Fig 1 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
The principal gradient and the two neocortical origins. A) Cortical parcellation based on the Allen Brain Mouse Atlas. B) The principal gradient of functional connectivity sampled on the surface mesh. Note that the gradients represent abstract dimensions and have no units. C) Mesh representations of the hippocampal area (archicortex, petrol ) and piriform area (paleocortex, rose ), that were used as source regions to calculate geodesic distance to the two neocortical origins. D) The cortical surface was divided into two zones based on the minimal distance of each surface node to either paleocortex or archicortex. Shown are the Gradient 1 values in either zone (Wilcoxon rank-sum = 69.6, P corr = 0.01). E) Map of the combined geodesic distance, obtained by subtracting distance to archicortex from distance to paleocortex. High negative values indicate proximity to paleocortex, high positive values indicate proximity to archicortex. Values close to zero mark that a surface node is equally distant from both origins.
Source publication
Understanding cortical organization is a fundamental goal of neuroscience that requires comparisons across species and modalities. Large-scale connectivity gradients have recently been introduced as a data-driven representation of the intrinsic organization of the cortex. We studied resting-state functional connectivity gradients in the mouse corte...
Contexts in source publication
Context 1
... EPI to anatomical, anatomical to study template, and study template to Allen Mouse CCF v3 registrations were combined into one set of forward and backward transforms ( ComposeMultiTransform ). The combined transforms were applied to the denoised fMRI data ( Fig. S1 ). ...
Context 2
... the dual origin analysis, we used volumetric masks of the piriform area and the hippocampal region from the Allen Mouse CCF v3 ( Fig. 1 C), and sampled these masks onto the surface mesh. We then created two distance maps by assigning to each isocortical surface node its shortest distance to any node within the hippocampal and piriform mask, respectively. ...
Context 3
... assess whether Gradient 1-6 values in the two zones were likely to be drawn from different distributions, we computed the Wilcoxon rank-sum statistic as implemented in Scipy ( Virtanen et al., 2020 ). We also created a combined distance map by subtracting the hippocampal distance map from the piriform distance map ( Fig. 1 E). The combined distance map captures a spatial gradient between the two origins at the extreme ends of the value scale. ...
Context 4
... investigate the intrinsic functional organization of the mouse cortex, high quality rsfMRI data were obtained in ventilated, lightlyanesthetized mice. After careful preprocessing ( Fig. S1 ), a functional connectivity matrix was calculated and decomposed into a set of onedimensional gradients, each capturing part of the variance in functional connectivity patterns across the cortex. Each gradient can be thought of as a spectrum of functional connectivity similarity: Cortical locations that have similar Gradient x values resemble each other in the aspect of functional connectivity that is captured in Gradient x . ...
Context 5
... principal gradient -or Gradient 1 -captures the highest amount of variance in the functional connectivity data ( Fig. S2 ). Upon visual examination, we noticed that its spatial distribution showed a marked similarity to previous representations of the dual origin organization in the mouse cortex ( Fig. 1 B, cf. Goulas et al., 2019b ). ...
Context 6
... et al., 2019b ). We therefore tested the hypothesis that the principal gradient reflects a spatial progression from the two postulated neocortical origins, the archicortex (hippocampus) and paleocortex (piriform area) ( Fig. 1 C). To this end, we computed the geodesic distance along the cortical surface from the hippocampus and piriform cortex. ...
Context 7
... distance along the cortical surface from the hippocampus and piriform cortex. First, we divided the cortex into a paleocortical and an archicortical zone, based on the minimal distance of every surface node to either origin. We found that Gradient 1 values differ significantly between the two zones (Wilcoxon rank-sum = 69.6, P corr = 0.01, Fig. 1 D). We then created a combined distance map by subtracting distance to archicortex from distance to paleocortex ( Fig 1 E). In this combined distance map, values close to zero indicate maximum distance to either origin, i.e. the transition between the two zones from the previous analysis. Large negative values indicate proximity to ...
