Browsing by Subject "MODELS"

Sort by: Order: Results:

Now showing items 1-20 of 230
  • Vikke, Heidi Storm; Vittinghus, Svend; Betzer, Martin; Giebner, Matthias; Kolmos, Hans Jorn; Smith, Karen; Castren, Maaret; Lindström, Veronica; Mäkinen, Marja; Harve, Heini; Mogensen, Christian Backer (2019)
    BackgroundHand hygiene (HH), a cornerstone in infection prevention and control, lacks quality in emergency medical services (EMS). HH improvement includes both individual and institutional aspects, but little is known about EMS providers' HH perception and motivations related to HH quality. Therefore, we aimed to investigate the HH perception and assess potential factors related to self-reported HH compliance among the EMS cohort.MethodsA cross-sectional, self-administered questionnaire consisting of 24 items (developed from the WHOs Perception Survey for Health-Care Workers) provided information on demographics, HH perceptions and self-reported HH compliance among EMS providers from Denmark.ResultsOverall, 457 questionnaires were answered (response rate 52%). Most respondents were advanced-care providers, males, had >5years of experience, and had received HH training
  • Sahlin, Ullrika; Helle, Inari; Perepolkin, Dmytro (2021)
    Failing to communicate current knowledge limitations, that is, epistemic uncertainty, in environmental risk assessment (ERA) may have severe consequences for decision making. Bayesian networks (BNs) have gained popularity in ERA, primarily because they can combine variables from different models and integrate data and expert judgment. This paper highlights potential gaps in the treatment of uncertainty when using BNs for ERA and proposes a consistent framework (and a set of methods) for treating epistemic uncertainty to help close these gaps. The proposed framework describes the treatment of epistemic uncertainty about the model structure, parameters, expert judgment, data, management scenarios, and the assessment's output. We identify issues related to the differentiation between aleatory and epistemic uncertainty and the importance of communicating both uncertainties associated with the assessment predictions (direct uncertainty) and the strength of knowledge supporting the assessment (indirect uncertainty). Probabilities, intervals, or scenarios are expressions of direct epistemic uncertainty. The type of BN determines the treatment of parameter uncertainty: epistemic, aleatory, or predictive. Epistemic BNs are useful for probabilistic reasoning about states of the world in light of evidence. Aleatory BNs are the most relevant for ERA, but they are not sufficient to treat epistemic uncertainty alone because they do not explicitly express parameter uncertainty. For uncertainty analysis, we recommend embedding an aleatory BN into a model for parameter uncertainty. Bayesian networks do not contain information about uncertainty in the model structure, which requires several models. Statistical models (e.g., hierarchical modeling outside the BNs) are required to consider uncertainties and variability associated with data. We highlight the importance of being open about things one does not know and carefully choosing a method to precisely communicate both direct and indirect uncertainty in ERA. Integr Environ Assess Manag 2020;00:1-12. (c) 2020 The Authors. Integrated Environmental Assessment and Management published by Wiley Periodicals LLC on behalf of Society of Environmental Toxicology & Chemistry (SETAC)
  • Fülöp, Ludovic; Jussila, Vilho; Aapasuo, Riina Maria; Vuorinen, Tommi Antton Tapani; Mäntyniemi, Päivi Birgitta (2020)
    We propose a ground-motion prediction equation (GMPE) for probabilistic seismic hazard analysis of nuclear installations in Finland. We collected and archived the acceleration recordings of 77 earthquakes from seismic stations on very hard rock (VHR, i.e., the shear-wave velocity in the upper 30 m of the geological profile = 2800 m/s according to the definition used in the nuclear industry) in Finland and Sweden since 2006 and computed the corresponding response spectra important for engineering evaluation. We augmented the narrow magnitude range of the local data by a subset of VHR recordings of 33 earthquakes from the Next Generation Attenuation for Central and Eastern North America (CENA) (NGA-East) database, mainly from eastern Canada. We adapted the backbone curves of the G16 equation proposed by Graizer (2016) for CENA. After the calibration, we evaluated the accuracy of the median prediction and the random error. We conclude that the GMPE developed can be used for predicting ground motions in Fennoscandia. Because of compatibility with the original G16 backbone curve and comparisons with the NGA-East GMPEs, we estimate that the formulation proposed is valid on VHR over the range of 2
