Effects of detectability on estimates of geographic range size in bignonieae

Extinction risk has not been evaluated for 96% of all described plant species. Given that the Global Strategy for Plant Conservation proposes preliminary conservation assessments of all described plant species by 2010, herbarium specimens (i.e., primary occurrence data) are increasingly being used to infer threat components from estimates of geographic range size. Nevertheless, estimates of range size based on herbarium data may be inaccurate due to collection bias associated with interspecific variation in detectability. We used data on 377 species of Bignonieae to test the hypothesis that there is a positive relationship between detectability and estimates of geographic range size derived from herbarium specimens. This relationship is expected if the proportion of the true geographic range size of a species that is documented by herbarium specimens is given by the product of the true geographic range size and the detectability of the species, assuming no relationship between true geographic range size and detectability. We developed 4 measures of detectability that can be estimated from herbarium data and examined the relationship between detectability and 2 types of estimates of geographic range size: area of occupancy and extent of occurrence. Our results from regressing estimates of extent of occurrence and area of occupancy on detectability across genera provided no support for this hypothesis. The same was true for regressions of estimated extent of occurrence on detectability across species within genera. Nevertheless, regressions of estimated area of occupancy on detectability across species within genera provided partial support for our hypothesis. We considered 3 possible explanations for this mixed outcome: violation of the assumption of no relationship between true geographic range size and detectability; the relationships between estimated geographic range size and detectability may be an artifact of a negative relationship between estimated area of occupancy and the sampling variance of detectability; detectability may have had 2 opposite effects on estimated species range sizes: one determines the proportion of the true range of a species documented by herbarium specimens and the other determines the distribution of true range size for the species actually observed with herbarium data. Our findings should help improve understanding of the potential biases incurred with the use of herbarium data.

Building binary-tree-based multiclass classifiers using separability measures

Various popular machine learning techniques, like support vector machines, are originally conceived for the solution of two-class (binary) classification problems. However, a large number of real problems present more than two classes. A common approach to generalize binary learning techniques to solve problems with more than two classes, also known as multiclass classification problems, consists of hierarchically decomposing the multiclass problem into multiple binary sub-problems, whose outputs are combined to define the predicted class. This strategy results in a tree of binary classifiers, where each internal node corresponds to a binary classifier distinguishing two groups of classes and the leaf nodes correspond to the problem classes. This paper investigates how measures of the separability between classes can be employed in the construction of binary-tree-based multiclass classifiers, adapting the decompositions performed to each particular multiclass problem. (C) 2010 Elsevier B.V. All rights reserved.

On the influence of imputation in classification: practical issues

The substitution of missing values, also called imputation, is an important data preparation task for many domains. Ideally, the substitution of missing values should not insert biases into the dataset. This aspect has been usually assessed by some measures of the prediction capability of imputation methods. Such measures assume the simulation of missing entries for some attributes whose values are actually known. These artificially missing values are imputed and then compared with the original values. Although this evaluation is useful, it does not allow the influence of imputed values in the ultimate modelling task (e.g. in classification) to be inferred. We argue that imputation cannot be properly evaluated apart from the modelling task. Thus, alternative approaches are needed. This article elaborates on the influence of imputed values in classification. In particular, a practical procedure for estimating the inserted bias is described. As an additional contribution, we have used such a procedure to empirically illustrate the performance of three imputation methods (majority, naive Bayes and Bayesian networks) in three datasets. Three classifiers (decision tree, naive Bayes and nearest neighbours) have been used as modelling tools in our experiments. The achieved results illustrate a variety of situations that can take place in the data preparation practice.

Influence of education and depressive symptoms on cognitive function in the elderly

Objective: The purpose of the present study was to investigate the influence that education and depression have on the performance of elderly people in neuropsychological tests. Methods: The study was conducted at the Institute of Psychiatry, University of Sao Paulo School of Medicine, Hospital das Clinicas. All of the individuals evaluated were aged 60 or older. The study sample consisted of 59 outpatients with depressive disorders and 51 healthy controls. We stratified the sample by level of education: low = 1-4 years of schooling; high = 5 or more years of schooling. Evaluations consisted of psychiatric assessment, cognitive assessment, laboratory tests and cerebral magnetic resonance imaging. Results: We found that level of education influenced all the measures of cognitive domains investigated (intellectual efficiency, processing speed, attention, executive function and memory) except the Digit Span Forward and Fuld Object Memory Evaluation (immediate and delayed recall), whereas depressive symptoms influenced some measures of memory, attention, executive function and processing speed. Although the combination of a low level of education and depression had a significant negative influence on Stroop Test part B, Trail Making Test part B and Logical Memory (immediate recall), we found no other significant effects of the interaction between level of education and depression. Conclusion: The results of this study underscore the importance of considering the level of education in the analysis of cognitive performance in depressed elderly patients, as well as the relevance of developing new cognitive function tests in which level of education has a reduced impact on the results.

