On Causal Mediation Analysis with a Survival Outcome

Suppose that having established a marginal total effect of a point exposure on a time-to-event outcome, an investigator wishes to decompose this effect into its direct and indirect pathways, also know as natural direct and indirect effects, mediated by a variable known to occur after the exposure and prior to the outcome. This paper proposes a theory of estimation of natural direct and indirect effects in two important semiparametric models for a failure time outcome. The underlying survival model for the marginal total effect and thus for the direct and indirect effects, can either be a marginal structural Cox proportional hazards model, or a marginal structural additive hazards model. The proposed theory delivers new estimators for mediation analysis in each of these models, with appealing robustness properties. Specifically, in order to guarantee ignorability with respect to the exposure and mediator variables, the approach, which is multiply robust, allows the investigator to use several flexible working models to adjust for confounding by a large number of pre-exposure variables. Multiple robustness is appealing because it only requires a subset of working models to be correct for consistency; furthermore, the analyst need not know which subset of working models is in fact correct to report valid inferences. Finally, a novel semiparametric sensitivity analysis technique is developed for each of these models, to assess the impact on inference, of a violation of the assumption of ignorability of the mediator.

GEOSTATISTICAL INFERENCE UNDER PREFERENTIAL SAMPLING

Geostatistics involves the fitting of spatially continuous models to spatially discrete data (Chil`es and Delfiner, 1999). Preferential sampling arises when the process that determines the data-locations and the process being modelled are stochastically dependent. Conventional geostatistical methods assume, if only implicitly, that sampling is non-preferential. However, these methods are often used in situations where sampling is likely to be preferential. For example, in mineral exploration samples may be concentrated in areas thought likely to yield high-grade ore. We give a general expression for the likelihood function of preferentially sampled geostatistical data and describe how this can be evaluated approximately using Monte Carlo methods. We present a model for preferential sampling, and demonstrate through simulated examples that ignoring preferential sampling can lead to seriously misleading inferences. We describe an application of the model to a set of bio-monitoring data from Galicia, northern Spain, in which making allowance for preferential sampling materially changes the inferences.

COVARIATE-ADJUSTED NONPARAMETRIC ANALYSIS OF MAGNETIC RESONANCE IMAGES USING MARKOV CHAIN MONTE CARLO

Permutation tests are useful for drawing inferences from imaging data because of their flexibility and ability to capture features of the brain that are difficult to capture parametrically. However, most implementations of permutation tests ignore important confounding covariates. To employ covariate control in a nonparametric setting we have developed a Markov chain Monte Carlo (MCMC) algorithm for conditional permutation testing using propensity scores. We present the first use of this methodology for imaging data. Our MCMC algorithm is an extension of algorithms developed to approximate exact conditional probabilities in contingency tables, logit, and log-linear models. An application of our non-parametric method to remove potential bias due to the observed covariates is presented.

Recurrent voluminous sector collapses at Volcán Barú, Panama

Two volcanic debris avalanche deposits (VDADs), both attributed to sector collapse at Volcán Barú, Panama, have been identified after an investigation of deposits that covered more than a thousand square kilometers. The younger Barriles Deposit is constrained by two radiocarbon ages that are ~9 ka; the older Caisán Deposit is at or beyond the radiocarbon range, >43,500 ybp. The total runout length of the Caisán Deposit was ~50 km and it covers 1190 km2. The Barriles Deposit extended to about 45 km and covered an area of 966 km2, overlapping most of the Caisán. The VDADs are blanketed by pyroclastic deposits and contain a predominance of andesitic material likely representing volcanic dome rock which accumulated above the active vent at Barú before collapsing. Despite heavy vegetation in the field area, over 4000 individual hummocks were digitized from aerial photography. Statistical analysis of hummock locations and geometries depict flow patterns of highly- fragmented material reflecting the effects of underlying topography and also help to define the limit of Barriles’ shorter termination. Barriles and Caisán are primarily unconfined, subaerial volcanic deposits that are among the world’s most voluminous. Calculated through two different geospatial processes, thickness values from field measurements and inferences yield volumes >30 km23 for both deposits. VDADs of comparable scale come from Mount Shasta, USA; Socompa, Chile/Argentina; and Shiveluch, Russia. Currently, the modern edifice is 200-400m lower than the pre-collapse Barriles and Caisán summits and only 16-25% of the former edifice has been replaced since the last failure.

