Browsing by Subject " classification"

Sort by: Order: Results:

Now showing items 1-20 of 31
  • Leppäniemi, A.; Tolonen, M.; Tarasconi, A.; Segovia-Lohse, H.; Gamberini, E.; Kirkpatrick, A.W.; Ball, C.G.; Parry, N.; Sartelli, M.; Wolbrink, D.; Van Goor, H.; Baiocchi, G.; Ansaloni, L.; Biffl, W.; Coccolini, F.; Di Saverio, S.; Kluger, Y.; Moore, E.; Catena, F. (2019)
    Although most patients with acute pancreatitis have the mild form of the disease, about 20-30% develops a severe form, often associated with single or multiple organ dysfunction requiring intensive care. Identifying the severe form early is one of the major challenges in managing severe acute pancreatitis. Infection of the pancreatic and peripancreatic necrosis occurs in about 20-40% of patients with severe acute pancreatitis, and is associated with worsening organ dysfunctions. While most patients with sterile necrosis can be managed nonoperatively, patients with infected necrosis usually require an intervention that can be percutaneous, endoscopic, or open surgical. These guidelines present evidence-based international consensus statements on the management of severe acute pancreatitis from collaboration of a panel of experts meeting during the World Congress of Emergency Surgery in June 27-30, 2018 in Bertinoro, Italy. The main topics of these guidelines fall under the following topics: Diagnosis, Antibiotic treatment, Management in the Intensive Care Unit, Surgical and operative management, and Open abdomen. © 2019 The Author(s).
  • Mandelblum, Jorge; Fischer, Naomi; Achiron, Asaf; Goldberg, Mordechai; Tuuminen, Raimo; Zunz, Eran; Spierer, Oriel (2020)
    Background: The aim of this study was to evaluate whether a simplified pre-operative nuclear classification score (SPONCS) was valid, both for clinical trials and real-world settings. Methods: Cataract classification was based on posterior nuclear color: 0 (clear), 1 (subcapsular/posterior cataract with clear nucleus), 2 (mild "green nucleus" with plus sign for yellow reflection of the posterior cortex), 3 (medium "yellow nucleus" with plus sign for brown/red posterior cortex reflection), 4 (advanced with 4 being "red/brown nucleus" and 4+ white nucleus), and 5 (hypermature/Morgagnian nucleus). Inter- and intra-observer validity was assessed by 30 Ophthalmologists for 15 cataract cases. The reliability of the cataract grading score in a surgical setting was evaluated. Correlation of nuclear scores was compared with phacoemulsification cumulative dissipated energy (CDE) in 596 patients. Results: Analysis of mean intra-observer Cohen kappa agreement was 0.55 with an inter-observer score of 0.54 for the first assessment and 0.49 for the repeat assessment one week later. When evaluating results by nuclear color alone, there was a substantial agreement for both the intra-observer (0.70) and inter-observer parameters: 0.70 for the first test, and 0.66 on repetition with randomization of the cases after a week. CDE levels were found to be significantly different between all SPONCS score groups (p <0.001), with a lower CDE related to a lower SPONCS score. A strong correlation was found between the SPONCS score and CDE (Spearman ' s rho = 0.8, p <0.001). Conclusion: This method of grading cataract hardness is both simple and repeatable. This system can be easily incorporated in randomized controlled trials to lower bias and confounding effects regarding nuclear density along with application in the clinical setting.
  • Yu, Xiaowei; Litkey, Paula; Hyyppa, Juha; Holopainen, Markus; Vastaranta, Mikko (2014)
  • Kokko, Jan (Helsingin yliopisto, 2019)
    In this thesis we present a new likelihood-free inference method for simulator-based models. A simulator-based model is a stochastic mechanism that specifies how data are generated. Simulator-based models can be as complex as needed, but they must allow exact sampling. One common difficulty with simulator-based models is that learning model parameters from observed data is generally challenging, because the likelihood function is typically intractable. Thus, traditional likelihood-based Bayesian inference is not applicable. Several likelihood-free inference methods have been developed to perform inference when a likelihood function is not available. One popular approach is approximate Bayesian computation (ABC), which relies on the fundamental principle of identifying parameter values for which summary statistics of simulated data are close to those of observed data. However, traditional ABC methods tend have high computational cost. The cost is largely due to the need to repeatedly simulate data sets, and the absence of knowledge of how to specify the discrepancy between the simulated and observed data. We consider speeding up the earlier method likelihood-free inference by ratio estimation (LFIRE) by replacing the computationally intensive grid evaluation with Bayesian optimization. The earlier method is an alternative to ABC that relies on transforming the original likelihood-free inference problem into a classification problem that can be solved using machine learning. This method is able to overcome two traditional difficulties with ABC: it avoids using a threshold value that controls the trade-off between computational and statistical efficiency, and combats the curse of dimensionality by offering an automatic selection of relevant summary statistics when using a large number of candidates. Finally, we measure the computational and statistical efficiency of the new method by applying it to three different real-world time series models with intractable likelihood functions. We demonstrate that the proposed method can reduce the computational cost by some orders of magnitude while the statistical efficiency remains comparable to the earlier method.