  • Sarviaho, R.; Hakosalo, O.; Tiira, K.; Sulkama, S.; Niskanen, J. E.; Hytonen, M. K.; Sillanpää, M. J.; Lohi, H. (2020)
    The complex phenotypic and genetic nature of anxieties hampers progress in unravelling their molecular etiologies. Dogs present extensive natural variation in fear and anxiety behaviour and could advance the understanding of the molecular background of behaviour due to their unique breeding history and genetic architecture. As dogs live as part of human families under constant care and monitoring, information from their behaviour and experiences are easily available. Here we have studied the genetic background of fearfulness in the Great Dane breed. Dogs were scored and categorised into cases and controls based on the results of the validated owner-completed behavioural survey. A genome-wide association study in a cohort of 124 dogs with and without socialisation as a covariate revealed a genome-wide significant locus on chromosome 11. Whole exome sequencing and whole genome sequencing revealed extensive regions of opposite homozygosity in the same locus on chromosome 11 between the cases and controls with interesting neuronal candidate genes such as MAPK9/JNK2, a known hippocampal regulator of anxiety. Further characterisation of the identified locus will pave the way for molecular understanding of fear in dogs and may provide a natural animal model for human anxieties.
  • Tanhuanpaa, Topi; Saarinen, Ninni; Kankare, Ville; Nurminen, Kimmo; Vastaranta, Mikko; Honkavaara, Eija; Karjalainen, Mika; Yu, Xiaowei; Holopainen, Markus; Hyyppa, Juha (Springer International Publishing AG, 2017)
    Lecture Notes in Geoinformation and Cartography
    During the past decade, airborne laser scanning (ALS) has established its status as the state-of-the-art method for detailed forest mapping and monitoring. Current operational forest inventory widely utilizes ALS-based methods. Recent advances in sensor technology and image processing have enabled the extraction of dense point clouds from digital stereo imagery (DSI). Compared with ALS data, the DSI-based data are cheap and the point cloud densities can easily reach that of ALS. In terms of point density, even the high-altitude DSI-based point clouds can be sufficient for detecting individual tree crowns. However, there are significant differences in the characteristics of ALS and DSI point clouds that likely affect the accuracy of tree detection. In this study, the performance of high-altitude DSI point clouds was compared with low-density ALS in detecting individual trees. The trees were extracted from DSI-and ALS-based canopy height models (CHM) using watershed segmentation. The use of both smoothed and unsmoothed CHMs was tested. The results show that, even though the spatial resolution of the DSI-based CHM was better, in terms of detecting the trees and the accuracy of height estimates, the low-density ALS performed better. However, utilizing DSI with shorter ground sample distance (GSD) and more suitable image matching algorithms would likely enhance the accuracy of DSI-based approach.
  • Molnar, Csaba; Jermyn, Ian H.; Kato, Zoltan; Rahkama, Vesa; Ostling, Paivi; Mikkonen, Piia; Pietiainen, Vilja; Horvath, Peter (2016)
    The identification of fluorescently stained cell nuclei is the basis of cell detection, segmentation, and feature extraction in high content microscopy experiments. The nuclear morphology of single cells is also one of the essential indicators of phenotypic variation. However, the cells used in experiments can lose their contact inhibition, and can therefore pile up on top of each other, making the detection of single cells extremely challenging using current segmentation methods. The model we present here can detect cell nuclei and their morphology even in high-confluency cell cultures with many overlapping cell nuclei. We combine the "gas of near circles" active contour model, which favors circular shapes but allows slight variations around them, with a new data model. This captures a common property of many microscopic imaging techniques: the intensities from superposed nuclei are additive, so that two overlapping nuclei, for example, have a total intensity that is approximately double the intensity of a single nucleus. We demonstrate the power of our method on microscopic images of cells, comparing the results with those obtained from a widely used approach, and with manual image segmentations by experts.
