Frontiers in neuroergonomics最新文献

Speech imagery brain-computer interfaces (SI-BCIs) aim to decode imagined speech from brain activity and have been successfully established using non-invasive brain measures such as electroencephalography (EEG). However, current EEG-based SI-BCIs predominantly rely on high-resolution systems with 64 or more electrodes, making them cumbersome to set up and impractical for real-world use. In this study, we evaluated several electrode reduction algorithms in combination with various feature extraction and classification methods across three distinct EEG-based speech imagery datasets to identify the optimal number and position of electrodes for SI-BCIs. Our results showed that, across all datasets, the original 64 channels could be reduced by 50% without a significant performance loss in classification accuracy. Furthermore, the relevant areas were not limited to the left hemisphere, widely known to be responsible for speech production and comprehension, but were distributed across the cortex. However, we could not identify a consistent set of optimal electrode positions across datasets, indicating that electrode configurations are highly subject-specific and should be individually tailored. Nonetheless, our findings support the move away from extensive and costly high-resolution systems toward more compact, user-specific setups, facilitating the transition of SI-BCIs from laboratory settings to real-world applications.

Neuroadaptive technologies are a type of passive Brain-computer interface (pBCI) that aim to incorporate implicit user-state information into human-machine interactions by monitoring neurophysiological signals. Evaluating machine learning and signal processing approaches represents a core aspect of research into neuroadaptive technologies. These evaluations are often conducted under controlled laboratory settings and offline, where exhaustive analyses are possible. However, the manner in which classifiers are evaluated offline has been shown to impact reported accuracy levels, possibly biasing conclusions. In the current study, we investigated one of these sources of bias, the choice of cross-validation scheme, which is often not reported in sufficient detail. Across three independent electroencephalography (EEG) n-back datasets and 74 participants, we show how metrics and conclusions based on the same data can diverge with different cross-validation choices. A comparison of cross-validation schemes in which train and test subset boundaries either respect the block-structure of the data collection or not, illustrated how the relative performance of classifiers varies significantly with the evaluation method used. By computing bootstrapped 95% confidence intervals of differences across datasets, we showed that classification accuracies of Riemannian minimum distance (RMDM) classifiers may differ by up to 12.7% while those of a Filter Bank Common Spatial Pattern (FBCSP) based linear discriminant analysis (LDA) may differ by up to 30.4%. These differences across cross-validation implementations may impact the conclusions presented in research papers, which can complicate efforts to foster reproducibility. Our results exemplify why detailed reporting on data splitting procedures should become common practice.

Introduction: This study aims to develop and evaluate support vector machines (SVMs) learning models for predicting cognitive workload (CWL) based on physiological data. The objectives include creating robust binary classifiers, expanding these to multiclass models for nuanced CWL prediction, and exploring the benefits of individualized models for enhanced accuracy. Cognitive workload assessment is critical for operator performance and safety in high-demand domains like aviation. Traditional CWL assessment methods rely on subjective reports or isolated metrics, which lack real-time applicability. Machine learning offers a promising solution for integrating physiological data to monitor and predict CWL dynamically. SVMs provide transparent and auditable decision-making pipelines, making them particularly suitable for safety-critical environments.

Methods: Physiological data, including electrocardiogram (ECG) and pupillometry metrics, were collected from three participants performing tasks with varying demand levels in a low-fidelity aviation simulator. Binary and multiclass SVMs were trained to classify task demand and subjective CWL ratings, with models tailored to individual and combined subject datasets. Feature selection approaches evaluated the impact of streamlined input variables on model performance.

Results: Binary SVMs achieved accuracies of 70.5% and 80.4% for task demand and subjective workload predictions, respectively, using all features. Multiclass models demonstrated comparable discrimination (AUC-ROC: 0.75-0.79), providing finer resolution across CWL levels. Individualized models outperformed combined-subject models, showing a 13% average improvement in accuracy. SVMs effectively predict CWL from physiological data, with individualized multiclass models offering superior granularity and accuracy.

Discussion: These findings emphasize the potential of tailored machine learning approaches for real-time workload monitoring in fields that can justify the added time and expense required for personalization. The results support the development of adaptive automation systems in aviation and other high-stakes domains, enabling dynamic interventions to mitigate cognitive overload and enhance operator performance and safety.

Introduction: The EmotiBit photoplethysmography (PPG) device allows user-owned data collection for measures of cardiovascular activity (CVA) and electrodermal activity (EDA) in naturalistic settings. The aim of this study was to evaluate the validity of this device for collecting high-quality data while participants experience varying levels of cognitive workload.

