Synthetic Assessment of Cognitive Load in Human-Machine Interaction Process

Assessing cognitive load on web search is useful for characterizing search system features and search tasks with respect to their demands on the searcher’s mental effort. It is also helpful for examining how individual differences among searchers (e.g. cognitive abilities) affect the search process. We examined cognitive load from the perspective of primary and secondary task performance. A controlled web search study was conducted with 48 participants. The primary task performance components were found to be significantly related to both the objective and the subjective task difficulty. However, the relationship between objective and subjective task difficulty and the secondary task performance measures was weaker than expected. The results indicate that the dual-task approach needs to be used with caution. INTRODUCTION AND BACKGROUND Web search behavior is affected by the task, system, and individual searcher characteristics. Search tasks and their characterization have been a subject of recent systematic studies [1-3]. In particular, researchers have focused on the effects of task complexity and difficulty on information search process [4-10]. One kind of difficulty encountered by searchers is related to mental, or cognitive, requirements imposed by the search system or the task itself. Understanding factors that contribute to user’s cognitive load on search tasks is crucial to identifying search system features and search tasks types that impose increased levels of load on users. As new interactive features are introduced into the information search systems we need to understand what determines their acceptance and why some evidently useful functions are not widely used. For example, user relevance feedback is a feature that has been reported to be avoided by users due to the heightened cognitive load [11]. Among other factors affecting search performance are the user’s cognitive characteristics (e.g., [12, 13]). Methods used to date in assessing cognitive load included searcher observation, self-reports (e.g., using questionnaires, think-aloud protocols, and post-search interviews), dual-task techniques [14, 16], and various approaches that employ external devices to collect additional data on users (e.g., eye-tracking, pressure-sensitive mouse and other physiological sensors [17]). The latter two groups of techniques have the advantage of enabling real-time, ontask data collection. However, the use of external devices can be expensive and impractical. Hence, the promise of dual-task (DT) method that allows for an indirect objective assessment of mental effort on the primary task. Only few studies employed this method to assess cognitive load in online search tasks (e.g., [14, 15, 18]). The article discusses the dual-task method as the technique for assessing cognitive load on web search tasks and presents research that *Address correspondence to this author at the Department of Library & Information Science, School of Communication and Information, Rutgers University, New Brunswick, NJ, USA; E-mail: WebCog@gwizdka.com contributes to better understanding of how objective task difficulty affects searchers’ behavior and their perception of task difficulty.

Automatic cognitive load evaluation using writing features: An exploratory study

A key for earcon design in public environments is to incorporate an individual’s perceived level of cognitive load for better communication. This study aimed to examine the cognitive load changes required to perform a melodic contour identification task (CIT). While healthy college students (N = 16) were presented with five CITs, behavioral (reaction time and accuracy) and cerebral hemodynamic responses were measured using functional near-infrared spectroscopy. Our behavioral findings showed a gradual increase in cognitive load from CIT1 to CIT3 followed by an abrupt increase between CIT4 (i.e., listening to two concurrent melodic contours in an alternating manner and identifying the direction of the target contour, p < 0.001) and CIT5 (i.e., listening to two concurrent melodic contours in a divided manner and identifying the directions of both contours, p < 0.001). Cerebral hemodynamic responses showed a congruent trend with behavioral findings. Specific to the frontopolar area (Brodmann’s area 10), oxygenated hemoglobin increased significantly between CIT4 and CIT5 (p < 0.05) while the level of deoxygenated hemoglobin decreased. Altogether, the findings indicate that the cognitive threshold for young adults (CIT5) and appropriate tuning of the relationship between timbre and pitch contour can lower the perceived cognitive load and, thus, can be an effective design strategy for earcon in a public environment.

Is Symbolic Gestural Interaction Better for the Visual Attention?