  • Agnelli, J. P.; Çöl, A.; Lassas, M.; Murthy, R.; Santacesaria, M.; Siltanen, S. (2020)
    Electrical impedance tomography (EIT) is an emerging non-invasive medical imaging modality. It is based on feeding electrical currents into the patient, measuring the resulting voltages at the skin, and recovering the internal conductivity distribution. The mathematical task of EIT image reconstruction is a nonlinear and ill-posed inverse problem. Therefore any EIT image reconstruction method needs to be regularized, typically resulting in blurred images. One promising application is stroke-EIT, or classification of stroke into either ischemic or hemorrhagic. Ischemic stroke involves a blood clot, preventing blood flow to a part of the brain causing a low-conductivity region. Hemorrhagic stroke means bleeding in the brain causing a high-conductivity region. In both cases the symptoms are identical, so a cost-effective and portable classification device is needed. Typical EIT images are not optimal for stroke-EIT because of blurriness. This paper explores the possibilities of machine learning in improving the classification results. Two paradigms are compared: (a) learning from the EIT data, that is Dirichlet-to-Neumann maps and (b) extracting robust features from data and learning from them. The features of choice are virtual hybrid edge detection (VHED) functions (Greenleaf et al 2018 Anal. PDE 11) that have a geometric interpretation and whose computation from EIT data does not involve calculating a full image of the conductivity. We report the measures of accuracy, sensitivity and specificity of the networks trained with EIT data and VHED functions separately. Computational evidence based on simulated noisy EIT data suggests that the regularized grey-box paradigm (b) leads to significantly better classification results than the black-box paradigm (a).
  • Ma, Liang; Chen, Zehua; Huang, Da Wei; Cisse, Ousmane H.; Rothenburger, Jamie L.; Latinne, Alice; Bishop, Lisa; Blair, Robert; Brenchley, Jason M.; Chabe, Magali; Deng, Xilong; Hirsch, Vanessa; Keesler, Rebekah; Kutty, Geetha; Liu, Yueqin; Margolis, Daniel; Morand, Serge; Pahar, Bapi; Peng, Li; Van Rompay, Koen K. A.; Song, Xiaohong; Song, Jun; Sukura, Antti; Thapar, Sabrina; Wang, Honghui; Weissenbacher-Lang, Christiane; Xu, Jie; Lee, Chao-Hung; Jardine, Claire; Lempicki, Richard A.; Cushion, Melanie T.; Cuomo, Christina A.; Kovacs, Joseph A. (2020)
    Pneumocystis, a major opportunistic pathogen in patients with a broad range of immunodeficiencies, contains abundant surface proteins encoded by a multicopy gene family, termed the major surface glycoprotein (Msg) gene superfamily. This superfamily has been identified in all Pneumocystis species characterized to date, highlighting its important role in Pneumocystis biology. In this report, through a comprehensive and in-depth characterization of 459 msg genes from 7 Pneurnocystis species, we demonstrate, for the first time, the phylogeny and evolution of conserved domains in Msg proteins and provide a detailed description of the classification, unique characteristics, and phylogenetic relatedness of five Msg families. We further describe, for the first time, the relative expression levels of individual msg families in two rodent Pneumocystis species, the substantial variability of the msg repertoires in P. coda from laboratory and wild rats, and the distinct features of the expression site for the classic msg genes in Pneumocystis from 8 mammalian host species. Our analysis suggests multiple functions for this superfamily rather than just conferring antigenic variation to allow immune evasion as previously believed. This study provides a rich source of information that lays the foundation for the continued experimental exploration of the functions of the Msg superfamily in Pneumocystis biology. IMPORTANCE Pneumocystis continues to be a major cause of disease in humans with immunodeficiency, especially those with HIV/AIDS and organ transplants, and is being seen with increasing frequency worldwide in patients treated with immunode-pleting monoclonal antibodies. Annual health care associated with Pneumocystis pneumonia costs similar to$475 million dollars in the United States alone. In addition to causing overt disease in immunodeficient individuals, Pneumocystis can cause subclinical infection or colonization in healthy individuals, which may play an important role in species preservation and disease transmission. Our work sheds new light on the diversity and complexity of the msg superfamily and strongly suggests that the versatility of this superfamily reflects multiple functions, including antigenic variation to allow immune evasion and optimal adaptation to host environmental conditions to promote efficient infection and transmission. These findings are essential to consider in developing new diagnostic and therapeutic strategies.
