Geostatistical Analyst has several thousands of users and they have very different backgrounds and interests. Unfortunately, many researchers use the software only to make maps. Many statisticians do not understand or appreciate the full utility of GIS for spatial data analysis. Other researchers are not educated in spatialstatistics, so they are unaware of techniques for modeling uncertainty, even though they realize that measuring and modeling without errors is impossible. At the same time, they readily use automatic “geoprocessing” tools that arithmetically add or average raster data values in the cells without taking into account the impact of error propagation on the results. After several such geoprocessing steps, the resulting data structure can be completely random and, consequently, decisions made from these results may be wrong. Still others use the software for analyses for which it was not designed. This problem often arises when users implement Geostatistical Analyst with aggregated data that are associated with spatially discrete units. For example, we have watched users discuss which interpolator, inverse squared distance weighting or splines, is better for mapping of proportions of females in burial populations. What is the best prediction of the proportion of females outside cemeteries? One would hope it is zero. There are other methods in spatialstatistics that are more suitable for this type of data. For example, a marked point pattern analysis that incorporates an attribute value recorded at each location could be used. The mark would be the occurrence (or not) of a female burial at each location, and a marked point pattern model could then be used to estimate and map the intensity of females in burial populations, assuming that data point locations are given by nature and not selected by the user. Hopefully, all of these problems can be solved by education. Case studies can help to show users how GIS can and should be used for more sophisticated statistical analysis and modeling. 3. IMPROVED SPATIAL DATA
In this work, we aimed at bridging the gap in previous literature by introducing a series of morphological and spatial analyses for the characterization of nanotubular architectures, culminating in the descriptive analysis of nanotubular surfaces towards a cohesive investigation of cell-nanotube interactions. To this end, we investigated three nanotubular arrays with variable nanotube diameters and a two-tiered honeycomb structure. Successively, we carried out morphological and spatial analysis (e.g., Voronoi entropy) to quantify the tubular geometry, arrangement as well as degree of order for these architectures. Our experimental data was validated by compu- tational simulations to provide greater insight into the role of morphological parameters and spatialstatistics. Subsequently, we evaluated the effects of the four surfaces on hMSC bioac- tivity (i.e., proliferative and morphological analyses) and osteogenic induction (e.g, commitment to osteogenic differen- tiation and bone mineral quality). Results from our study (i) highlight the importance of including additional morphological analyses and spatialstatistics in the characterization of nano- tubular surfaces for the purpose of enhancing the validity of cross-study comparisons, (ii) provide a comprehensive corre- lation between a multifactorial array of these parameters and hMSC activity extending from adhesion to bone mineral deposition, and lastly (iii) report the synergistic effects elicited by the HC architecture.
Lattice models are a way of representing spatial locations in a grid where each cell is in a certain state and evolves according to transition rules and rates dependent on a surrounding neighbourhood. These models are capable of describing many phenom- ena such as the simulation and growth of a forest fire front. These spatial simulation models as well as spatial descriptive statistics such as Ripley’s K-function have wide applicability in spatialstatistics but in general do not scale well for large datasets. Parallel computing (high performance computing) is one solution that can provide limited scalability to these applications. This is done using the message passing in- terface (MPI) framework implemented in R through the Rmpi package. Other useful techniques in spatialstatistics such as point pattern reconstruction and Markov Chain Monte Carlo (MCMC) methods are discussed from a parallel computing perspective as well. In particular, an improved point pattern reconstruction is given and imple- mented in parallel. Single chain MCMC methods are also examined and improved upon to give faster convergence using parallel computing. Optimizations, and compli- cations that arise from parallelizing existing spatialstatistics algorithms are discussed and methods are implemented in an accompanying R package, parspatstat.
The Gleason score of a prostatic carcinoma is generally considered as one of the most important prognostic parameters of this tumour type. In the present study, it was attempted to study the relation between the Gleason score and objective data of spatialstatistics, and to predict this score from such data. For this purpose, 25 T1 incidental prostatic carcinomas, 50 pT2N0, and 28 pT3N0 prostatic adenocarcinomas were characterized by a histological texture analysis based on principles of spatialstatistics. On sectional images, progression from low grade to high grade prostatic cancer in terms of the Gleason score is correlated with complex changes of the epithelial cells and their lumina with respect to their area, boundary length and Euler number per unit area. The central finding was a highly significant negative correlation between the Gleason score and the Euler number of the epithelial cell phase per unit area. The Gleason score of all individual cases was predicted from the spatial statistical variables by multivariate linear regression. This approach means to perform a multiclass pattern recognition, as opposed to the usual problem of binary pattern recognition. A prediction was considered as acceptable when its deviation from the human classification was no more than 1 point. This was achieved in 79 of these 103 cases when only the Euler number density was used as predictor variable. The accuracy could be risen slightly to 84 of the 103 cases, when 7 input variables were used for prediction of the Gleason score, which means an accuracy of 81.5%.
