scholarly journals Ensemble Deep Learning on Large, Mixed-Site fMRI Datasets in Autism and Other Tasks

2020 ◽  
Vol 30 (07) ◽  
pp. 2050012
Author(s):  
Matthew Leming ◽  
Juan Manuel Górriz ◽  
John Suckling
Keyword(s):  
Deep Learning ◽  
Autism Spectrum ◽  
Black Box ◽  
Typically Developing ◽  
Learning Models ◽  
Cross Sectional ◽  
Functional Connections ◽  
Independent Variable ◽  
The Right ◽  
And Task

Deep learning models for MRI classification face two recurring problems: they are typically limited by low sample size, and are abstracted by their own complexity (the “black box problem”). In this paper, we train a convolutional neural network (CNN) with the largest multi-source, functional MRI (fMRI) connectomic dataset ever compiled, consisting of 43,858 datapoints. We apply this model to a cross-sectional comparison of autism spectrum disorder (ASD) versus typically developing (TD) controls that has proved difficult to characterize with inferential statistics. To contextualize these findings, we additionally perform classifications of gender and task versus rest. Employing class-balancing to build a training set, we trained [Formula: see text] modified CNNs in an ensemble model to classify fMRI connectivity matrices with overall AUROCs of 0.6774, 0.7680, and 0.9222 for ASD versus TD, gender, and task versus rest, respectively. Additionally, we aim to address the black box problem in this context using two visualization methods. First, class activation maps show which functional connections of the brain our models focus on when performing classification. Second, by analyzing maximal activations of the hidden layers, we were also able to explore how the model organizes a large and mixed-center dataset, finding that it dedicates specific areas of its hidden layers to processing different covariates of data (depending on the independent variable analyzed), and other areas to mix data from different sources. Our study finds that deep learning models that distinguish ASD from TD controls focus broadly on temporal and cerebellar connections, with a particularly high focus on the right caudate nucleus and paracentral sulcus.

scholarly journals Utilization of MCH handbooks and attitudes of pregnant women regarding pregnancy and childbirth complications

2019 ◽  
Vol 27 (2) ◽  
pp. 71
Author(s):  
Dian Irawati ◽  
Agustin Dwi Syalfina
Keyword(s):  
Pregnant Women ◽  
P Value ◽  
Chi Square ◽  
Cross Sectional ◽  
Live Births ◽  
Pregnancy And Childbirth ◽  
Independent Variable ◽  
The Right ◽  
Chi Square Analysis ◽  
The Relationship

Objectives: Maternal Mortality Rate (MMR) is closely related to maternal care. MMR in Indonesia based on the 2015 IDHS is 359 per 100,000 live births. By increasing the utilization of MCH handbooks, MMR would be decline. Therefore, this research investigated the relationship between of the utilization of MCH handbooks and attitudes pregnant women regarding pregnancy and childbirth complications.Materials and Methods: We conducted a cross sectional research with 54 pregnant women during March - June 2018 at the Sooko Health Center, Mojokerto. The independent variable was the utilization of MCH handbooks and the dependent variable was the attitude of pregnan women regarding pregnancy and childbirth complications. Data analysis included descriptive and bivariate analysis.Results: The results showed that 79.6% of respondents used the MCH handbook well. The results of the chi square analysis test showed a p value of 0.027 (<0.05).Conclusion: MCH handbook utilization effected the attitude of pregnant women regarding the complication of pregnancy and childbirth. Pregnant women who read and utilize MCH handbook would be have better alertness about the risk of complication so they would make the right decision for their pregnancy.

scholarly journals The attitudes of typically developing adolescents towards their sibling with autism spectrum disorder

2017 ◽  
Vol 64 (1) ◽  
Author(s):  
Christine Van der Merwe ◽  
Juan Bornman ◽  
Dana Donohue ◽  
Michal Harty
Keyword(s):  
Autism Spectrum Disorder ◽  
Autism Spectrum ◽  
Sibling Relationship ◽  
Spectrum Disorder ◽  
Typically Developing ◽  
Cross Sectional ◽  
Mean Values ◽  
Friedman Analysis ◽  
Children With Asd ◽  
Positive Attitudes

