Higher Mutual Information Research Articles

Decoding MRI-informed brain age using mutual information

ObjectiveWe aimed to develop a standardized method to investigate the relationship between estimated brain age and regional morphometric features, meeting the criteria for simplicity, generalization, and intuitive interpretability.MethodsWe utilized T1-weighted magnetic resonance imaging (MRI) data from the Cambridge Centre for Ageing and Neuroscience project (N = 609) and employed a support vector regression method to train a brain age model. The pre-trained brain age model was applied to the dataset of the brain development project (N = 547). Kraskov (KSG) estimator was used to compute the mutual information (MI) value between brain age and regional morphometric features, including gray matter volume (GMV), white matter volume (WMV), cerebrospinal fluid (CSF) volume, and cortical thickness (CT).ResultsAmong four types of brain features, GMV had the highest MI value (8.71), peaking in the pre-central gyrus (0.69). CSF volume was ranked second (7.76), with the highest MI value in the cingulate (0.87). CT was ranked third (6.22), with the highest MI value in superior temporal gyrus (0.53). WMV had the lowest MI value (4.59), with the insula showing the highest MI value (0.53). For brain parenchyma, the volume of the superior frontal gyrus exhibited the highest MI value (0.80).ConclusionThis is the first demonstration that MI value between estimated brain age and morphometric features may serve as a benchmark for assessing the regional contributions to estimated brain age. Our findings highlighted that both GMV and CSF are the key features that determined the estimated brain age, which may add value to existing computational models of brain age.Critical relevance statementMutual information (MI) analysis reveals gray matter volume (GMV) and cerebrospinal fluid (CSF) volume as pivotal in computing individuals’ brain age.Key PointsMutual information (MI) interprets estimated brain age with morphometric features.Gray matter volume in the pre-central gyrus has the highest MI value for estimated brain age.Cerebrospinal fluid volume in the cingulate has the highest MI value.Regarding brain parenchymal volume, the superior frontal gyrus has the highest MI value.The value of mutual information underscores the key brain regions related to brain age.Graphical

Estimation of complete mutual information exploiting nonlinear magnitude-phase dependence: Application to spatial FNC for complex-valued fMRI data

BackgroundReal-valued mutual information (MI) has been used in spatial functional network connectivity (FNC) to measure high-order and nonlinear dependence between spatial maps extracted from magnitude-only functional magnetic resonance imaging (fMRI). However, real-valued MI cannot fully capture the group differences in spatial FNC from complex-valued fMRI data with magnitude and phase dependence. MethodsWe propose a complete complex-valued MI method according to the chain rule of MI. We fully exploit the dependence among magnitudes and phases of two complex-valued signals using second and fourth-order joint entropies, and propose to use a Gaussian copula transformation with a lower bound property to avoid inaccurate estimation of joint probability density function when computing the joint entropies. ResultsThe proposed method achieves more accurate MI estimates than the two histogram-based (normal and symbolic approaches) and kernel density estimation methods for simulated signals, and enhances group differences in spatial functional network connectivity for experimental complex-valued fMRI data. Comparison with existing methodsCompared with the simplified complex-valued MI and real-valued MI, the proposed method yields higher MI estimation accuracy, leading to 17.4 % and 145.5 % wider MI ranges, and more significant connectivity differences between healthy controls and schizophrenia patients. A unique connection between executive control network (EC) and right frontal parietal areas, and three additional connections mainly related to EC are detected than the simplified complex-valued MI. ConclusionsWith capability in quantifying MI fully and accurately, the proposed complex-valued MI is promising in providing qualified FNC biomarkers for identifying mental disorders such as schizophrenia.