weather-related processes has similar importance as temporal prediction because, as men- tioned by Cressie (1993), “spatial prediction is just as important as temporal prediction, because people living those cities and rural districts without monitoring stations have the same right to know how little or how much their water or their air is polluted.” Most of the spatial prediction of weather-related processes are based on data collected by high- performance sensors at meteorological stations or images captured by high-resolution cameras in satellites. Recently, with the advancement of mobile sensor-related technology, geo-tagged weather information is being collected by micro-sensors installed in mobile devices and gathered by mobile weather applications like AccuWeather, WeatherSignal etc. These datasets are often referred as ‘crowdsourced’ weather data as the information is coming from the mobile application users. Standard methodologies in geostatistics or spatialstatistics is not directly applicable to these mobile sensor-generated data as quality of the observations are often hampered in crowdsourced datatsets due to several factors: the low-quality of the sensors, indoor-outdoor user activity, influence of exter- nal and internal processes etc. to name a few. Developing data-driven robust as well as scalable methodologies to analyze noisy spatial data like crowdsourced weather data is broad focus of this dissertation.
While this study provides hints at strong effects underlying the spatial distribution of violent crime in Durham and how projects like HOPE VI may alter that distribution, for the most part they remain hints. Similar work with spatialstatistics to provide stronger, more robust understanding of these distributions would require both additional data sources, quantifying the many factors which might contribute to the changes in spatial distribution, as well as multivariate statistical techniques in addition to the univariate techniques used in this study. As discussed earlier, the most meaningful if potentially difficult additional factor which could be added to this study would be a spatially explicit representation of the vulnerable population. This conceivably could be developed as an activity model incorporating residential, employment, and retail data, as well as traffic counts and transportation models. Alternately, a spatial interpolation model could be built from empirically collected activity counts at key areas of the city. Even here, however, spatial interpolation becomes difficult, because activity counts do not vary smoothly along Euclidean distances, but follow rigid features of the urban landscape.
Background: Autism spectrum disorders (ASD) are associated with widespread alterations in white matter (WM) integrity. However, while a growing body of studies is shedding light on microstructural WM alterations in high-functioning adolescents and adults with ASD, literature is still lacking in information about whole brain structural connectivity in children and low-functioning patients with ASD. This research aims to investigate WM connectivity in ASD children with and without mental retardation compared to typically developing controls (TD). Methods: Diffusion tensor imaging (DTI) was performed in 22 young children with ASD (mean age: 5.54 years) and 10 controls (mean age: 5.25 years). Data were analysed both using the tract-based spatialstatistics (TBSS) and the tractography. Correlations were investigated between the WM microstructure in the identified altered regions and the productive language level.
Given the vast toolset spatial epidemiologists have at their disposal, including a variety of global and local clustering statistics as well as descriptive spatialstatistics, it is im- portant to establish appropriate evidence-based method- ologies. These analyses found that local spatialstatistics tended to generally agree on where clusters occurred, with regions within Alachua County, Florida, USA consistently showing clustering, regardless of minority status. Figures 3 and A1 (see Additional file 1) show smaller, more lo- calized cancer clusters among minority members within rural areas of the county than those who were non- minority members, with “hot spots” located in the west and southwest rural areas. Such findings may be used not only to direct ongoing and future outreach efforts, but may have policy implications within the area in terms of service assessment and resource allocation. Further, these findings are consistent with previous oncological litera- ture, suggesting minority populations, particularly in rural areas throughout the US have disproportionately high cancer morbidity and mortality. This may potentially be a result of increased barriers to obtaining preventa- tive screenings which can catch lesions before cancer progression . The results of these analyses therefore suggest that if patterns are present on the landscape, such local clustering techniques may have the appropri- ate resolution to consistently identify them, regardless of test statistic used.