  • Fleischer, Thomas; Klajic, Jovana; Aure, Miriam Ragle; Louhimo, Riku; Pladsen, Arne V.; Ottestad, Lars; Touleimat, Nizar; Laakso, Marko; Halvorsen, Ann Rita; Alnaes, Grethe I. Grenaker; Riis, Margit L. H.; Helland, Aslaug; Hautaniemi, Sampsa; Lonning, Per Eystein; Naume, Bjorn; Borresen-Dale, Anne-Lise; Tost, Joerg; Kristensen, Vessela N. (2017)
    Breast cancer patients with Luminal A disease generally have a good prognosis, but among this patient group are patients with good prognosis that are currently overtreated with adjuvant chemotherapy, and also patients that have a bad prognosis and should be given more aggressive treatment. There is no available method for subclassification of this patient group. Here we present a DNA methylation signature (SAM40) that segregates Luminal A patients based on prognosis, and identify one good prognosis group and one bad prognosis group. The prognostic impact of SAM40 was validated in four independent patient cohorts. Being able to subdivide the Luminal A patients may give the two-sided benefit of identifying one subgroup that may benefit from a more aggressive treatment than what is given today, and importantly, identifying a subgroup that may benefit from less treatment.
  • Silén, Yasmina; Sipilä, Pyry N; Raevuori, Anu; Mustelin, Linda; Marttunen, Mauri; Kaprio, Jaakko; Keski-Rahkonen, Anna (2020)
    OBJECTIVE: We aimed to assess the lifetime prevalence, 10-year incidence, and peak periods of onset for eating disorders as defined by the Fifth Diagnostic and Statistical Manual of Mental Disorders (DSM-5) among adolescents and young adults born in the 1980s in Finland. METHOD: Virtually all Finnish twins born in 1983-1987 (n = 5,600) were followed prospectively from the age of 12 years. A subsample of participants (n = 1,347) was interviewed using a semi-structured diagnostic interview in their early twenties. RESULTS: The prevalence of lifetime DSM-5 eating disorders was 17.9% for females and 2.4% for males (pooled across genders, 10.5%). The estimated lifetime prevalences for females and males, respectively, were 6.2 and 0.3% for anorexia nervosa (AN), 2.4 and 0.16% for bulimia nervosa (BN), 0.6 and 0.3% for binge-eating disorder (BED), 4.5 and 0.16% for other specified feeding or eating disorder (OSFED), and 4.5 and 1.6% for unspecified feeding or eating disorder (UFED). Among females, the prevalence of OSFED subcategories was as follows: atypical AN 2.1%, purging disorder 1.3%, BED of low frequency/limited duration 0.7%, and BN of low frequency/limited duration 0.4%. The 10-year incidence rate of eating disorders was 1,700 per 100,000 person-years among females (peak age of onset 16-19 years) and 220 per 100,000 person-years among males. DISCUSSION: Eating disorders are a common public health concern among youth and young adults, affecting one in six females and one in 40 males. Adequate screening efforts, prevention, and interventions are urgently needed.
