Browsing by Subject "MODEL SELECTION"

Sort by: Order: Results:

Now showing items 1-13 of 13
  • Mutanen, Tuomas P.; Metsomaa, Johanna; Liljander, Sara; Ilmoniemi, Risto J. (2018)
    Electroencephalography (EEG) and magnetoencephalography (MEG) often suffer from noise-and artifact-contaminated channels and trials. Conventionally, EEG and MEG data are inspected visually and cleaned accordingly, e.g., by identifying and rejecting the so-called "bad" channels. This approach has several shortcomings: data inspection is laborious, the rejection criteria are subjective, and the process does not fully utilize all the information in the collected data. Here, we present noise-cleaning methods based on modeling the multi-sensor and multi-trial data. These approaches offer objective, automatic, and robust removal of noise and disturbances by taking into account the sensor-or trial-specific signal-to-noise ratios. We introduce a method called the source-estimate-utilizing noise-discarding algorithm (the SOUND algorithm). SOUND employs anatomical information of the head to cross-validate the data between the sensors. As a result, we are able to identify and suppress noise and artifacts in EEG and MEG. Furthermore, we discuss the theoretical background of SOUND and show that it is a special case of the well-known Wiener estimators. We explain how a completely data-driven Wiener estimator (DDWiener) can be used when no anatomical information is available. DDWiener is easily applicable to any linear multivariate problem; as a demonstrative example, we show how DDWiener can be utilized when estimating event-related EEG/MEG responses. We validated the performance of SOUND with simulations and by applying SOUND to multiple EEG and MEG datasets. SOUND considerably improved the data quality, exceeding the performance of the widely used channel-rejection and interpolation scheme. SOUND also helped in localizing the underlying neural activity by preventing noise from contaminating the source estimates. SOUND can be used to detect and reject noise in functional brain data, enabling improved identification of active brain areas.
  • Peltola, Tomi; Marttinen, Pekka; Jula, Antti; Salomaa, Veikko; Perola, Markus; Vehtari, Aki (2012)
  • Yrjölä, Rauno A.; Tanskanen, Antti; Sarvanne, Hannu; Vickholm, Jorma; Lehikoinen, Aleksi (2018)
    Urbanization and other human activities can lead to decreasing animal populations in nearby areas. The impact of human activitiesmay vary depending on the characteristics of the areas and region or on the strength of the disturbance. We investigated forest bird population changes in an EU Natura 2000 area during the construction of the new Helsinki Vuosaari Harbour in southern Finland in 2002-2011 as part of an environmental impact assessment. We evaluated whether the changes observed were linked with the harbour construction work by comparing the populations at sites near the development with those corresponding values obtained from national common bird monitoring in southern Finland. Themean population changes of 23 boreal forest bird species that inhabited the Natura 2000 area and southern Finland were significantly and positively correlated, but the population inside the Natura 2000 study area also showed lower mean numbers (a mean decline of 9% occurred over the study period). Our case study emphasizes the importance of intensive monitoring before, during and after work at the construction site and in the surrounding areas to detect actual changes in the populations.
  • CORE Collaboration; Finelli, F.; Hindmarsh, M.; Kiiveri, K.; Väliviita, J.; Kurki-Suonio, H.; Lindholm, V. (2018)
    We forecast the scientific capabilities to improve our understanding of cosmic inflation of CORE, a proposed CMB space satellite submitted in response to the ESA fifth call for a medium-size mission opportunity. The CORE satellite will map the CMB anisotropies in temperature and polarization in 19 frequency channels spanning the range 60-600 GHz. CORE will have an aggregate noise sensitivity of 1.7 mu K.arcmin and an angular resolution of 5' at 200 GHz. We explore the impact of telescope size and noise sensitivity on the inflation science return by making forecasts for several instrumental configurations. This study assumes that the lower and higher frequency channels suffice to remove foreground contaminations and complements other related studies of component separation and systematic effects, which will be reported in other papers of the series "Exploring Cosmic Origins with CORE." We forecast the capability to determine key inflationary parameters, to lower the detection limit for the tensor-to-scalar ratio down to the 10(-3) level, to chart the landscape of single field slow-roll inflationary models, to constrain the epoch of reheating, thus connecting inflation to the standard radiation-matter dominated Big Bang era, to reconstruct the primordial power spectrum, to constrain the contribution from isocurvature perturbations to the 10(-3) level, to improve constraints on the cosmic string tension to a level below the presumptive GUT scale, and to improve the current measurements of primordial non-Gaussianities down to the f(NL)(local) <1 level. For all the models explored, CORE alone will improve significantly on the present constraints on the physics of inflation. Its capabilities will be further enhanced by combining with complementary future cosmological observations.