The Impact of Lexical Bundle Length on L2 Oral Proficiency

Lexical bundles (LBs) are crucial in L2 oral proficiency, yet their complexity in terms of length is under-researched. This study therefore examines the relationship between longer and shorter LBs and oral proficiency among 150 L2 learners of varying proficiency levels at a UK university. Through the analysis of oral presentation data (scores ranging from intermediate to advanced) and employing a combined text-internal and text-external approach (two- to five-word bundles), this study advances an innovative text-internal LB refinement procedure, thus isolating the unique contribution of LB length. Robust regression, dominance analysis, and random forest statistical techniques reveal the predictive power of bigram mutual information (MI) and longer three-to-five-word sequences on higher proficiency scores. Our results show that learners using higher MI score bigrams tend to perform better in their presentations, with a strong positive impact on scores (b = 14.38, 95% CI [8.01, 20.76], t = 4.42; dominance weight = 58.63%). Additionally, the use of longer three-to-five-word phrases also contributes to better performance, though to a lesser extent (dominance weight = 18.80%). These findings highlight the pedagogical potential of a nuanced approach to the strategic deployment of LBs, particularly bigram MI, to foster oral proficiency. Suggestions for future LB proficiency research are discussed in relation to L2 speech production models.

Verifying the security of a continuous variable quantum communication protocol via quantum metrology

Quantum mechanics offers the possibility of unconditionally secure communication between multiple remote parties. Security proofs for such protocols typically rely on bounding the capacity of the quantum channel in use. In a similar manner, Cramér-Rao bounds in quantum metrology place limits on how much information can be extracted from a given quantum state about some unknown parameters of interest. In this work we establish a connection between these two areas. We first demonstrate a three-party sensing protocol, where the attainable precision is dependent on how many parties work together. This protocol is then mapped to a secure access protocol, where only by working together can the parties gain access to some high security asset. Finally, we map the same task to a communication protocol where we demonstrate that a higher mutual information can be achieved when the parties work collaboratively compared to any party working in isolation.

A Corpus-based Ecosophical Analysis of Discourse Produced Around the China-Pakistan Economic Corridor (CPEC)

This research explores the Ecosophy of discourse produced around the China-Pakistan Economic Corridor (CPEC). The study starts with building a corpus of texts, taken from the official website of the CPEC. The list of 18 key words was made, based on three United Nations (UN) Sustainable Development Goals (SDGs) out of 17 goals. Theoretical framework of Stibbe’s (2015) ecocritical discourse perspective is used to develop the Ecosophy of the CPEC, while analytical framework is based on three UN sustainable development goals outlined in 2015. Corpus linguistics as a methodology is used to undergo quantitative and qualitative analysis. In quantitative analysis, collocates were analyzed by identifying Mutual Information (MI) score. The highest MI score (11.26013) is gained by the collocation named CPEC-Deforestation and the lowest is of CPEC-Water (0.87352) which show that the infrastructure project of the CPEC caused deforestation and less affected water.