Background: Understanding how the cognitive, emotional and behavioural components of sibling attitudes interact with one another at various stages of a sibling’s lifespan will allow clinicians to provide better support to children with autism spectrum disorder (ASD) and their families. However, no research exists which focusses on describing the attitudes of adolescent siblings of children with ASD within the South African context towards their sibling with an ASD. The primary aim of this study was to investigate how typically developing adolescents recall their past attitudes and describe their present attitudes towards their sibling with an ASD.Methods: Thirty typically developing adolescents who have siblings with ASD were selected to complete the survey instrument, the Lifespan Sibling Relationship Scale, using a cross-sectional design.Results: Results indicate that the measure has internal consistency within this sample. Wilcoxon signed-ranks tests were used to test for significant differences between the mean values for the two self-reported time periods. Friedman analysis of variances (ANOVAs) was used to test for significant differences in the three components of attitudes, namely affect, behaviour and cognition. Results indicate that participants held more positive attitudes towards their siblings with ASD as adolescents compared with when they were younger and that adolescents rated their current emotions towards and beliefs about their sibling with ASD to be more positive than their current interaction experiences.Conclusion: As siblings’ attitudes appear to change over time, clinicians should use a lifespan approach to sibling attitudes when designing and implementing supports for siblings of children with ASD.

scholarly journals Local Post-hoc Explainable Methods for Adversarial Text Attacks

2021 ◽  
Author(s):  
Yidong Chai ◽  
Ruicheng Liang ◽  
Hongyi Zhu ◽  
Sagar Samtani ◽  
Meng Wang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Natural Language Processing ◽  
Language Processing ◽  
Black Box ◽  
Learning Models ◽  
Two Phase ◽  
Sensitivity Estimation ◽  
Execution Phase ◽  
Adversarial Examples ◽  
Post Hoc

Deep learning models have significantly advanced various natural language processing tasks. However, they are strikingly vulnerable to adversarial text attacks, even in the black-box setting where no model knowledge is accessible to hackers. Such attacks are conducted with a two-phase framework: 1) a sensitivity estimation phase to evaluate each element’s sensitivity to the target model’s prediction, and 2) a perturbation execution phase to craft the adversarial examples based on estimated element sensitivity. This study explored the connections between the local post-hoc explainable methods for deep learning and black-box adversarial text attacks and proposed a novel eXplanation-based method for crafting Adversarial Text Attacks (XATA). XATA leverages local post-hoc explainable methods (e.g., LIME or SHAP) to measure input elements’ sensitivity and adopts the word replacement perturbation strategy to craft adversarial examples. We evaluated the attack performance of the proposed XATA on three commonly used text-based datasets: IMDB Movie Review, Yelp Reviews-Polarity, and Amazon Reviews-Polarity. The proposed XATA outperformed existing baselines in various target models, including LSTM, GRU, CNN, and BERT. Moreover, we found that improved local post-hoc explainable methods (e.g., SHAP) lead to more effective adversarial attacks. These findings showed that when researchers constantly advance the explainability of deep learning models with local post-hoc methods, they also provide hackers with weapons to craft more targeted and dangerous adversarial attacks.

scholarly journals Explainable Artificial Intelligence (xAI) Approaches and Deep Meta-Learning Models

2020 ◽  
Author(s):  
Evren Dağlarli
Keyword(s):  
Artificial Intelligence ◽  
Deep Learning ◽  
Black Box ◽  
Learning Models ◽  
Learning Methods ◽  
Data Set ◽  
Box Models ◽  
Explainable Artificial Intelligence ◽  
Artificial Neural ◽  
Black Box Models

The explainable artificial intelligence (xAI) is one of the interesting issues that has emerged recently. Many researchers are trying to deal with the subject with different dimensions and interesting results that have come out. However, we are still at the beginning of the way to understand these types of models. The forthcoming years are expected to be years in which the openness of deep learning models is discussed. In classical artificial intelligence approaches, we frequently encounter deep learning methods available today. These deep learning methods can yield highly effective results according to the data set size, data set quality, the methods used in feature extraction, the hyper parameter set used in deep learning models, the activation functions, and the optimization algorithms. However, there are important shortcomings that current deep learning models are currently inadequate. These artificial neural network-based models are black box models that generalize the data transmitted to it and learn from the data. Therefore, the relational link between input and output is not observable. This is an important open point in artificial neural networks and deep learning models. For these reasons, it is necessary to make serious efforts on the explainability and interpretability of black box models.

