Sample size and statistical power in the hierarchical analysis of variance: applications in morphometry of the nervous system.

Analysis of variance is commonly used in morphometry in order to ascertain differences in parameters between several populations. Failure to detect significant differences between populations (type II error) may be due to suboptimal sampling and lead to erroneous conclusions; the concept of statistical power allows one to avoid such failures by means of an adequate sampling. Several examples are given in the morphometry of the nervous system, showing the use of the power of a hierarchical analysis of variance test for the choice of appropriate sample and subsample sizes. In the first case chosen, neuronal densities in the human visual cortex, we find the number of observations to be of little effect. For dendritic spine densities in the visual cortex of mice and humans, the effect is somewhat larger. A substantial effect is shown in our last example, dendritic segmental lengths in monkey lateral geniculate nucleus. It is in the nature of the hierarchical model that sample size is always more important than subsample size. The relative weight to be attributed to subsample size thus depends on the relative magnitude of the between observations variance compared to the between individuals variance.

Paper 4: EUROCAT statistical monitoring: identification and investigation of ten year trends of congenital anomalies in Europe.

BACKGROUND: As part of EUROCAT's surveillance of congenital anomalies in Europe, a statistical monitoring system has been developed to detect recent clusters or long-term (10 year) time trends. The purpose of this article is to describe the system for the identification and investigation of 10-year time trends, conceived as a "screening" tool ultimately leading to the identification of trends which may be due to changing teratogenic factors.METHODS: The EUROCAT database consists of all cases of congenital anomalies including livebirths, fetal deaths from 20 weeks gestational age, and terminations of pregnancy for fetal anomaly. Monitoring of 10-year trends is performed for each registry for each of 96 non-independent EUROCAT congenital anomaly subgroups, while Pan-Europe analysis combines data from all registries. The monitoring results are reviewed, prioritized according to a prioritization strategy, and communicated to registries for investigation. Twenty-one registries covering over 4 million births, from 1999 to 2008, were included in monitoring in 2010.CONCLUSIONS: Significant increasing trends were detected for abdominal wall anomalies, gastroschisis, hypospadias, Trisomy 18 and renal dysplasia in the Pan-Europe analysis while 68 increasing trends were identified in individual registries. A decreasing trend was detected in over one-third of anomaly subgroups in the Pan-Europe analysis, and 16.9% of individual registry tests. Registry preliminary investigations indicated that many trends are due to changes in data quality, ascertainment, screening, or diagnostic methods. Some trends are inevitably chance phenomena related to multiple testing, while others seem to represent real and continuing change needing further investigation and response by regional/national public health authorities.

ROBETH: un logiciel pour les procédés statistiques robustes [ROBETH: a library for robust statistical procedures].

In the past 20 years the theory of robust estimation has become an important topic of mathematical statistics. We discuss here some basic concepts of this theory with the help of simple examples. Furthermore we describe a subroutine library for the application of robust statistical procedures, which was developed with the support of the Swiss National Science Foundation.

A statistical methodology for the comparison of blue gel pen inks analyzed by laser desorption/ionization mass spectrometry

A statistical methodology for the objective comparison of LDI-MS mass spectra of blue gel pen inks was evaluated. Thirty-three blue gel pen inks previously studied by RAMAN were analyzed directly on the paper using both positive and negative mode. The obtained mass spectra were first compared using relative areas of selected peaks using the Pearson correlation coefficient and the Euclidean distance. Intra-variability among results from one ink and inter-variability between results from different inks were compared in order to choose a differentiation threshold minimizing the rate of false negative (i.e. avoiding false differentiation of the inks). This yielded a discriminating power of up to 77% for analysis made in the negative mode. The whole mass spectra were then compared using the same methodology, allowing for a better DP in the negative mode of 92% using the Pearson correlation on standardized data. The positive mode results generally yielded a lower differential power (DP) than the negative mode due to a higher intra-variability compared to the inter-variability in the mass spectra of the ink samples.

