Cortical surface registration using unsupervised learning

Longitudinally consistent registration and parcellation of cortical surfaces using semi-supervised learning

Temporally consistent and accurate registration and parcellation of longitudinal cortical surfaces is of great importance in studying longitudinal morphological and functional changes of human brains. However, most existing methods are developed for registration or parcellation of a single cortical surface. When applying to longitudinal studies, these methods independently register/parcellate each surface from longitudinal scans, thus often generating longitudinally inconsistent and inaccurate results, especially in small or ambiguous cortical regions. Essentially, longitudinal cortical surface registration and parcellation are highly correlated tasks with inherently shared constraints on both spatial and temporal feature representations, which are unfortunately ignored in existing methods. To this end, we unprecedentedly propose a novel semi-supervised learning framework to exploit these inherent relationships from limited labeled data and extensive unlabeled data for more robust and consistent registration and parcellation of longitudinal cortical surfaces. Our method utilizes the spherical topology characteristic of cortical surfaces. It employs a spherical network to function as an encoder, which extracts high-level cortical features. Subsequently, we build two specialized decoders dedicated to the tasks of registration and parcellation, respectively. To extract more meaningful spatial features, we design a novel parcellation map similarity loss to utilize the relationship between registration and parcellation tasks, i.e., the parcellation map warped by the deformation field in registration should match the atlas parcellation map, thereby providing extra supervision for the registration task and augmented data for parcellation task by warping the atlas parcellation map to unlabeled surfaces. To enable temporally more consistent feature representation, we additionally enforce longitudinal consistency among longitudinal surfaces after registering them together using their concatenated features. Experiments on two longitudinal datasets of infants and adults have shown that our method achieves significant improvements on both registration/parcellation accuracy and longitudinal consistency compared to existing methods, especially in small and challenging cortical regions.

SUGAR: Spherical ultrafast graph attention framework for cortical surface registration

Cortical surface registration plays a crucial role in aligning cortical functional and anatomical features across individuals. However, conventional registration algorithms are computationally inefficient. Recently, learning-based registration algorithms have emerged as a promising solution, significantly improving processing efficiency. Nonetheless, there remains a gap in the development of a learning-based method that exceeds the state-of-the-art conventional methods simultaneously in computational efficiency, registration accuracy, and distortion control, despite the theoretically greater representational capabilities of deep learning approaches. To address the challenge, we present SUGAR, a unified unsupervised deep-learning framework for both rigid and non-rigid registration. SUGAR incorporates a U-Net-based spherical graph attention network and leverages the Euler angle representation for deformation. In addition to the similarity loss, we introduce fold and multiple distortion losses to preserve topology and minimize various types of distortions. Furthermore, we propose a data augmentation strategy specifically tailored for spherical surface registration to enhance the registration performance. Through extensive evaluation involving over 10,000 scans from 7 diverse datasets, we showed that our framework exhibits comparable or superior registration performance in accuracy, distortion, and test-retest reliability compared to conventional and learning-based methods. Additionally, SUGAR achieves remarkable sub-second processing times, offering a notable speed-up of approximately 12,000 times in registering 9,000 subjects from the UK Biobank dataset in just 32 min. This combination of high registration performance and accelerated processing time may greatly benefit large-scale neuroimaging studies.

A joint parcellation and boundary network with multi-rate-shared dilated graph attention for cortical surface parcellation

Cortical surface parcellation aims to segment the surface into anatomically and functionally significant regions, which are crucial for diagnosing and treating numerous neurological diseases. However, existing methods generally ignore the difficulty in learning labeling patterns of boundaries, hindering the performance of parcellation. To this end, this paper proposes a joint parcellation and boundary network (JPBNet) to promote the effectiveness of cortical surface parcellation. Its core is developing a multi-rate-shared dilated graph attention (MDGA) module and incorporating boundary learning into the parcellation process. The former, in particular, constructs a dilated graph attention strategy, extending the dilated convolution from regular data to irregular graph data. We fuse it with different dilated rates to extract context information in various scales by devising a shared graph attention layer. After that, a boundary enhancement module and a parcellation enhancement module based on graph attention mechanisms are built in each layer, forcing MDGA to capture informative and valuable features for boundary detection and parcellation tasks. Integrating MDGA, the boundary enhancement module, and the parcellation enhancement module at each layer to supervise boundary and parcellation information, an effective JPBNet is formed by stacking multiple layers. Experiments on the public dataset reveal that the proposed method outperforms comparison methods and performs well on boundaries for cortical surface parcellation.