Interest in applying cognitive load theory in health care simulation is growing. This line of inquiry requires measures that are sensitive to changes in cognitive load arising from different instructional designs. Recently, mental effort ratings and secondary task performance have shown promise as measures of cognitive load in health care simulation. We investigate the sensitivity of these measures to predicted differences in intrinsic load arising from variations in task complexity and learner expertise during simulation-based surgical skills training. We randomly assigned 28 novice medical students to simulation training on a simple or complex surgical knot-tying task. Participants completed 13 practice trials, interspersed with computer-based video instruction. On trials 1, 5, 9 and 13, knot-tying performance was assessed using time and movement efficiency measures, and cognitive load was assessed using subjective rating of mental effort (SRME) and simple reaction time (SRT) on a vibrotactile stimulus-monitoring secondary task. Significant improvements in knot-tying performance (F(1.04,24.95) =41.1, p<0.001 for movements; F(1.04,25.90) =49.9, p<0.001 for time) and reduced cognitive load (F(2.3,58.5) =57.7, p<0.001 for SRME; F(1.8,47.3) =10.5, p<0.001 for SRT) were observed in both groups during training. The simple-task group demonstrated superior knot tying (F(1,24) =5.2, p=0.031 for movements; F(1,24) =6.5, p=0.017 for time) and a faster decline in SRME over the first five trials (F(1,26) = 6.45, p=0.017) compared with their peers. Although SRT followed a similar pattern, group differences were not statistically significant. Both secondary task performance and mental effort ratings are sensitive to changes in intrinsic load among novices engaged in simulation-based learning. These measures can be used to track cognitive load during skills training. Mental effort ratings are also sensitive to small differences in intrinsic load arising from variations in the physical complexity of a simulation task. The complementary nature of these subjective and objective measures suggests their combined use is advantageous in simulation instructional design research.

A novel approach using transient chaotic neural network (TCNN) and self-organizing feature map (SOFM) process to solve large-scale combinatorial optimization problems has been proposed. With the clustering function of self-organizing feature map, the computational cost of a large-scale combinatorial optimization problem solved by TCNN is reduced. Numerical simulation of TSP shows that the proposed method is effective to solve large-scale optimization problems.

In recent decades, voltage instability has become an important contributor to large-scale power outages globally. Periodic voltage security assessments are key for the secure operation of power systems. There are various indicators which have been implemented in different types and complexity of systems to provide a measure of closeness to voltage collapse, but some of these algorithms may require advanced computational resources or may not perform well, depending on the prevailing conditions. Expanding interconnected power systems is needed like never before due to the increasing world demand for electricity, and voltage stability is therefore a worth consideration. The growing demand penetration of solar photovoltaic (PV) systems has some potential stability challenges due to their intermittent characteristics. Existing research evaluates the test systems of voltage stability and does not optimize using Intelligent models, which have variable power generation from Solar PV and Wind, as well as uncertainties in load demand and the initial step considering the evaluation of the model have to cover either inputs. A feed–forward back propagation artificial neural network (ANN) for voltage stability assessment to classify the operating scenario of a power system into one of two states, i.e. safe or unsafe. Neural network input data is generated from load flow analysis using Newton-Raphson (NR), and stability indices are computed over different scenarios. Line Stability Index (Lmn), counts the stability of a transmission line in power systems and provides an indication of its potential voltage collapse. The Fast Voltage Stability Index (FVSI), which also estimates line stability through reactive power and impedance characteristics. Voltage Collapse Proximity Index (VCPI) assesses system proximity to voltage collapse and identifies weak transmission lines. Values approaching 1 suggest voltage collapse conditions. The FVSI, VCPI, and Lmn are compared to a baseline NR stability index. Moreover, self-organizing map (SOM) neural network is used for stability results classification. Multi-objective grasshopper optimization algorithm (MOGOA) has been employed as an optimization network on the dual axis solar PV and wind turbine-based hybrid power generation system to analyze technical, economic, as well as environmental relations of energy conversion. The findings reveal that the adapted grid-tied solar-wind composite power system substantially mitigates the power deficiencies, reduces the real-power outages, diminishes the motion voltage small amendments, maximizes voltage prosperity, and improves power system resilience. The main findings show a significant reduction in real power loss, which went from 289,391.30 kW to only 5.43 kW, and a remarkable decrease in voltage deviation, which reduced from 0.56 p.u to 0.0493 p.u. Ensures Improved Service Quality and Enhanced Equipment Protection While gas generation costs increased by a small amount, it is important to note that this is an acceptable trade-off for a more robust and sustainable system, with costs per kWh of $0.142 vs $0.2881 per kWh, respectively. In addition, stability index was substantially decreased from 0.7666 to 0.0768, indicating better grid and resilience to disturbances. Findings emphasize the contribution of embedding MOGOA in conjunction with machine learning models including ANN and Self Organizing Maps (SOM) to fulfill the primal requisites of stability, efficiency and sustainability in modernized power systems, illustrating a coupled advancement in the realm of power systems engineering and optimization.

