Reproducibility and Sensitivity of Resting-State fMRI in Patients With Parkinson's Disease Using Cross Validation-Based Data Censoring.

BACKGROUND: Uncontrollable body movements are typical symptoms of Parkinson's disease (PD), which results in inconsistent findings regarding resting-state functional connectivity (rsFC) networks, especially for group difference clusters. Systematically identifying the motion-associated data was highly demanded. PURPOSE: To determine data censoring criteria using a quantitative cross validation-based data censoring (CVDC) method and to improve the detection of rsFC deficits in PD. STUDY TYPE: Prospective. SUBJECTS: Forty-one PD patients (68.63 ± 9.17 years, 44% female) and 20 healthy controls (66.83 ± 12.94 years, 55% female). FIELD STRENGTH/SEQUENCE: 3-T, T1-weighted gradient echo and EPI sequences. ASSESSMENT: Clusters with significant differences between groups were found in three visual networks, default network, and right sensorimotor network. Five-fold cross-validation tests were performed using multiple motion exclusion criteria, and the selected criteria were determined based on cluster sizes, significance values, and Dice coefficients among the cross-validation tests. As a reference method, whole brain rsFC comparisons between groups were analyzed using a FMRIB Software Library (FSL) pipeline with default settings. STATISTICAL TESTS: Group difference clusters were calculated using nonparametric permutation statistics of FSL-randomize. The family-wise error was corrected. Demographic information was evaluated using independent sample t-tests and Pearson's Chi-squared tests. The level of statistical significance was set at P < 0.05. RESULTS: With the FSL processing pipeline, the mean Dice coefficient of the network clusters was 0.411, indicating a low reproducibility. With the proposed CVDC method, motion exclusion criteria were determined as frame-wise displacement >0.55 mm. Group-difference clusters showed a mean P-value of 0.01 and a 72% higher mean Dice coefficient compared to the FSL pipeline. Furthermore, the CVDC method was capable of detecting subtle rsFC deficits in the medial sensorimotor network and auditory network that were unobservable using the conventional pipeline. DATA CONCLUSION: The CVDC method may provide superior sensitivity and improved reproducibility for detecting rsFC deficits in PD. LEVEL OF EVIDENCE: 1 TECHNICAL EFFICACY: Stage 2.

Systematic evaluation of head motion on resting-state functional connectivity MRI in the neonate.

Reliability and robustness of resting state functional connectivity MRI (rs-fcMRI) relies, in part, on minimizing the influence of head motion on measured brain signals. The confounding effects of head motion on functional connectivity have been extensively studied in adults, but its impact on newborn brain connectivity remains unexplored. Here, using a large newborn data set consisting of 159 rs-fcMRI scans acquired in the Developing Brain Institute at Children's National Hospital and 416 scans from The Developing Human Connectome Project (dHCP), we systematically investigated associations between head motion and rs-fcMRI. Head motion during the scan significantly affected connectivity at sensory-related networks and default mode networks, and at the whole brain scale; the direction of motion effects varied across the whole brain. Comparing high- versus low-head motion groups suggested that head motion can impact connectivity estimates across the whole brain. Censoring of high-motion volumes using frame-wise displacement significantly reduced the confounding effects of head motion on neonatal rs-fcMRI. Lastly, in the dHCP data set, we demonstrated similar persistent associations between head motion and network connectivity despite implementing a standard denoising strategy. Collectively, our results highlight the importance of using rigorous head motion correction in preprocessing neonatal rs-fcMRI to yield reliable estimates of brain activity.

CondiS: A conditional survival distribution-based method for censored data imputation overcoming the hurdle in machine learning-based survival analysis.

Data analyses by machine learning (ML) algorithms are gaining popularity in biomedical research. When time-to-event data are of interest, censoring is common and needs to be properly addressed. Most ML methods cannot conveniently and appropriately take the censoring information into consideration, potentially leading to inaccurate or biased results. We aim to develop a general-purpose method for imputing censored survival data, facilitating downstream ML analysis. In this study, we propose a novel method of imputing the survival times for censored observations. The proposal is based on their conditional survival distributions (CondiS) derived from Kaplan-Meier estimators. CondiS can replace censored observations with their best approximations from the statistical model, allowing for direct application of ML methods. When covariates are available, we extend CondiS by incorporating the covariate information through ML modeling (CondiS-X), which further improves the accuracy of the imputed survival time. Compared with existing methods with similar purposes, the proposed methods achieved smaller prediction errors and higher concordance with the underlying true survival times in extensive simulation studies. We also demonstrated the usage and advantages of the proposed methods through two real-world cancer datasets. The major advantage of CondiS is that it allows for the direct application of standard ML techniques for analysis once the censored survival times are imputed. We present a user-friendly R package to implement our method, which is a useful tool for ML-based biomedical research in this era of big data.

