Mapping Human Brain Pathways: Challenges and Opportunities in the Integration of Scales Free

The brain is composed of a complex network of pathways [Sporns and Zwi, 2004], and there is much enthusiasm in using neuroimaging techniques to study connectivity changes. Yet, there is a need to accurately trace the origins and terminations of pathways in order to understand how the brain generates behaviors and how it has changed across species [Axer and Amunts, 2022]. While many studies rely on previous reconstructions of human brain pathways as a basis for study, there is sometimes an implicit assumption that these neuroimaging techniques accurately track the origin and termination of human pathways. The wiring diagram of the human brain is still largely unknown, and we need to accurately map connections and pathways in the human brain as a mandatory first step to characterizing brain development, evolution, and disease states [Van Essen et al., 2014; Zeng, 2022].

There are a few techniques to map human brain connections. Diffusion magnetic resonance (MR) tractography captures the origins and terminations of pathways and its pathways in three dimensions (Fig. 1,–7). Several methods exist (e.g., Klinger method, polarized light microscopy) to probe the structure of the human brain, but these methods do not conclusively identify origins and terminations of pathways. Other methods (e.g., functional magnetic resonance imaging) probe coordinated brain signals to make inferences about connectivity profiles and so only indirectly infer connections [Smith et al., 2013]. The review discusses some of these methods with a focus on diffusion magnetic resonance imaging and its tractography because diffusion MR tractography is a high-throughput and noninvasive method that can be used to investigate neural pathways in adulthood, in aging, and over the course of development. It can be used to study brains ex vivo as well as in vivo [Fig. 1a; Mori and Zhang 2006; Qiu et al., 2015; Pietrasik et al., 2022]. Accordingly, diffusion MR tractography generates a rich amount of information across the brain.

Working to improve the accuracy of tools to map human brain pathways is important [Axer et al., 2011; Zingg et al., 2014; Daducci et al., 2016; Kebschull et al., 2016; Romero-Garcia et al., 2018; Shin et al., 2018; Chen et al., 2019; Winnubst et al., 2019; Barbas et al., 2022; Yan et al., 2022]. Novel statistical tools can serve to reconstruct brain pathways and may theoretically be superior to past ones, but the lack of ground truth to map human brain pathways will necessarily put the accuracy of these tractographies into question. There is recent work focused on the integration of information across different scales of biological organization to trace pathways and their modifications [Chen et al., 2019; Axer and Amunts, 2022; Charvet et al., 2022; Fig. 1]. This review first discusses the range of techniques available to study brain structure and pathways (Fig. 1). It subsequently delves into the potential misconnections in pathways with a focus on two cortico-cortical association pathways. The uncertainties in tractography motivate the need to broaden our toolkit to study diffusion MR tractography to more precisely phenotype a wide range of diseases [Schmahmann et al., 2007; Charvet et al., 2022; Zeng, 2022]. Challenges in mapping human brain pathways may be opportunities for comparative neuroscientists.

Multiple methods are available to study white matter pathways in the human brain (Fig. 1a–d). Some of these methods are based on light shining through histological information (e.g., the polarized light imaging; Fig. 1b), while others are based on gross white matter dissection techniques (e.g., the Klinger method; Fig. 1c). Each of these methods provides complimentary information to study white matter structure, but many of these methods probe the direction and composition of fibers and generally fall short of identifying the origins and terminations of pathways. We will review some of the advantages and the limitations of each of these methods.

In Europe, Josef Klingler (1888–1963) developed a novel method of dissection based on a freezing technique that revealed the white matter tracts of the brain. This technique was useful for surgeons because it provided 3-dimensional information about large white matter bundles in the human brain. It is possible to detect fibers connecting the cortex with subcortical structures as well as callosal and U fibers (Fig. 1c). This technique relies on freezing postmortem brains and sequentially dissecting tissues to visualize the underlying microstructure [Fig. 1c; Agrawal et al., 2011; Decramer et al., 2018; Wysiadecki et al., 2019]. The Klinger dissection is a valuable tool for understanding the organization of white matter fibers and can be used with other methods to study white matter pathways. However, the Klinger dissection method gives a gross overview of the direction of white matter pathways in the brain (Fig. 1c). Accordingly, the Klinger method does not conclusively resolve crossing pathways nor has it been combined with microscopes to visualize details of white matter pathways. Therefore, additional methods are needed to trace origins and terminations of pathways.