Context 8
... corr = 0.01, Fig. 1 D). We then created a combined distance map by subtracting distance to archicortex from distance to paleocortex ( Fig 1 E). In this combined distance map, values close to zero indicate maximum distance to either origin, i.e. the transition between the two zones from the previous analysis. ...
Context 9
... investigated the intrinsic functional organization of the mouse cortex through resting-state functional connectivity gradients. The most robust organizational feature was found to be a prominent gradient reflecting the spatial distance from the presumed origins of cortical evolution ( Fig. 1 & S5 ). Several stable gradients represent highly specialized sensory modalities in the mouse cortex ( Fig. 3 ). ...
Context 10
... found that the principal gradient of functional connectivity in the mouse cortex reflects the spatial organization predicted by the dual origin theory ( Fig. 1 ). This finding resonates with a recent study demonstrating a fundamental division of the mouse cortex into an archicortical and a paleocortical zone based on tract-tracing data ( Goulas et al., 2019b ). ...
Citations
... The mouse scans were preprocessed as described in Huntenburg et al. (2020). Briefly, the anatomical scans were corrected for the B1-field inhomogeneity (ANTs, N4BiasFieldCorrection), denoised (ANTs, DenoiseImage), brain-masked (ANTs, antsBrainExtraction.sh) and, via the study template, registered to the Allen reference template (resampled to a 0.2 mm 3 resolution, ANTs, antsRegistration). ...
The subdivisions of the extended cingulate cortex of the human brain are implicated in a number of high-level behaviors and affected by a range of neuropsychiatric disorders. Its anatomy, function, and response to therapeutics are often studied using non-human animals, including the mouse. However, the similarity of human and mouse frontal cortex, including cingulate areas, is still not fully understood. Some accounts emphasize resemblances between mouse cingulate cortex and human cingulate cortex while others emphasize similarities with human granular prefrontal cortex. We use comparative neuroimaging to study the connectivity of the cingulate cortex in the mouse and human, allowing comparisons between mouse ‘gold standard’ tracer and imaging data, and, in addition, comparison between the mouse and the human using comparable imaging data. We find overall similarities in organization of the cingulate between species, including anterior and midcingulate areas and a retrosplenial area. However, human cingulate contains subareas with a more fine-grained organization than is apparent in the mouse and it has connections to prefrontal areas not present in the mouse. Results such as these help formally address between-species brain organization and aim to improve the translation from preclinical to human results.
... Les données souris (provenant de dix femelles C57Bl6) ont été fournies par un collaborateur (Huntenburg et al. 2021) et acquises sous anesthésie par injection sous-cutanée de médétomidine et bromure de pancuronium combinée à une anesthésie gazeuse (isoflurane à 0,5%), comme décrit dans Grandjean et al. (2014). Les images fonctionnelles ont été acquises dans une machine Bruker Biospec 94/30 (Bruker BioSpin MRI, Ettlingen, Germany) à 9,4T, grâce à une séquence d'imagerie écho-planaire à écho de gradient (GE-EPI) : temps de répétition (TR) = 1000 ms, temps d'echo (TE) = 10 ms, angle de bascule = 90°, champ de vision = 23,7×14 mm, 12 coupes d'une épaisseur de 0,5 mm, résolution = 263×233 µm, durée d'aquisition = 3,5 min). ...
Les réseaux cérébraux au repos, visibles en IRM fonctionnelle, sont les reflets de l’activité intrinsèque du cerveau et donnent de précieuses informations sur la fonction cérébrale saine et pathologique. L’étude de ces réseaux dans des modèles animaux pourrait, en améliorant la compréhension des différences anatomiques et fonctionnelles existant entre les espèces de mammifères, être à la base d’avancées substantielles dans la compréhension des maladies neurologiques et des bases de la fonction cérébrale chez l’humain. Cependant, la comparaison interspécifique de ces réseaux cérébraux est ardue car les atlas cérébraux, les protocoles d’acquisition et la résolution des IRM sont trop dissemblables. Elle nécessiterait donc un cadre d’étude plus rigoureux, qui pour le moment fait défaut. En effet, de nombreuses études ont décrit des réseaux homologues dans différentes espèces mammifères, mais très peu d’études interspécifiques ont à ce jour été réalisées. Notre étude est donc la première à extraire et comparer des réseaux cérébraux au repos et leurs sous-réseaux chez l’humain, la souris et le primate microcèbe. Nous avons pu observer de nombreuses similitudes entre les réseaux extraits dans les différentes espèces, mais également de notables divergences qu’il convient d’analyser à la lumière de ce que nous savons sur les capacités cérébrales de chacune d’entre elles.
