representations, and white matter fascicles within human ventral occipito-temporal cortex.
The MFS Opens New Questions about the Evolution of Ventral Occipito-Temporal Cortex
While other non-human hominoids such as chimpanzees have an FG (Retzius, 1906; Parr et al., 2009; Chance et al., 2013), it is presently unknown if non-human hominoids also have an MFS. Thus, future morphological studies of the MFS in non-human hominoids will reveal if the MFS is present in both humans and non-human hominoids. This has important implications for understanding the evolution of ventral occipito-temporal cortex: if the MFS is also present in non-human hominoids, then it is a hominoid-specific structure; if it is not, then it is a human-specific structure. Either conclusion motivates understanding how the structure of the FG evolved its many specialized roles—such as face processing and reading—and also, how the large-scale maps and fine-scale clusters that are spatially laid out in an orderly fashion relative to the MFS contribute to those specialized roles.
Accurate Identification of the MFS Has Translational Applications
“The fusiform face area (FFA) within the FG could not be identified separately because neither gross anatomical landmarks nor cytoarchitectonic criteria have been established in the literature to identify the FFA within the FG in human post-mortem brains.” van Kooten et al., 2008, p. 989
The recent findings that the MFS is a functional and cytoarchitectonic landmark solves both of these issues. Functionally, a position relative to the MFS in post-mortem tissue can be related either to (1) large-scale functional maps since cortical locations lateral to the MFS are functionally distinct from cortical locations medial to the MFS and (2) to fine-scale functional regions since the anterior tip of the MFS predicts the location of a face-selective region (mFus-faces/FFA-2). Cytoarchitectonically, as the MFS predicts cytoarchitectonic transitions among four different areas, future studies can also use the cytoarchitectonic structure of areas FG1-4 as a baseline from which to assess typicality in different patient populations (Uppal et al., 2014).
The translational applications do not stop with post-mortem assessments of cytoarchitectonics. Indeed, future studies comparing the structural–functional organization of the brain in controls and patients with disorders that have been associated with the FG could quantify and compare (1) MFS morphology, (2) similarities and differences in the structural–functional coupling among the MFS, large-scale functional maps, and fine-scale functional regions, as well as (3) similarities and differences in the tripartite relationship among the MFS, white matter tracts, and functional regions. Regarding the latter, the study conducted by Gomez et al., (2015) that was mentioned in Section 5 (The relationship between anatomical connectivity and the MFS) and depicted in Figure 6B, also quantified similarities and differences between control participants and patients who could not perceive faces (Gomez et al., 2015). Both the controls and the patients had a similar topology of these tracts relative to cortical folding. However, white matter properties within the tracts lateral, but not medial, to the MFS that intersected with face-selective regions were different between the patients compared to controls. Thus, the large-scale topology of the tracts developed normally in the patients, but the tracts were functioning differently in the patients compared to controls. Of course, there are many more examples that one could give, but these examples already show some translational applications that acknowledging the MFS can serve.
Situating the MFS Relative to the Terms Accepted by the TNA
In the TNA (FIPAT, 2017; Ten Donkelaar et al., 2017, 2018), the approved label of the fourth temporal gyrus in the “US English” and “UK English” columns is fusiform gyrus (p. 64). In the “other” column of the TNA for this gyrus, the approved label is lateral occipitotemporal gyrus (or LOTG). These two terms date back to the original labeling of the FG by Huschke (1854) and a re-labeling of this macroanatomical structure to the LOTG by Pansch (1866) (see Weiner and Zilles, 2016 for review). Perhaps it is obvious to the reader that the MFS label fits more appropriately with the FG label as opposed to the LOTG label. For example, consider the following two macroanatomical definitions of the MFS: (1) the MFS is a longitudinal sulcus that divides the lateral FG from the medial FG, or (2) the MFS is a longitudinal sulcus that divides the lateral lateral occipito-temporal gyrus from the medial lateral occipito-temporal gyrus. The former is clearer than the latter. One could argue that should the MFS be accepted as the “US English” and “UK English” columns in future revisions of the TNA, then occipitotemporal sulcus or lateral occipitotemporal sulcus could be accepted in the “other” column. However, this would add confusion rather than clarity for two reasons. First, the TNA already acknowledges an occipitotemporal sulcus or lateral occipitotemporal sulcus that is not within the FG (FIPAT, 2017, p. 64), but forms the lateral boundary of the FG. Second, the OTS is morphologically distinct based on its depth compared to the MFS as discussed in the first section of this article (Fig. 1).
Thus, if the goal of the TNA is to prevent confusion, I suggest that (a) MFS is the least confusing label for this sulcus for the “US English” and “UK English” columns of the TNA table because it is already widely used in the neuroscience literature (Retzius, 1896; Bailey and von Bonin, 1951; Puce et al., 1996; Nobre et al., 1998; Allison et al., 1999; Weiner and Grill-Spector 2010, 2013; Nasr et al., 2011; Petrides 2012; Grill-Spector and Weiner, 2014; McGugin et al., 2014; Weiner et al., 2014; Yeatman et al., 2014; Gomez et al., 2015; McGugin et al., 2015; Lorenz et al., 2017; van den Hurk et al., 2015; Jacques et al., 2016; Kadipasaoglu et al., 2016; Natu et al., 2016; Weiner and Zilles 2016; Gomez et al., 2017; Weiner et al., 2017a) and (b) sulcus sagittalis gyri fusiformis (as originally proposed by Retzius (1896)) is the least confusing for the “other” column of the TNA table because it is also historically accurate and clearly differentiable from other sulcal names accepted by the TNA. Concomitantly, MFS and sulcus sagittalis gyri fusiformis seem to be the labels that will minimize confusion and maximize clarity among neuroanatomists spanning basic and applied research, as well as those in different medical fields.
CONCLUSIONS
The MFS (sulcus sagittalis gyri fusiformis) has been identified by the eyes of expert neuroanatomists since 1896 and by algorithms over 120 years later. And yet, the MFS is more than just identifiable in every hemisphere. It is also a landmark that identifies (a) cytoarchitectonic transitions among four different areas, (b) transitions among a multitude of large-scale functional maps, (c) the location of a fine-scale functional region that is causally implicated in visual perception, and (d) the cortical location of endpoints from vertical and longitudinal white matter fascicles. Altogether, this article serves as a comprehensive reference source regarding these anatomical and functional details of the MFS, as well as provides a growing number of reasons to include the MFS as a recognized neuroanatomical structure in future revisions of the TNA. Formal acknowledgement of the MFS by the TNA has benefits not only for basic research in neuroanatomy and neuroscience, but also for translational applications that could use the MFS for both potential diagnostic purposes as well as for improved understanding of structural–functional organization of the FG in health and disease across spatial scales—from cellular to areal and systems organization.
ACKNOWLEDGMENTS
This work was supported by start-up funds provided by the University of California, Berkeley and the Helen Wills Neuroscience Institute. I thank all of my collaborators who have contributed to the studies that are comprehensively reviewed in this article, especially Kalanit Grill-Spector, Karl Zilles, and Katrin Amunts.