Polarized light imaging can be used to trace the direction of neurons in biological tissues from postmortem histological material [Axer et al., 2011]. It is based on the principle that polarized light passes through a histological slice. The transmission of light can be used to estimate the fiber orientation and inclination angles across imaged sections [Fig. 1b; Larsen et al., 2007; Axer et al., 2011; Zilles et al., 2016; Axer et al., 2020; Axer and Amunts, 2022]. This method provides fine resolution to identify the direction of axons, but we have yet to reconstruct the origins and terminations of pathways from polarized light imaging. Such an approach is laborious but has much potential to trace pathways in the human brain.

There are several methods available to study tissue composition. One method called optical coherence tomography relies on low-coherence light to visualize the microstructure of biological tissues [Aumann et al., 2019; Araki et al., 2022]. This method is used by ophthalmologists to visualize retinal structure in a clinical setting [Kashani et al., 2017]. Another method called diffusion MR imaging relies on the principle of diffusion of water molecules to make inferences about the composition of neural structures [Fig. 1; Le Bihan et al., 1986; Cory and Garroway, 1990; Beaulieu, 2002; Mori and Van Zijl, 2002; Wedeen et al., 2008; Jones et al., 2013]. For example, fractional anisotropy measures the degree of anisotropy of water molecules to infer the extent to which fibers course together (Fig. 2). Other metrics have been developed to quantify the microstructure and to make inferences about neurite density and axonal diameter (e.g., AxCaliber, ActiveAx, neurite orientation dispersion and density imaging [NODDI]: NODDI-Watson, NODDI-Bingham, and NODDI-DTI models; Assaf et al., 2008; Alexander et al., 2010; Zhang et al., 2011; Xu et al., 2014; Raffelt et al., 2017; Genc et al., 2018; Huang et al., 2020; Dhollander et al., 2021). These methods focus on brain structure rather than the origins or terminations of pathways.

Diffusion MR tractography can be used to visualize pathways in adulthood as well as developmental processes, including fibers such as radial glia, which guide the migration of neuroblasts in the cortex [Fig. 1-7; Xu et al., 2014; Vasung et al., 2019; Wilson et al., 2021]. These reconstructed fibers are based on the diffusion of voxels with the insight that diffusion aligns with axons or fibers. Voxels are at a millimeter scale so that these tractographies average information across multiple axons, and multiple fibers may cross within a voxel. Several statistical procedures have been developed to track fibers. Those include diffusion tensor, high angular resolution diffusion, and diffusion spectrum imaging. These latter two techniques were developed in part to resolve tractographies in zones of crossing fibers [Tuch, 2004; Zhan and Yang, 2006; Wedeen et al., 2008; Catani et al., 2012; Yendiki et al., 2022]. While these are sophisticated procedures to reconstruct tractographies, they have not overcome the issue of crossing fibers [Catani et al., 2012; Wedeen et al., 2008; Yendiki et al., 2022], and additional reconstructive techniques are needed to accurately map human brain pathways.

The diffusion MR tractography of the adult human brain is from the Allen Brain Atlas [Ding et al., 2016]. Briefly, diffusion-weighted data of a 34-year-old postmortem female brain were acquired with a 3D steady-state free precession sequence (repetition time = 29.9 ms, α = 60°, echo time = 24.96 ms, isotropic resolution: 900 μm; 44 directions; b value = 3,686 s/mm²; 8 b = 0 images; Fig. 1a, 2, 4-7). The tracking algorithm is based on fiber assignment by continuous tracking. More information on the tractography parameters used for reconstruction is provided by Ding et al. [2016]. Although the use of diffusion MR scans from other brains or from different statistical procedures may generate variation in tractography reconstructions from those presented here, these scans were selected because the pathways resemble those reported in other atlases and because these scans are meant to be used as atlases for researchers [Catani and De Schotten, 2008].