... The overall diminished DMLN activity at 10 months also led to a breakdown of the LCN DMLN state flux reflected in the lack of presence of the LCN DMLN QPP in the zQ175DN HET group, alongside the decreased occurrence of DMLN. Human FC has been shown to rely on a particular hierarchy that is reflected in cortical gradients of the RSNs, which shows variation from unimodal regions on one end, such as S1Ctx and S2Ctx, which are hard-wired and have stronger structural-functional coupling, and transmodal regions on the other end, which underly multi-input integration, such as the DMN constituents, a major hub for incorporating cortical information [59][60][61][62] . This gradient cortical hierarchy was also found in mice, with similarities observed in both functional and long-range axonal properties 61,63,64 . ...
... Human FC has been shown to rely on a particular hierarchy that is reflected in cortical gradients of the RSNs, which shows variation from unimodal regions on one end, such as S1Ctx and S2Ctx, which are hard-wired and have stronger structural-functional coupling, and transmodal regions on the other end, which underly multi-input integration, such as the DMN constituents, a major hub for incorporating cortical information [59][60][61][62] . This gradient cortical hierarchy was also found in mice, with similarities observed in both functional and long-range axonal properties 61,63,64 . The transmodal DMLN regions play a central part in multimodal input, hence, they are highly adaptive to different types of information, therefore playing a crucial role in normal sensory processing and integration 62,63 . ...
Distinct resting-state networks (RSNs) are differentially altered in the course of Huntington’s disease (HD). However, these RSN changes are depicted using traditional functional connectivity analyses which ignore the dynamic brain states that constitute these RSNs and their time-dependent relationship. Dynamic states are represented by recurring spatiotemporal patterns of propagating cortical and subcortical brain activity observed in low-frequency BOLD fluctuations, called quasi-periodic patterns (QPPs). In this study, we used resting-state fMRI to investigate QPPs in the zQ175DN mouse model of HD at 3, 6 and 10 months of age. We identified age- and genotype-specific short (3s) QPPs, representative of the lateral cortical network (LCN) and the default mode-like network (DMLN), and a long (10s) QPP, the homolog of the human primary QPP, exhibiting the propagation of activity between the LCN and the DMLN. Hyperactivity was present in the caudate putamen and the somatosensory cortex in zQ175DN mice at 3 and 6 months of age. Moreover, DMLN-wide reduction in activation was observed at all ages, where at 6 and 10 months of age the reduced activity gradually advanced into a breakdown of the LCN-to-DMLN propagation. We then investigated the relationship in the timing of peak activity of six brain regions involved in the long QPPs and found that the retrosplenial cortex, a transmodal region which orchestrates multisensory integration, has a premature peak of BOLD activation in zQ175DN mice at 6 months of age, as compared to age-matched controls. Irrespective of either LCN or DMLN activation, this resulted in an asynchrony of the retrosplenial cortex in the peak timing relationships relative to other regions during the long (10s) QPPs. Finally, the normative, age-dependent, wild-type QPPs were significantly decreased in occurrence in the zQ175DN group at each age, indicating the presence of phenotypically-driven LCN and DMLN states as captured with QPPs. As BOLD-dependent variations result from neurovascular coupling, we assessed mutant huntingtin (mHTT) deposition in astrocytes and pericytes, known components of the neurogliovascular unit. These analyses showed increased cell-type dependent deposition starting at 6 months in the caudate putamen, somatosensory and motor cortex, regions that are prominently involved in HD pathology as seen in humans. Our findings provide meaningful insights into the development and progression of altered functional brain dynamics in this HD model, opening potential new avenues for its application in clinical HD research.