Joint cortical registration of geometry and function using semi-supervised learning

Brain surface-based image registration, an important component of brain image analysis, establishes spatial correspondence between cortical surfaces. Existing iterative and learning-based approaches focus on accurate registration of folding patterns of the cerebral cortex, and assume that geometry predicts function and thus functional areas will also be well aligned. However, structure/functional variability of anatomically corresponding areas across subjects has been widely reported. In this work, we introduce a learning-based cortical registration framework, JOSA, which jointly aligns folding patterns and functional maps while simultaneously learning an optimal atlas. We demonstrate that JOSA can substantially improve registration performance in both anatomical and functional domains over existing methods. By employing a semi-supervised training strategy, the proposed framework obviates the need for functional data during inference, enabling its use in broad neuroscientific domains where functional data may not be observed. The source code of JOSA will be released to the public at https://voxelmorph.net.

Automatic landmark detection and registration of brain cortical surfaces via quasi-conformal geometry and convolutional neural networks

In medical imaging, surface registration is extensively used for performing systematic comparisons between anatomical structures, with a prime example being the highly convoluted brain cortical surfaces. To obtain a meaningful registration, a common approach is to identify prominent features on the surfaces and establish a low-distortion mapping between them with the feature correspondence encoded as landmark constraints. Prior registration works have primarily focused on using manually labeled landmarks and solving highly nonlinear optimization problems, which are time-consuming and hence hinder practical applications. In this work, we propose a novel framework for the automatic landmark detection and registration of brain cortical surfaces using quasi-conformal geometry and convolutional neural networks. We first develop a landmark detection network (LD-Net) that allows for the automatic extraction of landmark curves given two prescribed starting and ending points based on the surface geometry. We then utilize the detected landmarks and quasi-conformal theory for achieving the surface registration. Specifically, we develop a coefficient prediction network (CP-Net) for predicting the Beltrami coefficients associated with the desired landmark-based registration and a mapping network called the disk Beltrami solver network (DBS-Net) for generating quasi-conformal mappings from the predicted Beltrami coefficients, with the bijectivity guaranteed by quasi-conformal theory. Experimental results are presented to demonstrate the effectiveness of our proposed framework. Altogether, our work paves a new way for surface-based morphometry and medical shape analysis.

SF2Former: Amyotrophic lateral sclerosis identification from multi-center MRI data using spatial and frequency fusion transformer

Amyotrophic Lateral Sclerosis (ALS) is a complex neurodegenerative disorder characterized by motor neuron degeneration. Significant research has begun to establish brain magnetic resonance imaging (MRI) as a potential biomarker to diagnose and monitor the state of the disease. Deep learning has emerged as a prominent class of machine learning algorithms in computer vision and has shown successful applications in various medical image analysis tasks. However, deep learning methods applied to neuroimaging have not achieved superior performance in classifying ALS patients from healthy controls due to insignificant structural changes correlated with pathological features. Thus, a critical challenge in deep models is to identify discriminative features from limited training data. To address this challenge, this study introduces a framework called SF2Former, which leverages the power of the vision transformer architecture to distinguish ALS subjects from the control group by exploiting the long-range relationships among image features. Additionally, spatial and frequency domain information is combined to enhance the network's performance, as MRI scans are initially captured in the frequency domain and then converted to the spatial domain. The proposed framework is trained using a series of consecutive coronal slices and utilizes pre-trained weights from ImageNet through transfer learning. Finally, a majority voting scheme is employed on the coronal slices of each subject to generate the final classification decision. The proposed architecture is extensively evaluated with multi-modal neuroimaging data (i.e., T1-weighted, R2*, FLAIR) using two well-organized versions of the Canadian ALS Neuroimaging Consortium (CALSNIC) multi-center datasets. The experimental results demonstrate the superiority of the proposed strategy in terms of classification accuracy compared to several popular deep learning-based techniques.

Computing personalized brain functional networks from fMRI using self-supervised deep learning

A novel self-supervised deep learning (DL) method is developed to compute personalized brain functional networks (FNs) for characterizing brain functional neuroanatomy based on functional MRI (fMRI). Specifically, a DL model of convolutional neural networks with an encoder-decoder architecture is developed to compute personalized FNs directly from fMRI data. The DL model is trained to optimize functional homogeneity of personalized FNs without utilizing any external supervision in an end-to-end fashion. We demonstrate that a DL model trained on fMRI scans from the Human Connectome Project can identify personalized FNs and generalizes well across four different datasets. We further demonstrate that the identified personalized FNs are informative for predicting individual differences in behavior, brain development, and schizophrenia status. Taken together, the self-supervised DL allows for rapid, generalizable computation of personalized FNs.

