NOTE: Element change to Human Factors & Behavioral Performance; previously Behavioral Health & Performance (Ed., 1/17/17)

NOTE: Project continues as "SCALE: Shared Cognitive Architectures for Long-term Exploration--80NSSC18K0221," due to Principal Investigator move to Northwestern University in Fall 2016.

In Study 1, we use meta-analysis to explore the conditions under which shared cognition is most important to crew performance. We explore three types of moderating variables, including team type (e.g., team diversity, autonomy, and lifespan), team context (e.g., virtuality, temporal and geographic dispersion, and organizational embeddedness), and team conditions (e.g., physical and psychological stressors). Results help us determine (1) the extent to which prior teams research is relevant to long distance space exploration, (2) factors which need to be modeled effectively in order to predict, diagnose, and prevent breakdowns in shared cognition, and (3) which antecedent and moderating conditions are most relevant to cognitive sharedness in long distance space exploration.

In Study 2, we developed an approach to using text analytics as an unobtrusive means of measuring shared cognition using the archived mission transcripts from the Skylab missions. We identified the text-based signatures of cognitive breakdown that ultimately led to conflict in Skylab 4. We also explored the dynamic conversational patterns in each mission that were indicative of cognitive sharedness or its breakdown. This study informs our broader goal of developing efficient and unobtrusive measures of shared cognition in multiteam systems.

In Study 3, we developed an agent based model (ABM) using empirical data collected from the HERA analog. We developed the ABM in two phases. First, we canvassed the relevant literature on shared cognition to identify its antecedent conditions. Second, we collected laboratory and analog data to estimate and then test the parameters of the agent based model. In particular, we explored personality factors, social relations, situational factors, and task stream characteristics that associate with cognitive sharedness. The ABM will be used in Year 3 to run “virtual experiments” wherein we can pose “what if” questions that will directly inform the development of interventions aimed at promoting shared cognition, ultimately informing our major goal associated with modeling shared cognition in teams and multiteam systems.

Studies 4-6 speak directly to our major goal of identifying antecedents conditions of cognitive sharedness. In Study 4, we collected data in the HERA analog using daily planning surveys and the Project RED task battery to explore the role of social connectedness, communication delay, and sleep deprivation on shared cognition. In Study 5, we conducted archival research on thirty documented Antarctic exploration missions to explore how social roles, leadership isomorphism, and group dynamics affected cognitive sharedness and ultimately team functioning and viability on these missions. Finally, in Study 6, we worked with Diego Urbina from Space Applications, Inc. to collect data in the Moonwalk analogs in Rio Tinto, Spain and Marseilles, France. In these analogs, we explored how information sharedness affects shared cognition.

Our research will provide fundamental, generalizable findings in the area of shared cognition in teams and multiteam systems. Researchers have bemoaned the inefficiencies associated with measuring shared cognition as historically such measurements have been intrusive and time-consuming. In this program of research, we are exploring the use of text analytics, social role network analyses, and surveys as efficient and less intrusive means of assessing cognitive sharedness within the team and broader multiteam system. Second, as teamwork becomes more prevalent across workplaces and occupations and its nature more variable, a broader array of antecedent, moderating, and mediating conditions associated with shared cognition come into play. We are conducting meta-analytic, archival, experimental, and analog studies designed to identify and understand the antecedents and consequences of shared cognition in teams and multiteam systems. Finally, breakdowns in shared cognition must be identified and addressed quickly before team functioning is irrevocably affected. We use data collected in these meta-analytic, archival, experimental, and analog studies to develop an agent based model which will be used in Year 3 to conduct virtual experiments and ultimately create a dashboard for diagnosing and developing interventions for breakdowns in shared cognition.