A sample of 33 experiments was extracted from the Web-of-Science database over a 5-year period (2016–2020) that used physiological measures to measure intrinsic cognitive load. Only studies that required participants to solve tasks of varying complexities using a within-subjects design were included. The sample identified a number of different physiological measures obtained by recording signals from four main body categories (heart and lungs, eyes, skin, and brain), as well as subjective measures. The overall validity of the measures was assessed by examining construct validity and sensitivity. It was found that the vast majority of physiological measures had some level of validity, but varied considerably in sensitivity to detect subtle changes in intrinsic cognitive load. Validity was also influenced by the type of task. Eye-measures were found to be the most sensitive followed by the heart and lungs, skin, and brain. However, subjective measures had the highest levels of validity. It is concluded that a combination of physiological and subjective measures is most effective in detecting changes in intrinsic cognitive load.

Distance Interpreting (DI) is a form of technology-mediated interpreting which has gained traction due to the high demand for multilingual conferences, live-streaming programs, and public service sectors. The current study synthesized the DI literature to build a framework that represents the construct and measurement of cognitive load in DI. Two major areas of research were identified, i.e., causal factors and methods of measuring cognitive load. A number of causal factors that can induce change in cognitive load in DI were identified and reviewed. These included factors derived from tasks (e.g., mode of presentation), environment (e.g., booth type), and interpreters (e.g., technology awareness). In addition, four methods for measuring cognitive load in DI were identified and surveyed: subjective methods, performance methods, analytical methods, and psycho-physiological methods. Together, the causal factors and measurement methods provide a multifarious approach to delineating and quantifying cognitive load in DI. This multidimensional framework can be applied as a tool for pedagogical design in interpreting programs at both the undergraduate and graduate levels. It can also provide implications for other fields of educational psychology and language learning and assessment.

Driving in monotonous road environments may cause drivers to become bored and impair their alertness. This inattention may increase the risk of road accidents. Drivers usually respond to visual signals on the road. A large body of research has shown that the level of a driver’s visual attention is closely related to traffic accidents. Visual attention is often studied by tracking drivers’ eye movements. However, few studies to date have explored eye-movement patterns in different road environments such as open roads and tunnel expressways. This dissertation aims to address that research gap by studying drivers’ eye movements and visual characteristics in different driving conditions in real and simulated environments. In the pilot study, drivers’ eye movement patterns were recorded during high-speed driving in tunnels and on expressways in Singapore. Twenty-two drivers participated in the study; they drove a total of 55 km, which included a 9-km section of tunnel. The results of the study showed that drivers experienced longer fixation duration when driving in tunnels than on the open road but experienced a higher number of fixations, higher instantaneous velocity, acceleration, and dispersion on open roads than in tunnels. These findings suggest that drivers’ eye movements are more concentrated and have longer fixation durations in tunnels than in open road driving, indicating that drivers have a greater mental workload when driving in tunnels. However, testing drivers in a real environment has limitations. For example, weather and road characteristics cannot be manipulated. To address this limitation, Experiment 1 was conducted in a driving simulator using virtual reality simulations in UC-win/road. Experiment 1 sought to investigate drivers’ eye movements in various driving environments. Eighteen drivers participated in the experiment and drove on normal roads, expressways, and through underground tunnels in both sunny and rainy weather. The comparisons of their eye-movement patterns (e.g., saccadic parameters and pupil size) in the simulated environments were consistent with those in real open-road environments and tunnels in sunny conditions. The findings suggest that the eye-movement patterns of drivers in a simulator are comparable to those in real environments, indicating its ecological validity. Previous literature has suggested a relationship between eye movement patterns and mental workload. To verify that relationship, mental workloads in Experiments 2 and 3 were directly measured using questionnaires. Dynamic and static hazardous scenarios were simulated to induce different levels of mental workload. Experiment 2 investigated eye movements, driving performance, and mental workload of drivers who encountered static and dynamic obstacles while manual driving. The results from 23 drivers found that their mental workload in dynamic conditions was higher than in static hazardous conditions; it was also higher before encountering hazards than afterwards. The results indicated that pupil size and saccade numbers are effective indicators of mental workload. Experiment 3 was conducted under conditionally automated driving conditions, where participants were involved in tasks related to or not related to driving and requested to resume control manually when their vehicle was in semi-automated driving mode. Drivers’ eye movements and mental workloads were measured when they were driving, especially during the transitions from semi-automated to manual driving. Similar to Experiment 2, the results from 32 drivers found that mental workload was higher for dynamic hazards than for static hazards and higher before hazards were encountered than after. The results also indicated that pupil size is the only effective indicator of mental workload. Taken together, the results of these experiments reveal that pupil size is the most reliable way to measure drivers’ mental workload. Furthermore, in line with the load theory of attention and De Waard (1996)’s theory of workload and performance, it can be concluded that mental workload (indicated by pupil size) and visual workload (by saccadic dispersion) are the key determinants of driving performance and that drivers can only respond to emergencies during periods of medium levels of mental and visual workload (not underload or overload) with available attentional capacity. The findings from this thesis on eye movement and mental workload may be used to monitor drivers’ real-time attention and workload and thus provide guidance for the design of safe-driving autonomous vehicles that consider drivers’ capacity for information processing when driving. Knowledge about the limitations and capacity of human drivers in tandem with automated systems may also help improve user acceptance of autonomous vehicles in the future.