Statistical analysis of multichannel scalp field data

High density spatial and temporal sampling of EEG data enhances the quality of results of electrophysiological experiments. Because EEG sources typically produce widespread electric fields (see Chapter 3) and operate at frequencies well below the sampling rate, increasing the number of electrodes and time samples will not necessarily increase the number of observed processes, but mainly increase the accuracy of the representation of these processes. This is namely the case when inverse solutions are computed. As a consequence, increasing the sampling in space and time increases the redundancy of the data (in space, because electrodes are correlated due to volume conduction, and time, because neighboring time points are correlated), while the degrees of freedom of the data change only little. This has to be taken into account when statistical inferences are to be made from the data. However, in many ERP studies, the intrinsic correlation structure of the data has been disregarded. Often, some electrodes or groups of electrodes are a priori selected as the analysis entity and considered as repeated (within subject) measures that are analyzed using standard univariate statistics. The increased spatial resolution obtained with more electrodes is thus poorly represented by the resulting statistics. In addition, the assumptions made (e.g. in terms of what constitutes a repeated measure) are not supported by what we know about the properties of EEG data. From the point of view of physics (see Chapter 3), the natural “atomic” analysis entity of EEG and ERP data is the scalp electric field

Are clustering effects accounted for in statistical analysis in leading dental specialty journals?

OBJECTIVES In dental research multiple site observations within patients or taken at various time intervals are commonplace. These clustered observations are not independent; statistical analysis should be amended accordingly. This study aimed to assess whether adjustment for clustering effects during statistical analysis was undertaken in five specialty dental journals. METHODS Thirty recent consecutive issues of Orthodontics (OJ), Periodontology (PJ), Endodontology (EJ), Maxillofacial (MJ) and Paediatric Dentristry (PDJ) journals were hand searched. Articles requiring adjustment accounting for clustering effects were identified and statistical techniques used were scrutinized. RESULTS Of 559 studies considered to have inherent clustering effects, adjustment for this was made in the statistical analysis in 223 (39.1%). Studies published in the Periodontology specialty accounted for clustering effects in the statistical analysis more often than articles published in other journals (OJ vs. PJ: OR=0.21, 95% CI: 0.12, 0.37, p<0.001; MJ vs. PJ: OR=0.02, 95% CI: 0.00, 0.07, p<0.001; PDJ vs. PJ: OR=0.14, 95% CI: 0.07, 0.28, p<0.001; EJ vs. PJ: OR=0.11, 95% CI: 0.06, 0.22, p<0.001). A positive correlation was found between increasing prevalence of clustering effects in individual specialty journals and correct statistical handling of clustering (r=0.89). CONCLUSIONS The majority of studies in 5 dental specialty journals (60.9%) examined failed to account for clustering effects in statistical analysis where indicated, raising the possibility of inappropriate decreases in p-values and the risk of inappropriate inferences.

Unresolved Problems with the "I", the "A" and the "T": Logical and Psychometric Critique of the Implicit Association Test (IAT)

The Implicit Association Test (IAT) had already gained the status of a prominent assessment procedure before its psychometric properties and underlying task structure were understood. The present critique addresses five major problems that arise when the IAT is used for diagnostic inferences: (1) the asymmetry of causal and diagnostic inferences; (2) the viability of the underlying association model; (3) the lack of a testable model underlying IAT-based inferences; (4) the difficulties of interpreting difference scores; and (5) the susceptibility of the IAT to deliberate faking and strategic processing. Based on a theoretical reflection of these issues, and a comprehensive survey of published IAT studies, it is concluded that a number of uncontrolled factors can produce (or reduce) significant IAT scores independently of the personality attribute that is supposed to be captured by the IAT procedure.

East Asian ethnolinguistic phylogeography

A polyphyletic understanding of Asian linguistic diversity was first propagated in 1823. Since 1901, various scholars have proposed larger linguistic phyla uniting two or more recognised Asian language families. The most recent proposal in this tradition, Starosta’s 2001 East Asian phylum, comprising the Trans-Himalayan, Hmong-Mien, Austroasiatic, Austronesian and Kradai language families, is reassessed in light of linguistic and non-linguistic evidence. Ethnolinguistically informed inferences based on Asian Y chromosomal phylogeography lead to a reconstruction of various episodes of ethnolinguistic prehistory which lie beyond the linguistic event horizon, i.e. at a time depth empirically inaccessible to historical linguistics. The Father Tongue correlation in population genetics, the evidence for refugia during the Last Glacial Maximum and the hypothesis of language families having arisen as the result of demographic bottlenecks in prehistory are shown to be crucial to an understanding of the ethnogenesis of East Asian linguistic phyla. The prehistory of several neighbouring Asian language families is discussed, and the Centripetal Migration model is opposed to the Farming Language Dispersal theory.

