Multivariate pattern analysis: a method and software to reveal, quantify, and visualize predictive association patterns in multicollinear data.

BACKGROUND: Strongly multicollinear covariates, such as those typically represented in metabolomics applications, represent a challenge for multivariate regression analysis. These challenges are commonly circumvented by reducing the number of covariates to a subset of linearly independent variables, but this strategy may lead to loss of resolution and thus produce models with poorer interpretative potential. The aim of this work was to implement and illustrate a method, multivariate pattern analysis (MVPA), which can handle multivariate covariates without compromising resolution or model quality. RESULTS: MVPA has been implemented in an open-source R package of the same name, mvpa. To facilitate the usage and interpretation of complex association patterns, mvpa has also been integrated into an R shiny app, mvpaShiny, which can be accessed on www.mvpashiny.org . MVPA utilizes a general projection algorithm that embraces a diversity of possible models. The method handles multicollinear and even linear dependent covariates. MVPA separates the variance in the data into orthogonal parts within the frame of a single joint model: one part describing the relations between covariates, outcome, and explanatory variables and another part describing the "net" predictive association pattern between outcome and explanatory variables. These patterns are visualized and interpreted in variance plots and plots for pattern analysis and ranking according to variable importance. Adjustment for a linear dependent covariate is performed in three steps. First, partial least squares regression with repeated Monte Carlo resampling is used to determine the number of predictive PLS components for a model relating the covariate to the outcome. Second, postprocessing of this PLS model by target projection provided a single component expressing the predictive association pattern between the outcome and the covariate. Third, the outcome and the explanatory variables were adjusted for the covariate by using the target score in the projection algorithm to obtain "net" data. We illustrate the main features of MVPA by investigating the partial mediation of a linearly dependent metabolomics descriptor on the association pattern between a measure of insulin resistance and lifestyle-related factors. CONCLUSIONS: Our method and implementation in R extend the range of possible analyses and visualizations that can be performed for complex multivariate data structures. The R packages are available on github.com/liningtonlab/mvpa and github.com/liningtonlab/mvpaShiny.

An approach to assess and adjust for the influence of multicollinear covariates on metabolomics association patterns-applied to a study of the associations between a comprehensive lipoprotein profile and the homeostatic model assessment of insulin resistance.

INTRODUCTION: Comprehensive lipoprotein profiling using proton nuclear magnetic resonance (NMR) spectroscopy of serum represents an alternative to the homeostatic model assessment of insulin resistance (HOMA-IR). Both adiposity and physical (in)activity associate to insulin resistance, but quantification of the influence of these two lifestyle related factors on the association pattern of HOMA-IR to lipoproteins suffers from lack of appropriate methods to handle multicollinear covariates. OBJECTIVES: We aimed at (i) developing an approach for assessment and adjustment of the influence of multicollinear and even linear dependent covariates on regression models, and (ii) to use this approach to examine the influence of adiposity and physical activity on the association pattern between HOMA-IR and the lipoprotein profile. METHODS: For 841 children, lipoprotein profiles were obtained from serum proton NMR and physical activity (PA) intensity profiles from accelerometry. Adiposity was measured as body mass index, the ratio of waist circumference to height, and skinfold thickness. Target projections were used to assess and isolate the influence of adiposity and PA on the association pattern of HOMA-IR to the lipoproteins. RESULTS: Adiposity explained just over 50% of the association pattern of HOMA-IR to the lipoproteins with strongest influence on high-density lipoprotein features. The influence of PA was mainly attributed to a strong inverse association between adiposity and moderate and high-intensity physical activity. CONCLUSION: The presented covariate projection approach to obtain net association patterns, made it possible to quantify and interpret the influence of adiposity and physical (in)activity on the association pattern of HOMA-IR to the lipoprotein features.

Structural validity of a test battery for assessment of fundamental movement skills in Norwegian 3-6-year-old children.

Fundamental movement skills (FMS) are building blocks of more advanced movements, including subdomains of locomotion, object control and balance skills, but limited evidence exists for this three-factor structure. The aim of this study was to examine the structural validity of a three-factor modified test battery of FMS across age and sex in two large samples of preschoolers aged 3-6 years (sample 1: n = 1213, mean age 4.8 (.09); sample 2: n = 1198, mean age 4.3 (.09)). We used a test battery of FMS consisting of movement tasks for locomotion (run, horizontal jump and hop) and object control (catch, overhand throw and kick) from the Test of Gross Motor Development and balance skills (single-leg standing, walking line backward and walking line forward) from the Preschoolers Gross Motor Quality Scale. Confirmatory factor analysis was used to analyse the data. We found support for both a three-factor and a two-factor structure. Measurement invariance testing showed invariance over age and partial scalar invariance over sex. We conclude that our modified test battery is an appropriate measure of young children's FMS across the domains of locomotion, object control and balance, but that locomotion and object control subdomains provide limited unique information.

