Artificial Intelligence applications in diagnostic medicine: a decade of expectations

2023 | Americas Conference on Information Systems | Citations: 0

Authors: Cabral, Bernardo; Castro, Leonardo; Fornazin, Marcelo; Penteado, Bruno; Silva, Sandro Luís Freire de Castro; Braga, Luiza Amara Maciel; Filho, Carlos Gilbert Conte; Mota, Fabio Batista

Abstract: Considered a new paradigm in healthcare, AI is expected to change diagnostic me ... Expand

Abstract: Considered a new paradigm in healthcare, AI is expected to change diagnostic medicine in the coming decades. However, its future is still uncertain -partly because it is an emerging technology being gradually applied to a mature field of medicine. We conducted a global cross-sectional survey with more than 1,400 authors of recent scientific publications indexed by the Web of Science to foresee the future of AI in diagnostic medicine. Most respondents expect AI to change diagnostic medicine in this decade radically. In this period, the two most likely outcomes are reduced screening costs and increased diagnostic reliability. X-ray diagnosis and heart rhythm interpretation are the two diagnostic tools most likely to be integrated with AI. The two main barriers are the difficulty of incorporating it into clinical practice and the ethical-regulatory issues. Respondents' expectations align with the literature and suggest that AI may substantially change diagnostic medicine within this decade. Collapse

Topics: artificial intelligence electronic mail database system

Methods: survey cross sectional survey cross sectional research literature study magnetic resonance imaging

AI-Assisted Diagnosis of Bone Tuberculosis: A Design Science Research Approach

2023 | International Conference on Information Systems | Citations: 0

Authors: Ding, Wenwen; Hoehle, Hartmut

Abstract: Bone Tuberculosis (TB) is a significant public health challenge requiring early ... Expand

Abstract: Bone Tuberculosis (TB) is a significant public health challenge requiring early and precise diagnosis for effective treatment. Traditional methods like radiography and biopsy are invasive and costly. Our study introduces a holistic AI-assisted orthopedic clinical diagnosis system developed through an Action Design Research approach. Unlike previous efforts focused solely on algorithmic design, our system is iteratively validated with real-world clinical data, ensuring both theoretical rigor and practical applicability. By fine-tuning AI algorithms to meet actual clinical needs, we bridge the gap between technological innovation and healthcare relevance. Our research offers innovative insights into the design and evaluation of AI-assisted systems, emphasizing the role of empirical data and diverse evaluation metrics. The study is expected to have broader implications for the adoption of AI in clinical settings, offering a more comprehensive and reliable solution for bone TB diagnosis. Collapse

Topics: artificial intelligence healthcare data innovation management digital innovation

Methods: computational algorithm action design research machine learning magnetic resonance imaging longitudinal research

Promoting Learning Through Explainable Artificial Intelligence: An Experimental Study in Radiology

2023 | International Conference on Information Systems | Citations: 0

Authors: Ellenrieder, Sara; Kallina, Emma; Pumplun, Luisa; Gawlitza, Joshua Felix; Ziegelmayer, Sebastian; Buxmann, Peter

Abstract: The deployment of machine learning (ML)-based decision support systems (DSSs) i ... Expand

Abstract: The deployment of machine learning (ML)-based decision support systems (DSSs) in high-risk environments such as radiology is increasing. Despite having achieved high decision accuracy, they are prone to errors. Thus, they are primarily used to assist radiologists in their decision making. However, collaborative decision making poses risks to the decision maker, e.g. automation bias and long-term performance degradation. To address these issues, we propose combining findings of the research streams of explainable artificial intelligence and education to promote human learning through interaction with ML-based DSSs. We provided radiologists with explainable vs nonexplainable decision support that was high-vs low-performing in a between-subject experimental study to support manual segmentation of 690 brain tumor scans. Our results show that explainable ML-based DSSs improved human learning outcomes and prevented false learning triggered by incorrect decision support. In fact, radiologists were able to learn from errors made by the low-performing explainable ML-based DSS. Collapse

Topics: decision support system decision support explainable artificial intelligence decision making artificial intelligence

Methods: experiment qualitative interview magnetic resonance imaging parametric test ANOVA

What makes a Tweet so sweet: The neurocognitive effects of hedonic information consumption

2023 | Information & Management | Citations: 0

Authors: Fadel, Kelly J.; Meservy, Thomas O.; Kirwan, C. Brock

Abstract: Hedonic information consumption (HIC) refers to the consumption of information f ... Expand

Abstract: Hedonic information consumption (HIC) refers to the consumption of information for non-task-oriented purposes such as entertainment or pleasure. With the explosive growth of information availability through online channels such as social media, HIC has become a common and growing phenomenon; however, most information systems research to date has focused on factors surrounding information consumption for utilitarian rather than hedonic purposes. In this paper, we take an initial step to inform theory in this area by applying a NeuroIS lens to explore HIC behavior. We postulate that information is more likely to be hedonically consumed when it is appraised as interesting, and that certain characteristics of the information involved in HIC, namely its novelty and valence, play a role in determining its interestingness. Moreover, we hypothesize that HIC is driven by the neurobiological reactions that take place in response to information novelty and valence. We report the results of a behavioral and an fMRI experiment designed to measure the relationship between these information dimensions and neu­ rocognitive responses in the context of a popular social media platform. Our results provide preliminary evidence for differential processing patterns related to novelty and valence that can inform ongoing theory building surrounding HIC behaviors. Collapse