A Multidimensional Data Fusion Model Based on Deep Learning for a Patient Similarity Network (Preprint)

2021 ◽  
Author(s):  
Mohamed Adel Serhani ◽  
Hadeel T. El Kassabi ◽  
Hadeel T. El Kassabi ◽  
Aberrahim Oulahaj ◽  
Khaled Khalil
Keyword(s):  
Deep Learning ◽  
Longitudinal Data ◽  
Data Model ◽  
Multidimensional Data ◽  
Learning Models ◽  
Fusion Model ◽  
Dynamic Data ◽  
Treatment Pathways ◽  
Similarity Network ◽  
The Right

BACKGROUND Precision medicine is a novel approach for patient care. It allows the prescription of the appropriate drug as well as suitable treatments to the right patient at the right time. It can be envisioned as the comparison of a new patient with existing patients having similar characteristics, which can be referred to as patient similarity. Several statistical, data mining, and deep learning models have been used to build and apply patient similarity network (PSN) for various purposes. However, the challenges associated with data heterogeneity and dimensionality make it difficult to use a single model that addresses both the challenges of reducing data dimensionality and capturing features of diverse data types, including contextual and longitudinal data. Furthermore, when applying multiple models, we can observe the additional challenges associated with the development of an optimum aggregation scheme that maintains high accuracy and preserves data veracity. OBJECTIVE In this study, we propose a multi-model PSN that considers heterogeneous data with static and dynamic characteristics for disease diagnosis for improving prediction accuracy. The static data model manages the data obtained from patient profiles, whereas the dynamic data model manages longitudinal data from patient treatment pathways and clinical data. METHODS We propose a combination of deep learning models and patient similarity network to obtain abundant clinical evidence and extract relevant information based on which similar patients can be explored and compared, thereby obtaining more accurate and comprehensive diagnosis and recommendations. We use the bidirectional encoder representations from transformers (BERT) to process and analyze the contextual data and generate word embedding, where semantic features are captured using a CNN. Dynamic data is analyzed using a long–short-term memory (LSTM)-based autoencoder, which reduces data dimensionality while preserving the temporal features of the data. Furthermore, we propose an aggregation-based fusion approach in which temporal data and clinical narrative data are combined for estimating the patient similarity. RESULTS We evaluated our proposed method through a series of experiments. The obtained results proved that our proposed deep learning-based PSN fusion model provides higher classification accuracy in determining various patient health outcomes when compared with other traditional classification algorithms. CONCLUSIONS Our multi-model highlights the intensity of the similarity between pairs of patients, thereby realizing precise diagnosis and recommendations for a new patient.

scholarly journals Organisational and task factors influencing teachers’ professional development at work

2016 ◽  
Vol 40 (1) ◽  
pp. 36-55 ◽  
Author(s):  
Arnoud T. Evers ◽  
Béatrice I. J. M. Van der Heijden ◽  
Karel Kreijns
Keyword(s):  
Professional Development ◽  
Social Support ◽  
Learning Climate ◽  
Cross Sectional ◽  
Content Type ◽  
Study Results ◽  
Learning Competencies ◽  
Teachers Professional Development ◽  
The Right ◽  
And Task

Purpose – The purpose of this paper is to investigate organisational (cultural and relational) and task factors which potentially enhance teachers’ professional development at work (TPD at Work). The development of lifelong learning competencies and, consequently, the careers of teachers, has become a permanent issue on the agenda of schools worldwide. The workplace is also increasingly regarded as the place to develop these competencies. Design/methodology/approach – A model incorporating the relationships between organisational and task factors as predictor variables and TPD at Work as the dependent variable, is presented and empirically tested by a quantitative (survey research) method. Findings – The study results indicated that learning climate, social support from one’s immediate supervisor, social support from close colleagues and learning value of the function can act as important job resources for TPD at Work. Work pressure and emotional demands, on the other hand, appeared to act as job demands for TPD at Work, but also have the potential to enhance TPD at Work. Research limitations/implications – The most important limitations of the study were the cross-sectional nature of the study and the use of self-ratings only, which may imply common method bias. Practical implications – To enhance TPD at Work, it is vital for actors inside and outside schools to focus on the right working conditions (as mentioned under findings) in schools, so that teachers can learn from their job. Originality/value – Knowledge in schools and empirical research about which factors at the organisational and task level are important to enhance TPD at Work seems scarce. This research contributes to this knowledge gap.

scholarly journals White matter microstructural and morphometric alterations in autism: Implications for intellectual and verbal capabilities

2021 ◽  
Author(s):  
Chun-Hung Yeh ◽  
Rung-Yu Tseng ◽  
Hsing-Chang Ni ◽  
Luca Cocchi ◽  
Jung-Chi Chang ◽  
...  
Keyword(s):  
White Matter ◽  
Corpus Callosum ◽  
Dorsolateral Prefrontal Cortex ◽  
Autism Spectrum ◽  
Typically Developing ◽  
Diffusion Weighted Mri ◽  
Fiber Density ◽  
Dorsolateral Prefrontal ◽  
Crus I ◽  
The Right