Textile wearable system for knee angle monitoring in three planes

Introduction
 Monitoring biomechanics is crucial in sports and rehabilitation, and frontal knee angle is of special interest in these applications. Current solutions – optical motion capture (OMC), or inertial measurement units suits – are costly, spatially constrained, and impractical for use in daily life. Textile-based wearable systems are a valuable alternative for unobtrusive movement monitoring. Textile-based wearables for knee angle monitoring have mostly been used for sagittal angle prediction, however, frontal knee angle measurement is more challenging. We investigated the design and performance of a smart garment for the detection of knee joint angles in three planes during different activities.
 Methods
 We equipped a pair of tight pants with four helical auxetic yarn capacitive strain sensors (Cuthbert et al., 2022) placed close to the knees. The exact positioning was optimized with an OMC study: markers were placed in potential sensor locations (Gholami et al., 2019) and the pairs of markers whose distance had the highest mutual information with knee angles were selected for sensor placement. A healthy participant performed walking and turning around, and knee ab/adduction activities wearing the sensorized prototype. The latter activity emphasized knee motion in the frontal and transverse planes. The capacitances from the sensors were recorded with a custom electronics board that transmitted data wirelessly to a smartphone. Multiple regression algorithms were implemented to predict knee angles from the strain sensors data, with the ground truth obtained from the OMC data recorded simultaneously during the experiments.
 Results
 The optimal sensor placements were above the kneecaps, orientated as the vastus medialis and the rectus femoris. Xgboost regression algorithm yielded best performance for walking with root mean square errors (RMSE) of 10.79°, 3.77°, and 2.49° for the sagittal, frontal, and transverse angles, respectively. Linear regression performed the best for knee ab/adduction with RMSEs of 8.96°, 6.33°, and 1.58° for the sagittal, frontal, and transverse angles (Fornaciari, 2023).
 Discussion/Conclusion
 The smart garment system was overall able to track the knee angle in three planes. The larger errors, compared with previous works (Gholami et al., 2019), reported for the walking and turning around movement are likely because of high variations in the movements of the participants during turning around. Additionally, the proposed system showed capability to monitor frontal and transverse angles with an average RMSE of 3.5°. The larger error values of the sagittal angles are likely because of higher range of motion in that plane. The proposed system allows for continuous and unobtrusive knee angle monitoring outside of the laboratory settings in the comfortable form factor of smart clothing.
 References
 Cuthbert, T. J., Hannigan, B. C., Roberjot, P., Shokurov, A. V., & Menon, C. (2022). HACS: Helical auxetic yarn capacitive strain sensors with sensitivity beyond the theoretical limit. Advanced materials, 35(10) Article 2209321. https://doi.org/10.1002/adma.202209321
 Fornaciari, A. (2023). Wearable technology for lower limb movement monitoring [Master’s thesis]. Politecnico di Milano.
 Gholami, M., Rezaei, A., Cuthbert, T. J., Napier, C., & Menon, C. (2019). Lower body kinematics monitoring in running using fabric-based wearable sensors and deep convolutional neural networks. Sensors, 19(23), Article 5325. https://doi.org/10.3390/s19235325

MUTUAL INFORMATION OF NEUROIMAGES AND GENETIC VARIANTS TO STUDY THE DETERMINANTS OF BRAIN AGING

Abstract Understanding the genetic determinants of brain aging is important because advanced biological age of the brain is a risk factor for Alzheimer’s disease (AD). Single-nucleotide polymorphisms (SNPs) are genetic variations with effects on brain aging; mapping their influence on brain structure can elucidate the genetic correlates of AD. We adopted an information theoretic approach to this by computing the mutual information (MI) between (A) 22 AD-risk SNPs and (B) the MRI intensities of the cortex in 6,000 UK Biobank participants (~50% females) aged 54 to 84 years. These cortex-wide calculations quantified how SNPs impact brain structure. As expected, the APOE-ε4 allele, the strongest genetic risk factor for AD, exhibited significantly (p < 0.025) higher MI with MRI intensity, when compared to the average MI expected for a null distribution model, in the medial parietal lobe, pars orbitalis, and left inferior occipital lobe, three brain structures responsible for memory recall, language processing, and visual perception. These findings reflect the memory loss, decreased speech, and reduced peripheral vision experienced by individuals with AD. The ABCA7 allele, a risk factor for late- onset AD, had significantly (p < 0.025) higher MI in the inferior temporal sulcus and angular gyrus. These structures are responsible for visual recognition and attention, two functions that are severely impaired by AD. Other findings reveal similar relationships between AD-risk SNPs and brain structure. Our findings illustrate how mapping the MI between MRI intensities and SNPs provides insight into genetic influences on brain structures linked to AD symptoms.