Hierarchical particle optimization for cortical shape correspondence in temporal lobe resection

BACKGROUND

Anterior temporal lobe resection is an effective treatment for temporal lobe epilepsy. The post-surgical structural changes could influence the follow-up treatment. Capturing post-surgical changes necessitates a well-established cortical shape correspondence between pre- and post-surgical surfaces. Yet, most cortical surface registration methods are designed for normal neuroanatomy. Surgical changes can introduce wide ranging artifacts in correspondence, for which conventional surface registration methods may not work as intended.

METHODS

In this paper, we propose a novel particle method for one-to-one dense shape correspondence between pre- and post-surgical surfaces with temporal lobe resection. The proposed method can handle partial structural abnormality involving non-rigid changes. Unlike existing particle methods using implicit particle adjacency, we consider explicit particle adjacency to establish a smooth correspondence. Moreover, we propose hierarchical optimization of particles rather than full optimization of all particles at once to avoid trappings of locally optimal particle update.

RESULTS

We evaluate the proposed method on 25 pairs of T1-MRI with pre- and post-simulated resection on the anterior temporal lobe and 25 pairs of patients with actual resection. We show improved accuracy over several cortical regions in terms of ROI boundary Hausdorff distance with 4.29 mm and Dice similarity coefficients with average value 0.841, compared to existing surface registration methods on simulated data. In 25 patients with actual resection of the anterior temporal lobe, our method shows an improved shape correspondence in qualitative and quantitative evaluation on parcellation-off ratio with average value 0.061 and cortical thickness changes. We also show better smoothness of the correspondence without self-intersection, compared with point-wise matching methods which show various degrees of self-intersection.

CONCLUSION

The proposed method establishes a promising one-to-one dense shape correspondence for temporal lobe resection. The resulting correspondence is smooth without self-intersection. The proposed hierarchical optimization strategy could accelerate optimization and improve the optimization accuracy. According to the results on the paired surfaces with temporal lobe resection, the proposed method outperforms the compared methods and is more reliable to capture cortical thickness changes.

G-RMOS: GPU-accelerated Riemannian Metric Optimization on Surfaces

Surface mapping is used in various brain imaging studies, such as for mapping gray matter atrophy patterns in Alzheimer's disease. Riemannian metrics on surface (RMOS) is a state-of-the-art surface mapping algorithm that optimizes Riemannian metrics to establish one-to-one correspondences between surfaces in the Laplace-Beltrami embedding space. However, owing to the complex calculation with accurate one-to-one correspondences, RMOS registration takes a long time. In this study, we propose G-RMOS, a graphics processing unit (GPU)-accelerated RMOS registration pipeline that uses three GPU kernel design strategies: 1. using GPU computing capability with a batch scheme; 2. using the cache in the GPU block to minimize memory latency in register and shared memory; and 3. maximizing the effective number of instructions per GPU cycle using instruction level parallelism. Using the experimental results, we compare the acceleration speed of the G-RMOS framework with that of RMOS using hippocampus and cortical surfaces, and show that G-RMOS achieves a significant speedup in surface mapping. We also compare the memory requirements for cortical surface mapping and show that G-RMOS uses less memory than RMOS.

Home language and literacy environment and its relationship to socioeconomic status and white matter structure in infancy

The home language and literacy environment (HLLE) in infancy has been associated with subsequent pre-literacy skill development and HLLE at preschool-age has been shown to correlate with white matter organization in tracts that subserve pre-reading and reading skills. Furthermore, childhood socioeconomic status (SES) has been linked with both HLLE and white matter organization. It is important to understand whether the relationships between environmental factors such as HLLE and SES and white matter organization can be detected as early as infancy, as this period is characterized by rapid brain development that may make white matter pathways particularly susceptible to these early experiences. Here, we hypothesized that HLLE (1) relates to white matter organization in pre-reading and reading-related tracts in infants, and (2) mediates a link between SES and white matter organization. To test these hypotheses, infants (mean age: 8.6 ± 2.3 months, N = 38) underwent diffusion-weighted imaging MRI during natural sleep. Image processing was performed with an infant-specific pipeline and fractional anisotropy (FA) was estimated from the arcuate fasciculus (AF) and superior longitudinal fasciculus (SLF) bilaterally using the baby automated fiber quantification method. HLLE was measured with the Reading subscale of the StimQ (StimQ-Reading) and SES was measured with years of maternal education. Self-reported maternal reading ability was also quantified and applied to our statistical models as a proxy for confounding genetic effects. StimQ-Reading positively correlated with FA in left AF and to maternal education, but did not mediate the relationship between them. Taken together, these findings underscore the importance of considering HLLE from the start of life and may inform novel prevention and intervention strategies to support developing infants during a period of heightened brain plasticity.