In adulthood, diffusion MR tractography identifies many pathways across the brain. Those include the corticospinal tract, corpus callosum, cingulate bundles, and many cortico-cortical association pathways (Fig. 1-8). A developmental series of diffusion MR scans shows that most, if not all, cortico-cortical association pathways are established before birth [Fig. 3; Keunen et al., 2017; Vasung et al., 2019; Thomason, 2020]. Human brains grow extensively prenatally and postnatally during the first 2 years of life [Dobbings and Sand, 1973] so that some of the growth in cortical association pathways should be achieved through a stretching mechanism. Another interesting observation is that the pathways in anterior regions of the cortex appear more mature than in posterior ones at postconception week 17, which aligns with past work relating anterior to posterior gradients in neurogenesis with structural variation in adulthood [Cahalane et al., 2014]. In fetal as well as in adult brains, pathways course through the white matter but it is seldom that the tractography accurately captures their termination in the cortical gray matter (Fig. 2, 4).

Diffusion MR tractography has frequently given implausible interpretations in tractography [Jones et al., 2013; Thomas et al., 2014; Schilling et al., 2018; Grisot et al., 2021; Maffei et al., 2022]. Accordingly, the tractography is considered to have limited utility to accurately reconstruct white matter pathways [Jones et al., 2013; Thomas et al., 2014; Maffei et al., 2022; Yendiki et al., 2022]. Although it is well known that tractography generates spurious reconstructions, there is an assumption that the number of streamlines can be used as an index of axon numbers [Sotiropoulos and Zalesky, 2019]. This assumption has not been systematically tested across the brain because we lack tools to track pathways and consequently quantify axonal numbers.

Diffusion MR tractography reliably identifies a range of white matter pathways in humans and model species. Examples include the corpus callosum, which consists of fibers connecting the two cerebral hemispheres, the fornix which connects limbic structures, and the optic radiation, which consists of fibers projecting from the lateral geniculate nucleus to the primary visual cortex [Fig. 2c; Catani and De Schotten, 2008; Schmahmann et al., 2007; Charvet et al., 2019]. Although the tractography identifies well-known pathways coursing through the white matter, it rarely tracks pathways in their entirety [Fig. 4, 6; Schmahmann et al., 2007; Charvet et al., 2022]. For example, callosal fibers course across the midline as expected, but these fibers typically terminate in the motor cortex and do not project into to the lateral cortex [Hofer and Frahm, 2006; Hofer et al., 2008; Phillips and Hopkins, 2012; Platas-Neri et al., 2015; Fig. 4, 6]. Lateral callosal projections are very much expected given a large body of work tract-tracer studies from many diverse model species. A common theme to emerge from evaluating well-characterized and conserved pathways across humans and model species is that diffusion MR tractography captures the overall direction of fibers in the white matter but that it does not systematically resolve crossing fibers and neglects to track the pathway from its origins and terminations.

Zones of crossing fibers are challenging for tractography reconstructions as they are sites where misconnections may occur. One well-known example of a pathway that courses through zones of fiber crossings is the corticospinal tract, which encounters crossing fibers ventral to the motor cortex and lateral to the cingulate cortex [Fig. 3, 6; Wedeen et al., 2008]. In this zone, the corpus callosum, corticospinal tract, and cortical association pathways cross ventral to the motor cortex and lateral to the cingulate cortex (Fig. 6, 7). Some pathways such as the corticospinal tract terminate abruptly, which highlight errors in tractography reconstructions. Here, tools (e.g., polarized light imaging) that capture fibers at a microscopic scale coupled with diffusion MR tractography would help resolve these zones of fiber crossings.