A prehistoric thoroughfare between the Ganges and the Himalayas

The greater Himalayan region, including the Tibetan plateau in the north and the Gangetic plain in the south, served as the principal prehistoric thoroughfare for the peopling of East and Southeast Asia. The descendants of ancient migrants through this region ultimately settled lands as far away as New Zealand, Madagascar and the Americas. Several of the keys to understanding the ethnogenesis of human diversity in Asia include the Father Tongue correlation, possible refugia during the Last Glacial Maximum and the hypothesis that language families may have arisen as the result of demographic bottlenecks in prehistory. Ethnolinguistically informed inferences based on Asian Y chromosomal phylogeography permit a reconstruction of episodes of ethnolinguistic prehistory which lie beyond the linguistic event horizon, i.e. beyond the time depth empirically accessible to historical linguistics. The origins of the language families which make up the hypothetical Uralo-Siberian and East Asian linguistic phyla are argued to have lain in the northeastern corner of the Indian subcontinent. Several other Asian language families are shown to be tied to the subcontinent. The Centripetal Migration model, which assumes that migrations in quest of a better life unfolded in both centrifugal and centripetal directions with respect to technologically more advanced centres of civilisation, is opposed to the Farming Language Dispersal theory, which assumes that all linguistic dispersals were driven by agricultural centrifugal migration.

Cognitive Differences in Chart Reading: A Comparison of Nurses and Physicians

This pilot study compares the mental models of a patient constructed by nurses and physicians while reading an electronic medical record. Preliminary results suggest that the participants' summaries were both quantitatively and qualitatively different. The physician made more inferences and focused on deeper relationships in the record, whereas the nurse focused on the descriptive surface structure of the record.

NETWORK TOPOLOGY IN HUMAN PROTEIN INTERACTION DATA PREDICTS FUNCTIONAL ASSOCIATION

High-throughput assays, such as yeast two-hybrid system, have generated a huge amount of protein-protein interaction (PPI) data in the past decade. This tremendously increases the need for developing reliable methods to systematically and automatically suggest protein functions and relationships between them. With the available PPI data, it is now possible to study the functions and relationships in the context of a large-scale network. To data, several network-based schemes have been provided to effectively annotate protein functions on a large scale. However, due to those inherent noises in high-throughput data generation, new methods and algorithms should be developed to increase the reliability of functional annotations. Previous work in a yeast PPI network (Samanta and Liang, 2003) has shown that the local connection topology, particularly for two proteins sharing an unusually large number of neighbors, can predict functional associations between proteins, and hence suggest their functions. One advantage of the work is that their algorithm is not sensitive to noises (false positives) in high-throughput PPI data. In this study, we improved their prediction scheme by developing a new algorithm and new methods which we applied on a human PPI network to make a genome-wide functional inference. We used the new algorithm to measure and reduce the influence of hub proteins on detecting functionally associated proteins. We used the annotations of the Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) as independent and unbiased benchmarks to evaluate our algorithms and methods within the human PPI network. We showed that, compared with the previous work from Samanta and Liang, our algorithm and methods developed in this study improved the overall quality of functional inferences for human proteins. By applying the algorithms to the human PPI network, we obtained 4,233 significant functional associations among 1,754 proteins. Further comparisons of their KEGG and GO annotations allowed us to assign 466 KEGG pathway annotations to 274 proteins and 123 GO annotations to 114 proteins with estimated false discovery rates of <21% for KEGG and <30% for GO. We clustered 1,729 proteins by their functional associations and made pathway analysis to identify several subclusters that are highly enriched in certain signaling pathways. Particularly, we performed a detailed analysis on a subcluster enriched in the transforming growth factor β signaling pathway (P<10-50) which is important in cell proliferation and tumorigenesis. Analysis of another four subclusters also suggested potential new players in six signaling pathways worthy of further experimental investigations. Our study gives clear insight into the common neighbor-based prediction scheme and provides a reliable method for large-scale functional annotations in this post-genomic era.

