Machine learning-based classifying of risk-takers and risk-aversive individuals using resting-state EEG data: A pilot feasibility study

BACKGROUND

Decision-making is vital in interpersonal interactions and a country's economic and political conditions. People, especially managers, have to make decisions in different risky situations. There has been a growing interest in identifying managers' personality traits (i.e., risk-taking or risk-averse) in recent years. Although there are findings of signal decision-making and brain activity, the implementation of an intelligent brain-based technique to predict risk-averse and risk-taking managers is still in doubt.

METHODS

This study proposes an electroencephalogram (EEG)-based intelligent system to distinguish risk-taking managers from risk-averse ones by recording the EEG signals from 30 managers. In particular, wavelet transform, a time-frequency domain analysis method, was used on resting-state EEG data to extract statistical features. Then, a two-step statistical wrapper algorithm was used to select the appropriate features. The support vector machine classifier, a supervised learning method, was used to classify two groups of managers using chosen features.

RESULTS

Intersubject predictive performance could classify two groups of managers with 74.42% accuracy, 76.16% sensitivity, 72.32% specificity, and 75% F1-measure, indicating that machine learning (ML) models can distinguish between risk-taking and risk-averse managers using the features extracted from the alpha frequency band in 10 s analysis window size.

CONCLUSIONS

The findings of this study demonstrate the potential of using intelligent (ML-based) systems in distinguish between risk-taking and risk-averse managers using biological signals.

Burn characterization using object-oriented hyperspectral image classification

This paper presents a new approach based on hyperspectral imaging combined with an object-oriented classification method that allows the generation of burn depth classification maps facilitating easier characterization of burns. Hyperspectral images of 14 patients diagnosed with burns on the upper and lower limbs were acquired using a pushbroom hyperspectral imaging system. The images were analyzed using an object-oriented classification approach that uses objects with specific spectral, textural and spatial attributes as the minimum unit for classifying information. The method performance was evaluated in terms of overall accuracy, sensitivity, precision and specificity computed from the confusion matrix. The results revealed that the approach proposed in this study performed well in differentiating burn classes with a high level of overall accuracy (95.99% ± 0.60%), precision (97.30% ± 2.46%), sensitivity (97.23% ± 3.02%) and specificity (98.02% ± 1.98%). In conclusion, the object-based approach for burns hyperspectral images classification can provide maps that can help surgeons identify with better precision different depths of burn wounds.

Investigation of molecular regulation mechanism under the pathophysiology of subarachnoid hemorrhage

This study aimed to investigate the molecular mechanism under the pathophysiology of subarachnoid hemorrhage (SAH) and identify the potential biomarkers for predicting the risk of SAH. Differentially expressed mRNAs (DEGs), microRNAs, and lncRNAs were screened. Protein-protein interaction (PPI), drug-gene, and competing endogenous RNA (ceRNA) networks were constructed to determine candidate RNAs. The optimized RNAs signature was established using least absolute shrinkage and selection operator and recursive feature elimination algorithms. A total of 124 SAH-related DEGs were identified, and were enriched in inflammatory response, TNF signaling pathway, and others. PPI network revealed 118 hub genes such as TNF, MMP9, and TLR4. Drug-gene network revealed that chrysin targeted more genes, such as TNF and MMP9. JMJD1C-AS-hsa-miR-204-HDAC4/SIRT1 and LINC01144-hsa-miR-128-ADRB2/TGFBR3 regulatory axes were found from ceRNA network. From these networks, 125 candidate RNAs were obtained. Of which, an optimal 38 RNAs signatures (2 lncRNAs, 1 miRNA, and 35 genes) were identified to construct a Support Vector Machine classifier. The predictive value of 38 biomarkers had an AUC of 0.990. Similar predictive performance was found in external validation dataset (AUC of 0.845). Our findings provided the potential for 38 RNAs to serve as biomarkers for predicting the risk of SAH. However, their application values should be further validated in clinical.

Detection and Classification of Myocardial Infarction with Support Vector Machine Classifier Using Grasshopper Optimization Algorithm

Background:

Providing a noninvasive, rapid, and cost-effective approach to diagnose of myocardial infarction (MI) is essential in the early stages of electrocardiogram (ECG) signaling. In this article, we proposed the new optimization method for support vector machine (SVM) classifier to MI classification.

