A Survey on Deep Learning for Neuroimaging-Based Brain Disorder Analysis

An objective quantitative diagnosis of depression using a local-to-global multimodal fusion graph neural network

This study developed an artificial intelligence (AI) system using a local-global multimodal fusion graph neural network (LGMF-GNN) to address the challenge of diagnosing major depressive disorder (MDD), a complex disease influenced by social, psychological, and biological factors. Utilizing functional MRI, structural MRI, and electronic health records, the system offers an objective diagnostic method by integrating individual brain regions and population data. Tested across cohorts from China, Japan, and Russia with 1,182 healthy controls and 1,260 MDD patients from 24 institutions, it achieved a classification accuracy of 78.75%, an area under the receiver operating characteristic curve (AUROC) of 80.64%, and correctly identified MDD subtypes. The system further discovered distinct brain connectivity patterns in MDD, including reduced functional connectivity between the left gyrus rectus and right cerebellar lobule VIIB, and increased connectivity between the left Rolandic operculum and right hippocampus. Anatomically, MDD is associated with thickness changes of the gray and white matter interface, indicating potential neuropathological conditions or brain injuries.

Extracting interpretable signatures of whole-brain dynamics through systematic comparison

The brain's complex distributed dynamics are typically quantified using a limited set of manually selected statistical properties, leaving the possibility that alternative dynamical properties may outperform those reported for a given application. Here, we address this limitation by systematically comparing diverse, interpretable features of both intra-regional activity and inter-regional functional coupling from resting-state functional magnetic resonance imaging (rs-fMRI) data, demonstrating our method using case-control comparisons of four neuropsychiatric disorders. Our findings generally support the use of linear time-series analysis techniques for rs-fMRI case-control analyses, while also identifying new ways to quantify informative dynamical fMRI structures. While simple statistical representations of fMRI dynamics performed surprisingly well (e.g., properties within a single brain region), combining intra-regional properties with inter-regional coupling generally improved performance, underscoring the distributed, multifaceted changes to fMRI dynamics in neuropsychiatric disorders. The comprehensive, data-driven method introduced here enables systematic identification and interpretation of quantitative dynamical signatures of multivariate time-series data, with applicability beyond neuroimaging to diverse scientific problems involving complex time-varying systems.

Combined graph convolutional networks with a multi-connection pattern to identify tremor-dominant Parkinson's disease and Essential tremor with resting tremor

Essential tremor with resting tremor (rET) and tremor-dominant Parkinson's disease (tPD) share many similar clinical symptoms, leading to frequent misdiagnoses. Functional connectivity (FC) matrix analysis derived from resting-state functional MRI (Rs-fMRI) offers a promising approach for early diagnosis and for exploring FC network pathogenesis in rET and tPD. However, methods relying solely on a single connection pattern may overlook the complementary roles of different connectivity patterns, resulting in reduced diagnostic differentiation. Therefore, we propose a multi-pattern connection Graph Convolutional Network (MCGCN) method to integrate information from various connection modes, distinguishing between rET and healthy controls (HC), tPD and HC, and rET and tPD. We constructed FC matrices using three different connectivity modes for each subject and used these as inputs to the MCGCN model for disease classification. The classification performance of the model was evaluated for each connectivity mode. Subsequently, gradient-weighted class activation mapping (Grad-CAM) was used to identify the most discriminative brain regions. The important brain regions identified were primarily distributed within cerebellar-motor and non-motor cortical networks. Compared with single-pattern GCN, our proposed MCGCN model demonstrated superior classification accuracy, underscoring the advantages of integrating multiple connectivity modes. Specifically, the model achieved an average accuracy of 88.0% for distinguishing rET from HC, 88.8% for rET from tPD, and 89.6% for tPD from HC. Our findings indicate that combining graph convolutional networks with multi-connection patterns can not only effectively discriminate between tPD, rET, and HC but also enhance our understanding of the functional network mechanisms underlying rET and tPD.

Transformer-based approaches for neuroimaging: an in-depth review of their role in classification and regression tasks

In the ever-evolving landscape of deep learning (DL), the transformer model emerges as a formidable neural network architecture, gaining significant traction in neuroimaging-based classification and regression tasks. This paper presents an extensive examination of transformer's application in neuroimaging, surveying recent literature to elucidate its current status and research advancement. Commencing with an exposition on the fundamental principles and structures of the transformer model and its variants, this review navigates through the methodologies and experimental findings pertaining to their utilization in neuroimage classification and regression tasks. We highlight the transformer model's prowess in neuroimaging, showcasing its exceptional performance in classification endeavors while also showcasing its burgeoning potential in regression tasks. Concluding with an assessment of prevailing challenges and future trajectories, this paper proffers insights into prospective research directions. By elucidating the current landscape and envisaging future trends, this review enhances comprehension of transformer's role in neuroimaging tasks, furnishing valuable guidance for further inquiry.

A multi-slice attention fusion and multi-view personalized fusion lightweight network for Alzheimer's disease diagnosis

OBJECTIVE

Alzheimer's disease (AD) is a type of neurological illness that significantly impacts individuals' daily lives. In the intelligent diagnosis of AD, 3D networks require larger computational resources and storage space for training the models, leading to increased model complexity and training time. On the other hand, 2D slices analysis may overlook the 3D structural information of MRI and can result in information loss.

APPROACH

We propose a multi-slice attention fusion and multi-view personalized fusion lightweight network for automated AD diagnosis. It incorporates a multi-branch lightweight backbone to extract features from sagittal, axial, and coronal view of MRI, respectively. In addition, we introduce a novel multi-slice attention fusion module, which utilizes a combination of global and local channel attention mechanism to ensure consistent classification across multiple slices. Additionally, a multi-view personalized fusion module is tailored to assign appropriate weights to the three views, taking into account the varying significance of each view in achieving accurate classification results. To enhance the performance of the multi-view personalized fusion module, we utilize a label consistency loss to guide the model's learning process. This encourages the acquisition of more consistent and stable representations across all three views.

MAIN RESULTS

The suggested strategy is efficient in lowering the number of parameters and FLOPs, with only 3.75M and 4.45G respectively, and accuracy improved by 10.5% to 14% in three tasks. Moreover, in the classification tasks of AD vs. CN, AD vs. MCI and MCI vs. CN, the accuracy of the proposed method is 95.63%, 86.88% and 85.00%, respectively, which is superior to the existing methods.

CONCLUSIONS

The results show that the proposed approach not only excels in resource utilization, but also significantly outperforms the four comparison methods in terms of accuracy and sensitivity, particularly in detecting early-stage AD lesions. It can precisely capture and accurately identify subtle brain lesions, providing crucial technical support for early intervention and treatment.

Common and unique brain aging patterns between females and males quantified by large-scale deep learning

There has been extensive evidence that aging affects human brain function. However, there is no complete picture of what brain functional changes are mostly related to normal aging and how aging affects brain function similarly and differently between males and females. Based on resting-state brain functional connectivity (FC) of 25,582 healthy participants (13,373 females) aged 49-76 years from the UK Biobank project, we employ deep learning with explainable AI to discover primary FCs related to progressive aging and reveal similarity and difference between females and males in brain aging. Using a nested cross-validation scheme, we conduct 4200 deep learning models to classify all paired age groups on the main data for females and males separately and then extract gender-common and gender-specific aging-related FCs. Next, we validate those FCs using additional 21,000 classifiers on the independent data. Our results support that aging results in reduced brain functional interactions for both females and males, primarily relating to the positive connectivity within the same functional domain and the negative connectivity between different functional domains. Regions linked to cognitive control show the most significant age-related changes in both genders. Unique aging effects in males and females mainly involve the interaction between cognitive control and the default mode, vision, auditory, and frontoparietal domains. Results also indicate females exhibit faster brain functional changes than males. Overall, our study provides new evidence about common and unique patterns of brain aging in females and males.