mricro/mricron/dcm2nii.html) was used to convert the raw DICOM files from the proprietary scanner format to the nifti format, ‘‘.image’’. The diffusion-weighted images were analysed using the Functional Magnetic Resonance Imaging of the Brain Library (FSL, Oxford, United Kingdom). Briefly, the FSL eddy-correct tool was used to register all diffuse images in the B0 image space. The FSL bet2 was then used to skull- strip the brain to ensure that only tensors inside the brain were calculated, rather than those in the surrounding air, using a threshold of 0.25. Finally, a FSL DTIFIT tool was applied to calculate the diffusion tensor model at each pixel and obtain a FA image map. After data processing, tract-based spatialstatistics (TBSS) 42 were used to explore
RESULTS: Tract-based spatialstatistics analysis applied to early-acquired scans (postmenstrual age of 30 –33 weeks) indicated a limited signiﬁcant positive association between motor skills and axial diffusivity and radial diffusivity values in the corpus callosum, internal and external/extreme capsules, and midbrain (P ⬍ .05, corrected). In contrast, for term scans (postmenstrual age of 37– 41 weeks), tract-based spatialstatistics analysis showed a signiﬁcant relationship between both motor and cognitive scores with fractional anisotropy in the corpus callosum and corticospinal tracts (P ⬍ .05, corrected). Tract-based spatialstatistics in a limited subset of neonates (n ⫽ 22) scanned at ⬍ 30 weeks did not signiﬁcantly predict neurodevelopmental outcomes.
DOI: 10.4236/ojs.2019.92015 211 Open Journal of Statistics to the growing demand for younger trees, thanks to intense selection of clones . This demand results in the systematic rotation of trees in Eucalyptus areas, a reason why probably ant nests are aggregated in borders, especially between native forest and Eucalyptus plantations or Cerrado and Eucalyptus plantations. With the rotation of trees, the number of leaves available to supply ant nests de- creases drastically and incipient nests can be rapidly eliminated, increasing the risk that ant populations will fail to persist. However, old nests easily persist to the next Eucalyptus planting, using other resources. A variety of resources can be obtained by workers living in old nests because they are capable of foraging for much longer distances than workers in incipient nests .
Because malaria treatment-seeking practices differ around the world [5-21], no universal strategy can be developed to tackle the issue of malaria incidence and treatment. Efforts to tailor malaria control programmes to local needs, requires an understanding of the factors that influence individual treatment-seeking practices. In this paper, spatial pattern analysis techniques and spatial regression are used to illustrate where national control programme services are well-utilized and where they are under-utilized, to identify the factors contributing to alternative treatment-seeking preferences, and to assess how the predictive strength of those factors change across the study area. Understanding where each factor is a strong predictor of treatment-seeking preferences can inform the design of targeted interventions aimed at increasing control programme utilization. Given the results of the spatial analysis presented, a variety of pos- sible intervention strategies are suggested.
A second set of registrations was then performed to register every indi- vidual FA map to the mean FA map. The aligned images were then used to create another mean FA map and a mean FA skeleton, which represented the centers of all tracts common to the group. This FA skeleton was thresholded at FA ⱖ 0.15 to exclude peripheral tracts with high intersubject variability and/orpartialvolumeeffectswithgraymatter.Eachsubjects’alignedFA,AD, and RD data were projected onto this mean FA skeleton. Voxelwise cross- subject statistics was performed to assess the relationship between FA, AD, and RD and performance scores of the BSITD-III, corrected for gestational age and postmenstrual age at the time of the scanning. The results were cor- rected for multiple comparisons by controlling the family-wise error rate following TFCE. 23
aged brain developed by the Montreal Neurological Institute). Sec- ond, the transformed FA images are averaged to create a mean FA image. Third, the mean FA is fed into the tract skeleton generation, which aims to represent all of the tracts that are “common” to all of the subjects. The skeleton will represent each such tract as a single line (or surface) running down the center of the tract. To achieve skeleton- ization, the local surface perpendicular direction is estimated (at all of the voxels in the image), and a nonmaximum suppression in this direction is performed. In other words, a search is made along all of the voxels in the local ‘‘tract perpendicular direction,” and the voxel with the highest FA is identified as the center of the tract. The esti- mated tract perpendicular direction is regularized to improve estima- tion robustness. Fourth, the center of each tract is found by compar- ing the FA value with the 2 closest neighbors on each side, in the direction of the tract perpendicular. If the FA value is greater than the neighboring values, then the voxel is marked as lying on the skeleton. Each subject’s aligned FA data were then projected onto this skel- eton (Fig 1), and the resulting data were fed into voxelwise cross- subject statistics. A randomization procedure (FSL’s randomize, Monte Carlo permutation test) was used to perform the group anal- ysis statistics. TBSS group maps were generated for the nonparamet- Table 1: Means, SDs, and values of t comparing means on psychological measures for control subjects and adults with dyslexia
Inference in the presence of missing data is a widely encountered and difficult problem in statistics. Imputation is often used to facilitate parameter estimation, which allows one to use the complete sample estimators on the imputed data set. In Chapter 2, We develop a parametric fractional imputation (PFI) method proposed by Kim (2011), which simplifies the computation associated with the EM algorithm for maximum likelihood estimation with missing data. We first consider the problem of parameter estimation for linear mixed models with non-ignorable missing values, which assumes that missingness depends on the missing values only through the random effects, leading to shared parameter models (Follmann and Wu,1995). In the M- step, the restricted or adjusted profiled maximum likelihood method is used to reduce the bias of maximum likelihood estimation of the variance components. Results from a limited simulation study are presented to compare the proposed method with the existing methods, which demonstrates that imputation can significantly reduce the non-response bias and the idea of adjusted profiled maximum likelihood works nicely in PFI for the bias correction in estimating the variance components. Variance estimation is also discussed. We next extend PFI to generalized linear mixed model and the flexibility of this method is illustrated by analyzing the infamous salamander mating data (McCullagh and Nelder, 1989).