Methods:

After preprocessing ECG signal and noise removal, three features such as Q-wave integral, T-wave integral, and QRS-complex integral have been extracted in this study. After that, different statistical tests have evaluated the matrix of these features. To more accurately detect and classify the MI disease, optimizing the SVM classification parameters using the grasshopper optimization algorithm (GOA) was first used in this study (that called SVM-GOA).

Results:

After applying the GOA on the SVM classifier for all three kernels, the final results of MI detection for sensitivity, specificity, and accuracy were 100% ± 0%, 100% ± 0%, and 100% ± 0%, respectively. The final results of different MI types' classification after applying the GOA on SVM for polynomial kernel were obtained 100% ± 0%, 97.37% ± 0%, and 94.2% ± 0.2% for sensitivity and specificity and accuracy, respectively. However, the results of both linear and RBF kernels that were used for the SVM classifier method have also shown a significant increase after using GOA.

Conclusion:

This article's results show the highly desirable effect of applying a GOA to optimize different kernel parameters used in the SVM classifier for accurate detection and classification of MI. The proposed algorithm's final results show that the proposed system has a relatively higher performance than other previous studies.

A Semi-Supervised Method for Tumor Segmentation in Mammogram Images

Background:

Breast cancer is one of the most common cancers in women. Mammogram images have an important role in the treatment of various states of this cancer. In recent years, machine learning methods have been widely used for tumor segmentation in mammogram images. Pixel-based segmentation methods have been presented using both supervised and unsupervised learning approaches. Supervised learning methods are usually fast and accurate, but they usually use a large number of labeled data. Besides, providing these samples is very hard and usually expensive. Unsupervised learning methods do not require the labels of the training data for decision making and they completely ignore the prior knowledge that may lead to a low performance. Semi-supervised learning methods which use a small number of labeled data solve the problem of providing the high number of samples in supervised methods, while they usually result in a higher accuracy in comparison to the unsupervised methods.

Methods:

In this study, we used a semisupervised method for tumor segmentation in which the pixel information is used for the classification. The static and gray level run length matrix features for each pixel are considered as the features, and Fisher discriminant analysis (FDA) is used for feature reduction. A cotraining algorithm based on support vector machine and Bayes classifiers is proposed for tumor segmentation on MIAS data set.

Results and Conclusion:

The results show that the proposed method outperforms both supervised methods.

A six‑gene support vector machine classifier contributes to the diagnosis of pediatric septic shock

Septic shock is induced by an uncontrolled inflammatory immune response to pathogens and the survival rate of patients with pediatric septic shock (PSS) is particularly low, with a mortality rate of 25‑50%. The present study explored the mechanisms of PSS using four microarray datasets (GSE26378, GSE26440, GSE13904 and GSE4607) that were obtained from the Gene Expression Omnibus database. Based on the MetaDE package, the consistently differentially expressed genes (DEGs) in the four datasets were screened. Using the WGCNA package, the disease‑associated modules and genes were identified. Subsequently, the optimal feature genes were further selected using the caret package. Finally, a support vector machine (SVM) classifier based on the optimal feature genes was built using the e1071 package. Initially, there were 2,699 consistent DEGs across the four datasets. From the 10 significantly stable modules across the datasets, four stable modules (including the magenta, purple, turquoise and yellow modules), in which the consistent DEGs were significantly enriched (P<0.05), were further screened. Subsequently, six optimal feature genes (including cysteine rich transmembrane module containing 1, S100 calcium binding protein A9, solute carrier family 2 member 14, stomatin, uridine phosphorylase 1 and utrophin) were selected from the genes in the four stable modules. Additionally, an effective SVM classifier was constructed based on the six optimal genes. The SVM classifier based on the six optimal genes has the potential to be applied for PSS diagnosis. This may improve the accuracy of early PSS diagnosis and suggest possible molecular targets for interventions.

Detection of focal electroencephalogram signals using higher-order moments in EMD-TKEO domain

Detection of epileptogenic focus based on electroencephalogram (EEG) signal screening is an important pre-surgical step to remove affected regions inside the human brain. Considering the fact above, in this work, a novel technique for detection of focal EEG signals is proposed using a combination of empirical mode decomposition (EMD) and Teager–Kaiser energy operator (TKEO). EEG signals belonging to focal (Fo) and non-focal (NFo) groups were at first decomposed into a set of intrinsic mode functions (IMFs) using EMD. Next, TKEO was applied on each IMF and two higher-order statistical moments namely skewness and kurtosis were extracted as features from TKEO of each IMF. The statistical significance of the selected features was evaluated using student's t-test and based on the statistical test, features from first three IMFs which show very high discriminative capability were selected as inputs to a support vector machine classifier for discrimination of Fo and NFo signals. It was observed that the classification accuracy of 92.65% is obtained in classifying EEG signals using a radial basis kernel function, which demonstrates the efficacy of proposed EMD-TKEO based feature extraction method for computer-based treatment of patients suffering from focal seizures.