Neuroimage analysis using artificial intelligence approaches: a systematic review

In the contemporary era, artificial intelligence (AI) has undergone a transformative evolution, exerting a profound influence on neuroimaging data analysis. This development has significantly elevated our comprehension of intricate brain functions. This study investigates the ramifications of employing AI techniques on neuroimaging data, with a specific objective to improve diagnostic capabilities and contribute to the overall progress of the field. A systematic search was conducted in prominent scientific databases, including PubMed, IEEE Xplore, and Scopus, meticulously curating 456 relevant articles on AI-driven neuroimaging analysis spanning from 2013 to 2023. To maintain rigor and credibility, stringent inclusion criteria, quality assessments, and precise data extraction protocols were consistently enforced throughout this review. Following a rigorous selection process, 104 studies were selected for review, focusing on diverse neuroimaging modalities with an emphasis on mental and neurological disorders. Among these, 19.2% addressed mental illness, and 80.7% focused on neurological disorders. It is found that the prevailing clinical tasks are disease classification (58.7%) and lesion segmentation (28.9%), whereas image reconstruction constituted 7.3%, and image regression and prediction tasks represented 9.6%. AI-driven neuroimaging analysis holds tremendous potential, transforming both research and clinical applications. Machine learning and deep learning algorithms outperform traditional methods, reshaping the field significantly.

Union is strength: the combination of radiomics features and 3D-deep learning in a sole model increases diagnostic accuracy in demented patients: a whole brain 18FDG PET-CT analysis

OBJECTIVE

FDG PET imaging plays a crucial role in the evaluation of demented patients by assessing regional cerebral glucose metabolism. In recent years, both radiomics and deep learning techniques have emerged as powerful tools for extracting valuable information from medical images. This article aims to provide a comparative analysis of radiomics features, 3D-deep learning convolutional neural network (CNN) and the fusion of them, in the evaluation of 18F-FDG PET whole brain images in patients with dementia and normal controls.

METHODS

18F-FDG brain PET and clinical score were collected in 85 patients with dementia and 125 healthy controls (HC). Patients were assigned to various form of dementia on the basis of clinical evaluation, follow-up and voxels comparison with HC using a two-sample Student's t -test, to determine the regions of brain involved. Radiomics analysis was performed on the whole brain after normalization to an optimized template. After selection using the minimum redundancy maximum relevance method and Pearson's correlation coefficients, the features obtained were added to a neural network model to find the accuracy in classifying HC and demented patients. Forty subjects not included in the training were used to test the models. The results of the three models (radiomics, 3D-CNN, combined model) were compared with each other.

RESULTS

Four radiomics features were selected. The sensitivity was 100% for the three models, but the specificity was higher with radiomics and combined one (100% vs. 85%). Moreover, the classification scores were significantly higher using the combined model in both normal and demented subjects.

CONCLUSION

The combination of radiomics features and 3D-CNN in a single model, applied to the whole brain 18FDG PET study, increases the accuracy in demented patients.

Extracting interpretable signatures of whole-brain dynamics through systematic comparison

The brain's complex distributed dynamics are typically quantified using a limited set of manually selected statistical properties, leaving the possibility that alternative dynamical properties may outperform those reported for a given application. Here, we address this limitation by systematically comparing diverse, interpretable features of both intra-regional activity and inter-regional functional coupling from resting-state functional magnetic resonance imaging (rs-fMRI) data, demonstrating our method using case-control comparisons of four neuropsychiatric disorders. Our findings generally support the use of linear time-series analysis techniques for rs-fMRI case-control analyses, while also identifying new ways to quantify informative dynamical fMRI structures. While simple statistical representations of fMRI dynamics performed surprisingly well (e.g., properties within a single brain region), combining intra-regional properties with inter-regional coupling generally improved performance, underscoring the distributed, multifaceted changes to fMRI dynamics in neuropsychiatric disorders. The comprehensive, data-driven method introduced here enables systematic identification and interpretation of quantitative dynamical signatures of multivariate time-series data, with applicability beyond neuroimaging to diverse scientific problems involving complex time-varying systems.

A within-subject voxel-wise constant-block partial least squares correlation method to explore MRI-based brain structure-function relationship

The brain structure-function relationship is crucial to how the human brain works under normal or diseased conditions. Exploring such a relationship is challenging when using the 3-dimensional magnetic resonance imaging (MRI) functional dataset which is temporal dynamic and the structural MRI which is static. Partial Least Squares Correlation (PLSC) is one of the classical methods for exploring the joint spatial and temporal relationship. The goal of PLSC is to identify covarying patterns via linear voxel-wise combinations in each of the structural and functional data sets to maximize the covariance. However, existing PLSC cannot adequately deal with the unmatched temporal dimensions between structural and functional data sets. We proposed a new alternative variant of the PLSC, termed within-subject, voxel-wise, and constant-block PLSC, to address this problem. To validate our method, we used two data sets with weak and strong relationships in simulated data. Additionally, the analysis of real brain data was carried out based on gray matter volume hubs derived from sMRI and whole-brain voxel-wise measures from resting-state fMRI for aging effect based on healthy subjects aged 16-85 years. Our results showed that our constant-block PLSC can detect weak structure-function relationships and has better robustness to noise. In fact, it adequately unearthed the true simulated number of significant and more accurate latent variables for the simulated data and more meaningful LVs for the real data, with covariance improvement from 16.19 to 41.48% (simulated) and 13.29-53.68% (real data), respectively. More interestingly in the real data analysis, our method identified simultaneously the well-known brain networks such as the default mode, sensorimotor, auditory, and dorsal attention networks both functionally and structurally, implying the hubs we derived from gray matter volumes are the basis of brain function, supporting diverse functions. Constant-block PLSC is a feasible tool for analyzing the brain structure-function relationship.

Understanding the Role of Self-Attention in a Transformer Model for the Discrimination of SCD From MCI Using Resting-State EEG

The identification of EEG biomarkers to discriminate Subjective Cognitive Decline (SCD) from Mild Cognitive Impairment (MCI) conditions is a complex task which requires great clinical effort and expertise. We exploit the self-attention component of the Transformer architecture to obtain physiological explanations of the model's decisions in the discrimination of 56 SCD and 45 MCI patients using resting-state EEG. Specifically, an interpretability workflow leveraging attention scores and time-frequency analysis of EEG epochs through Continuous Wavelet Transform is proposed. In the classification framework, models are trained and validated with 5-fold cross-validation and evaluated on a test set obtained by selecting 20% of the total subjects. Ablation studies and hyperparameter tuning tests are conducted to identify the optimal model configuration. Results show that the best performing model, which achieves acceptable results both on epochs' and patients' classification, is capable of finding specific EEG patterns that highlight changes in the brain activity between the two conditions. We demonstrate the potential of attention weights as tools to guide experts in understanding which disease-relevant EEG features could be discriminative of SCD and MCI.

A confounder controlled machine learning approach: Group analysis and classification of schizophrenia and Alzheimer's disease using resting-state functional network connectivity

Resting-state functional magnetic resonance imaging (rs-fMRI) has increasingly been used to study both Alzheimer's disease (AD) and schizophrenia (SZ). While most rs-fMRI studies being conducted in AD and SZ compare patients to healthy controls, it is also of interest to directly compare AD and SZ patients with each other to identify potential biomarkers shared between the disorders. However, comparing patient groups collected in different studies can be challenging due to potential confounds, such as differences in the patient's age, scan protocols, etc. In this study, we compared and contrasted resting-state functional network connectivity (rs-FNC) of 162 patients with AD and late mild cognitive impairment (LMCI), 181 schizophrenia patients, and 315 cognitively normal (CN) subjects. We used confounder-controlled rs-FNC and applied machine learning algorithms (including support vector machine, logistic regression, random forest, and k-nearest neighbor) and deep learning models (i.e., fully-connected neural networks) to classify subjects in binary and three-class categories according to their diagnosis labels (e.g., AD, SZ, and CN). Our statistical analysis revealed that FNC between the following network pairs is stronger in AD compared to SZ: subcortical-cerebellum, subcortical-cognitive control, cognitive control-cerebellum, and visual-sensory motor networks. On the other hand, FNC is stronger in SZ than AD for the following network pairs: subcortical-visual, subcortical-auditory, subcortical-sensory motor, cerebellum-visual, sensory motor-cognitive control, and within the cerebellum networks. Furthermore, we observed that while AD and SZ disorders each have unique FNC abnormalities, they also share some common functional abnormalities that can be due to similar neurobiological mechanisms or genetic factors contributing to these disorders' development. Moreover, we achieved an accuracy of 85% in classifying subjects into AD and SZ where default mode, visual, and subcortical networks contributed the most to the classification and accuracy of 68% in classifying subjects into AD, SZ, and CN with the subcortical domain appearing as the most contributing features to the three-way classification. Finally, our findings indicated that for all classification tasks, except AD vs. SZ, males are more predictable than females.

