Real-time performance modelling of a Sustained Attention to Response Task

Automated identification of fall-related injuries in unstructured clinical notes

Fall-related injuries (FRIs) are a major cause of hospitalizations among older patients, but identifying them in unstructured clinical notes poses challenges for large-scale research. In this study, we developed and evaluated natural language processing (NLP) models to address this issue. We utilized all available clinical notes from the Mass General Brigham health-care system for 2100 older adults, identifying 154 949 paragraphs of interest through automatic scanning for FRI-related keywords. Two clinical experts directly labeled 5000 paragraphs to generate benchmark-standard labels, while 3689 validated patterns were annotated, indirectly labeling 93 157 paragraphs as validated-standard labels. Five NLP models, including vanilla bidirectional encoder representations from transformers (BERT), the robustly optimized BERT approach (RoBERTa), ClinicalBERT, DistilBERT, and support vector machine (SVM), were trained using 2000 benchmark paragraphs and all validated paragraphs. BERT-based models were trained in 3 stages: masked language modeling, general boolean question-answering, and question-answering for FRIs. For validation, 500 benchmark paragraphs were used, and the remaining 2500 were used for testing. Performance metrics (precision, recall, F1 scores, area under the receiver operating characteristic curve [AUROC], and area under the precision-recall [AUPR] curve) were employed by comparison, with RoBERTa showing the best performance. Precision was 0.90 (95% CI, 0.88-0.91), recall was 0.91 (95% CI, 0.90-0.93), the F1 score was 0.91 (95% CI, 0.89-0.92), and the AUROC and AUPR curves were [both??] 0.96 (95% CI, 0.95-0.97). These NLP models accurately identify FRIs from unstructured clinical notes, potentially enhancing clinical-notes-based research efficiency.

A meta-analysis on the relationship between subjective cognitive failures as measured by the cognitive failures questionnaire (CFQ) and objective performance on executive function tasks

The Cognitive Failures Questionnaire (CFQ) has been widely used as a measure of subjective cognitive function in everyday life for decades. However, the evidence on how it relates to objective performance on executive function tasks is mixed. One possible reason for these mixed results is that the CFQ has selective relationships with some aspects of executive function and not others. Here, therefore, we classified tasks according to an influential framework of executive functions-switching, updating, inhibition, and we also considered the Sustained Attention to Response Task (SART) as a category because it was custom designed to gauge cognitive failures. We synthesized a large body of available evidence and performed four Bayesian meta-analyses on the relationship between CFQ scores and objective performance on executive function tasks in these four categories. Results suggested that CFQ scores were associated with objective performance on SART (18 effect sizes, μ = -.19, BF10 = 18.03, i.e., 18.03 times more evidence of a relationship versus no relationship), updating working memory (49 effect sizes, μ = -.06, BF10 = 17.80), and inhibition tasks (41 effect sizes, μ = -.07, BF10 = 15.40), whereas there was not definitive evidence regarding switching (34 effect sizes, μ = -.06, BF10 = .50, i.e., two times greater evidence for no relationship). This suggests that subjective cognitive function can predict objective performance on at least some executive function tasks. We discuss methodological and theoretical factors that constrain the maximum observable correlation and consider the relative insights that subjective measures versus task performance provide.

Decisional style, sleepiness, and online responsiveness

As sleep problems can impair quality of work, an online questionnaire was used to examine relationships between sleepiness and decision making while obtaining unobtrusive indices of performance. Participants (N = 344) completed the Insomnia Severity Index, Epworth Sleepiness Scale, and the Melbourne Decision Making Questionnaire in a Qualtrics survey while reporting mobile phone use. Qualtrics recorded the time and the number of clicks required to complete each page of the survey. Multiple regression indicated that insomnia was associated with daytime sleepiness and Hypervigilance, and mobile phone use before bed. Participants with moderate sleepiness required a greater number of clicks to complete the questionnaire. Greater sleepiness was associated with longer times to complete these self-assessment tasks. Clinically significant sleepiness produces changes in performance that can be detected from online responsivity. As sleepy individuals can be appreciably and quantitatively slower in performing subjective self-assessment tasks, this argues for objective measures of sleepiness and automated interventions and the design of systems that allow better quality sleep.Practitioner summary: Work can require processing of electronic messages, but 24/7 accessibility increases workload, causes fatigue and potentially creates security risks. Although most studies use people's self-reports, this study monitors time and clicks required to complete self-assessment rating scales. Sleepiness affected online responsivity, decreasing online accuracy and increasing response times and hypervigilance.

A review of visual sustained attention: neural mechanisms and computational models

Sustained attention is one of the basic abilities of humans to maintain concentration on relevant information while ignoring irrelevant information over extended periods. The purpose of the review is to provide insight into how to integrate neural mechanisms of sustained attention with computational models to facilitate research and application. Although many studies have assessed attention, the evaluation of humans' sustained attention is not sufficiently comprehensive. Hence, this study provides a current review on both neural mechanisms and computational models of visual sustained attention. We first review models, measurements, and neural mechanisms of sustained attention and propose plausible neural pathways for visual sustained attention. Next, we analyze and compare the different computational models of sustained attention that the previous reviews have not systematically summarized. We then provide computational models for automatically detecting vigilance states and evaluation of sustained attention. Finally, we outline possible future trends in the research field of sustained attention.