SUMMARY
Huntington’s disease (HD) is marked by irreversible loss of neuronal function for which currently no availability for disease-modifying treatment exists. Advances in the understanding of disease progression can aid biomarker development, which in turn can accelerate therapeutic discovery. We characterized the progression of altered dynamics of whole-brain network states in the zQ175DN mouse model of HD using a dynamic functional connectivity (FC) approach to resting-state fMRI and identified quasi-periodic patterns (QPPs) of brain activity constituting the most prominent resting-state networks. The occurrence of the normative QPPs, as observed in healthy controls, was reduced in the HD model as the phenotype progressed. This uncovered progressive cessation of synchronous brain activity with phenotypic progression, which is not observed with the conventional static FC approaches. This work opens new avenues in assessing the dynamics of whole brain states, through QPPs, in clinical HD research.
... In the thalamus, as with other cortical and subcortical structures, early circuit formation is scaffolded by the differential areal patterning of morphogenetic gradients during development [22][23][24] . These early developmental sequences are reflected by concerted variation of structural and functional properties along spatial axes in the adult cortex and subcortex [25][26][27][28][29][30][31][32] Indeed, studies have found evidence for gene expression gradients and variations in cytoarchitecture both across and within thalamic nuclei 13,14,20,21,33 . In the mouse, Phillips et al. observed that thalamic nuclei are arranged along an axis of gene expression that runs in a medial to lateral direction 14 . ...
The thalamus enables key sensory, motor, emotive, and cognitive processes via connections to the cortex. These projection patterns are traditionally considered to originate from discrete thalamic nuclei, however recent work showing gradients of molecular and connectivity features in the thalamus suggests the organisation of thalamocortical connections occurs along a continuous dimension. By performing a joint decomposition of densely sampled gene expression and non-invasive diffusion tractography in the adult human thalamus, we define a principal axis of genetic and connectomic variation along a medial-lateral thalamic gradient. Projections along this axis correspond to an anterior-posterior cortical pattern and are aligned with electrophysiological properties of the cortex. The medial-lateral axis demonstrates phylogenetic conservation, reflects transitions in neuronal subtypes, and shows associations with neurodevelopment and common brain disorders. This study provides evidence for a supra-nuclear axis of thalamocortical organisation characterised by a graded transition in molecular properties and anatomical connectivity.
... For instance, in the striatum, we report a rightward bias in the caudoputamen volumes, while there is a leftward bias for the putamen in humans. Beyond looking at structural markers, there is also evidence for functional asymmetries, including in autism spectrum disorders, for instance using functional gradients 44 , that would be interesting to investigate in corresponding animal models 45,46 . This could unlock additional comparisons between species and help to bring an understanding of functional asymmetries. ...
Hemispheric brain asymmetry is a basic organizational principle of the human brain and has been implicated in various psychiatric conditions, including autism spectrum disorder. Brain asymmetry is not a uniquely human feature and is observed in other species such as the mouse. Yet, asymmetry patterns are generally nuanced, and substantial sample sizes are required to detect these patterns. In this pre-registered study, we use a mouse dataset from the Province of Ontario Neurodevelopmental Network, which comprises structural MRI data from over 2000 mice, including genetic models for autism spectrum disorder, to reveal the scope and magnitude of hemispheric asymmetry in the mouse. Our findings demonstrate the presence of robust hemispheric asymmetry in the mouse brain, such as larger right hemispheric volumes towards the anterior pole and larger left hemispheric volumes toward the posterior pole, opposite to what has been shown in humans. This suggests the existence of species-specific traits. Further clustering analysis identified distinct asymmetry patterns in autism spectrum disorder models, a phenomenon that is also seen in atypically developing participants. Our study shows potential for the use of mouse models in studying the biological bases of typical and atypical brain asymmetry but also warrants caution as asymmetry patterns seem to differ between humans and mice.
