Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “canonical correlation analysis”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 91 records · Page 5Linked to original sources

A canonical correlation neural network for multicollinearity and functional data.

We review a recent neural implementation of Canonical Correlation Analysis and show, using ideas suggested by Ridge Regression, how to make the algorithm robust. The network is shown to operate on data sets which exhibit multicollinearity. We develop a second model which not only performs as well on multicollinear data but also on general data sets. This model allows us to vary a single parameter so that the network is capable of performing Partial Least Squares regression (at one extreme) to Canonical Correlation Analysis (at the other)and every intermediate operation between the two. On multicollinear data, the parameter setting is shown to be important but on more general data no particular parameter setting is required. Finally, we develop a second penalty term which acts on such data as a smoother in that the resulting weight vectors are much smoother and more interpretable than the weights without the robustification term. We illustrate our algorithms on both artificial and real data.

Child, Preschool↗

Continuity of individual differences in the mother-infant relationship from six to thirteen months.

The current study searched for continuities in mother-infant interaction observed at home, maternally perceived infant difficultness, and infant developmental competency. Also assessed were the possible moderator variables of maternal report of satisfaction with adjustment to the infant and background characteristics, for example, SES, presence of other young children, and infant gender. 128 dyads were assessed at ages 6 and 13 months. Many of the variables used to evaluate continuity were composites based on factor analysis. Bivariate correlations between 6- and 13-month variables produced a complex picture. In broad overview, as predicted by the prior literature, they indicated some degree of stability for mother variables and markedly less for the infant variables. Canonical correlation analysis yielded 3 significant canonical variates, producing a more integrated picture. The dimensions of continuity revealed by these variates were (1) Affectionate, Intellectually Stimulating Mothering, which had modest-to-moderate links to the indexes of infant competence; (2) Infant Temperamental Difficultness, which was determined mainly by consistency in mother report on 2 age-adjusted versions of the Infant Characteristics Questionnaire but which also contained loadings for objectively observed infant fussing, crying, and social demandingness; and (3) an unnamed dimension which resembles a pattern of intense involvement, both positive and negative, seen in older, clinically referred families.

Adult↗

The relationship between clinical signs and hypercoagulable state in toxemia of pregnancy.

The hypercoagulable state in patients with toxemia of pregnancy was investigated in comparison with normal pregnant women using new coagulation parameters, mainly thrombin-antithrombin III (TAT) complexes, alpha 2-antiplasmin-plasmin complexes (PIP), and D-dimer FDP. When the patients were categorized by the classification of the WHO Study Group (1985), significant increases of TAT complexes and alpha 2-PIP complexes with decreases of the ATIII level were observed in the groups with preeclampsia and severe gestational hypertensive disease as compared to normal pregnant women. A significant increase of D-dimer FDP was observed in a group with severe gestational hypertensive disease. Additionally, the relationship between clinical signs and the hypercoagulable state in the patients was analyzed using canonical correlation analysis as a multivariate analysis. The clinical signs and coagulation parameters had a significantly high correlation of lambda 1 = 0.7219, p less than 0.01. The results showed that clinical signs were associated with simultaneous coagulation abnormalities. The indices obtained from the results of canonical correlation analysis, which were called the clinical index and the coagulation index, should be useful in evaluating the efficacy of anticoagulation therapy.

Adult↗

[Factors influencing intelligence in elderly population in a community-based study in Taiyuan city].

OBJECTIVE: To explore the factors which influencing the intelligence in elderly in a community, so as to provide reference on primary prevention of dementia. METHODS: A door to door survey was conducted. A total number of 830 elderly were assessed using WAIS-RC, H-NTLA and data collected through questionnaires on lifestyles, family and social activity were analysed, using the factor analysis, stepwise regression and canonical correlation analysis. RESULTS: The main variables related to the verbal were eggs diet, blood pressure, fruit, visit their children, alcohol intake (standard coefficients of regression are -0.118, -0.079, -0.060, -0.036, -0.117, respectively). The main variables related to cognition were recreation, sports and hereditary history of mental disorders (standard coefficients of regression were 0.035, -0.127, respectively). The main variables related to memory were contact to chemical materials, age of their parents during delivery, alcohol intake and the relationship between husband and wife (standard coefficients of regression were -0.063, 0.055, -0.030, -0.037, respectively). The variables related to canonical variable V(1) would include education and occupation (canonical correlation = 0.5993, P = 0.0001) while V(2) would include cerebrovascular accident (canonical correlation = 0.3925, P = 0.0005). CONCLUSION: Intelligent work, family harmony, prevention of cerebrovascular diseases were the main areas to prevent intellectual disability in the elderly.