Terminology of cardiac troponin assays and data censoring.

We discuss the sensitivity terminology of cardiac troponin assays and its dependence on the selection of the reference population. In addition, the need for reasonable censoring of clinical laboratory test results is contrasted with potential loss of valuable clinical information.

Genetic analysis of sow longevity and sow lifetime reproductive traits using censored data.

Sow longevity is a key component for efficient and profitable pig farming; however, approximately 50% of sows are removed annually from a breeding herd. There is no consensus in the scientific literature regarding a definition for sow longevity; however, it has been suggested that it can be measured using several methods such as stayability and economic indicators such as lifetime piglets produced. Sow longevity can be improved by genetic selection; however, it is rarely included in genetic evaluations. One reason is elongated time intervals required to collect complete lifetime data. The effect of genetic parameter estimation software in handling incomplete data (censoring) and possible early indicator traits were evaluated analysing a 30% censored data set (12 725 pedigreed Landrace × Large White sows that included approximately 30% censored data) with DMU6, THRGIBBS1F90 and GIBBS2CEN. Heritability estimates were low for all the traits evaluated. The results show that the binary stayability traits benefited from being analysed with a threshold model compared to analysing with a linear model. Sires were ranked very similarly regardless if the program handled censoring when all available data were included. Accumulated born alive and stayability were good indicators for lifetime born alive traits. Number of piglets born alive within each parity could be used as an early indicator trait for sow longevity.

Negative values and variance functions: Implications for statistical analysis.

When reporting concentrations of substances in biological specimens it has been virtually universal practice to suppress negative results, initially by left-censoring negative results to zero and more recently by left-censoring to values such as limit of blank, limit of detection or even limit of quantification. Negative concentrations are obviously nonsensical and current reporting practices place proper emphasis on assisting the clinician. However, it is easily overlooked that negative concentrations are merely artefacts of data reduction and while adjusting them is sensible clinical practice there are potentially adverse consequences for statistical analysis, in particular for those parametric summaries and analyses which rely on reliable estimates of low-end uncertainty. This article puts a case for the availability of negative results, describes complications with respect to estimating variance functions and discusses practical workarounds.

A multidimensional artefact-reduction approach to increase robustness of first-level fMRI analyses: Censoring vs. interpolating.

BACKGROUND: This manuscript describes a new, multidimensional and data-driven approach to identify outlying datapoints from a first-level fMRI dataset. NEW METHOD: Using three different indicators of data corruption (the fast variance component of DVARS [Δ%D-var], scan-to-scan total displacement [STS], and each scan's overall explained variance [R2]), it identifies outlying datapoints while being balanced using Akaike'c corrected criterion (AIC C) to avoid overcorrection. We then explore the impact of censoring, interpolating, or both, to remove a bad scan's contribution to the final timeseries. RESULTS AND COMPARISON WITH EXISTING METHODS: Our results (using three real-life datasets and extensive simulations) show that motion-corrupted datapoints as well as non-motion related image artefacts are detected reliably. Using several indicators is shown to be an advantage over existing single-indicator solutions in different settings. As a result of using our algorithm, stronger activation (as detected by both T-value and number of activated voxels) and an increase in the temporal signal-to-noise ratio can be seen. The effects of censoring and interpolation are distinct and complex. CONCLUSIONS: The multidimensional approach described here is able to identify outlying datapoints in fMRI timeseries, with demonstrable positive effects on several outcome measures. While censoring datapoints may be preferable in many settings, the ultimate choice on which approach to choose may depend on the data at hand. Recommendations are provided for different scenarios.

The FDA review on data quality and conduct in vorapaxar trials: Much better than in PLATO, but still not perfect.

