T-distribution stochastic neighbor embedding for fine brain functional parcellation on rs-fMRI

Exploration and validation of key genes associated with early lymph node metastasis in thyroid carcinoma using weighted gene co-expression network analysis and machine learning

Background

Thyroid carcinoma (THCA), the most common endocrine neoplasm, typically exhibits an indolent behavior. However, in some instances, lymph node metastasis (LNM) may occur in the early stages, with the underlying mechanisms not yet fully understood.

Materials and methods

LNM potential was defined as the tumor's capability to metastasize to lymph nodes at an early stage, even when the tumor volume is small. We performed differential expression analysis using the 'Limma' R package and conducted enrichment analyses using the Metascape tool. Co-expression networks were established using the 'WGCNA' R package, with the soft threshold power determined by the 'pickSoftThreshold' algorithm. For unsupervised clustering, we utilized the 'ConsensusCluster Plus' R package. To determine the topological features and degree centralities of each node (protein) within the Protein-Protein Interaction (PPI) network, we used the CytoNCA plugin integrated with the Cytoscape tool. Immune cell infiltration was assessed using the Immune Cell Abundance Identifier (ImmuCellAI) database. We applied the Least Absolute Shrinkage and Selection Operator (LASSO), Support Vector Machine (SVM), and Random Forest (RF) algorithms individually, with the 'glmnet,' 'e1071,' and 'randomForest' R packages, respectively. Ridge regression was performed using the 'oncoPredict' algorithm, and all the predictions were based on data from the Genomics of Drug Sensitivity in Cancer (GDSC) database. To ascertain the protein expression levels and subcellular localization of genes, we consulted the Human Protein Atlas (HPA) database. Molecular docking was carried out using the mcule 1-click Docking server online. Experimental validation of gene and protein expression levels was conducted through Real-Time Quantitative PCR (RT-qPCR) and immunohistochemistry (IHC) assays.

Results

Through WGCNA and PPI network analysis, we identified twelve hub genes as the most relevant to LNM potential from these two modules. These 12 hub genes displayed differential expression in THCA and exhibited significant correlations with the downregulation of neutrophil infiltration, as well as the upregulation of dendritic cell and macrophage infiltration, along with activation of the EMT pathway in THCA. We propose a novel molecular classification approach and provide an online web-based nomogram for evaluating the LNM potential of THCA (http://www.empowerstats.net/pmodel/?m=17617_LNM). Machine learning algorithms have identified ERBB3 as the most critical gene associated with LNM potential in THCA. ERBB3 exhibits high expression in patients with THCA who have experienced LNM or have advanced-stage disease. The differential methylation levels partially explain this differential expression of ERBB3. ROC analysis has identified ERBB3 as a diagnostic marker for THCA (AUC=0.89), THCA with high LNM potential (AUC=0.75), and lymph nodes with tumor metastasis (AUC=0.86). We have presented a comprehensive review of endocrine disruptor chemical (EDC) exposures, environmental toxins, and pharmacological agents that may potentially impact LNM potential. Molecular docking revealed a docking score of -10.1 kcal/mol for Lapatinib and ERBB3, indicating a strong binding affinity.

Conclusion

In conclusion, our study, utilizing bioinformatics analysis techniques, identified gene modules and hub genes influencing LNM potential in THCA patients. ERBB3 was identified as a key gene with therapeutic implications. We have also developed a novel molecular classification approach and a user-friendly web-based nomogram tool for assessing LNM potential. These findings pave the way for investigations into the mechanisms underlying differences in LNM potential and provide guidance for personalized clinical treatment plans.

Brain imaging signatures of neuropathic facial pain derived by artificial intelligence

Advances in neuroimaging have permitted the non-invasive examination of the human brain in pain. However, a persisting challenge is in the objective differentiation of neuropathic facial pain subtypes, as diagnosis is based on patients' symptom descriptions. We use artificial intelligence (AI) models with neuroimaging data to distinguish subtypes of neuropathic facial pain and differentiate them from healthy controls. We conducted a retrospective analysis of diffusion tensor and T1-weighted imaging data using random forest and logistic regression AI models on 371 adults with trigeminal pain (265 classical trigeminal neuralgia (CTN), 106 trigeminal neuropathic pain (TNP)) and 108 healthy controls (HC). These models distinguished CTN from HC with up to 95% accuracy, and TNP from HC with up to 91% accuracy. Both classifiers identified gray and white matter-based predictive metrics (gray matter thickness, surface area, and volume; white matter diffusivity metrics) that significantly differed across groups. Classification of TNP and CTN did not show significant accuracy (51%) but highlighted two structures that differed between pain groups-the insula and orbitofrontal cortex. Our work demonstrates that AI models with brain imaging data alone can differentiate neuropathic facial pain subtypes from healthy data and identify regional structural indicates of pain.