Aged↗

Frontal and facial flatness of major human populations.

In the present study, the frontal and facial features of 112 populations from around the world are compared in terms of frontal and facial flatness measurements. Univariate analyses and canonical correlation analysis were applied to six indices representing flatness of frontal and facial bones. The deep infraglabellar notch, marked prognathism, and flat frontal bone show distinctive Australian/Melanesian characters among recent populations. Very flat faces in the transverse plane are the most common condition in eastern Asians. Some subSaharan Africans share similar characteristics with Australians in terms of marked prognathism and flat frontal bones in the sagittal plane on the one hand, and with eastern Asians on the other hand, for flat nasal and zygomaxillary regions. These results are not necessarily inconsistent with the evidence for regional continuity. The examination of relationships between frontal and facial flatness through canonical correlation analysis reveals a significant association between morphological features such as a deep infraglabellar notch, prognathism, flat frontal bone, and flat faces in the transverse plane. In this context, together with the generalized features of the late Pleistocene fossil record, the features of Australians having transversely projecting faces and of eastern Asians showing weak infraglabellar notches, ortho-/mosognathism, and rounded frontal bones can be interpreted as a differential retention of ancestral traits of anatomically modern humans. This may allow us to suppose that the frontal and facial flatness features treated herein can be explained by the hypothesis of a single origin of anatomically modern humans.

Asian People↗

Psychosocial correlates of job strain in a sample of working women.

BACKGROUND: This study identifies potential mediators of job strain effects on health by determining whether psychosocial factors known to predict an increased risk of cardiovascular disease and all-cause mortality are higher among women who report high levels of job strain. METHODS: Measures of job strain and other psychosocial risk factors were obtained in a sample of 152 female employees of a local corporation. Canonical correlation and analyses of covariance were used to assess relationships between job demands and decision latitude and other psychosocial risk factors. RESULTS: A significant (P = .002) solution to the canonical correlation analysis showed that high job demands and low decision latitude were correlated with a pattern of psychosocial factors consisting of (1) increased levels of negative emotions like anxiety, anger, depression, and hostility; (2) reduced levels of social support; and (3) a preponderance of negative compared with positive feelings in dealings with coworkers and supervisors. This pattern was confirmed by analyses of covariance that adjusted for demographic and specific job characteristics. CONCLUSIONS: The canonical correlation analysis results provide empirical support for the job strain construct. The most important finding is that health-damaging psychosocial factors like job strain, depression, hostility, anxiety, and social isolation tend to cluster in certain individuals.

Adult↗

Motor abilities in dance structure performance in female students.

The aim of the study was to analyze the relation between motor abilities and performance in folk dances originating from the island of Hvar and modern social dances. Two groups of variables were used in a sample of 78 female students of the Teacher Training College from Split: 7 motor variables as a battery of predictors, and performance evaluation of 4 dances (2 folk dances, i.e. ciciliona and pasavijen, and 2 social dances, i.e. cha-cha-cha and rock-'n'-roll) as criterion variables. Canonical correlation analysis between the groups of variables yielded two canonical correlations of 0.94 and 0.73, with a level of significance of p<0.001. The first canonical correlation was based on marked determination of coordination and ciciliona dance, and the second one on explosive strength of the running type with below-average coordination and the cha-cha-cha, rock-'n'-roll and pasavijen dances. Regression analysis indicated the battery of motor variables used to be a good predictor of performance in all study dances, with multiple correlation of 0.93 in ciciliona, 0.84 in pasavijen, 0.75 in rock-'n'-roll and 0.73 in cha-cha-cha. In ciciliona and pasavijen, the latent dance structure is predominantly explained by coordination, in rock-'n'-roll by explosive strength, and in cha-cha-cha by explosive strength and speed. Discriminative analysis revealed the general dance performance to mostly depend on coordination, then on explosive strength, and to a lesser extent on speed (movement frequency). Dance is an irreplaceable educational tool in kinesiologic education of female students, among others for its considerable contribution to the development and maintenance of basic motor abilities.

Adult↗

Determinants of reported health promotion behaviors in employed Mexican American women.