The ActiGraph counts processing and the assessment of vigorous activity.

The purpose of this study was to investigate the effect of different band-pass filters on the measurement bias with ActiGraph counts during high speed running and for estimating free-living vigorous physical activity (VPA). Two alternative band-pass filters were designed, extending the original frequency range from 0·29 to 1·66 Hz (AG) to 0·29-4 Hz (AC4) and 0·29-10 Hz (AC10). Sixty-two subjects in three age groups participated in a structured locomotion protocol consisting of multiple walking and running speeds. The time spent in free-living VPA using the three different band-pass filters were evaluated in 1121 children. Band-pass filter specific intensity cut-points from both linear regression and ROC analysis was identified from a calibration experiment using indirect calorimetry. The ActiGraph GT3X+ device recording raw acceleration at 30 Hz was used in all experiments. The linear association between counts and running speed was negative for AG but positive for AC4 and AC10 across all age groups. The time spent in free-living VPA was similar for all band-pass filters. Considering higher frequency information in the generation of ActiGraph counts with a hip/waist worn device reduces the measurement bias with running above 10 km·h-1 . However, additional developments are required to accurately capture all VPA, including intermittent activities.

Psychometric properties of the Norwegian version of the Kidscreen-27 questionnaire.

BACKGROUND: The Norwegian version of the Kidscreen-27, a measure of generic health-related quality of life, has not yet been validated. Thus, the aim of this study was to investigate the reliability and validity of the Norwegian Kidscreen-27, in 10 year-old children. METHODS: The Kidscreen-27 consists of five domains and was validated in a cross-sectional study of 1085 school children (52.5 % boys). In addition a subsample of 56 children also had repeated measures in order to study test-retest reliability. RESULTS: Cronbach's alpha values ranged from 0.73 to 0.83, while intraclass correlation values over time ranged from 0.71 to 0.81. The domains of physical well-being, psychological well-being and autonomy & parents improved over time (Ps < 0.05), while social support and school environment domains did not. Confirmatory factor analysis showed an acceptable overall model fit: X (2) = 707; df = 310; P <0.001, root mean squared error of approximation = 0.037, the comparative fit index = 0.96 and the Tucker-Lewis index = 0.95. All factor loading were > 0.40. The Kidscreen-27 domains were significantly associated with general life satisfaction as measured with the Cantrils Ladder (Spearman rank correlations ranged from 0.29 to 0.59, Ps < 0.05). CONCLUSION: The Norwegian version of Kidscreen-27 has good reliability and validity.

Associations between physical activity and quality of life outcomes in adults with severe obesity: a cross-sectional study prior to the beginning of a lifestyle intervention.

BACKGROUND: Severely obese individuals who seek lifestyle interventions have impaired quality of life (QoL). Research suggests that physical activity (PA) plays a role in weight reduction and improved health in this group, but knowledge about the association of PA with QoL outcomes is sparse and inconsistent. The aim of this study was to investigate whether a higher level of PA was independently associated with higher QoL in severely obese individuals prior to the beginning of a lifestyle intervention. METHODS: During 2010, a total of 49 severely obese individuals who began a lifestyle intervention programme in Western Norway agreed to participate in the study. Data were collected prior to the beginning of the intervention. QoL was measured by a one-item scale on life satisfaction and the SF-36, PA was measured by an accelerometer, and clinical data were collected by health staff. Linear regression analyses were used to determine the associations between PA and QoL outcomes (life satisfaction, physical functioning, and mental health), adjusting for age, gender, and body mass index (BMI). RESULTS: In the adjusted analyses, we found positive relationships between PA and life satisfaction (Stand. coeff. 0.39, p = 0.024) and physical functioning (Stand. coeff. 0.34, p = 0.025). There was no association between PA and mental health (Stand. coeff. 0.15, p = 0.376). CONCLUSION: This study detected associations between objectively measured PA and life satisfaction as well as physical functioning in a group of severely obese individuals before they began a lifestyle intervention programme.