  • Latva-Käyrä, Petri (Helsingfors universitet, 2012)
    The intensity and frequency of insect outbreaks have increased in Finland in the last decades and they are expected to increase even further in the future due to global climate change. In 1998-2001 Finland suffered the most severe insect outbreak ever recorded, over 500,000 hectares. The outbreak was caused by the common pine sawfly (Diprion pini L.). The outbreak has continued in the study area, Palokangas, ever since. To find a good method to monitor this type of outbreaks, the purpose of this study was to examine the efficacy of multitemporal ERS-2 and ENVISAT SAR imagery for estimating Scots pine defoliation. The study area, Palokangas, is located in Ilomantsi district, Eastern-Finland and consists mainly even-aged Scots pine forests on relatively dry soils. Most of the forests in the area are young or middle-aged managed forests. The study material was comprised of multi-temporal ERS-2 and ENVISAT synthetic aperture radar (SAR) data. The images had been taken between the years 2001 and 2008. The field data consisted 16 sample plots which had been measured seven times between the years 2002 and 2009. In addition, eight sample plots were added afterwards to places which were known to have had cuttings during the study period. Three methods were tested to estimate Scots pine defoliation: unsupervised k-means clustering, supervised linear discriminant analysis (LDA) and logistic regression. In addition, it was assessed if harvested areas could be differentiated from the defoliated forest using the same methods. Two different speckle filters were used to determine the effect of filtering on the SAR imagery and subsequent results. The logistic regression performed best, producing a classification accuracy of 81.6% (kappa 0.62) with two classes (no defoliation, >20% defoliation). LDA accuracy was with two classes at best 77.7% (kappa 0.54) and k-means 72.8 (0.46). In general, the largest speckle filter, 5 x 5 image window, performed best. When additional classes were added the accuracy was usually degraded on a step-by-step basis. The results were good, but because of the restrictions in the study they should be confirmed with independent data, before full conclusions can be made that results are reliable. The restrictions include the small size field data and, thus, the problems with accuracy assessment (no separate testing data) as well as the lack of meteorological data from the imaging dates.
  • Cajander, A. K. (Suomen metsätieteellinen seura, 1949)
  • Kyrö, Minna (Helsingfors universitet, 2011)
    FTIR spectroscopy (Fourier transform infrared spectroscopy) is a fast method of analysis. The use of interferometers in Fourier devices enables the scanning of the whole infrared frequency region in a couple of seconds. There is no need to elaborate sample preparation when the FTIR spectrometer is equipped with an ATR accessory and the method is therefore easy to use. ATR accessory facilitates the analysis of various sample types. It is possible to measure infrared spectra from samples which are not suitable for traditional sample preparation methods. The data from FTIR spectroscopy is frequently combined with statistical multivariate analysis techniques. In cluster analysis the data from spectra can be grouped based on similarity. In hierarchical cluster analysis the similarity between objects is determined by calculating the distance between them. Principal component analysis reduces the dimensionality of the data and establishes new uncorrelated principal components. These principal components should preserve most of the variation of the original data. The possible applications of FTIR spectroscopy combined with multivariate analysis have been studied a lot. For example in food industry its feasibility in quality control has been evaluated. The method has also been used for the identification of chemical compositions of essential oils and for the detection of chemotypes in oil plants. In this study the use of the method was evaluated in the classification of hog's fennel extracts. FTIR spectra of extracts from different plant parts of hog's fennel were compared with the measured FTIR spectra of standard substances. The typical absorption bands in the FTIR spectra of standard substances were identified. The wave number regions of the intensive absorption bands in the spectra of furanocoumarins were selected for multivariate analyses. Multivariate analyses were also performed in the fingerprint region of IR spectra, including the wave number region 1785-725 cm-1. The aim was to classify extracts according to the habitat and coumarin concentration of the plants. Grouping according to habitat was detected, which could mainly be explained by coumarin concentrations as indicated by analyses of the wave number regions of the selected absorption bands. In these analyses extracts mainly grouped and differed by their total coumarin concentrations. In analyses of the wave number region 1785-725 cm-1 grouping according to habitat was also detected but this could not be explained by coumarin concentrations. These groupings may have been caused by similar concentrations of other compounds in the samples. Analyses using other wave number regions were also performed, but the results from these experiments did not differ from previous results. Multivariate analyses of second-order derivative spectra in the fingerprint region did not reveal any noticeable changes either. In future studies the method could perhaps be further developed by investigating narrower carefully selected wave number regions of second-order derivative spectra.