Statistical discrimination and employers' recruitment : Practices for low-skilled workers

This paper deals with the recruitment strategies of employers in the low-skilled segment of the labour market. We focus on low-skilled workers because they are overrepresented among jobless people and constitute the bulk of the clientele included in various activation and labour market programmes. A better understanding of the constraints and opportunities of interventions in this labour market segment may help improve their quality and effectiveness. On the basis of qualitative interviews with 41 employers in six European countries, we find that the traditional signals known to be used as statistical discrimination devices (old age, immigrant status and unemployment) play a somewhat reduced role, since these profiles are overrepresented among applicants for low skill positions. However, we find that other signals, mostly considered to be indicators of motivation, have a bigger impact in the selection process. These tend to concern the channel through which the contact with a prospective candidate is made. Unsolicited applications and recommendations from already employed workers emit a positive signal, whereas the fact of being referred by the public employment office is associated with the likelihood of lower motivation.

Statistical Modeling of the Eye for Multimodal Treatment Planning for External Beam Radiation Therapy of Intraocular Tumors.

PURPOSE: Ocular anatomy and radiation-associated toxicities provide unique challenges for external beam radiation therapy. For treatment planning, precise modeling of organs at risk and tumor volume are crucial. Development of a precise eye model and automatic adaptation of this model to patients' anatomy remain problematic because of organ shape variability. This work introduces the application of a 3-dimensional (3D) statistical shape model as a novel method for precise eye modeling for external beam radiation therapy of intraocular tumors. METHODS AND MATERIALS: Manual and automatic segmentations were compared for 17 patients, based on head computed tomography (CT) volume scans. A 3D statistical shape model of the cornea, lens, and sclera as well as of the optic disc position was developed. Furthermore, an active shape model was built to enable automatic fitting of the eye model to CT slice stacks. Cross-validation was performed based on leave-one-out tests for all training shapes by measuring dice coefficients and mean segmentation errors between automatic segmentation and manual segmentation by an expert. RESULTS: Cross-validation revealed a dice similarity of 95% ± 2% for the sclera and cornea and 91% ± 2% for the lens. Overall, mean segmentation error was found to be 0.3 ± 0.1 mm. Average segmentation time was 14 ± 2 s on a standard personal computer. CONCLUSIONS: Our results show that the solution presented outperforms state-of-the-art methods in terms of accuracy, reliability, and robustness. Moreover, the eye model shape as well as its variability is learned from a training set rather than by making shape assumptions (eg, as with the spherical or elliptical model). Therefore, the model appears to be capable of modeling nonspherically and nonelliptically shaped eyes.

MDCT arthrography of the hip: value of the adaptive statistical iterative reconstruction technique and potential for radiation dose reduction.

OBJECTIVE: The purpose of this article is to assess the effect of the adaptive statistical iterative reconstruction (ASIR) technique on image quality in hip MDCT arthrography and to evaluate its potential for reducing radiation dose. SUBJECTS AND METHODS: Thirty-seven patients examined with hip MDCT arthrography were prospectively randomized into three different protocols: one with a regular dose (volume CT dose index [CTDIvol], 38.4 mGy) and two with a reduced dose (CTDIvol, 24.6 or 15.4 mGy). Images were reconstructed using filtered back projection (FBP) and four increasing percentages of ASIR (30%, 50%, 70%, and 90%). Image noise and contrast-to-noise ratio (CNR) were measured. Two musculoskeletal radiologists independently evaluated several anatomic structures and image quality parameters using a 4-point scale. They also jointly assessed acetabular labrum tears and articular cartilage lesions. RESULTS: With decreasing radiation dose level, image noise statistically significantly increased (p=0.0009) and CNR statistically significantly decreased (p=0.001). We also found a statistically significant reduction in noise (p=0.0001) and increase in CNR (p≤0.003) with increasing percentage of ASIR; in addition, we noted statistically significant increases in image quality scores for the labrum and cartilage, subchondral bone, overall diagnostic quality (up to 50% ASIR), and subjective noise (p≤0.04), and statistically significant reductions for the trabecular bone and muscles (p≤0.03). Regardless of the radiation dose level, there were no statistically significant differences in the detection and characterization of labral tears (n=24; p=1) and cartilage lesions (n=40; p≥0.89) depending on the ASIR percentage. CONCLUSION: The use of up to 50% ASIR in hip MDCT arthrography helps to reduce radiation dose by approximately 35-60%, while maintaining diagnostic image quality comparable to that of a regular-dose protocol using FBP.