This study based on the revised Pender Health Promotion Model (HPM) was conducted to determine the extent to which selected demographic modifying factors, health locus of control, self-efficacy, and current health status explained engaging in six health promotion practices in a sample of 397 employed Mexican American women aged 19 to 70 years who lived and worked in the Rio Grande Valley of Texas. Self-administered test packets contained measures of demographic characteristics, health locus of control, self-efficacy, health status, and the six health promotion practices. Canonical correlational analysis demonstrated two significant canonical variate pairs explaining 88% of variance in the dependent set, the subscale mean scores of the Health-Promoting Lifestyle Profile. Age; education; self-efficacy; internal and powerful others health locus of control; and prior, current, and future health status made statistically significant contributions. However, the demographic variables made only a minimal contribution, confirming the recent conclusions of researchers that the HPM as currently configured provides an inadequate explanation of the modifying factors' role in health promotion activity. Study results make an important addition to the understanding of how lifestyle factors contribute to Mexican American women's health and well-being.

Adult↗

Cocaine, Crime, Family History of Deviance-Are Psychosocial Correlates Related to These Phenomena in Male Cocaine Abusers?

This paper examines relationships among deviant behaviors such as cocaine abuse, crime, and family history of deviance; demographics; and psychological characteristics such as aggression, sensation seeking, confidence to resist taking cocaine and risk taking tendency to use cocaine in different situations; psychiatric symptoms (e.g., depression, anxiety, obsessive-compulsiveness, somatization, attention deficit); and social characteristics such as social adjustment. The sample consisted of men, mean age 35 years, who were veterans of the armed services. Canonical correlation analysis showed three significant canonical variates: The first indicated that verbal aggression, general sensation seeking, and a problem index for situations involving urges to use cocaine were positively related to number of times arrested and negatively related to age. The second revealed that experience seeking and problem indices for situations involving urges to use cocaine and for those involving pleasant times, were positively related to last cocaine use consumption level, total lifetime grams of cocaine used, and alcohol problems of the immediate family. The third showed that verbal aggression and a problem index for risk situations for use involving pleasant times were positively related, and experience seeking was negatively related, to number of times arrested, number of times incarcerated in the previous 30 days, age, and last cocaine use level. The implications of the associations found among these variables are discussed as they pertain to prevention, treatment, and future research.

Journal Article↗

[Study on the compatibility of composite herbal medicines of the lingguizhugan decoction].

OBJECTIVE: To study the compatibility of composite herbal medicines of the Ling Gui Zhu Gan Decoction. METHOD: Ethanol extract test solutions of the different combinations were prepared according to the orthogonal layout L16(4(5)). Pharmacologic experiments, such as the time of surviving of mice in shortage of oxygen in regular air pressure, the antagonizing effect on arrhythmia induced by chloroform and diuresis were carried out with the solutions. Variance analysis, canonical correlation and stepwise regression analysis were applied to interrelate the amount of each drug and the pharmacologic data. RESULT: The results confirmed that Fuling and Guizhi are the basis, while Baizhuand Gancao are the adjuvans, which is conformed to the theory of TCM. CONCLUSION: This study provides a significant try for studying the compatibility of composite herbal medicines.

Animals↗

[Study on the compatibility of composite herbal medicines of the Zishen pill].

OBJECTIVE: To study the compatibility of composite herbal medicines of the Zishen pill. METHOD: 50% ethanol extract test solution of the different combinations of composite herbal medicines of the Zishen pill were prepared. Pharmacologic experiments, such as anti-inflammatory, carbon particle clearance of RES were carried out with the solutions, and the corresponding pharmacological data were obtained. Variance analysis, canonical correlation and stepwise regression analysis were applied to interrelate the amount of each drug and the pharmacological data. RESULT: The results confirmed that Huangbo and Zhimu were the basis, while Rougui was the corrigent, which was conformed to the theory of TCM. CONCLUSION: This study provides a significant try for studying the compatibility of composite herbal medicines.

Anemarrhena↗

Selection of medical students at a medical faculty in South Africa.