  • Jong, Monica; Jonas, Jost B.; Wolffsohn, James S.; Berntsen, David A.; Cho, Pauline; Clarkson-Townsend, Danielle; Flitcroft, Daniel I.; Gifford, Kate L.; Haarman, Annechien E. G.; Pardue, Machelle T.; Richdale, Kathryn; Sankaridurg, Padmaja; Tedja, Milly S.; Wildsoet, Christine F.; Bailey-Wilson, Joan E.; Guggenheim, Jeremy A.; Hammond, Christopher J.; Kaprio, Jaakko; MacGregor, Stuart; Mackey, David A.; Musolf, Anthony M.; Klaver, Caroline C. W.; Verhoeven, Virginie J. M.; Vitart, Veronique; Smith, Earl L. (2021)
    PURPOSE. The International Myopia Institute (IMI) Yearly Digest highlights new research considered to be of importance since the publication of the first series of IMI white papers. METHODS. A literature search was conducted for articles on myopia between 2019 and mid-2020 to inform definitions and classifications, experimental models, genetics, interventions, clinical trials, and clinical management. Conference abstracts from key meetings in the same period were also considered. RESULTS. One thousand articles on myopia have been published between 2019 and mid-2020. Key advances include the use of the definition of premyopia in studies currently under way to test interventions in myopia, new definitions in the field of pathologicmyopia, the role of new pharmacologic treatments in experimental models such as intraocular pressure-lowering latanoprost, a large meta-analysis of refractive error identifying 336 new genetic loci, new clinical interventions such as the defocus incorporated multisegment spectacles and combination therapy with low-dose atropine and orthokeratology (OK), normative standards in refractive error, the ethical dilemma of a placebo control group when myopia control treatments are established, reporting the physical metric of myopia reduction versus a percentage reduction, comparison of the risk of pediatric OK wear with risk of vision impairment in myopia, the justification of preventing myopic and axial length increase versus quality of life, and future vision loss. CONCLUSIONS. Large amounts of research in myopia have been published since the IMI 2019 white papers were released. The yearly digest serves to highlight the latest research and advances in myopia.
  • Huong Thi Thanh Nguyen,; Trung Minh Doan,; Tomppo, Erkki; McRoberts, Ronald E. (2020)
    Information on land use and land cover (LULC) including forest cover is important for the development of strategies for land planning and management. Satellite remotely sensed data of varying resolutions have been an unmatched source of such information that can be used to produce estimates with a greater degree of confidence than traditional inventory estimates. However, use of these data has always been a challenge in tropical regions owing to the complexity of the biophysical environment, clouds, and haze, and atmospheric moisture content, all of which impede accurate LULC classification. We tested a parametric classifier (logistic regression) and three non-parametric machine learning classifiers (improved k-nearest neighbors, random forests, and support vector machine) for classification of multi-temporal Sentinel 2 satellite imagery into LULC categories in Dak Nong province, Vietnam. A total of 446 images, 235 from the year 2017 and 211 from the year 2018, were pre-processed to gain high quality images for mapping LULC in the 6516 km(2) study area. The Sentinel 2 images were tested and classified separately for four temporal periods: (i) dry season, (ii) rainy season, (iii) the entirety of the year 2017, and (iv) the combination of dry and rainy seasons. Eleven different LULC classes were discriminated of which five were forest classes. For each combination of temporal image set and classifier, a confusion matrix was constructed using independent reference data and pixel classifications, and the area on the ground of each class was estimated. For overall temporal periods and classifiers, overall accuracy ranged from 63.9% to 80.3%, and the Kappa coefficient ranged from 0.611 to 0.813. Area estimates for individual classes ranged from 70 km(2) (1% of the study area) to 2200 km(2) (34% of the study area) with greater uncertainties for smaller classes.