Background Pupillometry has been studied as a physiological marker for quantifying cognitive load since the early 1960s. It has been established that small changes in pupillary size can provide an index of the cognitive load of a participant as he/she performs a mental task. The utility of pupillometry as a measure of expertise is less well established, although recent research in the fields of education, medicine and psychology indicates that differences in pupillary size during domain-specific tasks allows differentiation between experts and novices in appropriately designed experiments. Purpose The goal of this review is to explore the existing body of evidence for the use of pupillometry as a measure of expertise and to identify its strengths and constraints within the context of expertise research in the medical sciences. Results Pupillometry is a robust metric that allows researchers to better understand cognitive load in medical practitioners with varying levels of expertise. In medical expertise research, it has been used to study surgeons, anesthetists and emergency physicians. Its strengths include its ability to provide quantitative and objective outputs, to be measured unobtrusively with new technology and to be precisely computed as cognitive load changes over the course of completion of a task. Constraints associated with this methodology include its potential inaccuracy with changes in ambient light and pupillary accommodation as well as the need for relatively expensive equipment. Conclusion With recent technological advances, pupillometry has become a simple and robust method for quantifying physiological changes attributable to cognitive load and is increasingly being utilized in medical education. It can be used as a reliable marker of mental effort and has been shown to differentiate levels of expertise in medical practitioners.

The effects of different interruption conditions on mental workload: an experimental study based on multimodal measurements

The purpose of this study was to investigate the efficiency of using magnesium oxide nanoparticles, twisted tapes with different and modified twist ratios, and the different rotation velocity of perforated twisted tapes in the evaluation of Nusselt number, overall heat transfer coefficient and pressure drop in a double pipe heat exchanger. Subsequently, the results obtained from the experimental tests of Nusselt number, overall heat transfer coefficient and pressure drop using a modified twisted tape in a double pipe heat exchanger and a back propagation artificial neural network with a multilayered perceptron structure show the existence of a relationship between heat transfer parameters and input data such as cold and hot fluid temperatures, volume fractions of nanofluid, twisted tapes of different twist ratios, the rotation velocity of twisted tapes, Reynolds number of hot and cold fluids, viscosity and thermal conductivity of the nanofluid. Also, the non-monitored artificial neural network selected 33 neurons for use in the multilayered perceptron neural network. The network included the sigmoid function in the hidden layer and the Levenberg–Marquardt training algorithm with a three-layer topology of 9–33-1 with the correlation coefficients of 0.9987, 0.9933 and 0.9944 and the mean squared error of 0.66535, 4547.68 and 0.008768. This network was best for predicting heat transfer parameters such as Nusselt number, overall heat transfer coefficient and pressure drop.

Influence of Cognitive Load on Voice Production: A Scoping Review

The rising popularity of commercial anatomy e-learning tools has been sustained, in part, due to increased annual enrollment and a reduction in laboratory hours across educational institutions. While e-learning tools continue to gain popularity, the research methodologies used to investigate their impact on learning remain imprecise. As new user interfaces are introduced, it is critical to understand how functionality can influence the load placed on a student's memory resources, also known as cognitive load. To study cognitive load, a dual-task paradigm wherein a learner performs two tasks simultaneously is often used, however, its application within educational research remains uncommon. Using previous paradigms as a guide, a dual-task methodology was developed to assess the cognitive load imposed by two commercial anatomical e-learning tools. Results indicate that the standard dual-task paradigm, as described in the literature, is insensitive to the cognitive load disparities across e-learning tool interfaces. Confounding variables included automation of responses, task performance tradeoff, and poor understanding of primary task cognitive load requirements, leading to unreliable quantitative results. By modifying the secondary task from a basic visual response to a more cognitively demanding task, such as a modified Stroop test, the automation of secondary task responses can be reduced. Furthermore, by recording baseline measures for the primary task as well as the secondary task, it is possible for task performance tradeoff to be detected. Lastly, it is imperative that the cognitive load of the primary task be designed such that it does not overwhelm the individual's ability to learn new material.