Methods: Using a standardized criterion validity protocol, recordings of 15 participants performing a cognitive workload task were compared for the EmotiBit and a reference electrocardiography (ECG) device (BITalino PsychoBit). Multiple preprocessing pipelines and a signal quality check were implemented. Parameters of interest including heart rate (HR), heart rate variability (HRV) measures, skin conductance level (SCL), and skin conductance response (SCR) measures were assessed using Bland-Altman plot and ratio (BAr) analyses, as well as cross-correlations of the EDA signal time series of both devices.

Results: BAr results indicated good agreement between devices regarding HR with an average difference of 1-2 beats per minute (bpm). HRV measures yielded an insufficient BAr, albeit most data points lay within a priori boundaries of agreement. EDA measures yielded insufficient agreement for comparing SCL and SCR number and amplitude.

Discussion: The results are comparable to the validation of similar wearable PPG devices and extend the validation of the EmotiBit by assessing the acquired signals during varying levels of cognitive workload. While the device may be used to collect HR for scientific data analysis, its quality regarding HRV and EDA measures is not comparable to a standard ECG.

Significance: This study provides the first systematic validation following a standardized protocol of the EmotiBit PPG device relative to an ECG when considering recordings collected during cognitive workload induction.

Introduction: The perception of physical comfort is one of the important workplace design parameters. Most comfort perception studies have mainly relied on subjective assessments and biomechanical techniques, with limited exploration of neural brain activity.

Methodology: The current study investigates this research gap by integrating the rating of perceiving physical comfort (RPPC) with brain network indices in an arm flexion task across different force levels. The applied arm forces, EEG-based neural responses, and the RPPC were measured, and the corresponding network theory indices were calculated. The following correlations were evaluated: (a) RPPC and applied forces, (b) network theory indices and applied forces, and (c) RPPC and network theory indices.

Results and discussion: Results for (a) revealed a significant negative correlation between RPPC and the applied force for the arm flexion task. This shows that as the exerted force difficulty increases to an extremely hard level, the perception of physical comfort decreases till it reaches no comfort level. Results for (b) showed a positive correlation between the applied forces and global efficiency for the alpha network coherence during an extremely hard task. In contrast, a negative correlation was found between applied forces and path length for beta coherence during a light task. Findings from (b) suggest that the brain is more efficient in transmitting information related to cognitive functioning during a highly demanding force exertion task than a light task. Results from (c) showed a negative correlation between RPPC and global efficiency for alpha coherence during an extremely hard force exertion task. Moreover, a positive correlation was observed between RPPC and local efficiency for beta coherence during a somewhat hard task. Findings from (c) also indicate that perceiving a low-comfort physical task might increase task alertness, with the corresponding neural network exhibiting a high level of internal brain organization.

Conclusions: The study results contribute valuable knowledge toward understanding the neural responses underlying the perception of physical comfort levels.

Introduction: Educational practice increasingly makes use of technology to improve teaching and learning. New wearable technology is being developed that measures mental states like attention and stress, through neurophysiological signals like electroencephalography (EEG), electrodermal activity (EDA) and heart rate. However, little is known about the ethical aspects of this technology.

Methodology: We provide an overview of current ethical considerations on such wearable technologies in classroom settings and analyze these critically. We distinguished three ethical angles to analyze new technologies: epistemic, principle-based, and Foucauldian. We focus on a Foucauldian analysis, outlining how such technologies affect power relationships and self-understanding, but also which responses people develop to evade power. In addition, a focus group of high school students was set up to identify young people's views on such wearable technology and to initiate a reflection on the theory-based ethical considerations.

Results: Our study shows that although wearables may provide information on learning and attention, and even though possible users are enthusiastic about the potential, there are several risks of applying such technologies in educational settings. These risks concern governance and surveillance, normalization and exclusion, placing technology before pedagogy, stimulating neoliberal values and quantified self-understanding, and possible negative impact on identity for those who think they are outside of the norm. High school students highlighted that people are not only subjected to new technologies, but also subject these technologies to their own goals.

Discussion: We end with a discussion on the perils of implementing new technologies, and provide an alternative to prohibition in the form of co-creating and educating. Any potential future implementation of mental state tracking technology is to be accompanied by normative discussions on legitimate aims, on rights, interests and needs of both pupils, teachers, and educational institutions, taking broader debates on what should count as a good pedagogical climate into account.