A range of tools are available in model species (e.g., mice, macaques, marmosets) and can be integrated with diffusion MR tractography [Aydogan et al., 2018]. Tract tracers require injections into living brains to label the origins and terminations of a small group of neurons. These tools are available to model species but not to humans because they are invasive [Dauguet et al., 2007]. Tract-tracing tools can be used to investigate anatomical connectivity with high accuracy, though they do have some limitations. First, there is a limit to the number of injections that can be performed on the same individual simultaneously, which hinders our ability to generate a complete view of an individual’s connectivity. Second, the tracers can overflow into surrounding brain regions, leading to inaccurate results. Third, the failure of tracers to show a connection between two regions does not necessarily mean a connection does not exist. Finally, the method relies on a few individuals to establish the existence of connections and is rarely amenable to large samples because it is laborious and invasive. Nevertheless, the combination of tract tracers with diffusion MR tractography is a powerful tool to visualize pathways in the mammalian brain.

One approach to mapping human brain pathways is to focus on those pathways that are conserved across model species (e.g., corticospinal tracts, cingulate bundles) to make inferences about connectivity profiles in humans. Past work has collated tract-tracer information across a range of model species, including mice, rats, cats, opossums, macaques, and marmosets, to make inferences about the origins and terminations of human brain pathways [Kawamura and Otani, 1970; Kawamura 1973a, 1973b, 1973c; Nudo and Masterdon, 1980; Nudo et al., 1995; Schmahmann and Pandya, 2007, 2009; (Yendiki et al, 2022)]. Yet, some brain pathways should be human specific. Working on expanding the toolkits for use with diffusion MR tractography is very much needed so that we can fully map human brain pathways (Fig. 5).

The next sections focus on two cortico-cortical pathways and issues that may cause potential inaccuracies in tractography reconstruction. Although the focus is on ex vivo data from a postmortem brain [Ding et al., 2016], the tractography reported here resembles that of many other studies [Catani and De Schotten, 2008; Caverzasi et al., 2014; Rollans et al., 2017; Vassal et al., 2018]. The review discusses the inferior fronto-occipital fasciculus (IFOF; Fig. 2), which is a reported long fiber bundle that projects across the anterior and posterior regions of the cortex. This fiber bundle courses along other well-established pathways. The review also discusses the aslant tract (Fig. 6), which has been reported to connect the motor and lateral frontal cortex within each hemisphere. There are multiple lines of evidence that suggest that these tractographies may be, at least to some extent, misconnected pathways.

The IFOF is a fiber bundle that courses across the anterior to posterior axis of the white matter. It is evident during fetal development as it is in adulthood (Fig. 2-4). There may be multiple definitions of the IFOF, but it is here defined as a large bundle of fibers that connects the ventral prefrontal cortex with the occipital cortex [Fig. 2b; Catani and De Schotten, 2008]. It is a rather large pathway in that it is composed of a relatively high number of streamlines [Catani and De Schotten, 2008], suggesting a strong connection between the ventral prefrontal and visual cortex in humans and nonhuman primates. Yet, a pathway connecting the most anterior and posterior regions of the cortex is elusive when sought from tract tracers [Barbas and Pandya, 1989; Cavada et al., 2000; Majka et al., 2020], which cast doubt on the existence of such a large fiber bundle connecting the most anterior and posterior regions of the cortex.

The IFOF overlaps with three other tracts that course across the anterior to posterior axis in the white matter. These include the optic radiation, which is a projection from the lateral geniculate nucleus to the primary visual cortex (Fig. 2, 4), the uncinate fasciculus, which connects the ventral prefrontal cortex with the temporal lobe (Fig. 2, 4), as well as the inferior longitudinal fasciculus, which is a long-range pathway that largely spans the anterior to posterior regions of the temporal lobe. In anterior regions of the human brain, the IFOF courses along the uncinate fasciculus (Fig. 2c). In posterior regions, the IFOF projects along the ILF as well as the optic radiation (Fig. 2b, d). The extensive overlap across fiber bundles may lead to misconnections during tractography reconstructions (Fig. 5, 6). At the level of the thalamus, the optic radiation and the uncinate fasciculus course along a similar direction. Mislabeled connections may occur in this region, giving rise to a pathway that would span the most anterior and posterior regions of the cortex and include fibers that make up the IFOF. The uncinate fasciculus courses across the anterior to posterior axis in the ventral white matter and makes an abrupt latero-ventral turn to penetrate the temporal lobe [Fig. 2d; Catani and De Schotten, 2008]. Settings in the tractography can filter out tracts that make sharp turns, leading to misconnections in the tractography [Ding et al., 2016].