  • Parsons, Michael T.; Tudini, Emma; Li, Hongyan; Hahnen, Eric; Wappenschmidt, Barbara; Feliubadaló, Lidia; Aalfs, Cora M.; Agata, Simona; Aittomäki, Kristiina; Alducci, Elisa; Alonso-Cerezo, María Concepción; Arnold, Norbert; Auber, Bernd; Austin, Rachel; Azzollini, Jacopo; Balmaña, Judith; Barbieri, Elena; Bartram, Claus R.; Blanco, Ana; Blümcke, Britta; Bonache, Sandra; Bonanni, Bernardo; Borg, Åke; Bortesi, Beatrice; Brunet, Joan; Bruzzone, Carla; Bucksch, Karolin; Cagnoli, Giulia; Caldés, Trinidad; Caliebe, Almuth; Caligo, Maria A.; Calvello, Mariarosaria; Capone, Gabriele L.; Caputo, Sandrine M.; Carnevali, Ileana; Carrasco, Estela; Caux-Moncoutier, Virginie; Cavalli, Pietro; Cini, Giulia; Clarke, Edward M.; Concolino, Paola; Cops, Elisa J.; Cortesi, Laura; Couch, Fergus J.; Darder, Esther; de la Hoya, Miguel; Dean, Michael; Debatin, Irmgard; del Valle, Jesús; Delnatte, Capucine; Derive, Nicolas; Diez, Orland; Ditsch, Nina; Domchek, Susan M.; Dutrannoy, Véronique; Eccles, Diana M.; Ehrencrona, Hans; Enders, Ute; Evans, D. Gareth; Faust, Ulrike; Felbor, Ute; Feroce, Irene; Fine, Miriam; Galvao, Henrique C.R.; Gambino, Gaetana; Gehrig, Andrea; Gensini, Francesca; Gerdes, Anne-Marie; Germani, Aldo; Giesecke, Jutta; Gismondi, Viviana; Gómez, Carolina; Gómez Garcia, Encarna B.; González, Sara; Grau, Elia; Grill, Sabine; Gross, Eva; Guerrieri-Gonzaga, Aliana; Guillaud-Bataille, Marine; Gutiérrez-Enríquez, Sara; Haaf, Thomas; Hackmann, Karl; Hansen, Thomas V.O.; Harris, Marion; Hauke, Jan; Heinrich, Tilman; Hellebrand, Heide; Herold, Karen N.; Honisch, Ellen; Horvath, Judit; Houdayer, Claude; Hübbel, Verena; Iglesias, Silvia; Izquierdo, Angel; James, Paul A.; Janssen, Linda A.M.; Jeschke, Udo; Kaulfuß, Silke; Keupp, Katharina; Kiechle, Marion; Kölbl, Alexandra; Krieger, Sophie; Kruse, Torben A.; Kvist, Anders; Lalloo, Fiona; Larsen, Mirjam; Lattimore, Vanessa L.; Lautrup, Charlotte; Ledig, Susanne; Leinert, Elena; Lewis, Alexandra L.; Lim, Joanna; Loeffler, Markus; López-Fernández, Adrià; Lucci-Cordisco, Emanuela; Maass, Nicolai; Manoukian, Siranoush; Marabelli, Monica; Matricardi, Laura; Meindl, Alfons; Michelli, Rodrigo D.; Moghadasi, Setareh; Moles-Fernández, Alejandro; Montagna, Marco; Montalban, Gemma; Monteiro, Alvaro N.; Montes, Eva; Mori, Luigi; Moserle, Lidia; Müller, Clemens R.; Mundhenke, Christoph; Naldi, Nadia; Nathanson, Katherine L.; Navarro, Matilde; Nevanlinna, Heli; Nichols, Cassandra B.; Niederacher, Dieter; Nielsen, Henriette R.; Ong, Kai-ren; Pachter, Nicholas; Palmero, Edenir I.; Papi, Laura; Pedersen, Inge Sokilde; Peissel, Bernard; Pérez-Segura, Pedro; Pfeifer, Katharina; Pineda, Marta; Pohl-Rescigno, Esther; Poplawski, Nicola K.; Porfirio, Berardino; Quante, Anne S.; Ramser, Juliane; Reis, Rui M.; Revillion, Françoise; Rhiem, Kerstin; Riboli, Barbara; Ritter, Julia; Rivera, Daniela; Rofes, Paula; Rump, Andreas; Salinas, Monica; Sánchez de Abajo, Ana María; Schmidt, Gunnar; Schoenwiese, Ulrike; Seggewiß, Jochen; Solanes, Ares; Steinemann, Doris; Stiller, Mathias; Stoppa-Lyonnet, Dominique; Sullivan, Kelly J.; Susman, Rachel; Sutter, Christian; Tavtigian, Sean V.; Teo, Soo H.; Teulé, Alex; Thomassen, Mads; Tibiletti, Maria Grazia; Tognazzo, Silvia; Toland, Amanda E.; Tornero, Eva; Törngren, Therese; Torres-Esquius, Sara; Toss, Angela; Trainer, Alison H.; van Asperen, Christi J.; van Mackelenbergh, Marion T.; Varesco, Liliana; Vargas-Parra, Gardenia; Varon, Raymonda; Vega, Ana; Velasco, Ángela; Vesper, Anne-Sophie; Viel, Alessandra; Vreeswijk, Maaike P.G.; Wagner, Sebastian A.; Waha, Anke; Walker, Logan C.; Walters, Rhiannon J.; Wang-Gohrke, Shan; Weber, Bernhard H.F.; Weichert, Wilko; Wieland, Kerstin; Wiesmüller, Lisa; Witzel, Isabell; Wöckel, Achim; Woodward, Emma R.; Zachariae, Silke; Zampiga, Valentina; Zeder-Göß, Christine; Investigators, KConFab; Lázaro, Conxi; De Nicolo, Arcangela; Radice, Paolo; Engel, Christoph; Schmutzler, Rita K.; Goldgar, David E.; Spurdle, Amanda B. (2019)