  • Mammola, Stefano; Aharon, Shlomi; Seifan, Merav; Lubin, Yael; Gavish-Regev, Efrat (2019)
    Caves are excellent model systems to study the effects of abiotic factors on species distributions due to their selective conditions. Different ecological factors have been shown to affect species distribution depending on the scale of analysis, whether regional or local. The interplay between local and regional factors in explaining the spatial distribution of cave-dwelling organisms is poorly understood. Using the troglophilic subterranean spider Artema nephilit (Araneae: Pholcidae) as a model organism, we investigated whether similar environmental predictors drive the species distribution at these two spatial scales. At the local scale, we monitored the abundance of the spiders and measured relevant environmental features in 33 caves along the Jordan Rift Valley. We then extended the analysis to a regional scale, investigating the drivers of the distribution using species distribution models. We found that similar ecological factors determined the distribution at both local and regional scales for A. nephilit. At a local scale, the species was found to preferentially occupy the outermost, illuminated, and warmer sectors of caves. Similarly, mean annual temperature, annual temperature range, and solar radiation were the most important drivers of its regional distribution. By investigating these two spatial scales simultaneously, we showed that it was possible to achieve an in-depth understanding of the environmental conditions that governs subterranean species distribution.
  • Lintusaari, Jarno; Gutmann, Michael U.; Dutta, Ritabrata; Kaski, Samuel; Corander, Jukka (2017)
    Bayesian inference plays an important role in phylogenetics, evolutionary biology, and in many other branches of science. It provides a principled framework for dealing with uncertainty and quantifying how it changes in the light of new evidence. For many complex models and inference problems, however, only approximate quantitative answers are obtainable. Approximate Bayesian computation (ABC) refers to a family of algorithms for approximate inference that makes a minimal set of assumptions by only requiring that sampling from a model is possible. We explain here the fundamentals of ABC, review the classical algorithms, and highlight recent developments.
  • Renk, Janina; Zumalacarregui, Miguel; Montanari, Francesco; Barreira, Alexandre (2017)
    Cosmological models with Galileon gravity are an alternative to the standard ACDM paradigm with testable predictions at the level of its self-accelerating solutions for the expansion history, as well as large-scale structure formation. Here, we place constraints on the full parameter space of these models using data from the cosmic microwave background (CMB) (including lensing), baryonic acoustic oscillations (BAO) and the Integrated Sachs Wolfe (ISW) effect. We pay special attention to the ISW effect for which we use the cross spectra, C-l(Tg), of CMB temperature maps and foreground galaxies from the WISE survey. The sign of C-l(Tg) is set by the time evolution of the lensing potential in the redshift range of the galaxy sample: it is positive if the potential decays (like in ACDM), negative if it deepens. We constrain three subsets of Galileon gravity separately known as the Cubic, Quartic and Quintic Galileons. The cubic Galileon model predicts a negative C-l(Tg) and exhibits a 7.8 sigma tension with the data, which effectively rules it out. For the quartic and quintic models the ISW data also rule out a significant portion of the parameter space but permit regions where the goodness-of-fit is comparable to ACDM. The data prefers a non zero sum of the neutrino masses (Sigma m(v) approximate to 0.5eV) with similar to 5 sigma significance in these models. The best-fitting models have values of Ho consistent with local determinations, thereby avoiding the tension that exists in ACDM. We also identify and discuss a similar to 2 sigma tension that Galileon gravity exhibits with recent BAO measurements. Our analysis shows overall that Galileon cosmologies cannot be ruled out by current data but future lensing, BAO and ISW data hold strong potential to do so.
  • DiLeo, Michelle F.; Husby, Arild; Saastamoinen, Marjo (2018)
    There is now clear evidence that species across a broad range of taxa harbor extensive heritable variation in dispersal. While studies suggest that this variation can facilitate demographic outcomes such as range expansion and invasions, few have considered the consequences of intraspecific variation in dispersal for the maintenance and distribution of genetic variation across fragmented landscapes. Here, we examine how landscape characteristics and individual variation in dispersal combine to predict genetic structure using genomic and spatial data from the Glanville fritillary butterfly. We used linear and latent factor mixed models to identify the landscape features that best predict spatial sorting of alleles in the dispersal-related gene phosphoglucose isomerase (Pgi). We next used structural equation modeling to test if variation in Pgi mediated gene flow as measured by F-st at putatively neutral loci. In a year when the population was recovering following a large decline, individuals with a genotype associated with greater dispersal ability were found at significantly higher frequencies in populations isolated by water and forest, and these populations showed lower levels of genetic differentiation at neutral loci. These relationships disappeared in the next year when metapopulation density was high, suggesting that the effects of individual variation are context dependent. Together our results highlight that (1) more complex aspects of landscape structure beyond just the configuration of habitat can be important for maintaining spatial variation in dispersal traits and (2) that individual variation in dispersal plays a key role in maintaining genetic variation across fragmented landscapes.