Neural coding of autonomic functions in different states of consciousness

Detecting signs of residual neural activity in patients with altered states of consciousness is a crucial issue for the customization of neurorehabilitation treatments and clinical decision-making. With this large observational prospective study, we propose an innovative approach to detect residual signs of consciousness via the assessment of the amount of autonomic information coded within the brain. The latter was estimated by computing the mutual information (MI) between preprocessed EEG and ECG signals, to be then compared across consciousness groups, together with the absolute power and an international qualitative labeling. One-hundred seventy-four patients (73 females, 42%) were included in the study (median age of 65 years [IQR = 20], MCS +: 29, MCS −: 23, UWS: 29). Electroencephalography (EEG) information content was found to be mostly related to the coding of electrocardiography (ECG) activity, i.e., with higher MI (p < 0.05), in Unresponsive Wakefulness Syndrome and Minimally Consciousness State minus (MCS −). EEG-ECG MI, besides clearly discriminating patients in an MCS − and +, significantly differed between lesioned areas (sides) in a subgroup of unilateral hemorrhagic patients. Crucially, such an accessible and non-invasive measure of residual consciousness signs was robust across electrodes and patient groups. Consequently, exiting from a strictly neuro-centric consciousness detection approach may be the key to provide complementary insights for the objective assessment of patients' consciousness levels and for the patient-specific planning of rehabilitative interventions.

Environmental factors influencing DDT–DDE spatial distribution in an agricultural drainage system determined by using machine learning techniques

The presence and persistence of pesticides in the environment are environmental problems of great concern due to the health implications for humans and wildlife. The persistence of DDT–DDE in a Mediterranean coastal plain where pesticides were widely used and were banned decades ago is the aim of this study. Different sources of analytical information from water and soil analysis and topography and geographical variables were combined with the purpose of analyzing which environmental factors are more likely to condition the spatial distribution of DDT–DDE in the drainage watercourses of the area. An approach combining machine learning techniques, such as Random Forest and Mutual Information (MI), for classifying DDT–DDE concentration levels based on other environmental predictive variables was applied. In addition, classification procedure was iteratively performed with different training/validation partitions in order to extract the most informative parameters denoted by the highest MI scores and larger accuracy assessment metrics. Distance to drain canals, soil electrical conductivity, and soil sand texture fraction were the most informative environmental variables for predicting DDT–DDE water concentration clusters.

Early deficits in GABA inhibition parallels an increase in L-type Ca2+ currents in the jaw motor neurons of SOD1G93A mouse model for ALS

Amyotrophic Lateral Sclerosis (ALS) involves protracted pre-symptomatic periods of abnormal motor neuron (MN) excitability occurring in parallel with central and peripheral synaptic perturbations. Focusing on inhibitory control of MNs, we first compared longitudinal changes in pre-synaptic terminal proteins for GABA and glycine neurotransmitters around the soma of retrogradely identified trigeminal jaw closer (JC) MNs and ChAT-labeled midbrain extraocular (EO) MNs in the SOD1G93A mouse model for ALS. Fluorescence immunocytochemistry and confocal imaging were used to quantify GAD67 and GlyT2 synaptic bouton density (SBD) around MN soma at pre-symptomatic ages ∼P12 (postnatal), ∼P50 (adult) and near disease end-stage (∼P135) in SOD1G93A mice and age-matched wild-type (WT) controls. We noted reduced GAD67 innervation in the SOD1G93A trigeminal jaw closer MNs around P12, relative to age-matched WT and no significant difference around P50 and P135. In contrast, both GAD67 and GlyT2 innervation were elevated in the SOD1G93A EO MNs at the pre-symptomatic time points. Considering trigeminal MNs are vulnerable in ALS while EO MNs are spared, we suggest that upregulation of inhibition in the latter might be compensatory. Notable contrast also existed in the innate co-expression patterns of GAD67 and GlyT2 with higher mutual information (co-dependency) in EO MNs compared to JC in both SOD1G93A and WT mice, especially at adult stages (P50 and P135). Around P12 when GAD67 terminals expression was low in the mutant, we further tested for persistent GABA inhibition in those MNs using in vitro patch-clamp electrophysiology. Our results show that SOD1G93A JC MNs have reduced persistent GABA inhibition, relative to WT. Pharmacological blocking of an underlying tonically active GABA conductance using the GABA-α5 subunit inverse agonist, L-655-708, disinhibited WT JC MNs and lowered their recruitment threshold, suggesting its role in the control of intrinsic MN excitability. Quantitative RT-PCR in laser dissected JC MNs further supported a reduction in GABA-α5 subunit mRNA expression in the mutant. In light of our previous report that JC MNs forming putative fast motor units have lower input threshold in the SOD1G93A mice, we suggest that our present result on reduced GABA-α5 tonic inhibition provides for a mechanism contributing to such imbalance. In parallel with reduced GABA inhibition, we noted an increase in voltage-gated L-type Ca2+ currents in the mutant JC MNs around P12. Together these results support that, early modifications in intrinsic properties of vulnerable MNs could be an adaptive response to counter synaptic deficits.

