Mental workload (MWL) monitoring with electroencephalography (EEG) is promising for adaptive systems, but practical use requires models that remain reliable across time, tasks, and users with feasible sensor layouts. This study investigated condition-associated EEG signatures in 15 healthy adults performing two complementary N-back paradigms: an N-back level manipulation (N-LEVELS: 1-back, 2-back, 3-back) and a stimulus-pacing manipulation at fixed 2-back (N-SPEED: slow, medium, fast). EEG was recorded using a compact 8-channel setup. Spectral bandpower features, composite workload indices, and N2/P3b ERP markers were analyzed. Linear classifiers were evaluated using within-subject TRAIN → TEST and leave-one-subject-out (LOSO) protocols, together with analyses of observation duration, time-on-task dynamics, temporal generalization, cross-paradigm transfer, and exploratory ordinal regression. N-LEVELS and N-SPEED produced distinct behavioral and subjective profiles: increasing N-back level was associated with greater task demand, whereas faster N-SPEED conditions showed greater perceived temporal demand and reduced response-time availability. Exploratory locally corrected spectral analyses suggested more frequent beta-band and engagement-index modulation in N-LEVELS and more frequent posterior/parietal alpha suppression and theta/alpha-related modulation in N-SPEED. ERP analyses showed reduced P3b amplitude under higher N-back levels and faster pacing. Classification performance was modest and context-dependent: N-SPEED three-class decoding was more consistent than N-LEVELS, and all three N-SPEED pairwise contrasts survived correction under both WITHIN and LOSO evaluation, whereas in N-LEVELS only the 1-back vs. 3-back contrast survived under LOSO. Post hoc artifact-sensitivity analyses supported the principal P3b findings and largely preserved reference-decoding and alpha time-on-task patterns. After delta-band removal, neither three-class LOSO result nor the N-SPEED slow-vs.-medium and medium-vs.-fast LOSO contrasts survived correction; N-SPEED slow vs. medium under LOSO also did not survive either high-pass variant. Exploratory ordinal regression showed a clearer ordered relationship for N-SPEED than N-LEVELS. Time-on-task analyses revealed selective non-stationarity, mainly progressive alpha build-up, whereas cross-paradigm transfer did not survive corrected inference. These findings indicate that compact EEG recordings contain condition-associated information, but the small sample, deterministic block order, modest decoding performance, and limited cross-paradigm generalization constrain causal attribution and applicability beyond the present protocol.
Operator state monitoring (OSM) increasingly combines physiological, behavioral, performance, and contextual evidence to inform adaptive human-machine systems. Yet detecting a change in operator condition does not establish its psychological meaning, its prospective operational consequences, or whether automation should intervene. Drawing on adaptive aiding, physiological computing, and human–autonomy teaming, this Conceptual Analysis proposes a six-layer architecture comprising multimodal sensing, signal-quality management, individualized baseline and context modeling, cognitive-state inference, forecasting and risk estimation, and adaptation policy. Its contribution is not a new sensing method, classifier, or adaptation algorithm, but an uncertainty-aware assurance chain in which application-specific evidentiary gates constrain progression from observation through inference and forecasting, while a defined policy boundary regulates action. Measurement limitations and contextual uncertainty remain visible downstream; functional-state hypotheses must justify their diagnostic value; and a separately validated construct-light pathway permits direct operational-risk forecasting when psychological distinctions are unsupported. Forecasts identify potential hazards under specified conditions but do not, by themselves, demonstrate that intervention will help or confer authority to act. Policy therefore evaluates permissible, bounded responses, including deliberate non-intervention and predefined loss-of-assurance fallbacks, against their anticipated benefits and costs. A hypothetical aviation approach illustrates how gate decisions can restrict evidence, withhold unsupported state labels, and permit a reversible intervention without altering flight-control authority. The central research proposition is whether explicit evidentiary gates and uncertainty propagation improve calibration, decision appropriateness, or joint-system outcomes relative to less constrained approaches. Until demonstrated, the architecture remains a conceptual framework for translating uncertain knowledge of operators into defensible machine behavior.