HTLV-1-associated genes as potential biomarkers for endometrial cancer

Endometrial carcinoma (EC) is a malignant neoplasm of the endometrial epithelium, which may be diagnosed by pathological investigations. The aim of the current study was to identify new markers for the diagnosis of EC using machine learning. The association between human T cell lymphotropic virus type 1 (HTLV-1) infection and endometrial cancer risk have not been widely reported. It remains ambiguous whether HTLV-1 infection is associated with several types of cancer. The present study investigated the association between HTLV-1 infection-associated genes and EC risk. RNA sequencing uterine cancer expression data were downloaded from The Cancer Genome Atlas (TCGA) database. Differentially expressed genes (DEGs) were identified between normal and matched tumor samples. A total of 41 genes were selected by an overlap between HTLV-1 infection pathway-associated genes and the DEGs. Two-way hierarchical clustering analysis (HCA) and a support vector machine (SVM) classifier were constructed using the 41 genes. The accuracy of the candidate genes in risk-stratifying the samples was 100%. The accuracy of the proposed SVM model was 100%. In addition, the classification power of the SVM model was validated using a merged dataset (TCGA and the Genotype-Tissue Expression project). This predictive feature achieved reliable outcomes with risk-stratifying samples of almost 99% in two-way HCA, and an accuracy yield of 98% of the SVM classifier. In conclusion, the 41 genes identified in the current study may be implicated in the development of EC and may be of prognostic value for the disease. The results obtained the current study suggest that HTLV-1 may be potentially associated with EC and highlight potential disease mechanisms.

Classification of micro-CT images using 3D characterization of bone canal patterns in human osteogenesis imperfecta

Few studies have analyzed the microstructural properties of bone in cases of Osteogenenis Imperfecta (OI), or 'brittle bone disease'. Current approaches mainly focus on bone mineral density measurements as an indirect indicator of bone strength and quality. It has been shown that bone strength would depend not only on composition but also structural organization. This study aims to characterize 3D structure of the cortical bone in high-resolution micro CT images. A total of 40 bone fragments from 28 subjects (13 with OI and 15 healthy controls) were imaged using micro tomography using a synchrotron light source (SRμCT). Minkowski functionals - volume, surface, curvature, and Euler characteristics - describing the topological organization of the bone were computed from the images. The features were used in a machine learning task to classify between healthy and OI bone. The best classification performance (mean AUC - 0.96) was achieved with a combined 4-dimensional feature of all Minkowski functionals. Individually, the best feature performance was seen using curvature (mean AUC - 0.85), which characterizes the edges within a binary object. These results show that quantitative analysis of cortical bone microstructure, in a computer-aided diagnostics framework, can be used to distinguish between healthy and OI bone with high accuracy.

Analysis of gene expression profile identifies potential biomarkers for atherosclerosis

The present study aimed to identify potential biomarkers for atherosclerosis via analysis of gene expression profiles. The microarray dataset no. GSE20129 was downloaded from the Gene Expression Omnibus database. A total of 118 samples from the peripheral blood of female patients was used, including 47 atherosclerotic and 71 non‑atherosclerotic patients. The differentially expressed genes (DEGs) in the atherosclerosis samples were identified using the Limma package. Gene ontology term and Kyoto Encyclopedia of Genes and Genomes pathway analyses for DEGs were performed using the Database for Annotation, Visualization and Integrated Discovery tool. The recursive feature elimination (RFE) algorithm was applied for feature selection via iterative classification, and support vector machine classifier was used for the validation of prediction accuracy. A total of 430 DEGs in the atherosclerosis samples were identified, including 149 up‑ and 281 downregulated genes. Subsequently, the RFE algorithm was used to identify 11 biomarkers, whose receiver operating characteristic curves had an area under curve of 0.92, indicating that the identified 11 biomarkers were representative. The present study indicated that APH1B, JAM3, FBLN2, CSAD and PSTPIP2 may have important roles in the progression of atherosclerosis in females and may be potential biomarkers for early diagnosis and prognosis as well as treatment targets for this disease.