In the current study, we sought to characterize WM diffusion properties associated with ASD in a sample of male and female preschool-aged children. We utilize DWI acquired during natural nocturnal sleep  to in- vestigate measures of FA, MD, RD, and AD across whole brain WM using a voxel-wise tract-based spatial statis- tics (TBSS) approach . We hypothesize that individ- uals with ASD will have significant differences in WM diffusion properties in tracts previously indicated in the condition, including the corpus callosum and superior longitudinal fasciculus. To our knowledge, our study represents the largest diffusion imaging study in terms of inclusion of preschool-aged females with ASD. Based on prior DWI findings from our group reporting signifi- cant sex differences in TD  and diagnosis-by-sex interaction effects in ASD , we anticipate both a sig- nificant main effect of sex and diagnosis-by-sex interac- tions in diffusion measures.
other one to identify the most representative one, and this image was used as the target image. This target image was then affine-aligned into MNI 152 standard space, and every image was transformed into 1 ⫻ 1 ⫻ 1 mm MNI 152 space by combining the nonlinear transform to the target FA image with the affine transform from the target native space to MNI 152 space. The mean FA image was created and thinned to create the mean FA skeleton, which represented the centers of all tracts common to the groups. This skeleton was thresholded at FA ⬎ 0.25. Each subject’s aligned FA map was then projected onto the skel- eton by assigning each point on the skeleton the maximum FA in a plane perpendicular to the local skeleton structure. The resulting skel- etons were fed into voxelwise statistics. The number of permutations was set to 5000. By using the TBSS results for the FA maps, we also analyzed maps of axial eigenvalues, radial eigenvalues, and ADC val- ues by TBSS analysis.
that the FA values of the bilateral PPN, bilateral PM, right orbitofrontal area and left SMA were lower in the ARWMC patients with FOG compared to those without. Their finding partly concurs the present result, which showed that the FA in the SLF close to the PM, and also in the CP close to the midbrain PPN, was correlated with FOG. The areas detected slightly differed probably because of the difference in the methodology, especially the spatial normalization procedure (TBSS versus ROI analysis). The present result suggests that multiple neural substrates related to the control of walking may be in- volved in FOG in consistent with the study by Youn et al. . This notion favors the view that FOG is likely to be a clinical condition resulting from multiple pathophysio- logical issues involving one or more nodes of the neural network regulating walking behavior. Since visual stimuli or some other extra cues can help or exacerbate FOG , the phenomenon might emerge when extra demands for sensory, cognitive, or emotional processing overburden the damaged neural substrates.
researchers who are using R to study health geographics data for outbreak management, public health planning, and other geographic-data intensive tasks to display their results directly onto dynamic maps. Thirdly, by creating a web application for running an R script, a statistician for example, can enable users who are entirely unfamil- iar with R to run the statistician’s R coded analysis of health geographics data. Users of web applications cre- ated by Rwui have no contact with the actual R script, which runs out of sight on the server; all that users see is the web pages of the application in their browser window. Fourthly, we envisage an educational role for such appli- cations. By leveraging the interactive features of Google dynamic maps, various aspects of spatial statistical analy- ses can be explored in a visual, intuitive way.