Speeding Up and Improving Image Quality in Glioblastoma MRI Protocol by Deep Learning Image Reconstruction

A fully diagnostic MRI glioma protocol is key to monitoring therapy assessment but is time-consuming and especially challenging in critically ill and uncooperative patients. Artificial intelligence demonstrated promise in reducing scan time and improving image quality simultaneously. The purpose of this study was to investigate the diagnostic performance, the impact on acquisition acceleration, and the image quality of a deep learning optimized glioma protocol of the brain. Thirty-three patients with histologically confirmed glioblastoma underwent standardized brain tumor imaging according to the glioma consensus recommendations on a 3-Tesla MRI scanner. Conventional and deep learning-reconstructed (DLR) fluid-attenuated inversion recovery, and T2- and T1-weighted contrast-enhanced Turbo spin echo images with an improved in-plane resolution, i.e., super-resolution, were acquired. Two experienced neuroradiologists independently evaluated the image datasets for subjective image quality, diagnostic confidence, tumor conspicuity, noise levels, artifacts, and sharpness. In addition, the tumor volume was measured in the image datasets according to Response Assessment in Neuro-Oncology (RANO) 2.0, as well as compared between both imaging techniques, and various clinical-pathological parameters were determined. The average time saving of DLR sequences was 30% per MRI sequence. Simultaneously, DLR sequences showed superior overall image quality (all p < 0.001), improved tumor conspicuity and image sharpness (all p < 0.001, respectively), and less image noise (all p < 0.001), while maintaining diagnostic confidence (all p > 0.05), compared to conventional images. Regarding RANO 2.0, the volume of non-enhancing non-target lesions (p = 0.963), enhancing target lesions (p = 0.993), and enhancing non-target lesions (p = 0.951) did not differ between reconstruction types. The feasibility of the deep learning-optimized glioma protocol was demonstrated with a 30% reduction in acquisition time on average and an increased in-plane resolution. The evaluated DLR sequences improved subjective image quality and maintained diagnostic accuracy in tumor detection and tumor classification according to RANO 2.0.

Pose-based tremor type and level analysis for Parkinson's disease from video

PURPOSE

Current methods for diagnosis of PD rely on clinical examination. The accuracy of diagnosis ranges between 73 and 84%, and is influenced by the experience of the clinical assessor. Hence, an automatic, effective and interpretable supporting system for PD symptom identification would support clinicians in making more robust PD diagnostic decisions.

METHODS

We propose to analyze Parkinson's tremor (PT) to support the analysis of PD, since PT is one of the most typical symptoms of PD with broad generalizability. To realize the idea, we present SPA-PTA, a deep learning-based PT classification and severity estimation system that takes consumer-grade videos of front-facing humans as input. The core of the system is a novel attention module with a lightweight pyramidal channel-squeezing-fusion architecture that effectively extracts relevant PT information and filters noise. It enhances modeling performance while improving system interpretability.

RESULTS

We validate our system via individual-based leave-one-out cross-validation on two tasks: the PT classification task and the tremor severity rating estimation task. Our system presents a 91.3% accuracy and 80.0% F1-score in classifying PT with non-PT class, while providing a 76.4% accuracy and 76.7% F1-score in more complex multiclass tremor rating classification task.

CONCLUSION

Our system offers a cost-effective PT classification and tremor severity estimation results as warning signs of PD for undiagnosed patients with PT symptoms. In addition, it provides a potential solution for supporting PD diagnosis in regions with limited clinical resources.

Artificial intelligence for risk analysis-A risk characterization perspective on advances, opportunities, and limitations

Artificial intelligence (AI) has seen numerous applications for risk analysis and provides ample opportunities for developing new and improved methods and models for this purpose. In the present article, we conceptualize the use of AI for risk analysis by framing it as an input-algorithm-output process and linking such a setup to three tasks in establishing a risk description: consequence characterization, uncertainty characterization, and knowledge management. We then give an overview of currently used concepts and methods for AI-based risk analysis and outline potential future uses by extrapolating beyond currently produced types of output. We end with a discussion of the limits of automation, both near-term limitations and a more fundamental question related to allowing AI to automatically prescribe risk management decisions. We conclude that there are opportunities for using AI for risk analysis to a greater extent than is commonly the case today; however, critical concerns about proper uncertainty representation and the need for risk-informed rather than risk-based decision-making also lead us to conclude that risk analysis and decision-making processes cannot be fully automated.

Greater accuracy of radiomics compared to deep learning to discriminate normal subjects from patients with dementia: a whole brain 18FDG PET analysis

METHODS

18F-FDG brain PET and clinical score were collected in 85 patients with dementia and 125 healthy controls (HC). Patients were assigned to various form of dementia on the basis of clinical evaluation, follow-up and voxels comparison with HC using a two-sample Student's t -test, to determine the regions of brain involved. Radiomic analysis was performed on the whole brain after normalization to an optimized template. After feature selection using the minimum redundancy maximum relevance method and Pearson's correlation coefficients, a Neural Network model was tested to find the accuracy to classify HC and demented patients. Twenty subjects not included in the training were used to test the models. The results were compared with those obtained by conventional CNN model.

RESULTS

Four radiomic features were selected. The validation and test accuracies were 100% for both models, but the probability scores were higher with radiomics, in particular for HC group ( P  = 0.0004).

CONCLUSION

Radiomic features extracted from standardized PET whole brain images seem to be more accurate than CNN to distinguish patients with and without dementia.

Deep learning-accelerated image reconstruction in back pain-MRI imaging: reduction of acquisition time and improvement of image quality

INTRODUCTION

Low back pain is a global health issue causing disability and missed work days. Commonly used MRI scans including T1-weighted and T2-weighted images provide detailed information of the spine and surrounding tissues. Artificial intelligence showed promise in improving image quality and simultaneously reducing scan time. This study evaluates the performance of deep learning (DL)-based T2 turbo spin-echo (TSE, T2DLR) and T1 TSE (T1DLR) in lumbar spine imaging regarding acquisition time, image quality, artifact resistance, and diagnostic confidence.

MATERIAL AND METHODS

This retrospective monocentric study included 60 patients with lower back pain who underwent lumbar spinal MRI between February and April 2023. MRI parameters and DL reconstruction (DLR) techniques were utilized to acquire images. Two neuroradiologists independently evaluated image datasets based on various parameters using a 4-point Likert scale.

RESULTS

Accelerated imaging showed significantly less image noise and artifacts, as well as better image sharpness, compared to standard imaging. Overall image quality and diagnostic confidence were higher in accelerated imaging. Relevant disk herniations and spinal fractures were detected in both DLR and conventional images. Both readers favored accelerated imaging in the majority of examinations. The lumbar spine examination time was cut by 61% in accelerated imaging compared to standard imaging.

CONCLUSION

In conclusion, the utilization of deep learning-based image reconstruction techniques in lumbar spinal imaging resulted in significant time savings of up to 61% compared to standard imaging, while also improving image quality and diagnostic confidence. These findings highlight the potential of these techniques to enhance efficiency and accuracy in clinical practice for patients with lower back pain.

