Building clinician trust in AI-assisted neurodiagnostics: A case-based evaluation

Abstract

Neurodiagnostics and AI: How Artificial Intelligence is changing the Future of Neurodiagnostics. Clinician trust is critical for successfully adopting AI-assisted tools in clinical settings. Here, we evaluate clinician trust in AI-based neurodiagnostic solutions using a simulated multicast-medicine scenario about stroke, neurodegenerative disease, and traumatic brain injury (TBI). Then, we study the reliability, explainability, and acceptance of the AI-generated diagnoses by both the deep learning models and XAI (explainable AI) methods. These results pinpoint significant predictors of clinician confidence and pathways for potential interventions to induce AI adaptation in neurodiagnostics. AI revolutionizes neurodiagnostics and provides more effective and cost-effective methods to detect neuro diseases. But clinicians’ trust in these tools is essential for the successful implementation of AI in clinical practice. This study evaluates clinician trust in AI neurodiagnostic tools in different clinical vignette cases based on those associated with reliability, interpretability, and agreement between AI models and expert judgments. We assess clinician responses (including those of IS, TBI, and neurodegenerative diseases) to AI-generated diagnoses using an AI-based decision-support system with deep learning and explainable AI methods. Our findings add to a growing understanding of factors that could influence clinician uptake and strategies to optimize AI-enabled diagnostic use in neurologic practice.

Ähnliche Arbeiten