The Sense of Embodiment (SoE) refers to the subjective experience of perceiving a non-biological body part as one's own. Virtual Reality (VR) provides a powerful platform to manipulate SoE, making it a crucial factor in immersive human-computer interaction. This becomes particularly relevant in Electroencephalography (EEG)-based Brain-Computer Interfaces (BCIs), especially motor imagery (MI)-BCIs, which harness brain activity to enable users to control virtual avatars in a self-paced manner. In such systems, a strong SoE can significantly enhance user engagement, control accuracy, and the overall effectiveness of the interface. However, SoE assessment remains largely subjective, relying on questionnaires, as no definitive EEG biomarkers have been established. Additionally, methodological inconsistencies across studies introduce biases that hinder biomarker identification. This study aimed to identify EEG-based SoE biomarkers by analyzing frequency band changes in a combined dataset of 41 participants under standardized experimental conditions. Participants underwent virtual SoE induction and disruption using multisensory triggers, with a validated questionnaire confirming the illusion. Results revealed a significant increase in Beta and Gamma power over the occipital lobe, suggesting these as potential EEG biomarkers for SoE. The findings underscore the occipital lobe's role in multisensory integration and sensorimotor synchronization, supporting the theoretical framework of SoE. However, no single frequency band or brain region fully explains SoE. Instead, it emerges as a complex, dynamic process evolving across time, frequency, and spatial domains, necessitating a comprehensive approach that considers interactions across multiple neural networks.

Introduction: Anger while driving is often provoked by on-road events like sudden cut-offs but can also arise from external factors, such as rumination of negative thoughts. With the rise of autonomous vehicles, drivers are expected to engage more in non-driving activities, potentially increasing the occurrence of anger stemming from non-driving-related sources. Given the well-established link between anger and aggressive driving behaviors, it is crucial to detect and understand the various origins of anger in autonomous driving contexts to enhance road safety.

Methods: This study investigates whether physiological (cardiac and respiratory activities) and ocular indicators of anger vary depending on its source (driving-related or external) in a simulated autonomous driving environment. Using a combination of autobiographical recall (AR) for external anger induction and driving-related scenarios (DS), 47 participants were exposed to anger and/or neutral conditions across four groups.

Results: The results revealed that combined anger induction (incorporating both external and driving-related sources) led to higher subjective anger ratings, more heart rate variability. However, when examined separately, individual anger sources did not produce significant differences in physiological responses and ocular strategies.

Discussion: These results suggest that the combination of anger-inducing events, rather than the specific source, is more likely to provoke a heightened state of anger. Consequently, future research should employ combined induction methods to effectively elicit anger in experimental settings. Moreover, anger detection systems should focus on the overall interplay of contributing factors rather than distinguishing between individual sources, as it is this cumulative dynamic that more effectively triggers significant anger responses.

Functional Near-Infrared Spectroscopy (fNIRS) has proven in recent time to be a reliable workload-detection tool, usable in real-time implicit Brain-Computer Interfaces. But what can be done in terms of application of neural measurements of the prefrontal cortex beyond mental workload? We trained and tested a first prototype example of a memory prosthesis leveraging a real-time implicit fNIRS-based BCI interface intended to present information appropriate to a user's current brain state from moment to moment. Our prototype implementation used data from two tasks designed to interface with different brain networks: a creative visualization task intended to engage the Default Mode Network (DMN), and a complex knowledge-worker task to engage the Dorsolateral Prefrontal Cortex (DLPFC). Performance of 71% from leave-one-out cross-validation across participants indicates that such tasks are differentiable, which is promising for the development of future applied fNIRS-based BCI systems. Further, analyses within lateral and medial left prefrontal areas indicates promising approaches for future classification.

Introduction: Recent advancements in augmented reality (AR) technology have opened up potential applications across various industries. In this study, we assess the effectiveness of psychomotor learning in AR compared to video-based training methods.

Methods: Thirty-three participants (17 males) trained on four selection-based AR interactions by either watching a video or engaging in hands-on practice. Both groups were evaluated by executing these learned interactions in AR.

Results: The AR group reported a higher subjective workload during training but showed significantly faster completion times during evaluation. We analyzed brain activation and functional connectivity using functional near-infrared spectroscopy during the evaluation phase. Our findings indicate that participants who trained in AR displayed more efficient brain networks, suggesting improved neural efficiency.

Discussion: Differences in sex-related activation and connectivity hint at varying neural strategies used during motor learning in AR. Future studies should investigate how demographic factors might influence performance and user experience in AR-based training programs.