Adaptive statistical iterative reconstruction reduces patient radiation dose in neuroradiology CT studies.

INTRODUCTION: Adaptive statistical iterative reconstruction (ASIR) can decrease image noise, thereby generating CT images of comparable diagnostic quality with less radiation. The purpose of this study is to quantify the effect of systematic use of ASIR versus filtered back projection (FBP) for neuroradiology CT protocols on patients' radiation dose and image quality. METHODS: We evaluated the effect of ASIR on six types of neuroradiologic CT studies: adult and pediatric unenhanced head CT, adult cervical spine CT, adult cervical and intracranial CT angiography, adult soft tissue neck CT with contrast, and adult lumbar spine CT. For each type of CT study, two groups of 100 consecutive studies were retrospectively reviewed: 100 studies performed with FBP and 100 studies performed with ASIR/FBP blending factor of 40 %/60 % with appropriate noise indices. The weighted volume CT dose index (CTDIvol), dose-length product (DLP) and noise were recorded. Each study was also reviewed for image quality by two reviewers. Continuous and categorical variables were compared by t test and free permutation test, respectively. RESULTS: For adult unenhanced brain CT, CT cervical myelography, cervical and intracranial CT angiography and lumbar spine CT both CTDIvol and DLP were lowered by up to 10.9 % (p < 0.001), 17.9 % (p = 0.005), 20.9 % (p < 0.001), and 21.7 % (p = 0.001), respectively, by using ASIR compared with FBP alone. Image quality and noise were similar for both FBP and ASIR. CONCLUSION: We recommend routine use of iterative reconstruction for neuroradiology CT examinations because this approach affords a significant dose reduction while preserving image quality.

Operational benefits and challenges of the use of fingerprint statistical models: a field study

Research projects aimed at proposing fingerprint statistical models based on the likelihood ratio framework have shown that low quality finger impressions left on crime scenes may have significant evidential value. These impressions are currently either not recovered, considered to be of no value when first analyzed by fingerprint examiners, or lead to inconclusive results when compared to control prints. There are growing concerns within the fingerprint community that recovering and examining these low quality impressions will result in a significant increase of the workload of fingerprint units and ultimately of the number of backlogged cases. This study was designed to measure the number of impressions currently not recovered or not considered for examination, and to assess the usefulness of these impressions in terms of the number of additional detections that would result from their examination.

Implementing statistical learning methods through Bayesian networks. Part 1: a guide to Bayesian parameter estimation using forensic science data

As a thorough aggregation of probability and graph theory, Bayesian networks currently enjoy widespread interest as a means for studying factors that affect the coherent evaluation of scientific evidence in forensic science. Paper I of this series of papers intends to contribute to the discussion of Bayesian networks as a framework that is helpful for both illustrating and implementing statistical procedures that are commonly employed for the study of uncertainties (e.g. the estimation of unknown quantities). While the respective statistical procedures are widely described in literature, the primary aim of this paper is to offer an essentially non-technical introduction on how interested readers may use these analytical approaches - with the help of Bayesian networks - for processing their own forensic science data. Attention is mainly drawn to the structure and underlying rationale of a series of basic and context-independent network fragments that users may incorporate as building blocs while constructing larger inference models. As an example of how this may be done, the proposed concepts will be used in a second paper (Part II) for specifying graphical probability networks whose purpose is to assist forensic scientists in the evaluation of scientific evidence encountered in the context of forensic document examination (i.e. results of the analysis of black toners present on printed or copied documents).