... The mouse scans were preprocessed as described in Huntenburg et al. (2020). Briefly, the anatomical scans were corrected for the B1-field inhomogeneity, denoised, brain-masked and, via the study template, registered to the Allen reference template (resampled to a 0.2 mm 3 resolution). ...
The subdivisions of the extended cingulate cortex of the human brain are both implicated in a number of high-level behaviors and affected by a range of neuropsychiatric disorders. Its anatomy, function, and response to therapeutics are often studied using non-human animals, including the mouse. However, the similarity of human and mouse frontal cortex, including cingulate areas, is still not fully understood. Some accounts emphasize resemblances between mouse cingulate cortex and human cingulate cortex while others emphasize similarities with human granular prefrontal cortex. We use comparative neuroimaging to study the connectivity of the cingulate cortex in the mouse and human, allowing comparisons between mouse gold standard tracer and imaging data, and, in addition, comparison between the mouse and the human using comparable imaging data. We find overall similarities in organization of the cingulate between species, including anterior and midcingulate areas and a retrosplenial area. However, human cingulate contains subareas with a more fine-grained organization than is apparent in the mouse and it has connections to prefrontal areas not present in the mouse. Results such as these help formally address between-species brain organization with an aim to improve the translation from preclinical to human results.
... Here, we investigate mouse cortex, because it was found to differ from primate cortical organization: First, there seems to be no clear global processing hierarchy from sensory to transmodal areas [28], as more of mouse cortex is used for sensory processing of specific modalities, instead of integrating information across modalities, or across longer timescales for cognitive processing. Second, although there exists evidence for gradients in interneuron numbers and intra-cortical connectivity from sensory to transmodal areas [29], the degree of interareal variation of microstructural properties in mice [30,31] is far less pronounced than in the highly differentiated primate cortex [32,33,34,35]. ...
... Thus, our results might also support a grouping of areas into two parallel yet hierarchically organized pathways. This was similarly found in a functional connectivity study that suggested a sensory-to-motor and a transmodal pathway [28]. ...
A core challenge for information processing in the brain is to integrate information across various timescales. This could be achieved by a hierarchical organization of temporal processing, as reported for primates; however, it is open whether this hierarchical organization generalizes to sensory processing across species. Here, we studied signatures of temporal processing along the anatomical hierarchy in the mouse visual system. We found that the intrinsic and information timescales of spiking activity, which serve as proxies for how long information is stored in neural activity, increased along the anatomical hierarchy. Using information theory, we also quantified the predictability of neural spiking. The predictability is expected to be higher for longer integration of past information, but low for redundancy reduction in an efficient code. We found that predictability decreases along the anatomical cortical hierarchy, which is in line with efficient coding, but in contrast to the expectation of higher predictability for areas with higher timescales. Mechanistically, we could explain these results in a basic network model, where the increase in timescales arises from increasing network recurrence, while recurrence also reduces predictability if the model's input is correlated. The model thus suggests that timescales are mainly a network-intrinsic effect, whereas information-theoretic predictability depends on other sources such as (correlated) sensory stimuli. This is supported by a comparison of experimental data from different stimulus conditions. Our results show a clear hierarchy across mouse visual cortex, and thus suggest that hierarchical temporal processing presents a general organization principle across mammals.
... Taken together, the diverging representations of other gradients highlight the apparently central nature of the SA-axis in the functional organization of the brain. Our findings suggest that this central role is seemingly the pinnacle of evolutionary and developmental processes that shape the functional organization across lifespan (Bethlehem et al., 2020;Dong et al., 2021b;Larivière et al., 2020b;Nenning et al., 2020;Xia et al., 2022) and species (Huntenburg et al., 2021;Xu et al., 2020). ...