Deep learning-accelerated image reconstruction in MRI of the orbit to shorten acquisition time and enhance image quality

BACKGROUND AND PURPOSE

This study explores the use of deep learning (DL) techniques in MRI of the orbit to enhance imaging. Standard protocols, although detailed, have lengthy acquisition times. We investigate DL-based methods for T2-weighted and T1-weighted, fat-saturated, contrast-enhanced turbo spin echo (TSE) sequences, aiming to improve image quality, reduce acquisition time, minimize artifacts, and enhance diagnostic confidence in orbital imaging.

METHODS

In a 3-Tesla MRI study of 50 patients evaluating orbital diseases from March to July 2023, conventional (TSES ) and DL TSE sequences (TSEDL ) were used. Two neuroradiologists independently assessed the image datasets for image quality, diagnostic confidence, noise levels, artifacts, and image sharpness using a randomized and blinded 4-point Likert scale.

RESULTS

TSEDL significantly reduced image noise and artifacts, enhanced image sharpness, and decreased scan time, outperforming TSES (p < .05). TSEDL showed superior overall image quality and diagnostic confidence, with relevant findings effectively detected in both DL-based and conventional images. In 94% of cases, readers preferred accelerated imaging.

CONCLUSION

The study proved that using DL for MRI image reconstruction in orbital scans significantly cut acquisition time by 69%. This approach also enhanced image quality, reduced image noise, sharpened images, and boosted diagnostic confidence.

Between neurons and networks: investigating mesoscale brain connectivity in neurological and psychiatric disorders

The study of brain connectivity has been a cornerstone in understanding the complexities of neurological and psychiatric disorders. It has provided invaluable insights into the functional architecture of the brain and how it is perturbed in disorders. However, a persistent challenge has been achieving the proper spatial resolution, and developing computational algorithms to address biological questions at the multi-cellular level, a scale often referred to as the mesoscale. Historically, neuroimaging studies of brain connectivity have predominantly focused on the macroscale, providing insights into inter-regional brain connections but often falling short of resolving the intricacies of neural circuitry at the cellular or mesoscale level. This limitation has hindered our ability to fully comprehend the underlying mechanisms of neurological and psychiatric disorders and to develop targeted interventions. In light of this issue, our review manuscript seeks to bridge this critical gap by delving into the domain of mesoscale neuroimaging. We aim to provide a comprehensive overview of conditions affected by aberrant neural connections, image acquisition techniques, feature extraction, and data analysis methods that are specifically tailored to the mesoscale. We further delineate the potential of brain connectivity research to elucidate complex biological questions, with a particular focus on schizophrenia and epilepsy. This review encompasses topics such as dendritic spine quantification, single neuron morphology, and brain region connectivity. We aim to showcase the applicability and significance of mesoscale neuroimaging techniques in the field of neuroscience, highlighting their potential for gaining insights into the complexities of neurological and psychiatric disorders.

Towards an accurate autism spectrum disorder diagnosis: multiple connectome views from fMRI data

Functional connectome has revealed remarkable potential in the diagnosis of neurological disorders, e.g. autism spectrum disorder. However, existing studies have primarily focused on a single connectivity pattern, such as full correlation, partial correlation, or causality. Such an approach fails in discovering the potential complementary topology information of FCNs at different connection patterns, resulting in lower diagnostic performance. Consequently, toward an accurate autism spectrum disorder diagnosis, a straightforward ambition is to combine the multiple connectivity patterns for the diagnosis of neurological disorders. To this end, we conduct functional magnetic resonance imaging data to construct multiple brain networks with different connectivity patterns and employ kernel combination techniques to fuse information from different brain connectivity patterns for autism diagnosis. To verify the effectiveness of our approach, we assess the performance of the proposed method on the Autism Brain Imaging Data Exchange dataset for diagnosing autism spectrum disorder. The experimental findings demonstrate that our method achieves precise autism spectrum disorder diagnosis with exceptional accuracy (91.30%), sensitivity (91.48%), and specificity (91.11%).

An interpretable model based on graph learning for diagnosis of Parkinson's disease with voice-related EEG

Parkinson's disease (PD) exhibits significant clinical heterogeneity, presenting challenges in the identification of reliable electroencephalogram (EEG) biomarkers. Machine learning techniques have been integrated with resting-state EEG for PD diagnosis, but their practicality is constrained by the interpretable features and the stochastic nature of resting-state EEG. The present study proposes a novel and interpretable deep learning model, graph signal processing-graph convolutional networks (GSP-GCNs), using event-related EEG data obtained from a specific task involving vocal pitch regulation for PD diagnosis. By incorporating both local and global information from single-hop and multi-hop networks, our proposed GSP-GCNs models achieved an averaged classification accuracy of 90.2%, exhibiting a significant improvement of 9.5% over other deep learning models. Moreover, the interpretability analysis revealed discriminative distributions of large-scale EEG networks and topographic map of microstate MS5 learned by our models, primarily located in the left ventral premotor cortex, superior temporal gyrus, and Broca's area that are implicated in PD-related speech disorders, reflecting our GSP-GCN models' ability to provide interpretable insights identifying distinctive EEG biomarkers from large-scale networks. These findings demonstrate the potential of interpretable deep learning models coupled with voice-related EEG signals for distinguishing PD patients from healthy controls with accuracy and elucidating the underlying neurobiological mechanisms.

Multi feature fusion network for schizophrenia classification and abnormal brain network recognition

Schizophrenia classification and abnormal brain network recognition have an important research significance. Researchers have proposed many classification methods based on machine learning and deep learning. However, fewer studies utilized the advantages of complementary information from multi feature to learn the best representation of schizophrenia. In this study, we proposed a multi-feature fusion network (MFFN) using functional network connectivity (FNC) and time courses (TC) to distinguish schizophrenia patients from healthy controls. DNN backbone was adopted to learn the feature map of functional network connectivity, C-RNNAM backbone was designed to learn the feature map of time courses, and Deep SHAP was applied to obtain the most discriminative brain networks. We proved the effectiveness of this proposed model using the combining two public datasets and evaluated this model quantitatively using the evaluation indexes. The results showed that the functional network connectivity generated by independent component analysis has advantage in schizophrenia classification by comparing static and dynamic functional connections. This method obtained the best classification accuracy (ACC=87.30%, SPE=89.28%, SEN=85.71%, F1 =88.23%, and AUC=0.9081), and it demonstrated the superiority of this proposed model by comparing state-of-the-art methods. Ablation experiment also demonstrated that multi feature fusion and attention module can improve classification accuracy. The most discriminative brain networks showed that default mode network and visual network of schizophrenia patients have aberrant connections in brain networks. In conclusion, this method can identify schizophrenia effectively and visualize the abnormal brain network, and it has important clinical application value.

Deep Learning Accelerated Image Reconstruction of Fluid-Attenuated Inversion Recovery Sequence in Brain Imaging: Reduction of Acquisition Time and Improvement of Image Quality

RATIONALE AND OBJECTIVES

Fluid-attenuated inversion recovery (FLAIR) imaging is playing an increasingly significant role in the detection of brain metastases with a concomitant increase in the number of magnetic resonance imaging (MRI) examinations. Therefore, the purpose of this study was to investigate the impact on image quality and diagnostic confidence of an innovative deep learning-based accelerated FLAIR (FLAIRDLR) sequence of the brain compared to conventional (standard) FLAIR (FLAIRS) imaging.

MATERIALS AND METHODS

Seventy consecutive patients with staging cerebral MRIs were retrospectively enrolled in this single-center study. The FLAIRDLR was conducted using the same MRI acquisition parameters as the FLAIRS sequence, except for a higher acceleration factor for parallel imaging (from 2 to 4), which resulted in a shorter acquisition time of 1:39 minute instead of 2:40 minutes (-38%). Two specialized neuroradiologists evaluated the imaging datasets using a Likert scale that ranged from 1 to 4, with 4 indicating the best score for the following parameters: sharpness, lesion demarcation, artifacts, overall image quality, and diagnostic confidence. Additionally, the image preference of the readers and the interreader agreement were assessed.