Fibers that take on stereotypical projection patterns from another pathway may indicate issues in tractography and possible misconnections in tracts. In the posterior cortex, the optic radiation is well characterized as it wraps around the posterior portions of the lateral ventricles (Fig. 2c). It has a recognizable projection pattern that widens posteriorly when viewed laterally and dorsally in humans as well as in nonhuman primates (Fig. 5). The fibers that project to the primary visual cortex in humans and nonhuman primates stand out in their density relative to the rest of the white matter so that it is possible to track the 3-dimensional route of the optic radiation [Sherbondy et al., 2008; Li et al., 2021]. As is characteristic of the optic radiation, the IFOF widens towards the posterior cortex and projects within and adjacent to the primary visual cortex. The IFOF takes on the shape of the well-characterized optic radiation, which suggests that at least some of the optic radiation is mislabeled as the IFOF [Fig. 2d; Schmahmann et al., 2007; Petrides and Pandya, 2006; 2012]. Considering the overlap in the organization, direction, and location of these pathways (Fig. 5), it is possible that all or parts of the optic radiation are mislabeled as the IFOF.

The shape and orientation of the IFOF, uncinate fasciculus, and optic radiation are similar in humans and nonhuman primates [Schmahmann et al., 2007]. Diffusion MR tractography identifies a relatively large fiber bundle across the anterior to posterior akin called the IFOF in chimpanzees, old world monkeys (e.g., vervet monkeys), and new world monkeys [Sarubbo et al., 2019; Bryant et al., 2020]. The observation that the IFOF pathway can be reconstructed across human and nonhuman primates might give credence to the existence of a fiber bundle connecting the ventral prefrontal cortex and the visual cortex. However, the shape and organization of brain pathways are very similar in human and nonhuman primates so that issues in tractography reconstructions relevant to human brains also pertain to nonhuman primates. The uncinate fasciculus and the optic radiation course across the anterior to posterior direction in human and nonhuman primates. The possibility that the uncinate fasciculus and the optic radiation misconnect at the level of the thalamus creating the IFOF is possible in nonhuman primates.

It is possible to integrate diffusion MR tractography with tract tracers directly in nonhuman primates. Tract tracers and lesion studies identify very few projections connecting the ventral prefrontal and occipital cortex in macaques as in marmosets [Petrides and Pandya, 2006; Majka et al., 2020]. The lack of concordance between tract tracer and diffusion MR tractography casts doubt on the existence of the IFOF. At present, there is no clear way to prove that these tractographies are correct or that the optic radiation is mislabeled as the IFOF. These inconsistencies and lack of concordance between diffusion MR tractography and other techniques motivate the need to develop alternative methods to trace pathways [Chen et al., 2019; Charvet et al., 2022].