  • Petaja, Liisa; Vaara, Suvi; Liuhanen, Sasu; Suojaranta-Ylinen, Raili; Mildh, Leena; Nisula, Sara; Korhonen, Anna-Maija; Kaukonen, Kirsi-Maija; Salmenpera, Markku; Pettila, Ville (2017)
    Objectives: Acute kidney injury (AKI) occurs frequently after cardiac surgery and is associated with increased mortality. The Kidney Disease: Improving Global Outcomes (KDIGO) criteria for diagnosing AKI include creatinine and urine output values. However, the value of the latter is debated. The authors aimed to evaluate the incidence of AKI after cardiac surgery and the independent association of KDIGO criteria, especially the urine output criterion, and 2.5-year mortality. Design: Prospective, observational, cohort study. Setting: Single-center study in a university hospital. Participants: The study comprised 638 cardiac surgical patients from September 1, 2011, to June 20, 2012. Interventions: None. Measurements and Main Results: Hourly urine output, daily plasma creatinine, risk factors for AKI, and variables for EuroSCORE II were recorded. AKI occurred in 183 (28.7%) patients. Patients with AKI diagnosed using only urine output had higher 2.5-year mortality than did patients without AKI (9/53 [17.0%] v 23/455 [5.1%], p = 0.001). AKI was associated with mortality (hazard ratios [95% confidence intervals]: 3.3 [1.8-6.1] for KDIGO I; 5.8 [2.7-12.1] for KDIGO 2; and 7.9 [3.5-17.6]) for KDIGO 3. KDIGO stages and AKI diagnosed using urine output were associated with mortality even after adjusting for mortality risk assessed using EuroSCORE II and risk factors for AKI. Conclusions: AKI diagnosed using only the urine output criterion without fulfilling the creatinine criterion and all stages of AKI were associated with long-term mortality. Preoperatively assessed mortality risk using EuroSCORE II did not predict this AKI-associated mortality. (C) 2017 Elsevier Inc. All rights reserved.
  • Arjas, Elja; Gasbarra, Dario (2022)
    Background: Adaptive designs offer added flexibility in the execution of clinical trials, including the possibilities of allocating more patients to the treatments that turned out more successful, and early stopping due to either declared success or futility. Commonly applied adaptive designs, such as group sequential methods, are based on the frequentist paradigm and on ideas from statistical significance testing. Interim checks during the trial will have the effect of inflating the Type 1 error rate, or, if this rate is controlled and kept fixed, lowering the power. Results: The purpose of the paper is to demonstrate the usefulness of the Bayesian approach in the design and in the actual running of randomized clinical trials during phase II and III. This approach is based on comparing the performance of the different treatment arms in terms of the respective joint posterior probabilities evaluated sequentially from the accruing outcome data, and then taking a control action if such posterior probabilities fall below a pre-specified critical threshold value. Two types of actions are considered: treatment allocation, putting on hold at least temporarily further accrual of patients to a treatment arm, and treatment selection, removing an arm from the trial permanently. The main development in the paper is in terms of binary outcomes, but extensions for handling time-to-event data, including data from vaccine trials, are also discussed. The performance of the proposed methodology is tested in extensive simulation experiments, with numerical results and graphical illustrations documented in a Supplement to the main text. As a companion to this paper, an implementation of the methods is provided in the form of a freely available R package 'barts'. Conclusion: The proposed methods for trial design provide an attractive alternative to their frequentist counterparts.
  • Ritella, Giuseppe; Loperfido, Fedela Feldia; De Giglio, Gianfranco; Scurani, Antonietta; Ligorio, Maria Beatrice (2022)
    This article explores how the adoption of educational robotics, cloud-based animation software, and simplified visual programming software can provide valuable opportunities for dialogic interaction and learning. The potentialities of this type of activity are often overlooked in dialogic investigations. Based on empirical illustration, we discuss how open-ended educational tasks involving the creation of material-digital artifacts can promote the expression of the students' voices and the emergence of a dialogic space in which both human and non-human Others, as well as chronotropic dynamics and materiality, play a crucial role. To provide a polyphonic account of the dialogical processes detected, we analyzed excerpts from two group interviews with seven lower secondary school students (aged 11-12) and excerpts taken from meetings with their teacher. Our qualitative analysis shows that the technology-mediated activity provided valuable opportunities for opening a dialogic space in which the students could express their own voice in interaction with both human and non-human Others. The material world (including the virtual materiality of computer-generated objects) seems to play a twofold role. First, the resistance of the virtual and material objects can contribute to the opening of a dialogical space between the child and the world; second, the chronotopic relations seem to have an impact on the dialogic learning process. These are valid opportunities for educationally relevant dialogic interaction. They should be cultivated and supported to further advance the pedagogical value of educational robotics and coding.
