Alteration of the Individual Metabolic Network of the Brain Based on Jensen-Shannon Divergence Similarity Estimation in Elderly Patients With Type 2 Diabetes Mellitus

Optimizing neural network algorithms for submerged membrane bioreactor: A comparative study of OVAT and RSM hyperparameter optimization techniques

Hyperparameter tuning is an important process to maximize the performance of any neural network model. This present study proposed the factorial design of experiment for screening and response surface methodology to optimize the hyperparameter of two artificial neural network algorithms. Feed-forward neural network (FFNN) and radial basis function neural network (RBFNN) are applied to predict the permeate flux of palm oil mill effluent. Permeate pump and transmembrane pressure of the submerge membrane bioreactor system are the input variables. Six hyperparameters of the FFNN model including four numerical factors (neuron numbers, learning rate, momentum, and epoch numbers) and two categorical factors (training and activation function) are used in hyperparameter optimization. RBFNN includes two numerical factors such as a number of neurons and spreads. The conventional method (one-variable-at-a-time) is compared in terms of optimization processing time and the accuracy of the model. The result indicates that the optimal hyperparameters obtained by the proposed approach produce good accuracy with a smaller generalization error. The simulation results show an improvement of more than 65% of training performance, with less repetition and processing time. This proposed methodology can be utilized for any type of neural network application to find the optimum levels of different parameters.

The promotion-like effect of the M1-STN hyperdirect pathway induced by ccPAS enhanced balance performances: From the perspective of brain connectivity

AIMS

The present study aimed to explore the effect of cortico-cortical paired-associative stimulation (ccPAS) in modulating hyperdirect pathway and its influence on balance performance.

METHODS

Forty healthy participants were randomly allocated to the active ccPAS group (n = 20) or the sham ccPAS group (n = 20). The primary motor cortex and subthalamic nucleus were stimulated sequentially with ccPAS. Unlike the active ccPAS group, one wing of coil was tilted to form a 90° angle with scalp of stimulation locations for the sham ccPAS group. Magnetic resonance imaging, functional reach test (FRT), timed up and go (TUG) test, and limit of stability (LOS) test were performed, and correlation between them was also analyzed.

RESULTS

Three participants in the sham ccPAS group were excluded because of poor quality of NIfTI images. The active group had strengthened hyperdirect pathway, increased functional connectivity (FC) between orbital part of frontal cortex and bilateral precuneus, and decreased FC among basal ganglia (all p < 0.05). Regional network properties of triangular and orbital parts of IFG, middle cingulate cortex, and hippocampus increased. The active group performed better in FRT and LOS (all p < 0.05). FRT positively correlated with FC of the hyperdirect pathway (r = 0.439, p = 0.007) and FCs between orbital part of frontal cortex and bilateral precuneus (all p < 0.05).

CONCLUSION

The ccPAS enhanced balance performance by promotion-like plasticity mechanisms through the hyperdirect pathway.

Metabolic Connectome and Its Role in the Prediction, Diagnosis, and Treatment of Complex Diseases

The interconnectivity of advanced biological systems is essential for their proper functioning. In modern connectomics, biological entities such as proteins, genes, RNA, DNA, and metabolites are often represented as nodes, while the physical, biochemical, or functional interactions between them are represented as edges. Among these entities, metabolites are particularly significant as they exhibit a closer relationship to an organism's phenotype compared to genes or proteins. Moreover, the metabolome has the ability to amplify small proteomic and transcriptomic changes, even those from minor genomic changes. Metabolic networks, which consist of complex systems comprising hundreds of metabolites and their interactions, play a critical role in biological research by mediating energy conversion and chemical reactions within cells. This review provides an introduction to common metabolic network models and their construction methods. It also explores the diverse applications of metabolic networks in elucidating disease mechanisms, predicting and diagnosing diseases, and facilitating drug development. Additionally, it discusses potential future directions for research in metabolic networks. Ultimately, this review serves as a valuable reference for researchers interested in metabolic network modeling, analysis, and their applications.