Reduction of Models in the Presence of Nuisance Parameters

In many statistical inference problems, there is interest in estimation of only some elements of the parameter vector that defines the adopted model. In general, such elements are associated to measures of location and the additional terms, known as nuisance parameters, to control the dispersion and asymmetry of the underlying distributions. To estimate all the parameters of the model and to draw inferences only on the parameters of interest. Depending on the adopted model, this procedure can be both algebraically is common and computationally very costly and thus it is convenient to reduce it, so that it depends only on the parameters of interest. This article reviews estimation methods in the presence of nuisance parameters and consider some applications in models recently discussed in the literature.

Analysis of response to enteral infusion of levodopa in patients with Parkinson's disease

Objective: We present a new evaluation of levodopa plasma concentrations and clinical effects during duodenal infusion of a levodopa/carbidopa gel (Duodopa ) in 12 patients with advanced Parkinson s disease (PD), from a study reported previously (Nyholm et al, Clin Neuropharmacol 2003; 26(3): 156-163). One objective was to investigate in what state of PD we can see the greatest benefits with infusion compared with corresponding oral treatment (Sinemet CR). Another objective was to identify fluctuating response to levodopa and correlate to variables related to disease progression. Methods: We have computed mean absolute error (MAE) and mean squared error (MSE) for the clinical rating from -3 (severe parkinsonism) to +3 (severe dyskinesia) as measures of the clinical state over the treatment periods of the study. Standard deviation (SD) of the rating was used as a measure of response fluctuations. Linear regression and visual inspection of graphs were used to estimate relationships between these measures and variables related to disease progression such as years on levodopa (YLD) or unified PD rating scale part II (UPDRS II).Results: We found that MAE for infusion had a strong linear correlation to YLD (r2=0.80) while the corresponding relation for oral treatment looked more sigmoid, particularly for the more advanced patients (YLD>18).

Computerized identification of motor complications in Parkinson's disease

Objective: To investigate whether spirography-based objective measures are able to effectively characterize the severity of unwanted symptom states (Off and dyskinesia) and discriminate them from motor state of healthy elderly subjects. Background: Sixty-five patients with advanced Parkinson’s disease (PD) and 10 healthy elderly (HE) subjects performed repeated assessments of spirography, using a touch screen telemetry device in their home environments. On inclusion, the patients were either treated with levodopa-carbidopa intestinal gel or were candidates for switching to this treatment. On each test occasion, the subjects were asked trace a pre-drawn Archimedes spiral shown on the screen, using an ergonomic pen stylus. The test was repeated three times and was performed using dominant hand. A clinician used a web interface which animated the spiral drawings, allowing him to observe different kinematic features, like accelerations and spatial changes, during the drawing process and to rate different motor impairments. Initially, the motor impairments of drawing speed, irregularity and hesitation were rated on a 0 (normal) to 4 (extremely severe) scales followed by marking the momentary motor state of the patient into 2 categories that is Off and Dyskinesia. A sample of spirals drawn by HE subjects was randomly selected and used in subsequent analysis. Methods: The raw spiral data, consisting of stylus position and timestamp, were processed using time series analysis techniques like discrete wavelet transform, approximate entropy and dynamic time warping in order to extract 13 quantitative measures for representing meaningful motor impairment information. A principal component analysis (PCA) was used to reduce the dimensions of the quantitative measures into 4 principal components (PC). In order to classify the motor states into 3 categories that is Off, HE and dyskinesia, a logistic regression model was used as a classifier to map the 4 PCs to the corresponding clinically assigned motor state categories. A stratified 10-fold cross-validation (also known as rotation estimation) was applied to assess the generalization ability of the logistic regression classifier to future independent data sets. To investigate mean differences of the 4 PCs across the three categories, a one-way ANOVA test followed by Tukey multiple comparisons was used. Results: The agreements between computed and clinician ratings were very good with a weighted area under the receiver operating characteristic curve (AUC) coefficient of 0.91. The mean PC scores were different across the three motor state categories, only at different levels. The first 2 PCs were good at discriminating between the motor states whereas the PC3 was good at discriminating between HE subjects and PD patients. The mean scores of PC4 showed a trend across the three states but without significant differences. The Spearman’s rank correlations between the first 2 PCs and clinically assessed motor impairments were as follows: drawing speed (PC1, 0.34; PC2, 0.83), irregularity (PC1, 0.17; PC2, 0.17), and hesitation (PC1, 0.27; PC2, 0.77). Conclusions: These findings suggest that spirography-based objective measures are valid measures of spatial- and time-dependent deficits and can be used to distinguish drug-related motor dysfunctions between Off and dyskinesia in PD. These measures can be potentially useful during clinical evaluation of individualized drug-related complications such as over- and under-medications thus maximizing the amount of time the patients spend in the On state.