Topics: Twitter social media social media site database system decision making

Methods: experiment functional magnetic resonance imaging theory development NeuroIS magnetic resonance imaging

Theories: theory of information information systems theory

Information Filtering in Electronic Networks of Practice: An fMRI Investigation of Expectation (Dis)confirmation

2022 | Journal of the Association for Information Systems | Citations: 0

Authors: Fadel, Kelly; Meservy, Thomas; Kirwan, C. Brock

Abstract: Online forums sponsored by electronic networks of practice (ENPs) have become a ... Expand

Abstract: Online forums sponsored by electronic networks of practice (ENPs) have become an important platform for technology-mediated knowledge exchange, yet relatively little is known about how ENP participants filter and evaluate the information they encounter on these forums. This study integrates perspectives from expectation confirmation theory, prospect theory, and neuroscience research to explore how ENP forum filtering judgments are influenced when expectations formed on the basis of contextual cues are confirmed or disconfirmed by the examination of solution quality. We summarize six different models of expectation confirmation explored in previous IS literature and report the results of a neuroimaging experiment using functional MRI (fMRI) that paired both positive and negative contextual cues with high-and low-quality solutions on a mock ENP forum interface. Results show that evaluation judgments are strongest in conditions where initial contextual cue judgments are confirmed by examination of solution quality except when the perceived expectation-experience gap is large, providing evidence for an assimilation-contrast model of expectation confirmation. We also found neural activation differences for expectation confirmation vs. disconfirmation and, consistent with prospect theory, differences in filtering behaviors with respect to unexpected gains vs. unexpected losses. has also been presented and recognized at several international information systems conferences.Thomas O. Meservy is an associate professor of information systems at Brigham Young University. He earned his PhD from the University of Arizona in 2007. He has three main research streams including human decision-making in online forums, automated deception detection, and large group collaboration. His research has appeared in Management Information Systems Quarterly, Information Systems Research, Journal of Management Information Systems, and other high-quality outlets and conferences. He enjoys teaching. He also maintains ties to industry doing consulting and training related to software development, testing, and cybersecurity with the Department of Homeland Security, the Department of Defense, and Fortune 100 and Fortune 500 companies.C. Brock Kirwan is a professor of psychology and neuroscience and director of the MRI Research Facility at Brigham Young University. He leads the Memory and Decision Making Lab, which studies memory-dependent decision making. His collaborative research with the NeuroSecurity Lab using neuroscience methods applied to computer security problems has been funded by industry and the National Science Foundation. He has been published in Science, Collapse

Topics: software developer information exchange technology adoption database system Java

Methods: functional magnetic resonance imaging experiment magnetic resonance imaging statistical hypothesis test experimental task

Theories: prospect theory expectation confirmation theory dual-system theory information systems theory

To download or not to download? Spatial and temporal neural dynamics across the brain regions when deciding to download an app

2022 | International Journal of Information Management | Citations: 0

Authors: Parmar, Harshit; Tahvildar, Ala; Ghasemi, Ehsan; Jung, Shinyoung; Davis, Fred; Walden, Eric

Abstract: Technology adoption is one of the most important research streams in the field o ... Expand

Abstract: Technology adoption is one of the most important research streams in the field of IS. However, it has largely been studied with endpoint models rather than process models. In this research, we attempt to develop a process model for technology adoption using fMRI (functional Magnetic Resonance Imaging) techniques. fMRI tech­ niques allow us to directly investigate the cognitive process of technology adoption instead of using mental representations such as perceived usefulness and perceived ease of use. We measure the brain activities of 22 students from a major western US university while they are making the decision to download or not download apps using a novel temporal neural correlate analysis. The research finds that the download decision is a complex process that includes the visual regions of the brain, the anterior cingulate cortex, the middle temporal gyrus and many other regions. The practical implication is that there are many points of intervention to influence the download decision. In particular, the visual features of app presentation seem to be a very important consideration. Collapse

Topics: mobile application decision making technology adoption packaged software perceived ease of use

Methods: functional magnetic resonance imaging longitudinal research time series analysis business process modeling magnetic resonance imaging

Theories: Lewin's field theory

Early Diagnosis of Mild Cognitive Impairment with 2-Dimensional Convolutional Neural Network Classification of Magnetic Resonance Images

2021 | HICSS | Citations: 0

Authors: Angelopoulos, Spyros

Abstract: We motivate and implement an Artificial Intelligence (AI) Computer Aided Diagnos ... Expand