RESULTS

The average age of the patients was 63 ± 11years. FLAIRDLR exhibited significantly less image noise than FLAIRS, with P-values of< .001 and< .05, respectively. The sharpness of the images and the ability to detect lesions were rated higher in FLAIRDLR, with a median score of 4 compared to a median score of 3 in FLAIRS (P-values of<.001 for both readers). In terms of overall image quality, FLAIRDLR was rated superior to FLAIRS, with a median score of 4 vs 3 (P-values of<.001 for both readers). Both readers preferred FLAIRDLR in 68/70 cases.

CONCLUSION

The feasibility of deep learning FLAIR brain imaging was shown with additional 38% reduction in examination time compared to standard FLAIR imaging. Furthermore, this technique has shown improvement in image quality, noise reduction, and lesion demarcation.

Neural Computation-Based Methods for the Early Diagnosis and Prognosis of Alzheimer's Disease Not Using Neuroimaging Biomarkers: A Systematic Review

Background

The growing number of older adults in recent decades has led to more prevalent geriatric diseases, such as strokes and dementia. Therefore, Alzheimer's disease (AD), as the most common type of dementia, has become more frequent too.

Background

Objective: The goals of this work are to present state-of-the-art studies focused on the automatic diagnosis and prognosis of AD and its early stages, mainly mild cognitive impairment, and predicting how the research on this topic may change in the future.

Methods

Articles found in the existing literature needed to fulfill several selection criteria. Among others, their classification methods were based on artificial neural networks (ANNs), including deep learning, and data not from brain signals or neuroimaging techniques were used. Considering our selection criteria, 42 articles published in the last decade were finally selected.

Results

The most medically significant results are shown. Similar quantities of articles based on shallow and deep ANNs were found. Recurrent neural networks and transformers were common with speech or in longitudinal studies. Convolutional neural networks (CNNs) were popular with gait or combined with others in modular approaches. Above one third of the cross-sectional studies utilized multimodal data. Non-public datasets were frequently used in cross-sectional studies, whereas the opposite in longitudinal ones. The most popular databases were indicated, which will be helpful for future researchers in this field.

Conclusions

The introduction of CNNs in the last decade and their superb results with neuroimaging data did not negatively affect the usage of other modalities. In fact, new ones emerged.

Improving Psychiatry Services with Artificial Intelligence: Opportunities and Challenges

Mental disorders are a critical global public health problem due to their increasing prevalence, rising costs, and significant economic burden. Despite efforts to increase the mental health workforce in Türkiye, there is a significant shortage of psychiatrists, limiting the quality and accessibility of mental health services. This review examines the potential of artificial intelligence (AI), especially large language models, to transform psychiatric care in the world and in Türkiye. AI technologies, including machine learning and deep learning, offer innovative solutions for the diagnosis, personalization of treatment, and monitoring of mental disorders using a variety of data sources, such as speech patterns, neuroimaging, and behavioral measures. Although AI has shown promising capabilities in improving diagnostic accuracy and access to mental health services, challenges such as algorithmic biases, data privacy concerns, ethical implications, and the confabulation phenomenon of large language models prevent the full implementation of AI in practice. The review highlights the need for interdisciplinary collaboration to develop culturally and linguistically adapted AI tools, particularly in the Turkish context, and suggests strategies such as fine-tuning, retrieval-augmented generation, and reinforcement learning from human feedback to increase AI reliability. Advances suggest that AI can improve mental health care by increasing diagnostic accuracy and accessibility while preserving the essential human elements of medical care. Current limitations need to be addressed through rigorous research and ethical frameworks for effective and equitable integration of AI into mental health care. Keywords: Artificial İntelligence, Health, Large Language Model, Machine Learning, Psychiatry.

Verifying and refining early statuses in Alzheimer's disease progression: a possibility from deep feature comparison

Defining the early status of Alzheimer's disease is challenging. Theoretically, the statuses in the Alzheimer's disease continuum are expected to share common features. Here, we explore to verify and refine candidature early statuses of Alzheimer's disease with features learned from deep learning. We train models on brain functional networks to accurately classify between amnestic and non-amnestic mild cognitive impairments and between healthy controls and mild cognitive impairments. The trained models are applied to Alzheimer's disease and subjective cognitive decline groups to suggest feature similarities among the statuses and identify informative subpopulations. The amnestic mild cognitive impairment vs non-amnestic mild cognitive impairments classifier believes that 71.8% of Alzheimer's disease are amnestic mild cognitive impairment. And 73.5% of subjective cognitive declines are labeled as mild cognitive impairments, 88.8% of which are further suggested as "amnestic mild cognitive impairment." Further multimodal analyses suggest that the amnestic mild cognitive impairment-like Alzheimer's disease, mild cognitive impairment-like subjective cognitive decline, and amnestic mild cognitive impairment-like subjective cognitive decline exhibit more Alzheimer's disease -related pathological changes (elaborated β-amyloid depositions, reduced glucose metabolism, and gray matter atrophy) than non-amnestic mild cognitive impairments -like Alzheimer's disease, healthy control-like subjective cognitive decline, and non-amnestic mild cognitive impairments -like subjective cognitive decline. The test-retest reliability of the subpopulation identification is fair to good in general. The study indicates overall similarity among subjective cognitive decline, amnestic mild cognitive impairment, and Alzheimer's disease and implies their progression relationships. The results support "deep feature comparison" as a potential beneficial framework to verify and refine early Alzheimer's disease status.

A common spectrum underlying brain disorders across lifespan revealed by deep learning on brain networks

Brain disorders in the early and late life of humans potentially share pathological alterations in brain functions. However, the key neuroimaging evidence remains unrevealed for elucidating such commonness and the relationships among these disorders. To explore this puzzle, we build a restricted single-branch deep learning model, using multi-site functional magnetic resonance imaging data (N = 4,410, 6 sites), for classifying 5 different early- and late-life brain disorders from healthy controls (cognitively unimpaired). Our model achieves 62.6 ± 1.9% overall classification accuracy and thus supports us in detecting a set of commonly affected functional subnetworks, including default mode, executive control, visual, and limbic networks. In the deep-layer representation of data, we observe young and aging patients with disorders are continuously distributed, which is in line with the clinical concept of the "spectrum of disorders." The relationships among brain disorders from the revealed spectrum promote the understanding of disorder comorbidities and time associations in the lifespan.

Topological Data Analysis for Multivariate Time Series Data

Over the last two decades, topological data analysis (TDA) has emerged as a very powerful data analytic approach that can deal with various data modalities of varying complexities. One of the most commonly used tools in TDA is persistent homology (PH), which can extract topological properties from data at various scales. The aim of this article is to introduce TDA concepts to a statistical audience and provide an approach to analyzing multivariate time series data. The application's focus will be on multivariate brain signals and brain connectivity networks. Finally, this paper concludes with an overview of some open problems and potential application of TDA to modeling directionality in a brain network, as well as the casting of TDA in the context of mixed effect models to capture variations in the topological properties of data collected from multiple subjects.

