OBJECTIVE: The present study developed and validated a stochastic model of overt attention within a visual workspace. BACKGROUND: Technical specifications and recommended practices for the design of visual warning systems emphasize the role of alert salience. Task demands and display context can modulate alert noticeability, however, meaning that salience alone does not guarantee attention capture. METHOD: A stochastic model integrated elements from existing models of visual attention to predict attentional behavior in dynamic environments.Validation studies tested the predictions of the new model against scanning data from a high-fidelity simulator study and behavioral data from an alert detection experiment. RESULTS: The model accurately predicted the steady-state distribution of attention within a simulated cockpit as well as the effects of color similarity, eccentricity, and dynamic visual noise on miss rates and response times in the alert detection task. CONCLUSION: The model successfully predicts attentional behavior in complex visual workspaces with the use of parameter values selected by either the modeler or a subject matter expert. APPLICATION: The model provides a tool to test the effectiveness of visual alerts in various display configurations and with varying task demands.
OBJECTIVE: The present study developed and validated a stochastic model of overt attention within a visual workspace. BACKGROUND: Technical specifications and recommended practices for the design of visual warning systems emphasize the role of alert salience. Task demands and display context can modulate alert noticeability, however, meaning that salience alone does not guarantee attention capture. METHOD: A stochastic model integrated elements from existing models of visual attention to predict attentional behavior in dynamic environments.Validation studies tested the predictions of the new model against scanning data from a high-fidelity simulator study and behavioral data from an alert detection experiment. RESULTS: The model accurately predicted the steady-state distribution of attention within a simulated cockpit as well as the effects of color similarity, eccentricity, and dynamic visual noise on miss rates and response times in the alert detection task. CONCLUSION: The model successfully predicts attentional behavior in complex visual workspaces with the use of parameter values selected by either the modeler or a subject matter expert. APPLICATION: The model provides a tool to test the effectiveness of visual alerts in various display configurations and with varying task demands.