Cross-modality comparison between structural and metabolic networks in individual brain based on the Jensen-Shannon divergence method: a healthy Chinese population study

The study aimed to investigate the consistency and diversity between metabolic and structural brain networks at individual level constructed with divergence-based method in healthy Chinese population. The ¹⁸F-FDG PET and T1-weighted images of brain were collected from 209 healthy participants. The Jensen-Shannon divergence (JSD) was used to calculate metabolic or structural connectivities between any pair of brain regions and then individual brain networks were constructed. The global and regional topological properties of both networks were analyzed with graph theoretical analysis. Regional properties including nodal efficiency, degree, and betweenness centrality were used to define hub regions of networks. Cross-modality similarity of brain connectivity was analyzed with differential power (DP) analysis. The default mode network (DMN) had the largest number of brain connectivities with high DP values. The small-worldness indexes of metabolic and structural networks in all participants were greater than 1. The structural network showed higher assortativity and local efficiency than metabolic network, while hierarchy and global efficiency were greater in the metabolic network (all P < 0.001). Most of hubs in both networks were symmetrically spatial distributed in the regions of the DMN and subcortical nuclei including thalamus and amygdala, etc. The human brain presented small-world architecture both in perspective of individual metabolic and structural networks. There was a structural substrate that supported the brain to globally and efficiently integrate and process metabolic interaction across brain regions. The cross-modality cooperation or specialization in both networks might imply mechanisms of achieving higher-order brain functions.

Alternation of the Rich-Club Organization of Individual Brain Metabolic Networks in Parkinson’s Disease

Objective

The diagnosis of Parkinson’s disease (PD) remains challenging. Although ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography (PET) has revealed the metabolic abnormalities associated with PD at systemic levels, the underlying rich-club organization of the metabolic connectome in these patients remains largely unknown.

Materials and Methods

The data of 49 PD patients and 49 well-matched healthy controls (HCs) were retrieved and assessed. An individual metabolic connectome based on the standard uptake value (SUV) was built using the Jensen-Shannon Divergence Similarity Estimation (JSSE) method to compare the rich-club properties between PD patients and HC.

Results

Our results showed the rich-club organization of metabolic networks (normalized rich-club coefficients > 1) in the PD and HC group were within a range of thresholds. Further, patients with PD demonstrated lower strength and degree in rich-club connections compared with HCs (strength: HCs = 55.70 ± 8.52, PDs = 52.03 ± 10.49, p = 0.028; degree: HCs = 56.55 ± 8.60, PDs = 52.85 ± 10.62, p = 0.029), but difference between their feeder and local connections was not significant.

Conclusion

Individual metabolic networks combined with rich club analysis indicated that PD patients had decreased rich club connections but similar feeder and local connections compared with HCs, indicating rich club connections as a promising marker for early diagnosis of PD.

Age-Related Changes in Topological Properties of Individual Brain Metabolic Networks in Rats

Normal aging causes profound changes of structural degeneration and glucose hypometabolism in the human brain, even in the absence of disease. In recent years, with the extensive exploration of the topological characteristics of the human brain, related studies in rats have begun to investigate. However, age-related alterations of topological properties in individual brain metabolic network of rats remain unknown. In this study, a total of 48 healthy female Sprague–Dawley (SD) rats were used, including 24 young rats and 24 aged rats. We used Jensen-Shannon Divergence Similarity Estimation (JSSE) method for constructing individual metabolic networks to explore age-related topological properties and rich-club organization changes. Compared with the young rats, the aged rats showed significantly decreased clustering coefficient (Cp) and local efficiency (Eloc) across the whole-brain metabolic network. In terms of changes in local network measures, degree (D) and nodal efficiency (Enod) of left posterior dorsal hippocampus, and Enod of left olfactory tubercle were higher in the aged rats than in the young rats. About the rich-club analysis, the existence of rich-club organization in individual brain metabolic networks of rats was demonstrated. In addition, our findings further confirmed that rich-club connections were susceptible to aging. Relative to the young rats, the overall strength of rich-club connections was significantly reduced in the aged rats, while the overall strength of feeder and local connections was significantly increased. These findings demonstrated the age-related reorganization principle of the brain structure and improved our understanding of brain alternations during aging.