A fully supervised universal adversarial perturbations and the progressive optimization

Universal Adversarial Perturbations(UAPs), which are image-agnostic adversarial perturbations, have been demonstrated to successfully deceive computer vision models. Proposed UAPs in the case of data-dependent, use the internal layers’ activation or the output layer’s decision values as supervision. In this paper, we use both of them to drive the supervised learning of UAP, termed as fully supervised UAP(FS-UAP), and design a progressive optimization strategy to solve the FS-UAP. Specifically, we define an internal layers supervised objective relying on multiple major internal layers’ activation to estimate the deviations of adversarial examples from legitimate examples. We also define an output layer supervised objective relying on the logits of output layer to evaluate attacking degrees. In addition, we use the UAP found by previous stage as the initial solution of the next stage so as to progressively optimize the UAP stage-wise. We use seven networks and ImageNet dataset to evaluate the proposed FS-UAP, and provide an in-depth analysis for the latent factors affecting the performance of universal attacks. The experimental results show that our FS-UAP (i) has powerful capability of cheating CNNs (ii) has superior transfer-ability across models and weak data-dependent (iii) is appropriate for both untarget and target attacks.

A novel 5D brain parcellation approach based on spatio-temporal encoding of resting fMRI data from deep residual learning

OBJECTIVE

Brain parcellation is an essential aspect of computational neuroimaging research and deals with segmenting the brain into (possibly overlapping) sub-regions employed to study brain anatomy or function. In the context of functional parcellation, brain organization which is often measured via temporal metrics such as coherence, is highly dynamic. This dynamic aspect is ignored in most research, which typically applies anatomically based, fixed regions for each individual, and can produce misleading results.

METHODS

In this work, we propose a novel spatio-temporal-network (5D) brain parcellation scheme utilizing a deep residual network to predict the probability of each voxel belonging to a brain network at each point in time.

RESULTS

We trained 53 4D brain networks and evaluate the ability of these networks to capture spatial and temporal dynamics as well as to show sensitivity to individual or group-level variation (in our case with age).

CONCLUSION

The proposed system generates informative spatio-temporal networks that vary not only across individuals but also over time and space.

SIGNIFICANCE

The dynamic 5D nature of the developed approach provides a powerful framework that expands on existing work and has potential to identify novel and typically ignored findings when studying the healthy and disordered brain.

Image Features of Resting-State Functional Magnetic Resonance Imaging in Evaluating Poor Emotion and Sleep Quality in Patients with Chronic Pain under Artificial Intelligence Algorithm

The balanced iterative reducing and clustering using hierarchies (BIRCH) method was adopted to optimize the results of the resting-state functional magnetic resonance imaging (RS-fMRI) to analyze the changes in the brain function of patients with chronic pain accompanied by poor emotion or abnormal sleep quality in this study, so as to provide data support for the prevention and treatment of clinical chronic pain with poor emotion or sleep quality. 159 patients with chronic pain who visited the hospital were selected as the research objects, and they were grouped according to the presence or absence of abnormalities in emotion and sleep. The patients without poor emotion and sleep quality were set as the control group (60 cases), and the patients with the above symptoms were defined in the observation group (90 cases). The brain function was detected by RS-fMRI technology based on the BIRCH algorithm. The results showed that the rand index (RI), adjustment of RI (ARI), and Fowlkes–Mallows index (FMI) results in the k-means, flow cytometry (FCM), and BIRCH algorithms were 0.82, 0.71, and 0.88, respectively. The scores of Hamilton Depression Scale (HAHD), Hamilton Anxiety Scale (HAMA), and Pittsburgh Sleep Quality Index (PSQI) were 7.26 ± 3.95, 7.94 ± 3.15, and 8.03 ± 4.67 in the observation group and 4.03 ± 1.95, 5.13 ± 2.35, and 4.43 ± 2.07 in the control group; the higher proportion of RS-fMRI was with abnormal brain signal connections. A score of 7 or more meant that the number of brain abnormalities was more than 90% and that of less than 7 was less than 40%, showing a statistically obvious difference in contrast (P < 0.05). Therefore, the BIRCH clustering algorithm showed reliable value in the optimization of RS-fMRI images, and RS-fMRI showed high application value in evaluating the emotion and sleep quality of patients with chronic pain.