This research will be useful for investigating shared cognition in any modern-day organization facing complex collaborative challenges, such as NASA space exploration, large scientific consortia (e.g., CERN), cybersecurity teams, healthcare systems, and the military. Furthermore, findings could then be leveraged to develop system-wide interventions that increase overall work efficiency and resilience in safety-critical systems. In this review period, we have collected data across multiple platforms (please see subsequent sections of this report for details) to begin to understand the antecedent, consequent, and moderating conditions associated with shared cognition. In addition to providing Earth Benefits in the area of shared cognition in teams and multiteam systems, we are also testing our hypotheses in analogs that recreate the unique contexts in which astronauts operate (i.e., extreme, isolated/confined environments). We will compare results across research paradigms to see how effects of isolation/confinement may moderate shared cognition’s role in team functioning, thus providing unique findings that will inform NASA’s operations aboard the International Space Station (ISS) and on Long-Duration Space Exploration (LDSE) missions.

NOTE: Project continues as "SCALE: Shared Cognitive Architectures for Long-term Exploration--80NSSC18K0221," due to Principal Investigator move to Northwestern University in Fall 2016. See that project for subsequent reporting.

NOTE: Element change to Human Factors & Behavioral Performance; previously Behavioral Health & Performance (Ed., 1/17/17)

DATA COLLECTION

A. HERA Campaign Participation and Data Collection: We participated in HERA Campaign 3 Mission 1, in partial fulfillment of our objective to collect team task switching data in an analog environment. To capture the full representative structure of a NASA mission to Mars, the HERA crew served as an isolated/confined crew in an environment similar to one that may travel beyond low-Earth orbit, and eight undergraduate participants at Georgia Tech served as Mission Control teams supporting the HERA crew on their mission. We ran three 2-hour sessions (3-hour sessions for participants at Georgia Tech who received 1 hour of training) in which we collected unobtrusive team task-switching data, as well as performance data on the overall task. We also collected team process data through pop-up surveys administered throughout sessions. Additionally, we implemented a time delay within our own software platform to align with the time delay manipulation implemented in HERA during our second session. We are pleased to report that all data collection efforts across all three 2-hour HERA sessions were successful.

B. Project RED Platform, and Additional Multiteam System Experimental Sessions - Control Sessions: The platform used to collect this data was the computer-based multi-team task platform that we developed at Georgia Tech, called Project Red planet Exploration and Development (Project RED). In Project RED, four teams of 3 members each (i.e., a Multiteam System; MTS) work on solving the complex problem of designing and implementing a well infrastructure on Mars that will support future inhabitants of the planet. We are pleased to report that all data collection efforts across all three 2-hour HERA sessions were successful. We have also designed four additional control experiment sessions (one completed, three to be run by the end of the annual review period) to obtain baseline measures against which we can compare our analog data of crew members in an isolated/confined environment. One manipulation will be the distribution of the MTS. In the distributed condition, 4 participants will serve as the “crew” located in the SONIC lab at Northwestern University, and the remaining 8 “Mission Control” participants will be located in the DELTA lab at Georgia Tech. For the non-distributed MTS condition, all participants will be located at the DELTA lab at Georgia Tech. The second manipulation will be the presence or absence of a time delay in communication.

C. Project RED Relay - Mapping Shared Cognitive Architecture: To understand the knowledge that individuals have about who knows whom, we implemented a knowledge task called “Project RED Relay,” to assess how well people recall information about their group’s network structure, and how well they understand other people’s knowledge of the group’s network. This task provides insight into how accurate or convergent knowledge of networks relates to team performance and outcomes during an LDSE mission where individuals work with a variety of others. Project Red Relay is administered to participants following the Project RED Design task as a network game that people access via laptops. This task asks people to direct messages containing files that need to be sent to a specified target in as few steps as possible by routing the message to other people who may be connected to the intended final target. Participants can only choose a small number of contacts to use to direct messages, and to route the messages efficiently, they must be able to know whom other participants selected as their contacts. During this network game, we collected measures of cognitive awareness and perceived individual, team, and MTS efficacy at routing the messages. Therefore, this knowledge task allows us to measure the accuracy, efficiency, and perceptions of people’s cognitive structures.