Low-dimensional representations are increasingly used to study meaningful organizational principles within the human brain. Most notably, the sensorimotor-association axis consistently explains the most variance in the human connectome as its so-called principal gradient, suggesting that it represents a fundamental organizational principle. While recent work indicates these low dimensional representations are relatively robust, they are limited by modeling only certain aspects of the functional connectivity structure. To date, the majority of studies have restricted these approaches to the strongest connections in the brain, treating weaker or negative connections as noise despite evidence of meaningful structure among them. The present work examines connectivity gradients of the human connectome across a full range of connectivity strengths and explores the implications for outcomes of individual differences, identifying potential dependencies on thresholds and opportunities to improve prediction tasks. Interestingly, the sensorimotor-association axis emerged as the principal gradient of the human connectome across the entire range of connectivity levels. Moreover, the principal gradient of connections at intermediate strengths encoded individual differences, better followed individual-specific anatomical features, and was also more predictive of intelligence. Taken together, our results add to evidence of the sensorimotor-association axis as a fundamental principle of the brain's functional organization, since it is evident even in the connectivity structure of more lenient connectivity thresholds. These more loosely coupled connections further appear to contain valuable and potentially important information that could be used to improve our understanding of individual differences, diagnosis, and the prediction of treatment outcomes.
... Given that both interhemispheric signal coupling (Fig. 1D) and the propagation of complexity drops (Fig. 2, E to G) intrinsically followed the principal unimodal-totransmodal hierarchy, we explicitly estimated the corresponding gradient loadings from the FC data and related them to the topology of complexity states. Furthermore, these gradient loadings have been shown to be spatially correlated with cortical myeloarchitecture as a proxy of anatomical hierarchy (22,36,37), yielding an important structure-function relationship within the connectome (38) that also partly extends to nonprimate mammalian brains (39,40). Thus, we estimated cortical myelination as the T1-weighted/T2weighted image ratio and related this myelin distribution to the functional gradient loadings and complexity states. ...
The human brain operates in large-scale functional networks. These networks are an expression of temporally correlated activity across brain regions, but how global network properties relate to the neural dynamics of individual regions remains incompletely understood. Here, we show that the brain's network architecture is tightly linked to critical episodes of neural regularity, visible as spontaneous "complexity drops" in functional magnetic resonance imaging signals. These episodes closely explain functional connectivity strength between regions, subserve the propagation of neural activity patterns, and reflect interindividual differences in age and behavior. Furthermore, complexity drops define neural activity states that dynamically shape the connectivity strength, topological configuration, and hierarchy of brain networks and comprehensively explain known structure-function relationships within the brain. These findings delineate a principled complexity architecture of neural activity-a human "complexome" that underpins the brain's functional network organization.
... In the thalamus, as with other cortical and subcortical structures, early circuit formation is scaffolded by the differential areal patterning of morphogenetic gradients during development [22][23][24] . These early developmental sequences are reflected by concerted variation of structural and functional properties along spatial axes in the adult cortex and subcortex [25][26][27][28][29][30][31][32] Indeed, studies have found evidence for gene expression gradients and variations in cytoarchitecture both across and within thalamic nuclei 13,14,20,21,33 . In the mouse, Phillips et al. observed that thalamic nuclei can be arranged along an axis of gene expression running along a medial to lateral direction 14 . ...
The thalamus enables key sensory, motor, emotive, and cognitive processes via connections to the cortex. These projection patterns are traditionally considered to originate from discrete thalamic nuclei, however recent work showing gradients of molecular and connectivity features in the thalamus suggests the organisation of thalamocortical connections occurs along a continuous dimension. By performing a joint decomposition of densely sampled gene expression and non-invasive diffusion tractography in the adult human thalamus, we define a principal axis of genetic and connectomic variation along a medial-lateral thalamic axis. Projections along this axis correspond to an anterior-posterior cortical pattern and are aligned with electrophysiological properties of the cortex. The medial-lateral axis demonstrates phylogenetic conservation, reflects transitions in neuronal subtypes, and shows associations with neurodevelopment and common brain disorders. This study provides evidence for a supra-nuclear axis of thalamocortical organisation characterised by a graded transition in molecular properties and anatomical connectivity.