Synergistic effects of multiword sequences structure, function, frequency and association on raters’ evaluations of essay quality

Despite accumulated research findings confirming the link of multiword sequences (MWSs) structures and functions to essay quality, as well as the connection between MWSs statistical features (e.g., their frequency and association strengths in BNC/COCA) and writing quality, to date no study integrated these two separate lines of investigations. It remains to investigate whether and how MWSs structures, functions and their statistical features jointly affect writing quality. Drawing on 900 rated argumentative essays composed by Chinese grade 12 students in National Matriculation Test, the present study employed CollGram to automatically identify the nativelike 4-word sequences in these essays and to analyze their frequency and Mutual Information (MI) scores in COCA. The structures and functions of frequent nativelike 4-word sequences were also analyzed manually. A serial of linear mixed-effect models was constructed to investigate their main effects as well as interaction effects on essay scores. The best fit model revealed the links of higher essay scores to higher MI scores, to more noun-phrase sequences, to more stance sequences, as well as to fewer referential sequences. Additionally, the interaction of prepositional phrase sequences and their frequency in COCA affected essay scores, so did the interaction of verb phrase sequences and their MI in COCA, as well as the interaction of noun phrase sequences and their MI in COCA. The findings provide new insights into the complex interaction between MWSs structures, functions and their statistical features, as well as their joint effects on writing quality.

Beamforming Design for Integrated Sensing and Communication Systems With Finite Alphabet Input

Joint beamforming for integrated sensing and communication (ISAC) is an efficient way to combine two functionalities in a system at the waveform level. This letter proposes a beamforming design for ISAC systems with finite alphabet signaling. We formulate a problem to maximize the minimum Euclidean distance (MMED) among noise-free received signal vectors under a sensing constraint and a given power constraint. To tackle the formulated optimization problem, we transform it into a semi-definite programming (SDP) and solve it by the semi-definite relaxation (SDR) method. Comprehensive comparisons with existing schemes show that our proposed beamforming offers lower symbol error rate (SER), higher mutual information, and also better sensing performance than existing ISAC beamforming designs.

The coherence function and its information content for optical metrology

Abstract The coherence function offers new possibilities for optical metrology that are not available with conventional wave field sensing. Its measurement involves a spatio-temporal sampling of the wave fields modulated by the object under investigation. Temporal sampling is well known e. g. by means of White Light Interferometry (WLI) and spatial sampling can e. g. performed by Computational Shear Interferometry (CoSI). The present paper describes an approach that combines both temporal and spatial sampling using a robust common-path setup. While the evaluation of the coherence function is more elaborate than approaches that either sample the temporal or the spatial domain, an information theoretical treatment shows that it also delivers more information about the object under investigation. Our approach is based on the mutual information that represents the reduction of uncertainty about the object as a consequence of the measurements performed. Using a simplified measurement case, we calculate the mutual information for different measurement situations and demonstrate that spatio-temporal sampling of the coherence function results in a higher mutual information as compared to classical approaches. Based on the proposed approach, we identify further open research tasks for an efficient information extraction from the coherence function to surpass current limitations of optical metrology.

Automatic Muscle Artifacts Identification and Removal from Single-Channel EEG Using Wavelet Transform with Meta-Heuristically Optimized Non-Local Means Filter.