Abstract: We motivate and implement an Artificial Intelligence (AI) Computer Aided Diagnosis (CAD) framework, to assist clinicians in the early diagnosis of Mild Cognitive Impairment (MCI) and Alzheimer’s Disease (AD). Our framework is based on a Convolutional Neural Network (CNN) trained and tested on functional Magnetic Resonance Images datasets. We contribute to the literature on AI-CAD frameworks for AD by using a 2D CNN for early diagnosis of MCI. Contrary to current efforts, we do not attempt to provide an AI-CAD framework that will replace clinicians, but one that can work in synergy with them. Our framework is cheaper and faster as it relies on small datasets without the need of high-performance computing infrastructures. Our work contributes to the literature on digital transformation of healthcare, health Information Systems, and NeuroIS, while it opens novel avenues for further research on the topic. Collapse

Topics: artificial intelligence digital transformation health information system Python cloud computing

Methods: magnetic resonance imaging NeuroIS functional neuroimaging Keras machine learning

Comparative study of CNN models for brain tumor classification: Computational efficiency versus accuracy

2021 | Americas Conference on Information Systems | Citations: 0

Authors: Basyal, Ganga; Zeng, David; Bishop, Dave; Rimal, Bhaskar P.

Abstract: The application of Convolutional Neural Networks (CNNs) for medical image class ... Expand

Abstract: The application of Convolutional Neural Networks (CNNs) for medical image classification and segmentation tasks is well-known and highly utilized in deep learning architecture. Digital transformation of healthcare services and technology has presented a huge opportunity for new technologies to be investigated. However, analysis of these CNN models for task-specific selection presents an opportunity for additional research. Further, given the plethora of cloud computing-based services, the computational cost has become a crucial factor for model selection. In this paper, we compare the state-of-the-art CNN models in terms of accuracy and cost. A two-stage adaptive transfer learning model framework is designed based on Design Science principles. Our experimental results show that the ResNet50 CNN model has performed well and yielded 80.26% validation accuracy and 47.13% validation loss. The framework could be useful as a decision support tool for medical professionals in medical image classification. KeywordsConvolutional neural networks, design science, transfer learning, medical image classification. Comparative Study of CNN Models for Brain Tumor Classification Collapse

Topics: health information system decision support system digital transformation problem solving

Methods: design artifact residual neural network design science ImageNet literature study

High-Performance Detection of Mild Cognitive Impairment Using Resting-State EEG Signals Located in Broca’s Area: A Machine Learning Approach

2021 | Americas Conference on Information Systems | Citations: 0

Authors: Gross, Jan; Groiss, Niklas; Rieg, Thilo; Baumgartl, Hermann; Buettner, Ricardo

Abstract: Dementia and Alzheimer's disease represent one of the biggest medical challenge ... Expand

Abstract: Dementia and Alzheimer's disease represent one of the biggest medical challenges of our century, manifesting the risk to individuals of losing their language or selfmanagement skills. It is estimated that by 2050 around 1.2% of the world's population will suffer from these diseases. Since there are no effective treatment options available, it is of great impor tance to detect dementia tendencies at the earliest possible stage. Mild cognitive impairment represents a preclinical stage of Alzheimer's disease, allowing detecting dementia tendencies prematurely. As related work either relies on sen sitive test results, does not manifest a sophisticated level of accuracy, or relies on costly techniques, further research is required. In this study, we propose a machine learning approach using restingstate EEG recordings of channels located in the Broca's area. We achieved a benchmark accuracy of 90.91% in differentiating nonMCI and MCI individuals based on the assessment of phonemic verbal fluency. Collapse

Topics: personality database system health information system data quality

Methods: electroencephalography computational algorithm machine learning magnetic resonance imaging descriptive statistic

Is AI Ground Truth Really True? The Dangers of Training and Evaluating AI Tools Based on Experts’ Know-What

2021 | Management Information Systems Quarterly | Citations: 0

Authors: Lebovitz, Sarah; Levina, Natalia; Lifshitz-Assaf, Hila

Abstract: Organizational decision-makers need to evaluate AI tools in light of increasing ... Expand

Abstract: Organizational decision-makers need to evaluate AI tools in light of increasing claims that such tools outperform human experts. Yet, measuring the quality of knowledge work is challenging, raising the question of how to evaluate AI performance in such contexts. We investigate this question through a field study of a major U.S. hospital, observing how managers evaluated five different machine-learning (ML) based AI tools. Each tool reported high performance according to standard AI accuracy measures, which were based on ground truth labels provided by qualified experts. Trying these tools out in practice, however, revealed that none of them met expectations. Searching for explanations, managers began confronting the high uncertainty of experts' know-what knowledge captured in ground truth labels used to train and validate ML models. In practice, experts address this uncertainty by drawing on rich know-how practices, which were not incorporated into these ML-based tools. Discovering the disconnect between AI's know-what and experts' know-how enabled managers to better understand the risks and benefits of each tool. This study shows dangers of treating ground truth labels used in ML models objectively when the underlying knowledge is uncertain. We outline implications of our study for developing, training, and evaluating AI for knowledge work. Collapse

Topics: artificial intelligence knowledge work explainable artificial intelligence analytical information system crowdsourcing

Methods: machine learning qualitative interview theory development field study archival research