Charting the Progress of Epilepsy Classification: Navigating a Shifting Landscape

Epilepsy, a neurological disorder characterized by recurrent seizures, has witnessed a remarkable transformation in its classification paradigm, driven by advances in clinical understanding, neuroimaging, and molecular genetics. This narrative review navigates the dynamic landscape of epilepsy classification, offering insights into recent developments, challenges, and the promising horizon. Historically, epilepsy classification relied heavily on clinical observations, categorizing seizures based on their phenomenology and presumed etiology. However, the field has profoundly shifted from a symptom-based approach to a more refined, multidimensional system. One pivotal aspect of this evolution is the integration of neuroimaging techniques, particularly magnetic resonance imaging (MRI) and functional imaging modalities. These tools have unveiled the intricate neural networks implicated in epilepsy, facilitating the identification of distinct brain abnormalities and the categorization of epilepsy subtypes based on structural and functional findings. Furthermore, the role of genetics has become increasingly prominent in epilepsy classification. Genetic discoveries have not only unraveled the molecular underpinnings of various epileptic syndromes but have also provided valuable diagnostic and prognostic insights. This narrative review delves into the expanding realm of genetic testing and its impact on tailoring treatment strategies to individual patients. As the classification landscape evolves, there are accompanying challenges. The narrative review underscores the transformative potential of artificial intelligence and machine learning in epilepsy classification. These technologies hold promise in automating the analysis of complex neuroimaging and genetic data, offering enhanced accuracy and efficiency in epilepsy diagnosis and classification. In conclusion, navigating the shifting landscape of epilepsy classification is a journey marked by progress, complexity, and the prospect of improved patient care. We are charting a course toward more precise diagnoses and tailored treatments by embracing advanced neuroimaging, genetics, and innovative technologies. As the field continues to evolve, collaborative efforts and a holistic understanding of epilepsy's diverse manifestations will be instrumental in harnessing the full potential of this dynamic landscape.

Multimodal Autism Spectrum Disorder Diagnosis Method Based on DeepGCN

Multimodal data play an important role in the diagnosis of brain diseases. This study constructs a whole-brain functional connectivity network based on functional MRI data, uses non-imaging data with demographic information to complement the classification task for diagnosing subjects, and proposes a multimodal and across-site WL-DeepGCN-based method for classification to diagnose autism spectrum disorder (ASD). This method is used to resolve the existing problem that deep learning ASD identification cannot efficiently utilize multimodal data. In the WL-DeepGCN, a weight-learning network is used to represent the similarity of non-imaging data in the latent space, introducing a new approach for constructing population graph edge weights, and we find that it is beneficial and robust to define pairwise associations in the latent space rather than the input space. We propose a graph convolutional neural network residual connectivity approach to reduce the information loss due to convolution operations by introducing residual units to avoid gradient disappearance and gradient explosion. Furthermore, an EdgeDrop strategy makes the node connections sparser by randomly dropping edges in the raw graph, and its introduction can alleviate the overfitting and oversmoothing problems in the DeepGCN training process. We compare the WL-DeepGCN model with competitive models based on the same topics and nested 10-fold cross-validation show that our method achieves 77.27% accuracy and 0.83 AUC for ASD identification, bringing substantial performance gains.

Deep learning in neuroimaging of epilepsy

In recent years, artificial intelligence, particularly deep learning (DL), has demonstrated utility in diverse areas of medicine. DL uses neural networks to automatically learn features from the raw data while this is not possible with conventional machine learning. It is helpful for the assessment of patients with epilepsy and whilst most published studies have been aimed at the automatic detection and prediction of seizures from electroencephalographic records, there is a growing number of investigations that use neuroimaging modalities (structural and functional magnetic resonance imaging, diffusion-weighted imaging and positron emission tomography) as input data. We review the application of DL to neuroimaging (sMRI, fMRI, DWI and PET) of focal epilepsy, specifically presurgical evaluation of drug-refractory epilepsy. First, a brief theoretical overview of artificial neural networks and deep learning is presented. Next, we review applications of deep learning to neuroimaging of epilepsy: diagnosis and lateralization, automated detection of lesion, presurgical evaluation and prediction of postsurgical outcome. Finally, the limitations, challenges and possible future directions in the application of these methods in the study of epilepsies are discussed. This approach could become an essential tool in clinical practice, particularly in the evaluation of images considered negative by visual inspection, in individualized treatments, and in the approach to epilepsy as a network disorder. However, greater multicenter collaboration is required to achieve the collection of sufficient data with the required quality together with the open access availability of the developed codes and tools.

Neuroimaging and machine learning for studying the pathways from mild cognitive impairment to alzheimer's disease: a systematic review

BACKGROUND

This systematic review synthesizes the most recent neuroimaging procedures and machine learning approaches for the prediction of conversion from mild cognitive impairment to Alzheimer's disease dementia.

METHODS

We systematically searched PubMed, SCOPUS, and Web of Science databases following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) systematic review guidelines.

RESULTS

Our search returned 2572 articles, 56 of which met the criteria for inclusion in the final selection. The multimodality framework and deep learning techniques showed potential for predicting the conversion of MCI to AD dementia.

CONCLUSION

Findings of this systematic review identified that the possibility of using neuroimaging data processed by advanced learning algorithms is promising for the prediction of AD progression. We also provided a detailed description of the challenges that researchers are faced along with future research directions. The protocol has been registered in the International Prospective Register of Systematic Reviews- CRD42019133402 and published in the Systematic Reviews journal.

Challenges of implementing computer-aided diagnostic models for neuroimages in a clinical setting

Advances in artificial intelligence have cultivated a strong interest in developing and validating the clinical utilities of computer-aided diagnostic models. Machine learning for diagnostic neuroimaging has often been applied to detect psychological and neurological disorders, typically on small-scale datasets or data collected in a research setting. With the collection and collation of an ever-growing number of public datasets that researchers can freely access, much work has been done in adapting machine learning models to classify these neuroimages by diseases such as Alzheimer's, ADHD, autism, bipolar disorder, and so on. These studies often come with the promise of being implemented clinically, but despite intense interest in this topic in the laboratory, limited progress has been made in clinical implementation. In this review, we analyze challenges specific to the clinical implementation of diagnostic AI models for neuroimaging data, looking at the differences between laboratory and clinical settings, the inherent limitations of diagnostic AI, and the different incentives and skill sets between research institutions, technology companies, and hospitals. These complexities need to be recognized in the translation of diagnostic AI for neuroimaging from the laboratory to the clinic.

Estimating explainable Alzheimer's disease likelihood map via clinically-guided prototype learning

Identifying Alzheimer's disease (AD) involves a deliberate diagnostic process owing to its innate traits of irreversibility with subtle and gradual progression. These characteristics make AD biomarker identification from structural brain imaging (e.g., structural MRI) scans quite challenging. Using clinically-guided prototype learning, we propose a novel deep-learning approach through eXplainable AD Likelihood Map Estimation (XADLiME) for AD progression modeling over 3D sMRIs. Specifically, we establish a set of topologically-aware prototypes onto the clusters of latent clinical features, uncovering an AD spectrum manifold. Considering this pseudo map as an enriched reference, we employ an estimating network to approximate the AD likelihood map over a 3D sMRI scan. Additionally, we promote the explainability of such a likelihood map by revealing a comprehensible overview from clinical and morphological perspectives. During the inference, this estimated likelihood map served as a substitute for unseen sMRI scans for effectively conducting the downstream task while providing thorough explainable states.

Challenges, opportunities, and advances related to COVID-19 classification based on deep learning

The novel coronavirus disease, or COVID-19, is a hazardous disease. It is endangering the lives of many people living in more than two hundred countries. It directly affects the lungs. In general, two main imaging modalities: - computed tomography (CT) and chest x-ray (CXR) are used to achieve a speedy and reliable medical diagnosis. Identifying the coronavirus in medical images is exceedingly difficult for diagnosis, assessment, and treatment. It is demanding, time-consuming, and subject to human mistakes. In biological disciplines, excellent performance can be achieved by employing artificial intelligence (AI) models. As a subfield of AI, deep learning (DL) networks have drawn considerable attention than standard machine learning (ML) methods. DL models automatically carry out all the steps of feature extraction, feature selection, and classification. This study has performed comprehensive analysis of coronavirus classification using CXR and CT imaging modalities using DL architectures. Additionally, we have discussed how transfer learning is helpful in this regard. Finally, the problem of designing and implementing a system using computer-aided diagnostic (CAD) to find COVID-19 using DL approaches is highlighted a future research possibility.

