An evaluation of an adaptive automation system using a cognitive vigilance task

Area of Science:

• Human-Computer Interaction
• Cognitive Psychology
• Neuroscience

Background:

• Adaptive automation systems adjust their support based on user state.
• Cognitive vigilance tasks assess sustained attention.
• Engagement Index (EI) derived from EEG measures cognitive workload.

Purpose of the Study:

• To evaluate adaptive automation system performance using a cognitive vigilance task.
• To investigate the impact of different feedback conditions (positive and negative) on vigilance.
• To determine if EEG-derived engagement influences adaptive automation behavior.

Main Methods:

• Participants performed a 40-min vigilance task detecting a target stimulus.
• Electroencephalography (EEG) recorded brain activity to derive an Engagement Index (EI).
• Two feedback conditions manipulated array size based on EI: negative (inverse relationship) and positive (direct relationship). Yoked controls received identical array changes.

Main Results:

• A vigilance decrement was observed exclusively in the positive feedback experimental group.
• The negative feedback group and yoked controls did not exhibit a significant vigilance decrement.
• EEG-derived Engagement Index varied with task demands and feedback conditions.

Conclusions:

• The type of feedback in adaptive automation significantly impacts vigilance performance.
• Positive feedback, where increased engagement leads to increased task complexity, may impair sustained attention.
• Adaptive automation design should consider feedback mechanisms to optimize user performance and prevent vigilance decrements.