  • Smit, Peter; Virpioja, Sami; Kurimo, Mikko (2021)
    We describe a novel way to implement subword language models in speech recognition systems based on weighted finite state transducers, hidden Markov models, and deep neural networks. The acoustic models are built on graphemes in a way that no pronunciation dictionaries are needed, and they can be used together with any type of subword language model, including character models. The advantages of short subword units are good lexical coverage, reduced data sparsity, and avoiding vocabulary mismatches in adaptation. Moreover, constructing neural network language models (NNLMs) is more practical, because the input and output layers are small. We also propose methods for combining the benefits of different types of language model units by reconstructing and combining the recognition lattices. We present an extensive evaluation of various subword units on speech datasets of four languages: Finnish, Swedish, Arabic, and English. The results show that the benefits of short subwords are even more consistent with NNLMs than with traditional n-gram language models. Combination across different acoustic models and language models with various units improve the results further. For all the four datasets we obtain the best results published so far. Our approach performs well even for English, where the phoneme-based acoustic models and word-based language models typically dominate: The phoneme-based baseline performance can be reached and improved by 4% using graphemes only when several grapheme-based models are combined. Furthermore, combining both grapheme and phoneme models yields the state-of-the-art error rate of 15.9% for the MGB 2018 dev17b test. For all four languages we also show that the language models perform reasonably well when only limited training data is available.
  • Villa, Ana; Eckersten, H.; Gaiser, Thomas; Ahrends, Hella Ellen; Lewan, E. (2022)
    Predicting areas of severe biomass loss and increased N leaching risk under climate change is critical for applying appropriate adaptation measures to support more sustainable agricultural systems. The frequency of annual severe biomass loss for winter wheat and its coincidence with an increase in N leaching in a temperate region in Germany was estimated including the error from using soil and climate input data at coarser spatial scales, using the soil-crop model CoupModel. We ran the model for a reference period (1980–2010) and used climate data predicted by four climate model(s) for the Representative Concentration Pathways (RCP) 2.6, 4.5 and 8.5. The annual median biomass estimations showed that for the period 2070–2100, under the RCP8.5 scenario, the entire region would suffer from severe biomass loss almost every year. Annual incidence of severe biomass loss and increased N leaching was predicted to increase from RCP4.5 to the 8.5 scenario. During 2070–2100 for RCP8.5, in more than half of the years an area of 95% of the region was projected to suffer from both severe biomass loss and increased N leaching. The SPEI3 predicted a range of 32 (P3 RCP4.5) to 55% (P3 RCP8.5) of the severe biomass loss episodes simulated in the climate change scenarios. The simulations predicted more severe biomass losses than by the SPEI index which indicates that soil water deficits are important in determining crop losses in future climate scenarios. There was a risk of overestimating the area where “no severe biomass loss + increased N leaching” occurred when using coarser aggregated input data. In contrast, underestimation of situations where “severe biomass loss + increased N leaching” occurred when using coarser aggregated input data. Larger annual differences in biomass estimations compared to the finest resolution of input data occurred when aggregating climate input data rather than soil data. The differences were even larger when aggregating both soil and climate input data. In half of the region, biomass could be erroneously estimated in a single year by more than 40% if using soil and climate coarser input data. The results suggest that a higher spatial resolution of especially climate input data would be needed to predict reliably annual estimates of severe biomass loss and N leaching under climate change scenarios.