Fast cortical surface reconstruction from MRI using deep learning

Reconstructing cortical surfaces from structural magnetic resonance imaging (MRI) is a prerequisite for surface-based functional and anatomical image analyses. Conventional algorithms for cortical surface reconstruction are computationally inefficient and typically take several hours for each subject, causing a bottleneck in applications when a fast turnaround time is needed. To address this challenge, we propose a fast cortical surface reconstruction (FastCSR) pipeline by leveraging deep machine learning. We trained our model to learn an implicit representation of the cortical surface in volumetric space, termed the “level set representation”. A fast volumetric topology correction method and a topology-preserving surface mesh extraction procedure were employed to reconstruct the cortical surface based on the level set representation. Using 1-mm isotropic T1-weighted images, the FastCSR pipeline was able to reconstruct a subject’s cortical surfaces within 5 min with comparable surface quality, which is approximately 47 times faster than the traditional FreeSurfer pipeline. The advantage of FastCSR becomes even more apparent when processing high-resolution images. Importantly, the model demonstrated good generalizability in previously unseen data and showed high test–retest reliability in cortical morphometrics and anatomical parcellations. Finally, FastCSR was robust to images with compromised quality or with distortions caused by lesions. This fast and robust pipeline for cortical surface reconstruction may facilitate large-scale neuroimaging studies and has potential in clinical applications wherein brain images may be compromised.

Unified Deep Learning-Based Mouse Brain MR Segmentation: Template-Based Individual Brain Positron Emission Tomography Volumes-of-Interest Generation Without Spatial Normalization in Mouse Alzheimer Model

Although skull-stripping and brain region segmentation are essential for precise quantitative analysis of positron emission tomography (PET) of mouse brains, deep learning (DL)-based unified solutions, particularly for spatial normalization (SN), have posed a challenging problem in DL-based image processing. In this study, we propose an approach based on DL to resolve these issues. We generated both skull-stripping masks and individual brain-specific volumes-of-interest (VOIs—cortex, hippocampus, striatum, thalamus, and cerebellum) based on inverse spatial normalization (iSN) and deep convolutional neural network (deep CNN) models. We applied the proposed methods to mutated amyloid precursor protein and presenilin-1 mouse model of Alzheimer’s disease. Eighteen mice underwent T2-weighted MRI and ¹⁸F FDG PET scans two times, before and after the administration of human immunoglobulin or antibody-based treatments. For training the CNN, manually traced brain masks and iSN-based target VOIs were used as the label. We compared our CNN-based VOIs with conventional (template-based) VOIs in terms of the correlation of standardized uptake value ratio (SUVR) by both methods and two-sample t-tests of SUVR % changes in target VOIs before and after treatment. Our deep CNN-based method successfully generated brain parenchyma mask and target VOIs, which shows no significant difference from conventional VOI methods in SUVR correlation analysis, thus establishing methods of template-based VOI without SN.

Anatomically constrained squeeze-and-excitation graph attention network for cortical surface parcellation

In order to understand the organizational structures of healthy cerebral cortex and the abnormalities in neurological and psychiatric diseases, it is significant to parcellate the cortical surface. The cortical surface, however, is a highly folded complex geometric structure which challenges automatic cortical surface parcellation. Nowadays, the parcellation methods of cerebral cortex are mostly based on geometric simplification, i.e., iteratively inflating and mapping the cortical surface to a spherical surface for processing, which is time-consuming and cannot make full use of the intrinsic structural information of the original cortical surface. In this study, we proposed an anatomically constrained squeeze-and-excitation graph attention network (ASEGAT) for an end-to-end brain cortical surface parcellation on the original cortical surface manifold. The ASEGAT is formed by two graph attention modules and a squeeze-and-excitation module that incorporate self-attention and head attention for rendering features of each node. Furthermore, we designed an anatomic constraint loss to introduce the anatomical priori of regional adjacency relationships, which could improve the consistency of region labeling. We evaluated our model on a public dataset of 100 manually labeled brain surfaces. Compared with several advanced methods, the results showed that our proposed approach achieved state-of-the-art performance, obtaining an accuracy of 90.65% and a dice score of 89.00%.