The aslant tract is thought to connect the lateral inferior frontal cortex with the motor cortex, and it is thought to play a role in various cognitive processes [Dick et al., 2019; La Corte et al., 2021]. The aslant tract courses through zones of crossing fibers in the white matter ventral to the motor cortex and lateral to the cingulate cortex (Fig. 6, 7). Misconnections may occur at zones of fiber crossings and likely generate errors in the pathway reconstructions. Specifically, fibers course across the medial to lateral direction (which include the corpus callosum), the anterior to posterior direction (which include cortical association fibers), and across the dorsal to ventral direction (which include the corticospinal tract; Fig. 6). A 3-dimensional sphere was set to capture fibers coursing through these zones of crossing fibers (Fig. 6b, c) and terminate within this zone (Fig. 6c). Fibers that abruptly terminate in this zone include callosal, cortico-cortical association fibers, as well as the corticospinal tract (Fig. 6, 7). It would be expected that callosal fibers extend laterally, but many reconstructions fail to reconstruct callosal fibers extending laterally [Schwartz and Goldman-Rakic, 1984]. The early terminations of all these pathways in these examples highlight that the accuracy of pathway reconstructions is compromised.

The aslant tract, or at least many of these fibers that make up the aslant tract, likely represent the lateral extension of callosal fibers. These collosal fibers should be present but they are not observed in many diffusion MR tractography whole brain reconstructions [Hofer and Frahm, 2006; Hofer et al., 2008; Phillips and Hopkins, 2012; Platas-Neri et al., 2015; Fig. 6]. Similar to humans, nonhuman primates possess a zone of crossing fibers in the white matter ventral to the motor cortex and lateral to the cingulate cortex. Although a pathway connecting the medial to lateral frontal cortex is observed in nonhuman primates, tract tracers show rather modest connections between the lateral and motor cortex in nonhuman primates [Luppino et al., 1993; Majka et al., 2020]. The lack of concordance between tractography from diffusion MR tractography and tract tracers casts doubt on the existence of the aslant tract.

Integrating across scales and species can enhance our ability to map human brain pathways [Lemaitre et al., 2023]. In the following section, we discuss some of the latest metrics used for quantifying pathways with a focus on the integration of scales of biological organization. The integration of polarized light imaging, cytoarchitecture, and gene expression holds much promise to map human brain pathways and detect differences in projection patterns across species [Fig. 5; Charvet et al., 2022; Axer and Amunts, 2022].

There is currently uncertainty as to whether the number of tracts, a commonly used metric for measuring connectivity strength in diffusion MR tractography studies, provides biologically meaningful information. First, it is unclear whether there is a correlation between the number of tracts and the number of axons in the human brain. Second, diffusion MR tractography fails to accurately identify precise origins and terminations so that the number of tracts terminating within the gray matter is unlikely to be a proxy for axon numbers [Calabrese et al., 2015; Girard et al., 2020; Charvet et al., 2022]. Given these observations, it is a question as to whether labeling pathways based on their origins or terminations in parcellated cortical areas capture biologically relevant information. Instead, there is an important need to develop quantitative metrics that are adapted to the limitations of diffusion MR tractography.

We developed methods to quantify pathways to overcome some of the limitations associated with diffusion MR tractography [Hendy et al., 2020]. Our method relies on randomly selecting voxels in regions of interest within the white matter and classifying pathways based on their orientation and direction toward the gray matter. For example, some pathways were classified as cortico-cortical association pathways or cortico-subcortical structures and we remained agnostic as to the precise terminations of fibers within the gray matter [Hendy et al., 2020; Charvet et al., 2022]. Our analyses revealed that primates possess disproportionately more cortical association pathways than rodents. Integrative approaches are a powerful lens to identify species differences in projection patterns.

Integrating information across scales of study has much potential to resolve tractography of crossing fibers. One promising approach is the integration of polarized light imaging with diffusion MR tractography [Axer and Amunts, 2022]. Unlike diffusion MR tractography, which visualizes fibers at a millimeter scale, polarized light imaging captures axons at a microscopic resolution. By integrating information across these different scales, it may be possible to more accurately resolve fiber crossings in the cortical white matter. For example, combining polarized light imaging with diffusion MR tractography can provide detailed information on the origins and terminations of callosal projections, the aslant tract, and the corticospinal tract zones of white matter crossings in zones of fiber crossing. These new methods for integrating information across scales of biological organization offer exciting opportunities to generate a more comprehensive map of human brain pathways.