Electroencephalogram (EEG) signals may get easily contaminated by muscle artifacts, which may lead to wrong interpretation in the brain–computer interface (BCI) system as well as in various medical diagnoses. The main objective of this paper is to remove muscle artifacts without distorting the information contained in the EEG. A novel multi-stage EEG denoising method is proposed for the first time in which wavelet packet decomposition (WPD) is combined with a modified non-local means (NLM) algorithm. At first, the artifact EEG signal is identified through a pre-trained classifier. Next, the identified EEG signal is decomposed into wavelet coefficients and corrected through a modified NLM filter. Finally, the artifact-free EEG is reconstructed from corrected wavelet coefficients through inverse WPD. To optimize the filter parameters, two meta-heuristic algorithms are used in this paper for the first time. The proposed system is first validated on simulated EEG data and then tested on real EEG data. The proposed approach achieved average mutual information (MI) as 2.9684 ± 0.7045 on real EEG data. The result reveals that the proposed system outperforms recently developed denoising techniques with higher average MI, which indicates that the proposed approach is better in terms of quality of reconstruction and is fully automatic.

Mobile Two-Way Molecular Communication via Diffusion Using Amplify-and-Forward and Analog Network Coding.

Mobile molecular communication via diffusion (MCvD) has attracted lots of attentions due to its time-varying channels. In this paper, we investigate a mobile two-way MCvD model, which consists of two mobile source nanomachines and a mobile relay nanomachine. The amplify-and-forward (AF) relaying and analog network coding (ANC) are utilized to implement the exchange of information between two source nanomachines in this model. To explore the performance of the mobile two-way MCvD system, we first adopt the depleted molecule shift keying (D-MoSK) modulation, and then the mathematical expressions of symbol error probability (SEP) and mutual information are derived by using AF and ANC scheme on the basis of maximum a posteriori (MAP) probability detection. Finally, we present the numerical and simulation results. Compared with the AF-No-ANC scheme which is without use of ANC scheme, the scheme of combining AF and ANC can significantly improve the performance of SEP and mutual information. Moreover, the D-MoSK modulation outperforms the molecule shift keying (MoSK) modulation for this mobile two-way MCvD system. In addition, we propose the evaluation and discussion about the impacts of several important parameters on the performance of this system. These results can be used to design mobile two-way MCvD system with lower SEP and higher mutual information.

Optimal vocabulary selection approaches for privacy-preserving deep NLP model training for information extraction and cancer epidemiology.

With the use of artificial intelligence and machine learning techniques for biomedical informatics, security and privacy concerns over the data and subject identities have also become an important issue and essential research topic. Without intentional safeguards, machine learning models may find patterns and features to improve task performance that are associated with private personal information. The privacy vulnerability of deep learning models for information extraction from medical textural contents needs to be quantified since the models are exposed to private health information and personally identifiable information. The objective of the study is to quantify the privacy vulnerability of the deep learning models for natural language processing and explore a proper way of securing patients' information to mitigate confidentiality breaches. The target model is the multitask convolutional neural network for information extraction from cancer pathology reports, where the data for training the model are from multiple state population-based cancer registries. This study proposes the following schemes to collect vocabularies from the cancer pathology reports; (a) words appearing in multiple registries, and (b)words that have higher mutual information. We performed membership inference attacks on the models in high-performance computing environments. The comparison outcomes suggest that the proposed vocabulary selection methods resulted in lower privacy vulnerability while maintaining the same level of clinical task performance.

Application of multidomain sensor image fusion and training data augmentation for enhanced CNN image classification

Convolutional neural networks (CNNs) provide the sensing and detection community with a discriminative approach for classifying images. However, one of the largest limitations of deep CNN image classifiers is the need for extensive training datasets containing a variety of image representations. While current methods, such as generative adversarial network data augmentation, additions of noise, rotations, and translations, can allow CNNs to better associate new images and their feature representations to ones of a learned image class, many fail to provide new contexts of ground truth feature information. To expand the association of critical class features within CNN image training datasets, an image pairing and training dataset augmentation paradigm via a multi-sensor domain image data fusion algorithm is proposed. This algorithm uses a mutual information (MI) and merit-based feature selection subroutine to pair highly correlated cross-domain images from multiple sensor domain image datasets. It then re-augments the corresponding cross-domain image pairs into the opposite sensor domain’s feature set via a highest MI, cross sensor domain, and image concatenation function. This augmented image set then acts to retrain the CNN to recognize greater generalizations of image class features via cross domain, mixed representations. Experimental results indicated an increased ability of CNNs to generalize and discriminate between image classes during testing of class images from synthetic aperture radar vehicle, solar cell device reliability screening, and lung cancer detection image datasets.