  • Vastaranta, Mikko; Yrttimaa, Tuomas; Saarinen, Ninni; Yu, Xiaowei; Karjalainen, Mika; Nurminen, Kimmo; Karila, Kirsi; Kankare, Ville; Luoma, Ville; Pyörälä, Jiri; Junttila, Samuli; Tanhuanpaa, Topi; Kaartinen, Harri; Kukko, Antero; Honkavaara, Eija; Jaakkola, Anttoni; Liang, Xinlian; Wang, Yunsheng; Vaaja, Matti; Hyyppä, Hannu; Katoh, Masato; Wulder, Michael A.; Holopainen, Markus; Hyyppä, Juha (2018)
    The objective of this study is to better understand the relationship between forest structure and point cloud features generated from certain airborne and space borne sensors. Point cloud features derived from airborne laser scanning (ALS), aerial imagery (AI), WorldView-2 imagery (WV2), TerraSAR-X, and Tandem-X (TDX) data were classified as features characterizing forest height and density as well as variation in tree height. Correlations between these features and field-measured attributes describing forest height, density and tree height variation were investigated at plot scale. From the field-measured attributes, basal area (G) and the number of trees per unit area (N) were used as forest density indicators whereas maximum tree height (H-max) and standard deviation in tree height (H-std) were used as indicators for forest height and tree height variation, respectively. In the analyses, field observations from 91 sample plots (32 m x 32 m) located in southern Finland were used. Even though ALS was found to be the most accurate data source in characterizing forest structure, AI, WV2, and TDX were also capable of characterizing forest height at plot scale with correlation coefficients stronger than 0.85. However, ALS was the only data source capable of providing separate features for characterizing also the variation in tree height and forest density. Features related to forest height, generated from the other data sources besides ALS, also provided strongest correlation with the forest density attributes and variation in tree height, in addition to H-max. Due to these more diverse characterization capabilities, forest structural attributes can be predicted more accurately by using ALS, also in the areas where the relation between the attributes of interest is not solely dependent on forest height, compared to the other investigated 3D remote sensing data sources.
  • Heinaro, Einari; Tanhuanpaa, Topi; Yrttimaa, Tuomas; Holopainen, Markus; Vastaranta, Mikko (2021)
    Fallen trees decompose on the forest floor and create habitats for many species. Thus, mapping fallen trees allows identifying the most valuable areas regarding biodiversity, especially in boreal forests, enabling well-focused conservation and restoration actions. Airborne laser scanning (ALS) is capable of characterizing forests and the underlying topography. However, its potential for detecting and characterizing fallen trees under varying boreal forest conditions is not yet well understood. ALS-based fallen tree detection methods could improve our understanding regarding the spatiotemporal characteristics of dead wood over large landscapes. We developed and tested an automatic method for mapping individual fallen trees from an ALS point cloud with a point density of 15 points/m2. The presented method detects fallen trees using iterative Hough line detection and delineates the trees around the detected lines using region growing. Furthermore, we conducted a detailed evaluation of how the performance of ALS-based fallen tree detection is impacted by characteristics of fallen trees and the structure of vegetation around them. The results of this study showed that large fallen trees can be detected with a high accuracy in old-growth forests. In contrast, the detection of fallen trees in young managed stands proved challenging. The presented method was able to detect 78% of the largest fallen trees (diameter at breast height, DBH > 300 mm), whereas 30% of all trees with a DBH over 100 mm were detected. The performance of the detection method was positively correlated with both the size of fallen trees and the size of living trees surrounding them. In contrast, the performance was negatively correlated with the amount of undergrowth, ground vegetation, and the state of decay of fallen trees. Especially undergrowth and ground vegetation impacted the performance negatively, as they covered some of the fallen trees and lead to false fallen tree detections. Based on the results of this study, ALS-based collection of fallen tree information should be focused on old-growth forests and mature managed forests, at least with the current operative point densities.