    Abstract The multifactorial likelihood analysis method has demonstrated utility for quantitative assessment of variant pathogenicity for multiple cancer syndrome genes. Independent data types currently incorporated in the model for assessing BRCA1 and BRCA2 variants include clinically calibrated prior probability of pathogenicity based on variant location and bioinformatic prediction of variant effect, co-segregation, family cancer history profile, co-occurrence with a pathogenic variant in the same gene, breast tumor pathology, and case-control information. Research and clinical data for multifactorial likelihood analysis were collated for 1395 BRCA1/2 predominantly intronic and missense variants, enabling classification based on posterior probability of pathogenicity for 734 variants: 447 variants were classified as (likely) benign, and 94 as (likely) pathogenic; 248 classifications were new or considerably altered relative to ClinVar submissions. Classifications were compared to information not yet included in the likelihood model, and evidence strengths aligned to those recommended for ACMG/AMP classification codes. Altered mRNA splicing or function relative to known non-pathogenic variant controls were moderately to strongly predictive of variant pathogenicity. Variant absence in population datasets provided supporting evidence for variant pathogenicity. These findings have direct relevance for BRCA1 and BRCA2 variant evaluation, and justify the need for gene-specific calibration of evidence types used for variant classification. This article is protected by copyright. All rights reserved.
  • Jaakkola, A (Aalto University, 2015)
    Aalto University publication series DOCTORAL DISSERTATIONS
  • Jaskari, Joel; Myllärinen, Janne; Leskinen, Markus; Rad, Ali Bahrami; Hollmen, Jaakko; Andersson, Sture; Särkkä, Simo (2020)
    Preterm birth is the leading cause of mortality in children under the age of five. In particular, low birth weight and low gestational age are associated with an increased risk of mortality. Preterm birth also increases the risks of several complications, which can increase the risk of death, or cause long-term morbidities with both individual and societal impacts. In this work, we use machine learning for prediction of neonatal mortality as well as neonatal morbidities of bronchopulmonary dysplasia, necrotizing enterocolitis, and retinopathy of prematurity, among very low birth weight infants. Our predictors include time series data and clinical variables collected at the neonatal intensive care unit of Children's Hospital, Helsinki University Hospital. We examine 9 different classifiers and present our main results in AUROC, similar to our previous studies, and in F1-score, which we propose for classifier selection in this study. We also investigate how the predictive performance of the classifiers evolves as the length of time series is increased, and examine the relative importance of different features using the random forest classifier, which we found to generally perform the best in all tasks. Our systematic study also involves different data preprocessing methods which can be used to improve classifier sensitivities. Our best classifier AUROC is 0.922 in the prediction of mortality, 0.899 in the prediction of bronchopulmonary dysplasia, 0.806 in the prediction of necrotizing enterocolitis, and 0.846 in the prediction of retinopathy of prematurity. Our best classifier F1-score is 0.493 in the prediction of mortality, 0.704 in the prediction of bronchopulmonary dysplasia, 0.215 in the prediction of necrotizing enterocolitis, and 0.368 in the prediction of retinopathy of prematurity.
  • Broesby-Olsen, Sigurd; Dybedal, Ingunn; Gulen, Theo; Kristensen, Thomas K.; Moller, Michael B.; Ackermann, Leena; Saaf, Maria; Karlsson, Maria; Agertoft, Lone; Brixen, Kim; Hermann, Pernille; Stylianou, Eva; Mortz, Charlotte G.; Torfing, Trine; Havelund, Troels; Sander, Birgitta; Bergstrom, Anna; Bendix, Marie; Garvey, Lene H.; Weis Bjerrum, Ole; Valent, Peter; Bindslev-Jensen, Carsten; Nilsson, Gunnar; Vestergaard, Hanne; Hagglund, Hans (2016)
    Mastocytosis is a heterogeneous group of diseases defined by an increased number and accumulation of mast cells, and often also by signs and symptoms of mast cell activation. Disease subtypes range from indolent to rare aggressive forms. Mastocytosis affects people of all ages and has been considered rare; however, it is probably underdiagnosed with potential severe implications. Diagnosis can be challenging and symptoms may be complex and involve multiple organ-systems. In general it is advised that patients should be referred to centres with experience in the disease offering an individualized, multidisciplinary approach. We present here consensus recommendations from a Nordic expert group for the diagnosis and general management of patients with mastocytosis.