A framework for transportation infrastructure cost prediction: a support vector regression approach

ABSTRACT Highway agencies and Departments of Transportation (DOTs) often face a challenge in estimating annual maintenance expenditures (AMEXs) to allocate available funds to various divisions in their jurisdictions. Few AMEX models were developed in the past, which are based on traditional statistical methods. This study introduces a support vector regression (SVR) approach to predict AMEX using data on the U.S. rural interstate highway pavement performance. Results show that the best SVR model is based on the Gaussian kernel and sequential minimal optimization algorithm. The developed SVR model can reliably predict the AMEX at the state level with a root mean squared error of 1.28. Entropy analysis results indicate the existence of mutual information (MI) between all the considered variables and the AMEX, with the highest MI value of about 79% between the AMEX and the number of lane miles with IRI values greater than 1.5 m/km.

Obstructive sleep apnea predicts 10-year cardiovascular disease-related mortality in the Sleep Heart Health Study: a machine learning approach.

Obstructive sleep apnea (OSA) is considered to be an important risk factor for the development of cardiovascular disease (CVD). This study aimed to develop and evaluate a machine learning approach with a set of features for assessing the 10-year CVD mortality risk of the OSA population. This study included 2,464 patients with OSA who met study inclusion criteria and were selected from the Sleep Heart Health Study. We evaluated the importance of potential features by mutual information. The top 9 features were selected to develop a random forest model. We evaluated the model performance on a test set (n = 493) using the area under the receiver operating curve with 95% confidence interval and confusion matrix. A random forest model awarded the highest area under the receiver operating curve of 0.84 (95% confidence interval: 0.78-0.89). The specificity and sensitivity were 73.94% and 81.82%, respectively. Sixty-three years old was a threshold for increased risk of 10-year CVD mortality. Persons with severe OSA had higher risk than those with mild OSA. This study demonstrated that a random forest model can provide a quick assessment of the risk of 10-year CVD mortality. Our model may be more informative for patients with OSA in determining their future CVD mortality risk. Li A, Roveda JM, Powers LS, Quan SF. Obstructive sleep apnea predicts 10-year cardiovascular disease-related mortality in the Sleep Heart Health Study: a machine learning approach. J Clin Sleep Med. 2022;18(2):497-504.

Amplify-and-forward relaying in mobile multi-hop molecular communication via diffusion

In this paper, we investigate amplify-and-forward (AF) relaying scheme for mobile three-dimensional (3D) multi-hop molecular communication via diffusion (MCvD) in the stochastic channel by using different type of molecules (DTM) and same type of molecules (STM) in each hop, respectively. Under the DTM scheme, both the inter-symbol interference (ISI) and the noise of each link are taken into account Compared with the DTM scheme, an additional factor should be considered for the STM scheme, which is the self-interference (SI) effect. Under DTM and STM, we derive the mathematical expressions of optimal detection thresholds at receiver node in mobile two-hop MCvD system by using the maximum-a-posterior (MAP) decision method. Then we extend the analysis to mobile multi-hop MCvD system. Furthermore, the performances of bit error probability and mutual information are evaluated via particle-based simulation of the Brownian motion. The numerical and simulation results reveal that the performance under DTM outperforms that under STM, which indicates the potential of AF relaying scheme under DTM to improve the overall performance of this system. Besides, the comparison results between mobile and static multi-hop link under DTM and STM are given. The obtained results are expected to provide guidance significance for designing mobile multi-hop MCvD system with lower bit error probability and higher mutual information.