BACKGROUND/AIMS: Vorapaxar, a novel antiplatelet thrombin PAR-1 inhibitor, has been evaluated in TRA2P and TRACER trials. The drug is currently approved for post-myocardial infarction and peripheral artery disease indications with concomitant use of clopidogrel and/or aspirin. The FDA ruled that the overall vorapaxar data quality was acceptable, but conducted the sensitivity analyses for potential censoring. This was unusual, intriguing, and directly related to the challenged quality of ticagrelor dataset in PLATO in the previous New Drug Application for an oral antiplatelet agent submitted to the same Agency. METHODS: Hence, we compared the FDA-confirmed evidence of conduct and data quality in vorapaxar (TRA2P, and TRACER) with those of ticagrelor (PLATO) trials. RESULTS: The FDA provides a detailed report on information censoring, and follow-up completeness for 3 trials. TRA2P and TRACER used independent CRO for site monitoring, exhibit no heterogeneity in trial results dependent on geography, and consistent adjudication results with much less censoring than in PLATO. CONCLUSION: The data quality and trial conduct in vorapaxar trials were better than testing ticagrelor in PLATO, however, there is still some room for improvement especially with regard to follow-up completeness, and less information censoring.

Natural transformation of chlordecone into 5b-hydrochlordecone in French West Indies soils: statistical evidence for investigating long-term persistence of organic pollutants.

Chlordecone (CLD) was an organochlorine insecticide whose previous use resulted in an extensive pollution of the environment with severe health effects and social consequences. A closely related compound, 5b-hydrochlordecone (5b-hydroCLD), has been searched for and often detected in environmental matrices from the geographical area where CLD was applied. The current consensus considered that its presence was not the result of a biotic or abiotic dechlorination of CLD in these matrices but rather the consequence of its presence as impurity (synthesis by-product) in the CLD released into the environment. The aim of the present study was to determine if and to what extent degradation of CLD into 5b-hydroCLD occurred in the field. To test this hypothesis, the ratios of 5b-hydroCLD and CLD concentrations in a dataset of 810 soils collected between 2006 and 2012 in Martinique were compared to the ratios measured in 3 samples of the CLD dust commercial formulations applied in the banana fields of French West Indies (FWI) and 1 sample of the technical-grade CLD corresponding to the active ingredient used in such formulations. Soil data were processed with a hierarchical Bayesian model to account for random measurement errors and data censoring. Any pathway of CLD transformation into 5b-hydroCLD occurring over the long term in FWI soils would indeed change the ratio of 5b-hydroCLD/CLD compared to what it was in the initially applied formulations. Results showed a significant increase of the 5b-hydroCLD/CLD ratio in the soils-25 times greater in soil than in commercial formulations-which suggested that natural CLD transformation into 5b-hydroCLD over the long term occurred in these soils. Results from this study may impact future decisions for the remediation of the polluted areas.

20 years of long-term atrazine monitoring in a shallow aquifer in western Germany.

Atrazine was banned in Germany in 1991 due to findings of atrazine concentrations in ground- and drinking waters exceeding threshold values. Monitoring of atrazine concentrations in the groundwater since then provides information about the resilience of the groundwater quality to changing agricultural practices. In this study, we present results of a monitoring campaign of atrazine concentrations in the Zwischenscholle aquifer. This phreatic aquifer is exposed to intensive agricultural land use and susceptible to contaminants due to a shallow water table. In total 60 observation wells (OWs) have been monitored since 1991, of which 15 are sampled monthly today. Descriptive statistics of monitoring data were derived using the "regression on order statistics" (ROS) data censoring approach, estimating values for nondetects. The monitoring data shows that even 20 years after the ban of atrazine, the groundwater concentrations of sampled OWs remain on a level close to the threshold value of 0.1 µg l(-1) without any considerable decrease. The spatial distribution of atrazine concentrations is highly heterogeneous with OWs exhibiting permanently concentrations above the regulatory threshold on the one hand and OWs were concentrations are mostly below the limit of quantification (LOQ) on the other hand. A deethylatrazine-to-atrazine ratio (DAR) was used to distinguish between diffuse - and point-source contamination, with a global mean value of 0.84 indicating mainly diffuse contamination. Principle Component Analysis (PCA) of the monitoring dataset demonstrated relationships between the metabolite desisopropylatrazine, which was found to be exclusively associated with the parent compound simazine but not with atrazine, and between deethylatrazine, atrazine, nitrate, and the specific electrical conductivity. These parameters indicate agricultural impacts on groundwater quality. The findings presented in this study point at the difficulty to estimate mean concentrations of contamination for entire aquifers and to evaluate groundwater quality based on average parameters. However, analytical data of monthly sampled single observation wells provide adequate information to characterize local contamination and evolutionary trends of pollutant concentration.