A new way of quantifying diagnostic information from multilead electrocardiogram for cardiac disease classification

A new measure for quantifying diagnostic information from a multilead electrocardiogram (MECG) is proposed. This diagnostic measure is based on principal component (PC) multivariate multiscale sample entropy (PMMSE). The PC analysis is used to reduce the dimension of the MECG data matrix. The multivariate multiscale sample entropy is evaluated over the PC matrix. The PMMSE values along each scale are used as a diagnostic feature vector. The performance of the proposed measure is evaluated using a least square support vector machine classifier for detection and classification of normal (healthy control) and different cardiovascular diseases such as cardiomyopathy, cardiac dysrhythmia, hypertrophy and myocardial infarction. The results show that the cardiac diseases are successfully detected and classified with an average accuracy of 90.34%. Comparison with some of the recently published methods shows improved performance of the proposed measure of cardiac disease classification.

In silico comparative genomic analysis of two non-small cell lung cancer subtypes and their potentials for cancer classification

BACKGROUND/AIM

Lung adenocarcinoma (AC) and squamous cell lung carcinoma (SCC) are two main subtypes of non-small cell lung cancer. In order to understand their biological differences, we conducted an in silico comparative genomic analysis of their expression profiles.

MATERIALS AND METHODS

We utilized the published microarray data of 18 SCC samples and 40 AC samples to discriminate genes differentially expressed in SCC and AC. Genes were employed to construct a functional module network and build a support vector machine classifier. Another set of published non-small cell lung cancer microarray data was used to test the predictive accuracy of support vector machine classifier.

RESULTS

Our analysis showed that SCC shows an elevated expression of genes related to cell division and DNA replication while AC presents an elevated expression of the genes related to protein transport and cell junction. ROC analysis demonstrates that the support vector machine classifier has a high classification accuracy for AC and SCC.

CONCLUSION

AC and SCC are distinctively different in certain biological network modules. This proposes different pathological mechanisms involved in these two non-small cell lung cancer subtypes.

Real-time Detection of Precursors to Epileptic Seizures: Non-Linear Analysis of System Dynamics

We propose a novel approach for detecting precursors to epileptic seizures in intracranial electroencephalograms (iEEG), which is based on the analysis of system dynamics. In the proposed scheme, the largest Lyapunov exponent of the discrete wavelet packet transform (DWPT) of the segmented EEG signals is considered as the discriminating features. Such features are processed by a support vector machine (SVM) classifier to identify whether the corresponding segment of the EEG signal contains a precursor to an epileptic seizure. When consecutive EEG segments contain such precursors, a decision is made that a precursor is in fact detected. The proposed scheme is applied to the Freiburg dataset, and the results show that seizure precursors are detected in a time frame that unlike other existing schemes is very much convenient to patients, with sensitivity of 100% and negligible false positive detection rates.

ZebraZoom: an automated program for high-throughput behavioral analysis and categorization

The zebrafish larva stands out as an emergent model organism for translational studies involving gene or drug screening thanks to its size, genetics, and permeability. At the larval stage, locomotion occurs in short episodes punctuated by periods of rest. Although phenotyping behavior is a key component of large-scale screens, it has not yet been automated in this model system. We developed ZebraZoom, a program to automatically track larvae and identify maneuvers for many animals performing discrete movements. Our program detects each episodic movement and extracts large-scale statistics on motor patterns to produce a quantification of the locomotor repertoire. We used ZebraZoom to identify motor defects induced by a glycinergic receptor antagonist. The analysis of the blind mutant atoh7 revealed small locomotor defects associated with the mutation. Using multiclass supervised machine learning, ZebraZoom categorized all episodes of movement for each larva into one of three possible maneuvers: slow forward swim, routine turn, and escape. ZebraZoom reached 91% accuracy for categorization of stereotypical maneuvers that four independent experimenters unanimously identified. For all maneuvers in the data set, ZebraZoom agreed with four experimenters in 73.2-82.5% of cases. We modeled the series of maneuvers performed by larvae as Markov chains and observed that larvae often repeated the same maneuvers within a group. When analyzing subsequent maneuvers performed by different larvae, we found that larva-larva interactions occurred as series of escapes. Overall, ZebraZoom reached the level of precision found in manual analysis but accomplished tasks in a high-throughput format necessary for large screens.