The selection procedure at one medical faculty in the RSA was investigated. Students were admitted to the M.B. Ch.B. course on the basis of school results. Four groups were studied: matriculants; those selected after completing the 1st year of a B.Sc. course; those who had completed a full B.Sc. course; and those who had experience and other qualifications after leaving school. The matriculants were by far the largest group and the analysis of their marks received emphasis. Students who registered for their 1st year in 1980-1983 were followed up until the middle of their 3rd year. In total 635 students were assessed. The students' school results were compared with their preclinical marks. Chi-square tests for independence, multiple regression analysis, Pearson correlation coefficients and canonical correlation analysis were carried out. Although the results reflected a significant relationship between good selection marks and preclinical achievement, the strength of the association was too small for selection marks to have any reliable predictive value for an individual student. There was a low drop-out rate (5.4%) during the 1st year of study, but this rate increased threefold to 17.2% during the remaining preclinical period. This may indicate that the selection criteria were not specific enough to avoid this relatively large loss of students.

Educational Measurement↗

[Comprehension of synesthetic expressions: cross-modal modifications of sense adjectives].

This study examined (a) acceptability of phrases whose modifier adjective and head noun refer to different sense modalities, and (b) such cross-modal adjective properties by means of similarity judgment of these phrases. In Experiment 1, 58 subjects rated 60 sense adjectives as to their acceptability in modifying nine modality-denoting nouns (touch, taste, smell; color, sound; memory, mood, idea, personality). The phrases were rated highly acceptable when the adjectives denoting lower-modal qualities were combined with the nouns denoting higher-modal contents (e.g., soft sound, sweet mood, bright memory). In Experiment 2, two groups of 30 subjects judged the similarity of the highly acceptable phrases for each of the nine modality-denoting nouns. Multidimensional scaling of these data revealed two- and three-dimensional configurations of the adjectives for each modality-denoting noun. Multiple regression analysis indicated that two independently rated properties, pleasantness and intensity, provided a satisfactory interpretation of each configuration. Canonical correlation analysis indicated that most of the first- and second-order canonical correlations among the nine three-dimensional configurations were significant. These results suggest that there is a common dimension of sense adjectives across different modalities.

Adult↗

Severity of acute pancreatitis: a multivariate analysis of oxidative stress markers and modified Glasgow criteria.

BACKGROUND: It is unknown whether measurement of markers of oxidative stress can improve the prediction of severity of acute pancreatitis. METHODS: Consecutive patients admitted with a diagnosis of acute pancreatitis were divided into mild (n = 62) and severe (n = 23) groups based on the Atlanta classification. Plasma oxidative stress markers were measured within 24 h of admission and included ascorbic acid (endogenous antioxidant), protein carbonyl (a marker of protein oxidation), thiobarbituric acid reactive substances (a marker of lipid peroxidation) and myeloperoxidase (a neutrophil enzyme that produces oxidants). Canonical correlation analysis was used to describe the relationship between these markers and the modified Glasgow criteria. Canonical variate analysis was used to define the best variables that could discriminate mild and severe pancreatitis. RESULTS: There was a significant correlation between markers of oxidative stress and the modified Glasgow criteria (first canonical correlation 0.69, P < 0.0001, Wilk's lambda test). Blood urea, serum albumin and white cell count were the best variables that discriminated mild and severe acute pancreatitis, and all were better than the oxidative stress markers. CONCLUSION: The markers of oxidative stress were highly correlated with the severity of pancreatitis. They are unlikely to be better than the modified Glasgow criteria in predicting it.

Acute Disease↗

Spirituality of African-American women: correlations to health-promoting behaviors.

The purpose of this study was to investigate how spirituality relates to health-promoting behaviors in African-American women. Using Burkhart's theoretical framework for spirituality, a descriptive cross-sectional correlational design was used. A group of 260 (N=260) women completed Rosenbergh's Self-Esteem Scale, the Health Promoting Lifestyle Profile II, the Spiritual Perspective Scale, the Brief Block 2000 Food Frequency Questionnaire (FFQ). These women also provided the researchers with their socio-demographic data. Canonical correlation analysis identified a significant pair of canonical variables which indicated that those individuals with good nutrition (.95), physical activity (.79), and healthy eating (.42) were positively associated with stress management (.88), health responsibility (.67), spiritual growth (.66), interpersonal relations (.50), education (.49), and self-esteem (.33). This set of variables explained 56% of the variability (p < .001). Practitioners should incorporate the message of spirituality by focusing on strategies to improve health responsibility, interpersonal relations, and self-esteem, along with health-promoting behaviors.

Adolescent↗

Adaptive analysis of fMRI data.