A web-based system for visualizing upper limb motor performance of Parkinson’s disease patients

Objective To design, develop and set up a web-based system for enabling graphical visualization of upper limb motor performance (ULMP) of Parkinson’s disease (PD) patients to clinicians. Background Sixty-five patients diagnosed with advanced PD have used a test battery, implemented in a touch-screen handheld computer, in their home environment settings over the course of a 3-year clinical study. The test items consisted of objective measures of ULMP through a set of upper limb motor tests (finger to tapping and spiral drawings). For the tapping tests, patients were asked to perform alternate tapping of two buttons as fast and accurate as possible, first using the right hand and then the left hand. The test duration was 20 seconds. For the spiral drawing test, patients traced a pre-drawn Archimedes spiral using the dominant hand, and the test was repeated 3 times per test occasion. In total, the study database consisted of symptom assessments during 10079 test occasions. Methods Visualization of ULMP The web-based system is used by two neurologists for assessing the performance of PD patients during motor tests collected over the course of the said study. The system employs animations, scatter plots and time series graphs to visualize the ULMP of patients to the neurologists. The performance during spiral tests is depicted by animating the three spiral drawings, allowing the neurologists to observe real-time accelerations or hesitations and sharp changes during the actual drawing process. The tapping performance is visualized by displaying different types of graphs. Information presented included distribution of taps over the two buttons, horizontal tap distance vs. time, vertical tap distance vs. time, and tapping reaction time over the test length. Assessments Different scales are utilized by the neurologists to assess the observed impairments. For the spiral drawing performance, the neurologists rated firstly the ‘impairment’ using a 0 (no impairment) – 10 (extremely severe) scale, secondly three kinematic properties: ‘drawing speed’, ‘irregularity’ and ‘hesitation’ using a 0 (normal) – 4 (extremely severe) scale, and thirdly the probable ‘cause’ for the said impairment using 3 choices including Tremor, Bradykinesia/Rigidity and Dyskinesia. For the tapping performance, a 0 (normal) – 4 (extremely severe) scale is used for first rating four tapping properties: ‘tapping speed’, ‘accuracy’, ‘fatigue’, ‘arrhythmia’, and then the ‘global tapping severity’ (GTS). To achieve a common basis for assessment, initially one neurologist (DN) performed preliminary ratings by browsing through the database to collect and rate at least 20 samples of each GTS level and at least 33 samples of each ‘cause’ category. These preliminary ratings were then observed by the two neurologists (DN and PG) to be used as templates for rating of tests afterwards. In another track, the system randomly selected one test occasion per patient and visualized its items, that is tapping and spiral drawings, to the two neurologists. Statistical methods Inter-rater agreements were assessed using weighted Kappa coefficient. The internal consistency of properties of tapping and spiral drawing tests were assessed using Cronbach’s α test. One-way ANOVA test followed by Tukey multiple comparisons test was used to test if mean scores of properties of tapping and spiral drawing tests were different among GTS and ‘cause’ categories, respectively. Results When rating tapping graphs, inter-rater agreements (Kappa) were as follows: GTS (0.61), ‘tapping speed’ (0.89), ‘accuracy’ (0.66), ‘fatigue’ (0.57) and ‘arrhythmia’ (0.33). The poor inter-rater agreement when assessing “arrhythmia” may be as a result of observation of different things in the graphs, among the two raters. When rating animated spirals, both raters had very good agreement when assessing severity of spiral drawings, that is, ‘impairment’ (0.85) and irregularity (0.72). However, there were poor agreements between the two raters when assessing ‘cause’ (0.38) and time-information properties like ‘drawing speed’ (0.25) and ‘hesitation’ (0.21). Tapping properties, that is ‘tapping speed’, ‘accuracy’, ‘fatigue’ and ‘arrhythmia’ had satisfactory internal consistency with a Cronbach’s α coefficient of 0.77. In general, the trends of mean scores of tapping properties worsened with increasing levels of GTS. The mean scores of the four properties were significantly different to each other, only at different levels. In contrast from tapping properties, kinematic properties of spirals, that is ‘drawing speed’, ‘irregularity’ and ‘hesitation’ had a questionable consistency among them with a coefficient of 0.66. Bradykinetic spirals were associated with more impaired speed (mean = 83.7 % worse, P < 0.001) and hesitation (mean = 77.8% worse, P < 0.001), compared to dyskinetic spirals. Both these ‘cause’ categories had similar mean scores of ‘impairment’ and ‘irregularity’. Conclusions In contrast from current approaches used in clinical setting for the assessment of PD symptoms, this system enables clinicians to animate easily and realistically the ULMP of patients who at the same time are at their homes. Dynamic access of visualized motor tests may also be useful when observing and evaluating therapy-related complications such as under- and over-medications. In future, we foresee to utilize these manual ratings for developing and validating computer methods for automating the process of assessing ULMP of PD patients.

Growth and Inequality : A study of Swedish municipalities

This paper explores the relationship between the growth rate of the average income and income inequality using data at the municipal level in Sweden for the period 1992-2007. We estimate a fixed effects panel data growth model where the within-municipality income inequality is one of the explanatory variables. Different inequality measures (Gini coefficient, top income shares, and measures of inequality in the lower and upper ends of the income distribution) are also examined. We find a positive and significant relationship between income growth and income inequality, measured as the Gini coefficient and top income shares, respectively. In addition, while inequality at the upper end of the income distribution is positively associated with the income growth rate, inequality at the lower end of the income distribution seems to be negatively related to the growth rate. Our findings also suggest that increased income inequality enhances growth more in municipalities with a high level of average income than in those with a low level of average income.