  • Mattsson, Markus; Hailikari, Telle; Parpala, Anna (2020)
    Quantitative research into the nature of academic emotions has thus far been dominated by factor analyses of questionnaire data. Recently, psychometric network analysis has arisen as an alternative method of conceptualizing the composition of psychological phenomena such as emotions: while factor models view emotions as underlying causes of affects, cognitions and behavior, in network models psychological phenomena are viewed as arising from the interactions of their component parts. We argue that the network perspective is of interest to studies of academic emotions due to its compatibility with the theoretical assumptions of the control value theory of academic emotions. In this contribution we assess the structure of a Finnish questionnaire of academic emotions using both network analysis and exploratory factor analysis on cross-sectional data obtained during a single course. The global correlational structure of the network, investigated using the spinglass community detection analysis, differed from the results of the factor analysis mainly in that positive emotions were grouped in one community but loaded on different factors. Local associations between pairs of variables in the network model may arise due to different reasons, such as variable A causing variation in variable B or vice versa, or due to a latent variable affecting both. We view the relationship between feelings of self-efficacy and the other emotions as causal hypotheses, and argue that strengthening the students' self-efficacy may have a beneficial effect on the rest of the emotions they experienced on the course. Other local associations in the network model are argued to arise due to unmodeled latent variables. Future psychometric studies may benefit from combining network models and factor models in researching the structure of academic emotions.
  • Sampaio, Larissa; Ferraz, Dnilson Oliveira; Moreira da Costa, Ana Carolina; Aleixo, Alexandre; Cerqueira, Pablo Vieira; Araripe, Juliana; do Rego, Pericles Sena (2020)
    The present study aimed to confirm the occurrence of a hybridization event between the band-tailed manakin (Pipra fasciicauda) and the crimson-hooded manakin (Pipra aureola), based on the existence of a specimen that presents morphological traits of both taxa. We analyzed 297 taxidermized skins of adult males of the two species, including the potential hybrid. We also analyzed the mitochondrial (ND2, ND3 e COI) and nuclear (FGB-I5, MB-I2 e GAPDH-I3) genes of 12 adult specimens of the two taxa, diagnosed phenotypically, in addition to the potential hybrid. The analyses of the plumage indicated that the potential hybrid has an intermediate pattern of white banding on the tail that is less extensive than that found in Pipra fasciicauda, but that its other phenotypic traits are characteristic of Pipra aureola. The molecular topologies revealed two clades, one that groups P. aureola together with the potential hybrid, and the other that corresponds to P. fasciicauda. These findings allowed us to confirm the occurrence of a process of hybridization and potential introgression through secondary events in the P. aureola lineage.
  • Ovaskainen, Otso; de Camargo, Ulisses Moliterno; Somervuo, Panu (2018)
    Automated audio recording offers a powerful tool for acoustic monitoring schemes of bird, bat, frog and other vocal organisms, but the lack of automated species identification methods has made it difficult to fully utilise such data. We developed Animal Sound Identifier (ASI), a MATLAB software that performs probabilistic classification of species occurrences from field recordings. Unlike most previous approaches, ASI locates training data directly from the field recordings and thus avoids the need of pre-defined reference libraries. We apply ASI to a case study on Amazonian birds, in which we classify the vocalisations of 14 species in 194504 one-minute audio segments using in total two weeks of expert time to construct, parameterise, and validate the classification models. We compare the classification performance of ASI (with training templates extracted automatically from field data) to that of monitoR (with training templates extracted manually from the Xeno-Canto database), the results showing ASI to have substantially higher recall and precision rates.
  • Jaatinen, Kim; Moller, Anders P.; Ost, Markus (2019)
    The direction of predator-mediated selection on brain size is debated. However, the speed and the accuracy of performing a task cannot be simultaneously maximized. Large-brained individuals may be predisposed to accurate but slow decision-making, beneficial under high predation risk, but costly under low risk. This creates the possibility of temporally fluctuating selection on brain size depending on overall predation risk. We test this idea in nesting wild eider females (Somateria mollissima), in which head volume is tightly linked to brain mass (r(2) = 0.73). We determined how female relative head volume relates to survival, and characterized the seasonal timing of predation. Previous work suggests that relatively large-brained and small-brained females make slow versus fast nest-site decisions, respectively, and that predation events occur seasonally earlier when predation is severe. Large-brained, late-breeding females may therefore have higher survival during high-predation years, but lower survival during safe years, assuming that predation disproportionately affects late breeders in such years. Relatively large-headed females outsurvived smaller-headed females during dangerous years, whereas the opposite was true in safer years. Predation events occurred relatively later during safe years. Fluctuations in the direction of survival selection on relative brain size may therefore arise due to brain-size dependent breeding phenology.