This article introduces novel and fundamental improvements of fMRI data analysis. Central is a technique termed constrained canonical correlation analysis, which can be viewed as a natural extension and generalization of the popular general linear model method. The concept of spatial basis filters is presented and shown to be a very successful way of adaptively filtering the fMRI data. A general method for designing suitable hemodynamic response models is also proposed and incorporated into the constrained canonical correlation approach. Results that demonstrate how each of these parts significantly improves the detection of brain activity, with a computation time well within limits for practical use, are provided.

Algorithms↗

Survival prediction for clear cell renal cell carcinoma based on deep multimodal synergistic survival network.

Objective.To propose a deep multimodal synergistic survival analysis framework (Deep Multimodal Synergistic Survival Network, DMSSN) to achieve accurate prognostic analysis for clear cell renal cell carcinoma (ccRCC).Methods.This study (DMSSN) utilized matched multimodal data from the Cancer Genome Atlas-KIRC database, including CT imaging data, whole slide images, copy number variation (CNV) features, and clinical data. Deep Canonical Correlation Analysis was employed to map heterogeneous modalities into a shared latent space. Contrastive learning was introduced to enhance semantic consistency across multimodal features, and a gating network was utilized for the adaptive fusion of multimodal information to achieve precise survival risk prediction for patients.Results.Experimental results demonstrated that DMSSN achieved a Concordance Index (C-index) of 0.8153 &#xb1; 0.0994, with a Log-rank testp-value of 1.6553&#xd7;10-11. DMSSN exhibited significant performance advantages over traditional statistical methods like Log-rank-Cox (0.7055 &#xb1; 0.0670) and machine learning methods such as Random Survival Forest (RSF) (0.6836 &#xb1; 0.1048). Furthermore, in comparison with similar deep learning approaches, DMSSN outperformed late fusion strategies (0.7493 &#xb1; 0.1211) and discrete-time survival models such as DeepHit (0.7655 &#xb1; 0.1041) and Nnet-surv (0.7694 &#xb1; 0.0635). Notably, DMSSN still achieved the best predictive performance when compared to the classic deep survival model DeepSurv (0.7919 &#xb1; 0.0978) and advanced state-of-the-art multimodal fusion frameworks like Context-Aware Transformer (0.7735 &#xb1; 0.0818) and Multimodal Co-Attention Transformer (0.8102 &#xb1; 0.0972). Ablation studies showed that removing any single modality led to a decline in performance, with the largest numerical decrease occurring after removing CT imaging features (C-index decreased to 0.7327), validating the complementarity of multimodal data and the pivotal role of radiomic features in prognostic assessment. Module ablation experiments further confirmed the effectiveness of the core components.Conclusion:By effectively integrating imaging, pathology, genomic, and clinical features, the DMSSN framework demonstrates superior performance and robustness in the survival prediction of ccRCC.

Carcinoma, Renal Cell↗

Comparison of Short Form-36 and Migraine Disability Assessment questionnaire in patients with migraine.

BACKGROUND: The Short Form-36 (SF-36) and Migraine Disability Assessment (MIDAS) questionnaires are two of the most commonly used tools to measure outcomes in people suffering from headaches. Nevertheless, little is known about their interrelationship in patients with headache. OBJECTIVES: The aim of this study was to investigate the interrelationship between SF-36 and MIDAS questionnaires in patients with migraine. METHODS: We enrolled 231 patients with migraine (male/female: 43/188, mean age 35.3+/-8.1) who visited our headache clinic. They completed the SF-36, MIDAS, and a headache intake form. RESULTS: The correlation coefficients between the MIDAS score and 8 domains of the SF-36 ranged from -0.30 for the mental health domain to -0.53 for the social functioning domain (P<0.01). Canonical correlation analysis showed that the overall overlap between the 2 instruments was moderately strong (canonical correlation coefficients r=0.707 and 0.572). The overall measured redundancies for MIDAS and SF-36 scales in this study were 35.4% and 11.5%, respectively. The stepwise linear regression model showed that the social functioning domain alone explained 27.9% of variance in the MIDAS scores. Bodily pain, physical functioning, and general health domains added an additional 11.4% of the explained variance in the regression model. CONCLUSION: Despite the fact that these two measures were considerably correlated, the MIDAS and SF-36 were found to measure different aspects of the impact of headache for the sample investigated. The MIDAS questionnaire does not cover the emotional domain; therefore, an accompanying psychological questionnaire might help assess the outcome for headache studies.

Adult↗