IntroductionCybersickness is a common issue in immersive virtual reality (VR), causing symptoms such as nausea, dizziness, and eye strain. In demanding VR training scenarios, stress and mental workload may influence cybersickness severity. However, further research requires datasets that capture these factors alongside subjective and objective measures. This paper presents a multimodal dataset collected during a VR driving simulation to investigate cybersickness in relation to stress and mental workload.MethodsSixty participants completed a phase-based driving task in immersive VR. All participants performed three initial driving phases with progressively increasing difficulty, followed by a condition-specific final phase designed to elicit mental workload, stress, or a combination of both (20 participants per condition). Self-report questionnaires assessing cybersickness, mental workload, and stress-related symptoms were administered at multiple time points during the experiment. Physiological and behavioral data were continuously recorded using an instrumented VR headset and wearable sensors, including electroencephalography (EEG), cardiovascular signals, respiration, and head movement.ResultsRepresentative analyses of subjective and objective measures were conducted to validate the dataset. The analyses demonstrate changes in cybersickness symptoms across driving phases and experimental conditions, illustrating the potential of the collected data to examine cybersickness alongside stress and mental workload.DiscussionThis multimodal dataset provides a resource for neuroergonomics research on cybersickness in immersive VR driving scenarios. By integrating self-reported symptoms with physiological and behavioral measurements, it supports future investigations into the relationships between cybersickness, stress, and mental workload, as well as the development of approaches for monitoring and mitigating cybersickness in demanding VR applications.
Human performance often fails not because latent capacity is absent, but because access to high-functioning cognitive states is unstable under fatigue, distraction, overload, and pressure. This hypothesis paper proposes closed-loop cognitive state control—informally, supercognition—as a neuroengineering objective: rapid entry into, stabilization of, and recovery of an individual's highest available task-specific cognitive state. The framework treats cognition as a partially observable dynamical system whose latent states would be inferred from multimodal behavioral, physiological, contextual, and neural signals, and specifies a five-layer architecture comprising multimodal sensing, latent-state inference, predictive transition modeling, closed-loop intervention, and a longitudinal personal neural twin. The proposal is situated against existing closed-loop and neuroadaptive architectures, with an explicit statement of what is and is not new. The state ontology is given as a set of falsifiable assumptions rather than an intuitive taxonomy. Candidate computational implementations are specified for each layer, including hidden semi-Markov and switching state-space models, Bayesian and Kalman filtering, sequence models, hazard models, contextual bandits, model-predictive control, and offline reinforcement learning. The personal neural twin receives an operational specification covering its data requirements, update schedule, calibration and drift-monitoring criteria, and failure modes. Predictive and causal claims are separated, with the evidentiary requirements stated for each. The paper engages directly with the literature that challenges the framework's feasibility, including biomarker reliability, mixed neurofeedback findings, cross-task and cross-subject generalization failure, and the ecological-validity gap in wearable neurotechnology. Operational definitions, application scenarios, a validation protocol with power and diversity requirements, an expanded ethics and governance analysis, and a staged research roadmap are provided. No component of the architecture has been implemented or evaluated here: no simulation, clinical trial, or empirical efficacy claim is presented, and the validation programme described is prospective. The contribution is a falsifiable systems-engineering hypothesis for cognitive-state control.
IntroductionIn dynamic, high-stakes environments, such as medical emergencies, effective teamwork and team performance rely on team adaptation. Team Behavioral Interaction Patterns (i.e., recurring sequences of team behaviors; TBIPs) have been proposed as markers of adaptive functioning, while physiological coordination among team members may offer an unobtrusive, real-time proxy for such behavioral dynamics. The current paper presents an investigation of whether TBIPs are (1) related to self-reported team adaptation and rated performance, and (2) reflected in physiological coordination. The paper advances the potential use of wearable technology for (near-) real-time team monitoring.MethodsWe conducted a quasi-experimental study with 19 medical teams (N = 150) during simulation-based medical emergency team training. Teams were video-recorded and wore wrist-worn sensors capturing electrodermal activity (EDA) and heart rate (HR). Based on video data, we extracted TBIPs via T-pattern detection. We assessed the complexity of TBIPs and calculated physiological coordination (synchrony and regularity) for each pattern. Team adaptation was measured via self-report, and team performance was rated by trainers. We related physiological coordination to team adaptation and team performance, while accounting for the complexity of TBIPs.ResultsRegression analyses showed no association between TBIP complexity and team performance. Multilevel models revealed that the relationship between physiological synchrony (HR and EDA) and team adaptation, and that between EDA regularity and team performance, depended on TBIP complexity. Specifically, better team adaptation was related to higher HR synchrony during low-complexity TBIPs and lower EDA synchrony during high-complexity TBIPs, whereas better team performance was linked to higher EDA regularity during high-complexity TBIPs.ConclusionOur findings suggest that behavioral pattern complexity moderates the relationship between physiological coordination and team outcomes. This interplay between behavioral and physiological data highlights the value of combining both modalities to understand and potentially support adaptive team functioning.