Broad learning for early diagnosis of Alzheimer's disease using FDG-PET of the brain

Alzheimer's disease (AD) is a progressive neurodegenerative disease, and the development of AD is irreversible. However, preventive measures in the presymptomatic stage of AD can effectively slow down deterioration. Fluorodeoxyglucose positron emission tomography (FDG-PET) can detect the metabolism of glucose in patients' brains, which can help to identify changes related to AD before brain damage occurs. Machine learning is useful for early diagnosis of patients with AD using FDG-PET, but it requires a sufficiently large dataset, and it is easy for overfitting to occur in small datasets. Previous studies using machine learning for early diagnosis with FDG-PET have either involved the extraction of elaborately handcrafted features or validation on a small dataset, and few studies have explored the refined classification of early mild cognitive impairment (EMCI) and late mild cognitive impairment (LMCI). This article presents a broad network-based model for early diagnosis of AD (BLADNet) through PET imaging of the brain; this method employs a novel broad neural network to enhance the features of FDG-PET extracted via 2D CNN. BLADNet can search for information over a broad space through the addition of new BLS blocks without retraining of the whole network, thus improving the accuracy of AD classification. Experiments conducted on a dataset containing 2,298 FDG-PET images of 1,045 subjects from the ADNI database demonstrate that our methods are superior to those used in previous studies on early diagnosis of AD with FDG-PET. In particular, our methods achieved state-of-the-art results in EMCI and LMCI classification with FDG-PET.

Generalisability of fetal ultrasound deep learning models to low-resource imaging settings in five African countries

Most artificial intelligence (AI) research and innovations have concentrated in high-income countries, where imaging data, IT infrastructures and clinical expertise are plentiful. However, slower progress has been made in limited-resource environments where medical imaging is needed. For example, in Sub-Saharan Africa, the rate of perinatal mortality is very high due to limited access to antenatal screening. In these countries, AI models could be implemented to help clinicians acquire fetal ultrasound planes for the diagnosis of fetal abnormalities. So far, deep learning models have been proposed to identify standard fetal planes, but there is no evidence of their ability to generalise in centres with low resources, i.e. with limited access to high-end ultrasound equipment and ultrasound data. This work investigates for the first time different strategies to reduce the domain-shift effect arising from a fetal plane classification model trained on one clinical centre with high-resource settings and transferred to a new centre with low-resource settings. To that end, a classifier trained with 1792 patients from Spain is first evaluated on a new centre in Denmark in optimal conditions with 1008 patients and is later optimised to reach the same performance in five African centres (Egypt, Algeria, Uganda, Ghana and Malawi) with 25 patients each. The results show that a transfer learning approach for domain adaptation can be a solution to integrate small-size African samples with existing large-scale databases in developed countries. In particular, the model can be re-aligned and optimised to boost the performance on African populations by increasing the recall to [Formula: see text] and at the same time maintaining a high precision across centres. This framework shows promise for building new AI models generalisable across clinical centres with limited data acquired in challenging and heterogeneous conditions and calls for further research to develop new solutions for the usability of AI in countries with fewer resources and, consequently, in higher need of clinical support.

Interpretable deep learning for the prognosis of long-term functional outcome post-stroke using acute diffusion weighted imaging

Advances in deep learning can be applied to acute stroke imaging to build powerful and explainable prediction models that could supersede traditionally used biomarkers. We aimed to evaluate the performance and interpretability of a deep learning model based on convolutional neural networks (CNN) in predicting long-term functional outcome with diffusion-weighted imaging (DWI) acquired at day 1 post-stroke. Ischemic stroke patients (n = 322) were included from the ASTER and INSULINFARCT trials as well as the Pitié-Salpêtrière registry. We trained a CNN to predict long-term functional outcome assessed at 3 months with the modified Rankin Scale (dichotomized as good [mRS ≤ 2] vs. poor [mRS ≥ 3]) and compared its performance to two logistic regression models using lesion volume and ASPECTS. The CNN contained an attention mechanism, which allowed to visualize the areas of the brain that drove prediction. The deep learning model yielded a significantly higher area under the curve (0.83 95%CI [0.78-0.87]) than lesion volume (0.78 [0.73-0.83]) and ASPECTS (0.77 [0.71-0.83]) (p < 0.05). Setting all classifiers to the specificity as the deep learning model (i.e., 0.87 [0.82-0.92]), the CNN yielded a significantly higher sensitivity (0.67 [0.59-0.73]) than lesion volume (0.48 [0.40-0.56]) and ASPECTS (0.50 [0.41-0.58]) (p = 0.002). The attention mechanism revealed that the network learned to naturally attend to the lesion to predict outcome.

Prediction of gender from longitudinal MRI data via deep learning on adolescent data reveals unique patterns associated with brain structure and change over a two-year period

Deep learning algorithms for predicting neuroimaging data have shown considerable promise in various applications. Prior work has demonstrated that deep learning models that take advantage of the data's 3D structure can outperform standard machine learning on several learning tasks. However, most prior research in this area has focused on neuroimaging data from adults. Within the Adolescent Brain and Cognitive Development (ABCD) dataset, a large longitudinal development study, we examine structural MRI data to predict gender and identify gender-related changes in brain structure. Results demonstrate that gender prediction accuracy is exceptionally high (>97%) with training epochs > 200 and that this accuracy increases with age. Brain regions identified as the most discriminative in the task under study include predominantly frontal areas and the temporal lobe. When evaluating gender predictive changes specific to a two-year increase in age, a broader set of visual, cingulate, and insular regions are revealed. Our findings show a robust gender-related structural brain change pattern, even over a small age range. This suggests that it might be possible to study how the brain changes during adolescence by looking at how these changes are related to different behavioral and environmental factors.

Application of Artificial Intelligence in Image Processing of Neurodegenerative Disorders: A Review Study

: Neurodegenerative diseases can make life difficult and lead to death in many cases. They also can be difficult, time-consuming, and costly to diagnose with enough accuracy/certainty. Artificial intelligence (AI) has shown promise in tackling some of the challenges present in medical imaging and is anticipated to become a crucial tool in health care applications in the near future. In particular, deep learning methods have displayed great performance in various subfields of image processing, including but not limited to image segmentation, image synthesis, and image reconstruction. In this paper, many state-of-the-art applications of deep learning models in image processing were reviewed.

Incorporating multi-stage diagnosis status to mine associations between genetic risk variants and the multi-modality phenotype network in major depressive disorder

Depression (major depressive disorder, MDD) is a common and serious medical illness. Globally, it is estimated that 5% of adults suffer from depression. Recently, imaging genetics receives growing attention and become a powerful strategy for discoverying the associations between genetic variants (e.g., single-nucleotide polymorphisms, SNPs) and multi-modality brain imaging data. However, most of the existing MDD imaging genetic research studies conducted by clinicians usually utilize simple statistical analysis methods and only consider single-modality brain imaging, which are limited in the deeper discovery of the mechanistic understanding of MDD. It is therefore imperative to utilize a powerful and efficient technology to fully explore associations between genetic variants and multi-modality brain imaging. In this study, we developed a novel imaging genetic association framework to mine the multi-modality phenotype network between genetic risk variants and multi-stage diagnosis status. Specifically, the multi-modality phenotype network consists of voxel node features and connectivity edge features from structural magnetic resonance imaging (sMRI) and resting-state functional magnetic resonance imaging (rs-fMRI). Thereafter, an association model based on multi-task learning strategy was adopted to fully explore the relationship between the MDD risk SNP and the multi-modality phenotype network. The multi-stage diagnosis status was introduced to further mine the relation among the multiple modalities of different subjects. A multi-modality brain imaging data and genotype data were collected by us from two hospitals. The experimental results not only demonstrate the effectiveness of our proposed method but also identify some consistent and stable brain regions of interest (ROIs) biomarkers from the node and edge features of multi-modality phenotype network. Moreover, four new and potential risk SNPs associated with MDD were discovered.