  • Sohrab, Fahad; Raitoharju, Jenni; Iosifidis, Alexandros; Gabbouj, Moncef (Elsevier, 2021)
    Pattern Recognition 110
    In this paper, we propose a novel method for projecting data from multiple modalities to a new subspace optimized for one-class classification. The proposed method iteratively transforms the data from the original feature space of each modality to a new common feature space along with finding a joint compact description of data coming from all the modalities. For data in each modality, we define a separate transformation to map the data from the corresponding feature space to the new optimized subspace by exploiting the available information from the class of interest only. We also propose different regularization strategies for the proposed method and provide both linear and non-linear formulations. The proposed Multimodal Subspace Support Vector Data Description outperforms all the competing methods using data from a single modality or fusing data from all modalities in four out of five datasets.
  • Muukkonen, Ilkka (Helsingin yliopisto, 2018)
    Objectives: Faces provide an ideal platform to look into the ways in which our brains process multidimensional information. In order to still recognize an individual when their expression changes, our brain must be able to separate two overlapping sources of information. Previous fMRI-studies have found several brain areas involved in face processing, especially fusiform face area (FFA), occipital face area (OFA), and superior temporal sulcus (STS). EEG- and MEG-studies have also pointed out face-specific temporal components, mainly P1, N170, and N250. However, only few studies have varied both expressions and identities at the same time, or combined spatially precise fMRI with temporally precise M/EEG. Methods: In separate experiments, EEG and fMRI were measured while participants (n=17) viewed morphed faces varying in their expression (neutral, happy, fearful and angry) and in identity. Classification accuracies were calculated using support vector machine (SVM), both from different spatial locations in fMRI and from different timepoints in EEG. In addition, the classification information in fMRI and EEG were combined using representational similarity analysis (RSA). Results: In EEG, we found support for very early processing of expressions (at 110 ms), later processing of identities (at 250 ms) than expressions, and more sustained decoding of angry faces than faces with other expressions. In fMRI, coding of expressions were found on a broad area containing early visual areas and face processing areas OFA, FFA, and STS. Results for identities, although less clear, showed FFA and middle frontal gyrus (MFG). RSA combining both EEG and fMRI showed progression of information from early visual areas at 130 ms to FFA at 150 ms, and to FFA and STS at 200 ms. Conclusions: Our results showed that with multivariate data analysis methods, temporal and spatial neural representations of faces can be studied simultaneously. Consistent with neural models of face processing, our results suggest partially separate processing of expressions and identities in spatially distributed brain network.
  • Veilahti, Antti Veikko Petteri; Kovarskis, Levas; Cowley, Benjamin Ultan (2021)
    Neurofeedback for attention deficit/hyperactivity disorder (ADHD) has long been studied as an alternative to medication, promising non-invasive treatment with minimal side-effects and sustained outcome. However, debate continues over the efficacy of neurofeedback, partly because existing evidence for efficacy is mixed and often non-specific, with unclear relationships between prognostic variables, patient performance when learning to self-regulate, and treatment outcomes. We report an extensive analysis on the understudied area of neurofeedback learning. Our data comes from a randomised controlled clinical trial in adults with ADHD (registered trial ISRCTN13915109; N=23; 13:10 female:male; age 25-57). Patients were treated with either theta-beta ratio or sensorimotor-rhythm regimes for 40 one-hour sessions. We classify 11 learners vs 12 non-learners by the significance of random slopes in a linear mixed growth-curve model. We then analyse the predictors, outcomes, and processes of learners vs non-learners, using these groups as mutual controls. Significant predictive relationships were found in anxiety disorder (GAD), dissociative experience (DES), and behavioural inhibition (BIS) scores obtained during screening. Low DES, but high GAD and BIS, predicted positive learning. Patterns of behavioural outcomes from Test Of Variables of Attention, and symptoms from adult ADHD Self-Report Scale, suggested that learning itself is not required for positive outcomes. Finally, the learning process was analysed using structural-equations modelling with continuous-time data, estimating the short-term and sustained impact of each session on learning. A key finding is that our results support the conceptualisation of neurofeedback learning as skill acquisition, and not merely operant conditioning as originally proposed in the literature.