A Coastal Transect of McMurdo Dry Valleys (Antarctica) Snow and Firn: Marine and Terrestrial Influences on Glaciochemistry

Samples of snow and firn from accumulation zones on Clark, Commonwealth, Blue and Victoria Upper Glaciers in the McMurdo Dry Valleys (similar to 77-78 degrees S, 161-164 degrees E), Antarctica, are evaluated chemically and isotopically to determine the relative importance of local (site-specific) factors vs regional-scale influences in defining glaciochemistry. Spatial variation in snow and firn chemistry confirms documented trends within individual valleys regarding major-ion deposition relative to elevation and to distance from the coast. Sodium and methylsulfonate (MS-), for example, follow a decreasing gradient with distance from the coast along the axis of Victoria Valley (350-119 mu gL(-1) for Na+; 33-14 mu gL(-1) for MS-); a similar pattern exists between Commonwealth and Newall Glaciers in the Asgaard Range. When comparing major-ion concentrations (e.g. Na-+,Na- MS-, Ca2+) or trace metals (e.g. Al, Fe) among different valleys, however, site-specific exposures to marine and local terrestrial chemical sources play a dominant role. Because chemical signals at all sites respond to particulates with varying mixtures of marine and terrestrial sources, each of these influences on site glaciochemistry must be considered when drawing temporal climate inferences on regional scales.

Propositional and situational representations of text

Two informationally equivalent texts were constructed which described a fictitious town, emphasizing its spatial layout. In one version (Survey text), spatial information was in geographic terms, while in the other version (Route text), the equivalent information was provided in the form of directions for driving through the town. Subjects recalled these texts and verified old as well as inference statements. In Experiment I, subjects were able to recall the texts quite well, while showing little ability to use the information they had acquired to make inferences about spatial relations in the town which had not been directly stated in the text. With simpler texts, subjects in Experiment II were able to make infereces, especially when the form of the question corresponded to the version of the text they had read. It was concluded that free recall depended on the construction of a propositional textbase during comprehension, while inferences required a situation model, either in the form of a mental map or a procedural representation of the town. It could be shown that the form of the situation model depended on both the representation invited by the text and subject biases.

Die mentale Repräsentation von verbalen Ortsbeschreibungen

Studied the mental representation of verbally described spatial layouts. Human subjects: 36 normal Swiss adolescents and adults. Two informationally equivalent texts describing the same fictitious town were constructed and presented to different groups. Spatial information was given in geographic terms in the survey text and in the form of directions for driving through the town in the route text. After learning the texts Ss had to verify the route (forward vs backward) and survey inferences.

Palynological richness and evenness: insights from the taxa accumulation curve

Palynology provides the opportunity to make inferences on changes in diversity of terrestrial vegetation over long time scales. The often coarse taxonomic level achievable in pollen analysis, differences in pollen production and dispersal, and the lack of pollen source boundaries hamper the application of diversity indices to palynology. Palynological richness, the number of pollen types at a constant pollen count, is the most robust and widely used diversity indicator for pollen data. However, this index is also influenced by the abundance distribution of pollen types in sediments. In particular, where the index is calculated by rarefaction analysis, information on taxonomic richness at low abundance may be lost. Here we explore information that can be extracted from the accumulation of taxa over consecutive samples. The log-transformed taxa accumulation curve can be broken up into linear sections with different slope and intersect parameters, describing the accumulation of new taxa within the section. The breaking points may indicate changes in the species pool or in the abundance of high versus low pollen producers. Testing this concept on three pollen diagrams from different landscapes, we find that the break points in the taxa accumulation curves provide convenient zones for identifying changes in richness and evenness. The linear regressions over consecutive samples can be used to inter- and extrapolate to low or extremely high pollen counts, indicating evenness and richness in taxonomic composition within these zones. An evenness indicator, based on the rank-order-abundance is used to assist in the evaluation of the results and the interpretation of the fossil records. Two central European pollen diagrams show major changes in the taxa accumulation curves for the Lateglacial period and the time of human induced land-use changes, while they do not indicate strong changes in the species pool with the onset of the Holocene. In contrast, a central Swedish pollen diagram shows comparatively little change, but high richness during the early Holocene forest establishment. Evenness and palynological richness are related for most periods in the three diagrams, however, sections before forest establishment and after forest clearance show high evenness, which is not necessarily accompanied by high palynological richness, encouraging efforts to separate the two.