Machine learning in neuroimaging: from research to clinical practice

Neuroimaging is critical in clinical care and research, enabling us to investigate the brain in health and disease. There is a complex link between the brain's morphological structure, physiological architecture, and the corresponding imaging characteristics. The shape, function, and relationships between various brain areas change during development and throughout life, disease, and recovery. Like few other areas, neuroimaging benefits from advanced analysis techniques to fully exploit imaging data for studying the brain and its function. Recently, machine learning has started to contribute (a) to anatomical measurements, detection, segmentation, and quantification of lesions and disease patterns, (b) to the rapid identification of acute conditions such as stroke, or (c) to the tracking of imaging changes over time. As our ability to image and analyze the brain advances, so does our understanding of its intricate relationships and their role in therapeutic decision-making. Here, we review the current state of the art in using machine learning techniques to exploit neuroimaging data for clinical care and research, providing an overview of clinical applications and their contribution to fundamental computational neuroscience.

Mapping the subcortical connectivity of the human default mode network

The default mode network (DMN) mediates self-awareness and introspection, core components of human consciousness. Therapies to restore consciousness in patients with severe brain injuries have historically targeted subcortical sites in the brainstem, thalamus, hypothalamus, basal forebrain, and basal ganglia, with the goal of reactivating cortical DMN nodes. However, the subcortical connectivity of the DMN has not been fully mapped, and optimal subcortical targets for therapeutic neuromodulation of consciousness have not been identified. In this work, we created a comprehensive map of DMN subcortical connectivity by combining high-resolution functional and structural datasets with advanced signal processing methods. We analyzed 7 Tesla resting-state functional MRI (rs-fMRI) data from 168 healthy volunteers acquired in the Human Connectome Project. The rs-fMRI blood-oxygen-level-dependent (BOLD) data were temporally synchronized across subjects using the BrainSync algorithm. Cortical and subcortical DMN nodes were jointly analyzed and identified at the group level by applying a novel Nadam-Accelerated SCAlable and Robust (NASCAR) tensor decomposition method to the synchronized dataset. The subcortical connectivity map was then overlaid on a 7 Tesla 100 µm ex vivo MRI dataset for neuroanatomic analysis using automated segmentation of nuclei within the brainstem, thalamus, hypothalamus, basal forebrain, and basal ganglia. We further compared the NASCAR subcortical connectivity map with its counterpart generated from canonical seed-based correlation analyses. The NASCAR method revealed that BOLD signal in the central lateral nucleus of the thalamus and ventral tegmental area of the midbrain is strongly correlated with that of the DMN. In an exploratory analysis, additional subcortical sites in the median and dorsal raphe, lateral hypothalamus, and caudate nuclei were correlated with the cortical DMN. We also found that the putamen and globus pallidus are negatively correlated (i.e., anti-correlated) with the DMN, providing rs-fMRI evidence for the mesocircuit hypothesis of human consciousness, whereby a striatopallidal feedback system modulates anterior forebrain function via disinhibition of the central thalamus. Seed-based analyses yielded similar subcortical DMN connectivity, but the NASCAR result showed stronger contrast and better spatial alignment with dopamine immunostaining data. The DMN subcortical connectivity map identified here advances understanding of the subcortical regions that contribute to human consciousness and can be used to inform the selection of therapeutic targets in clinical trials for patients with disorders of consciousness.

S3Reg: Superfast Spherical Surface Registration Based on Deep Learning

Cortical surface registration is an essential step and prerequisite for surface-based neuroimaging analysis. It aligns cortical surfaces across individuals and time points to establish cross-sectional and longitudinal cortical correspondences to facilitate neuroimaging studies. Though achieving good performance, available methods are either time consuming or not flexible to extend to multiple or high dimensional features. Considering the explosive availability of large-scale and multimodal brain MRI data, fast surface registration methods that can flexibly handle multimodal features are desired. In this study, we develop a Superfast Spherical Surface Registration (S3Reg) framework for the cerebral cortex. Leveraging an end-to-end unsupervised learning strategy, S3Reg offers great flexibility in the choice of input feature sets and output similarity measures for registration, and meanwhile reduces the registration time significantly. Specifically, we exploit the powerful learning capability of spherical Convolutional Neural Network (CNN) to directly learn the deformation fields in spherical space and implement diffeomorphic design with "scaling and squaring" layers to guarantee topology-preserving deformations. To handle the polar-distortion issue, we construct a novel spherical CNN model using three orthogonal Spherical U-Nets. Experiments are performed on two different datasets to align both adult and infant multimodal cortical features. Results demonstrate that our S3Reg shows superior or comparable performance with state-of-the-art methods, while improving the registration time from 1 min to 10 sec.