  • Cardona, Wilmar; Durrer, Ruth; Kunz, Martin; Montanari, Francesco (2016)
    We demonstrate the importance of including the lensing contribution in galaxy clustering analyses with large galaxy redshift surveys. It is well known that radial cross-correlations between different redshift bins of galaxy surveys are dominated by lensing. But we show here that also neglecting lensing in the autocorrelations within one bin severely biases cosmological parameter estimation with redshift surveys. It leads to significant shifts for several cosmological parameters, most notably the scalar spectral index and the neutrino mass scale. Especially the latter parameter is one of the main targets of future galaxy surveys.
  • Jike, Wuhe; Li, Mingai; Zadra, Nicola; Barbaro, Enrico; Sablok, Gaurav; Bertorelle, Giorgio; Rota-Stabelli, Omar; Varotto, Claudio (2020)
    Polyploidization is a frequent phenomenon in plants, which entails the increase from one generation to the next by multiples of the haploid number of chromosomes. While tetraploidization is arguably the most common and stable outcome of polyploidization, over evolutionary time triploids often constitute only a transient phase, or a "triploid bridge", between diploid and tetraploid levels. In this study, we reconstructed in a robust phylogenomic and statistical framework the evolutionary history of polyploidization inArundo, a small genus from the Poaceae family with promising biomass, bioenergy and phytoremediation species. Through the obtainment of 10 novel leaf transcriptomes forArundoand outgroup species, our results prove that recurrent demiduplication has likely been a major driver of evolution in this species-poor genus. Molecular dating further demonstrates that the species originating by demiduplication stalled in the "triploid bridge" for evolutionary times in the order of millions of years without undergoing tetratploidization. Nevertheless, we found signatures of molecular evolution highlighting some of the processes that accompanied the genus radiation. Our results clarify the complex nature ofArundoevolution and are valuable for future gene functional validation as well as reverse and comparative genomics efforts in theArundogenus and other Arundinoideae.
  • Zou, Yuan; Roos, Teemu (Springer International Publishing AG, 2016)
    Lecture Notes in Computer Science
    Modeling interactions in regression models poses both computational as well as statistical challenges: the computational resources and the amount of data required to solve them increases sharply with the size of the problem. We focus on logistic regression with categorical variables and propose a method for learning dependencies that are ex- pressed as general Boolean formulas. The computational and statistical challenges are solved by applying a technique called transformed Lasso, which involves a matrix transformation of the original covariates. We compare the method to an earlier related method, LogicReg, and show that our method scales better in terms of the number of covariates as well as the order and complexity of the interactions.
  • Dengel, S.; Zona, D.; Sachs, T.; Aurela, M.; Jammet, M.; Parmentier, F. J. W.; Oechel, W.; Vesala, T. (2013)
  • Öst, Markus; Lindén, Andreas; Karell, Patrik; Ramula, Satu; Kilpi, Mikael (2018)
    Intermittent breeding may be adaptive for long-lived species subjected to large accessory reproductive costs, but it may also reflect reduced adaptation to the environment, reducing population growth. Nevertheless, environmental influences on breeding propensity, particularly that of predation risk, remain poorly understood and difficult to study, because non-breeders are typically not identified. Female eiders Somateria mollissima from the Baltic Sea provide an excellent testbed, because nesting females have been exposed to intensifying predation and growing male bias that may increase female harassment. We based our study on long-term data (14 years) on females captured and marked at the nest, and females individually identified at sea irrespective of capture status. We hypothesized that breeding propensity decreases with increasing predation risk and male bias, and increases with breeder age. Consistent with our hypotheses, females nesting on islands with higher nest predation risk were more likely to skip breeding, and breeding probability increased with age. In contrast, the steep temporal decline in breeding propensity could not be reliably attributed to annual adult sex ratio or to the abundance of white-tailed sea eagles (Haliaeetus albicilla), the main predator on females, at the nearby Hanko Bird Observatory. Breeding probability showed significant consistent individual variation. Our results demonstrate that spatiotemporal variation in predation risk affects the decision to breed and high incidence of non-breeding was associated with low fledging success. The increased frequency of intermittent breeding in this declining population should be explicitly considered in demographic models, and emphasis placed on understanding the preconditions for successful reproduction.