Using Real-Time Lexical Indicators to Detect Performance Decrements in Spaceflight Teams: A Methodology to Dynamically Monitor Cognitive, Emotional, and Social Mechanisms That Influence Performance

Future exploratory long-duration missions will incorporate a crew of six on a mission length of approximately 2.5 years. Challenges include the requirement for the crew to function autonomously, under significant communication delays, and with the potential for increased crew and interpersonal friction or tension. The specific aims of this research are to (1) develop a methodology to assess cognitive and emotional state at a distance though analysis of spontaneous verbal output in real-time communications and (2) test the feasibility of a real-time assessment tool, STRESSnet, to detect cognitive performance deficits, stress, fatigue, anxiety, and depression in the spaceflight operational setting.

Because the health and well-being of crew members directly affects mission success, it is important to track cognitive/emotional changes that may indicate a deficit. One problem with many existing assessment methods is that most require direct observation of behavior or performance or self-assessment by a pen and paper-type instrument. The requirement to assess individual and team functioning at a distance suggests the potential efficacy of a methodology to assess cognitive and emotional state in real-time from ongoing or spontaneous verbal output. The basic premise of this work is that spontaneous verbal output provides a natural and valid indicator of basic cognitive processes. Natural word use is not prone to the typical limitations of self-report measurements. That is, natural language use is less subject to social desirability bias, and can be derived in real-time without interfering with the cognitive processes being measured, and without interrupting crew performance. Moreover, natural word use is reliable and consistent across time and context, and can be meaningfully measured in individuals and teams.

STRESSnet is a lexical analysis tool designed to provide a non-obstrusive means of detecting stress and related deficits in long-duration spaceflight through the assessment of spontaneous verbal output in real-time crew communications. The research builds on existing work on text and sentiment analysis; however, STRESSnet is unique in that (1) it is specifically designed to assess stress and related cognitive/emotional states, (2) we draw on existing astronaut communications and mission logs to develop a lexicon that includes terms unique to this environment, and (3) we developed STRESSnet with the specific goal of application as a tool to assess user state and provide automatic schedule recommendations for crew work/leisure activities to counter identified deficits.

This report summarizes data derived from testing STRESSnet in the Human Exploration Research Analog (HERA) 3 campaign. In summary, the HERA 3 data results are consistent with and support the results observed in the HERA 1 and HERA 2 campaigns. Our overall research goals are to: (1) Demonstrate that the lexical indicator measure (STRESSnet) correlates with existing cognate pen-and-paper measures of stress. (2) Demonstrate that the lexical indicator measure (STRESSnet) is sensitive to differences in operational variability, such as levels of workload or stress. (3) Demonstrate that the lexical indicator measure (STRESSnet) is sensitive to other variables of interest, such as within-group or between-group differences.

The HERA 3 data described in the following sections indicate that the primary STRESSnet lexical measures of Social Impairment, Anxiety, Cognitive Load, and Attentional Focus were consistent with pen-and-paper measures of these same constructs. Negative affect as measured by Positive and Negative Affect Schedule (PANAS) was very low across Campaign 3 missions and this lack of variability resulted in inconsistent relationships with lexical measures of negative emotion. Second, STRESSnet measures for each primary facet consistently differentiate between high and low workload days in HERA. Third, STRESSnet measures were sensitive to differences in the type of communications (e.g., task-oriented vs. non-task oriented communications). Fourth, using STRESSnet measures of anger and cognitive anxiety, we were able to identify a reported elevated episode of anger in one mission, demonstrating the utility of this approach for diagnosing crew stress and well-being during long duration spaceflight.

More »