ABSTRACTBackgroundNeuroimage literature of autism spectrum disorder (ASD) has a moderate-to-high risk of bias, partially because those combined with intellectual impairment (II) and/or minimally verbal (MV) status are generally ignored. We aimed to provide more comprehensive insights into white matter alterations of ASD, inclusive of individuals with II (ASD-II-Only) or MV expression (ASD-MV).MethodsSixty-five participants with ASD (ASD-Whole; 16.6±5.9 years; comprising 34 intellectually able youth, ASD-IA, and 31 intellectually impaired youth, ASD-II, including 24 ASD-II-Only plus 7 ASD-MV) and 38 demographic-matched typically developing controls (TDC; 17.3±5.6 years) were scanned in accelerated diffusion-weighted MRI. Fixel-based analysis was undertaken to investigate the categorical differences in fiber density (FD), fiber cross-section (FC), and a combined index (FDC), and brain-symptom/cognition associations.ResultsASD-Whole had reduced FD in the anterior and posterior corpus callosum and left cerebellum Crus I, and smaller FDC in right cerebellum Crus II, compared to TDC. ASD-II, relative to TDC, showed almost identical alterations to those from ASD-Whole vs. TDC. ASD-II-Only had greater FD/FDC in the isthmus-splenium of callosum than ASD-MV. Autistic severity negatively correlated with FC in right Crus I. Non-verbal full-scale IQ positively correlated with FC/FDC in cerebellum VI. FD/FDC of the right dorsolateral prefrontal cortex showed a diagnosis-by-executive function interaction.ConclusionsASD-associated white matter alterations appear driven by individuals combined with II and/or MV. Results suggest that changes in the corpus callosum and cerebellum are key for psychopathology and cognition associated with ASD. Our work highlights an essential to include understudied sub-populations on the spectrum in research.

scholarly journals Analysis of Verbal Fluency Output on Semantic Categories of ‘Food’ and ‘Vehicle’ in Typically Developing Malayalam Speaking Children

2018 ◽  
Vol 22 (1) ◽  
pp. 328-353
Author(s):  
John Sunila ◽  
Bellur Rajashekhar ◽  
Vasudeva Guddattu
Keyword(s):  
Verbal Fluency ◽  
Cross Sectional Study ◽  
Positive Influence ◽  
Distinct Pattern ◽  
Developmental Trend ◽  
Typically Developing ◽  
Semantic Fluency ◽  
Typically Developing Children ◽  
Cross Sectional ◽  
And Task

Abstract In the wake of limited knowledge on verbal fluency performance in typically developing children, the present study aims at investigating the semantic fluency performance of Malayalam speaking children across age, gender and tasks. Using a cross-sectional study design, semantic fluency performance (on food and vehicle fluency tasks) was investigated in 1015 Malayalam speaking typically developing children aged 5 to 15 years. The findings revealed the positive influence of age and task with no substantial difference between gender groups, with good inter-rater and intra-rater reliability. The study outcomes depicted a distinct pattern of continuous and linear developmental trend in organizational strategies, with no specific age band showing any dramatic increase in performance. Semantic fluency as a task has great potential within the developmental context for understanding the highly language, culture, and task based word retrieval mechanism.

scholarly journals Performance Evaluation of DLSARS Framework in Intelligent Product Recommendation Systems

2020 ◽  
pp. 165-175
Keyword(s):  
Performance Evaluation ◽  
Deep Learning ◽  
Recommendation System ◽  
The Other ◽  
Program Model ◽  
Learning Models ◽  
Technical Approach ◽  
Huge Number ◽  
Intelligent Product ◽  
The Right

The recommendation framework is vital tool for efficient E-commerce contacts between customers and retailers. Efficient and friendly contacts to find the right product have a huge effect on the sales results. In the basis of a technical approach, four of the program model guidelines are: collective filtering, content-based and demographic filtering. Collaborative filtering is considered superior to other methods in the list. Of necessity, in terms of fortuity, novelty and precision, it provides advantages. The DLSARS Framework is a deep learning-based sentiment analysis for the DLSARS recommendation system that uses deep learning models for a proposed system. The dataset selected for this research is synthetic dataset which consists of huge number of reviews for every product. The proposed models display superiorities and compare the findings with other existing models. The proposed DLSARS frame with bigram approach is superior to the other domain on the E-commerce domain.

Sign in / Sign up

Export Citation Format

Share Document