Embedding Functional Brain Networks in Low Dimensional Spaces Using Manifold Learning Techniques

Background: fMRI data is inherently high-dimensional and difficult to visualize. A recent trend has been to find spaces of lower dimensionality where functional brain networks can be projected onto manifolds as individual data points, leading to new ways to analyze and interpret the data. Here, we investigate the potential of two powerful non-linear manifold learning techniques for functional brain networks representation: (1) T-stochastic neighbor embedding (t-SNE) and (2) Uniform Manifold Approximation Projection (UMAP) a recent breakthrough in manifold learning. Methods: fMRI data from the Human Connectome Project (HCP) and an independent study of aging were used to generate functional brain networks. We used fMRI data collected during resting state data and during a working memory task. The relative performance of t-SNE and UMAP were investigated by projecting the networks from each study onto 2D manifolds. The levels of discrimination between different tasks and the preservation of the topology were evaluated using different metrics. Results: Both methods effectively discriminated the resting state from the memory task in the embedding space. UMAP discriminated with a higher classification accuracy. However, t-SNE appeared to better preserve the topology of the high-dimensional space. When networks from the HCP and aging studies were combined, the resting state and memory networks in general aligned correctly. Discussion: Our results suggest that UMAP, a more recent development in manifold learning, is an excellent tool to visualize functional brain networks. Despite dramatic differences in data collection and protocols, networks from different studies aligned correctly in the embedding space.

Post-Stroke Cognitive Impairment: Pathophysiological Insights into Brain Disconnectome from Advanced Neuroimaging Analysis Techniques

The neurological symptoms of stroke have traditionally provided the foundation for functional mapping of the brain. However, there are many unresolved aspects in our understanding of cerebral activity, especially regarding high-level cognitive functions. This review provides a comprehensive look at the pathophysiology of post-stroke cognitive impairment in light of recent findings from advanced imaging techniques. Combining network neuroscience and clinical neurology, our research focuses on how changes in brain networks correlate with post-stroke cognitive prognosis. More specifically, we first discuss the general consequences of stroke lesions due to damage of canonical resting-state large-scale networks or changes in the composition of the entire brain. We also review emerging methods, such as lesion-network mapping and gradient analysis, used to study the aforementioned events caused by stroke lesions. Lastly, we examine other patient vulnerabilities, such as superimposed amyloid pathology and blood-brain barrier leakage, which potentially lead to different outcomes for the brain network compositions even in the presence of similar stroke lesions. This knowledge will allow a better understanding of the pathophysiology of post-stroke cognitive impairment and provide a theoretical basis for the development of new treatments, such as neuromodulation.

On the ability to exploit signal fluctuations in pseudocontinuous arterial spin labeling for inferring the major flow territories from a traditional perfusion scan

In arterial spin labeling (ASL) a magnetic label is applied to the flowing blood in feeding arteries allowing depiction of cerebral perfusion maps. The labeling efficiency depends, however, on blood velocity and local field inhomogeneities and is, therefore, not constant over time. In this work, we investigate the ability of statistical methods used in functional connectivity research to infer flow territory information from traditional pseudo-continuous ASL (pCASL) scans by exploiting artery-specific signal fluctuations. By applying an additional gradient during labeling the minimum amount of signal fluctuation that allows discrimination of the main flow territories is determined. The following three approaches were tested for their performance on inferring the large vessel flow territories of the brain: a general linear model (GLM), an independent component analysis (ICA) and t-stochastic neighbor embedding. Furthermore, to investigate the effect of large vessel pathology, standard ASL scans of three patients with a unilateral stenosis (>70%) of one of the internal carotid arteries were retrospectively analyzed using ICA and t-SNE. Our results suggest that the amount of natural-occurring variation in labeling efficiency is insufficient to determine large vessel flow territories. When applying additional vessel-encoded gradients these methods are able to distinguish flow territories from one another, but this would result in approximately 8.5% lower perfusion signal and thus also a reduction in SNR of the same magnitude.