Cross-species differences in diffusion MR tractography can be integrated with quantification of neuronal populations across the depth of the cortex. Cell types can be distinguished by their birth order, position, gene expression, as well as their patterns of connectivity [Rakic, 1974; Zeng et al., 2012; Charvet et al., 2017]. Neuron bodies that are preferentially located in superficial layers project within and across cortical areas, and those located in lower layers preferentially project to subcortical structures. Given these geometric patterns, variation in the relative number of neuronal soma across the depth of the cortex can be used to make inferences about species differences in connectivity. Past work has integrated information from the gray and white matter to infer species differences in connectivity patterns. Specifically, comparative analyses of cortical neuronal populations show a disproportionate expansion in layer II–IV neurons in primates relative to rodents [Charvet et al., 2017a]. Layer II–IV neuronal populations form long-range and short-range cortico-cortical projections, and this is of interest for understanding evolutionary modifications in pathway types. The expansion in layer II–IV neuron numbers is concomitant with an expansion in long-range cortico-cortical association in primates as assessed with diffusion MR tractography [Charvet et al., 2017a, b; Charvet et al., 2019; Hendy et al., 2020]. The use of histological information has some limitations in that different neuronal populations are clumped together. This is the case for layer II and layer III neurons. Other tools will provide an exciting opportunity to integrate information about cell populations with diffusion MR tractography. Nonetheless, the integration of histological information with diffusion MR tractography provides a strong toolkit to study group differences in connectivity patterns.

Single-cell RNA sequencing can be used to identify a range of different cell populations in some detail and integrated to enhance our visualization of white matter pathways in the human brain. Single-cell RNA sequencing can be used to distinguish different pyramidal neurons across layers, and it will be useful to integrate these data with diffusion MR tractography [Hendy et al., 2020; Zeisel et al., 2015; Zeng, 2022]. Although we have yet to agree on clustering approaches to define cell populations [Petegrosso et al., 2020; Kiselev et al., 2019; Zeng, 2022], the integration of single-cell RNA sequencing with histology provides a broad toolkit to study species differences in pathway types.

Another exciting venue focuses on integrating gene expression with structure to transcriptionally define neuronal populations. Past work has correlated gene expression with structure over the course of development with the aim to define neuronal populations [Fig. 9; Charvet et al., 2017a, b; 2019, 2022]. These markers can be used in conjunction with diffusion MR tractography to assess connectivity differences across populations [Charvet et al., 2022]. These analyses have highlighted similarities in frontal cortex circuits across human and nonhuman primates [Charvet et al., 2022]. We have yet to work out several details. For example, gene expression is extracted from frozen tissue but diffusion MR tractography is based on living individuals or formaldehyde-fixed tissue, restricting the range of integrative analyses that are possible. Nevertheless, approaches such as these have much potential to fully map the human brain projections and identify group differences in connectivity patterns in health, disease, and across species.

Broadening the toolkits to map human brain pathways is a useful endeavor as it will enhance our ability to more accurately trace the evolution of the human brain and identify the neurobiological bases of diseases. The issues with tractography motivate the need to use a range of neuroimaging tools to study the evolution of pathways.

Ashley Moore and Tom Spencer made some of the figure illustrations. www.Biorender.com was used to make some of the figures.

This is a review article. The section is not applicable.

The author has no conflicts of interest to declare.

Grant sponsor: NIGMS INBRE pilot grant to C.J.C from center grant, Grant No. P20GM103446; Grant sponsor: NIGMS COBRE, Grant No. 5P20GM103653; Grant sponsor: NICHD R21, Grant No. 1R21-HD-101964-01A1 and 7R21-HD101964-02 (to C.J.C.); and Auburn Startup funds to C.J.C. Some images are taken from biorender.com. Some of the MR scans used in this work are available at http://www.brainspan.org, with work supported by the National Institute of Health Contract HHSN-271-2008-00 047-C to the Allen Institute for Brain Science.

C.J.C. wrote the paper.

This work does not include data.

Christine Charvet is a member of the J.B. Johnston Club.