  • Trusbak Haumann, Niels; Hansen, Brian; Huotilainen, Minna; Vuust, Peter; Brattico, Elvira (2020)
    Background The accuracy of electroencephalography (EEG) and magnetoencephalography (MEG) in measuring neural evoked responses (ERs) is challenged by overlapping neural sources. This lack of accuracy is a severe limitation to the application of ERs to clinical diagnostics. New method We here introduce a theory of stochastic neuronal spike timing probability densities for describing the large-scale spiking activity in neural assemblies, and a spike density component analysis (SCA) method for isolating specific neural sources. The method is tested in three empirical studies with 564 cases of ERs to auditory stimuli from 94 humans, each measured with 60 EEG electrodes and 306 MEG sensors, and a simulation study with 12,300 ERs. Results The first study showed that neural sources (but not non-encephalic artifacts) in individual averaged MEG/EEG waveforms are modelled accurately with temporal Gaussian probability density functions (median 99.7 %–99.9 % variance explained). The following studies confirmed that SCA can isolate an ER, namely the mismatch negativity (MMN), and that SCA reveals inter-individual variation in MMN amplitude. Finally, SCA reduced errors by suppressing interfering sources in simulated cases. Comparison with existing methods We found that gamma and sine functions fail to adequately describe individual MEG/EEG waveforms. Also, we observed that principal component analysis (PCA) and independent component analysis (ICA) does not consistently suppress interference from overlapping brain activity in neither empirical nor simulated cases. Conclusions These findings suggest that the overlapping neural sources in single-subject or patient data can be more accurately separated by applying SCA in comparison to PCA and ICA.
  • Xu, Yongjun; Liu, Xin; Cao, Xin; Huang, Changping; Liu, Enke; Qian, Sen; Liu, Xingchen; Wu, Yanjun; Dong, Fengliang; Qiu, Cheng-Wei; Qiu, Junjun; Hua, Keqin; Su, Wentao; Wu, Jian; Xu, Huiyu; Han, Yong; Fu, Chenguang; Yin, Zhigang; Liu, Miao; Roepman, Ronald; Dietmann, Sabine; Virta, Marko; Kengara, Fredrick; Zhang, Ze; Zhang, Lifu; Zhao, Taolan; Dai, Ji; Yang, Jialiang; Lan, Liang; Luo, Ming; Liu, Zhaofeng; An, Tao; Zhang, Bin; He, Xiao; Cong, Shan; Liu, Xiaohong; Zhang, Wei; Lewis, James P.; Tiedje, James M.; Wang, Qi; An, Zhulin; Wang, Fei; Zhang, Libo; Huang, Tao; Lu, Chuan; Cai, Zhipeng; Wang, Fang; Zhang, Jiabao (2021)
    Y Artificial intelligence (AI) coupled with promising machine learning (ML) techniques well known from computer science is broadly affecting many aspects of various fields including science and technology, industry, and even our day-to-day life. The ML techniques have been developed to analyze high-throughput data with a view to obtaining useful insights, categorizing, predicting, and making evidence-based decisions in novel ways, which will promote the growth of novel applications and fuel the sustainable booming of AI. This paper undertakes a comprehensive survey on the development and application of AI in different aspects of fundamental sciences, including information science, mathematics, medical science, materials science, geoscience, life science, physics, and chemistry. The challenges that each discipline of science meets, and the potentials of AI techniques to handle these challenges, are discussed in detail. Moreover, we shed light on new research trends entailing the integration of AI into each scientific discipline. The aim of this paper is to provide a broad research guideline on fundamental sciences with potential infusion of AI, to help motivate researchers to deeply understand the state-of-the-art applications of AI-based fundamental sciences, and thereby to help promote the continuous development of these fundamental sciences.