An explainable deep learning framework for characterizing and interpreting human brain states

Deep learning approaches have been widely adopted in the medical image analysis field. However, a most of existing deep learning approaches focus on achieving promising performances such as classification, detection, and segmentation, and much less effort is devoted to the explanation of the designed models. Similarly, in the brain imaging field, many deep learning approaches have been designed and applied to characterize and predict human brain states. However, these models lack interpretation. In response, we propose a novel domain knowledge informed self-attention graph pooling-based (SAGPool) graph convolutional neural network to study human brain states. Specifically, the dense individualized and common connectivity-based cortical landmarks system (DICCCOL, structural brain connectivity profiles) and holistic atlases of functional networks and interactions system (HAFNI, functional brain connectivity profiles) are integrated with the SAGPool model to better characterize and interpret the brain states. Extensive experiments are designed and carried out on the large-scale human connectome project (HCP) Q1 and S1200 dataset. Promising brain state classification performances are observed (e.g., an average of 93.7% for seven-task classification and 100% for binary classification). In addition, the importance of the brain regions, which contributes most to the accurate classification, is successfully quantified and visualized. A thorough neuroscientific interpretation suggests that these extracted brain regions and their importance calculated from self-attention graph pooling layer offer substantial explainability.

Better characterization of attention and hyperactivity/impulsivity in children with ADHD: The key to understanding the underlying white matter microstructure

The apparent increase in the prevalence of the attention deficit hyperactivity disorder (ADHD) diagnosis raises many questions regarding the variability of the subjective diagnostic method. This comprehensive review reports findings in studies assessing white matter (WM) bundles in diffusion MRI and symptom severity in children with ADHD. These studies suggested the involvement of the connections between the frontal, parietal, and basal ganglia regions. This review discusses the limitations surrounding diffusion tensor imaging (DTI) and suggests novel imaging techniques allowing for a more reliable representation of the underlying biology. We propose a more inclusive approach to studying ADHD that includes known endophenotypes within the ADHD diagnosis. Aligned with the Research Domain Criteria Initiative, we also propose to investigate attentional capabilities and impulsive behaviours outside of the borders of the diagnosis. We support the existing hypothesis that ADHD originates from a developmental error and propose that it could lead to an accumulation in time of abnormalities in WM microstructure and pathways. Finally, state-of-the-art diffusion processing and novel artificial intelligence approaches would be beneficial to fully understand the pathophysiology of ADHD.

Interpretable brain disease classification and relevance-guided deep learning

Deep neural networks are increasingly used for neurological disease classification by MRI, but the networks' decisions are not easily interpretable by humans. Heat mapping by deep Taylor decomposition revealed that (potentially misleading) image features even outside of the brain tissue are crucial for the classifier's decision. We propose a regularization technique to train convolutional neural network (CNN) classifiers utilizing relevance-guided heat maps calculated online during training. The method was applied using T1-weighted MR images from 128 subjects with Alzheimer's disease (mean age = 71.9 ± 8.5 years) and 290 control subjects (mean age = 71.3 ± 6.4 years). The developed relevance-guided framework achieves higher classification accuracies than conventional CNNs but more importantly, it relies on less but more relevant and physiological plausible voxels within brain tissue. Additionally, preprocessing effects from skull stripping and registration are mitigated. With the interpretability of the decision mechanisms underlying CNNs, these results challenge the notion that unprocessed T1-weighted brain MR images in standard CNNs yield higher classification accuracy in Alzheimer's disease than solely atrophy.

Modern views of machine learning for precision psychiatry

In light of the National Institute of Mental Health (NIMH)'s Research Domain Criteria (RDoC), the advent of functional neuroimaging, novel technologies and methods provide new opportunities to develop precise and personalized prognosis and diagnosis of mental disorders. Machine learning (ML) and artificial intelligence (AI) technologies are playing an increasingly critical role in the new era of precision psychiatry. Combining ML/AI with neuromodulation technologies can potentially provide explainable solutions in clinical practice and effective therapeutic treatment. Advanced wearable and mobile technologies also call for the new role of ML/AI for digital phenotyping in mobile mental health. In this review, we provide a comprehensive review of ML methodologies and applications by combining neuroimaging, neuromodulation, and advanced mobile technologies in psychiatry practice. We further review the role of ML in molecular phenotyping and cross-species biomarker identification in precision psychiatry. We also discuss explainable AI (XAI) and neuromodulation in a closed human-in-the-loop manner and highlight the ML potential in multi-media information extraction and multi-modal data fusion. Finally, we discuss conceptual and practical challenges in precision psychiatry and highlight ML opportunities in future research.

Deep learning in fNIRS: a review

Significance: Optical neuroimaging has become a well-established clinical and research tool to monitor cortical activations in the human brain. It is notable that outcomes of functional near-infrared spectroscopy (fNIRS) studies depend heavily on the data processing pipeline and classification model employed. Recently, deep learning (DL) methodologies have demonstrated fast and accurate performances in data processing and classification tasks across many biomedical fields. Aim: We aim to review the emerging DL applications in fNIRS studies. Approach: We first introduce some of the commonly used DL techniques. Then, the review summarizes current DL work in some of the most active areas of this field, including brain-computer interface, neuro-impairment diagnosis, and neuroscience discovery. Results: Of the 63 papers considered in this review, 32 report a comparative study of DL techniques to traditional machine learning techniques where 26 have been shown outperforming the latter in terms of the classification accuracy. In addition, eight studies also utilize DL to reduce the amount of preprocessing typically done with fNIRS data or increase the amount of data via data augmentation. Conclusions: The application of DL techniques to fNIRS studies has shown to mitigate many of the hurdles present in fNIRS studies such as lengthy data preprocessing or small sample sizes while achieving comparable or improved classification accuracy.

Machine learning for autism spectrum disorder diagnosis using structural magnetic resonance imaging: Promising but challenging

Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder that affects approximately 1% of the population and causes significant burdens. ASD's pathogenesis remains elusive; hence, diagnosis is based on a constellation of behaviors. Structural magnetic resonance imaging (sMRI) studies have shown several abnormalities in volumetric and geometric features of the autistic brain. However, inconsistent findings prevented most contributions from being translated into clinical practice. Establishing reliable biomarkers for ASD using sMRI is crucial for the correct diagnosis and treatment. In recent years, machine learning (ML) and specifically deep learning (DL) have quickly extended to almost every sector, notably in disease diagnosis. Thus, this has led to a shift and improvement in ASD diagnostic methods, fulfilling most clinical diagnostic requirements. However, ASD discovery remains difficult. This review examines the ML-based ASD diagnosis literature over the past 5 years. A literature-based taxonomy of the research landscape has been mapped, and the major aspects of this topic have been covered. First, we provide an overview of ML's general classification pipeline and the features of sMRI. Next, representative studies are highlighted and discussed in detail with respect to methods, and biomarkers. Finally, we highlight many common challenges and make recommendations for future directions. In short, the limited sample size was the main obstacle; Thus, comprehensive data sets and rigorous methods are necessary to check the generalizability of the results. ML technologies are expected to advance significantly in the coming years, contributing to the diagnosis of ASD and helping clinicians soon.

Convolutional Recurrent Neural Network for Dynamic Functional MRI Analysis and Brain Disease Identification

Dynamic functional connectivity (dFC) networks derived from resting-state functional magnetic resonance imaging (rs-fMRI) help us understand fundamental dynamic characteristics of human brains, thereby providing an efficient solution for automated identification of brain diseases, such as Alzheimer's disease (AD) and its prodromal stage. Existing studies have applied deep learning methods to dFC network analysis and achieved good performance compared with traditional machine learning methods. However, they seldom take advantage of sequential information conveyed in dFC networks that could be informative to improve the diagnosis performance. In this paper, we propose a convolutional recurrent neural network (CRNN) for automated brain disease classification with rs-fMRI data. Specifically, we first construct dFC networks from rs-fMRI data using a sliding window strategy. Then, we employ three convolutional layers and long short-term memory (LSTM) layer to extract high-level features of dFC networks and also preserve the sequential information of extracted features, followed by three fully connected layers for brain disease classification. Experimental results on 174 subjects with 563 rs-fMRI scans from the Alzheimer's Disease Neuroimaging Initiative (ADNI) demonstrate the effectiveness of our proposed method in binary and multi-category classification tasks.