Need for cognition moderates the impairment of decision making caused by nightshift work in nurses

The current study explores the effect of nightshift work on the decision-making competence and performance of the Iowa Gambling Task (IGT) and analyzes whether individual differences in the need for cognition (NFC) can moderate this effect. A total of 107 female nurses were recruited to complete the decision-making competence scale and IGT at two times, after a night shift and after a day shift. The results revealed that the IGT scores and decision-making competence of nurses after nightshift work significantly declined, and also that the decrease in decision-making competence was related to the nurses’ performance of the IGT. Additionally, the decreasing degree of IGT and decision-making competence scores of the high-NFC group were significantly lower than those of the low-NFC group after nightshift work. In can be concluded that the decrease in decision-making competence which was related with poor decision-making due to nightshift work. NFC moderated the effect of nightshift work on decision-making.

Drowsiness or mind-wandering? Fluctuations in ocular parameters during attentional lapses

Two independent lines of evidence suggest that drowsiness and mind-wandering share common neurocognitive processes indexed by ocular parameters (e.g., eyeblink frequency and pupil dynamics). Mind-wandering and drowsiness frequently co-occur, however, such that it remains unclear whether observed oculometric variations are related to mind-wandering, drowsiness, or a mix of both. To address this issue, we assessed fluctuations in mind-wandering and sleepiness during a sustained attention task while ocular parameters were recorded. Results showed that oculometric variations during mind-wandering were fully explained by increased sleepiness. However, mind-wandering and sleepiness had additive deleterious effects on performance that were not fully explained by ocular parameters. These findings suggest that oculometric variations during task performance reflect increased drowsiness rather than processes specifically involved in mind-wandering, and that the neurocognitive processes indexed by oculometric parameters (e.g., regulatory processes of the locus coeruleus norepinephrine system) do not fully explain how mind-wandering and sleepiness cause attentional lapses.

Towards continuous and real-time attention monitoring at work: reaction time versus brain response

Continuous and objective measurement of the user attention state still represents a major challenge in the ergonomics research. Recently available wearable electroencephalography (EEG) opens new opportunities for objective and continuous evaluation of operators' attention, which may provide a new paradigm in ergonomics. In this study, wearable EEG was recorded during simulated assembly operation, with the aim to analyse P300 event-related potential component, which provides reliable information on attention processing. In parallel, reaction times (RTs) were recorded and the correlation between these two attention-related modalities was investigated. Negative correlation between P300 amplitudes and RTs has been observed on the group level (p < .001). However, on the individual level, the obtained correlations were not consistent. As a result, we propose the P300 amplitude for accurate attention monitoring in ergonomics research. On the other hand, no significant correlation between RTs and P300 latency was found on group, neither on individual level. Practitioner Summary: Ergonomic studies of assembly operations mainly investigated physical aspects, while mental states of the assemblers were not sufficiently addressed. Presented study aims at attention tracking, using realistic workplace replica. It is shown that drops in attention could be successfully traced only by direct brainwave observation, using wireless electroencephalographic measurements.

Rapid changes in arousal states of healthy volunteers during robot-assisted gait training: a quantitative time-series electroencephalography study

Background

Robot-assisted gait training (RAGT) is expected to be an effective rehabilitative intervention for patients with gait disturbances. However, the monotonous gait pattern provided by robotic guidance tends to induce sleepiness, and the resultant decreased arousal during RAGT may negatively affect gait training progress. This study assessed electroencephalography (EEG)-based, objective sleepiness during RAGT and examined whether verbal or nonverbal warning sounds are effective stimuli for counteracting such sleepiness.

Methods

Twelve healthy men walked on a treadmill for 6 min, while being guided by a Gait-Assistance Robot, under 3 experimental conditions: with sine-wave sound stimulation (SS), verbal sound stimulation (VS), and no sound stimulation (NS). The volunteers were provided with warning sound stimulation at 4 min (ST1), 4 min 20 s (ST2), 4 min 40 s (ST3), and 5 min (ST4) after the start of RAGT. EEGs were recorded at the central (Cz) and occipital (O1 and O2) regions (International 10–20 system) before and during RAGT, and 4-s segments of EEG data were extracted from the filtered data during the 8 experimental periods: middle of the eyes-closed condition; middle of the eyes-open condition; beginning of RAGT; immediately before ST1; immediately after ST1, ST2, ST3, and ST4. According to the method used in the Karolinska drowsiness test, the power densities of the theta, alpha 1, and alpha 2 bands were calculated as indices of objective sleepiness.

Results

Comparisons of the findings between baseline and before ST showed that the power densities of the alpha 1 and 2 bands tended to increase, whereas the theta power density increased significantly (P < .05). During NS, the power densities remained at a constant high level until after ST4. During SS and VS, the power densities were attenuated immediately to the same degree and maintained at a constant low level until after ST4.

Conclusions

This study is the first to demonstrate that EEG-measured arousal levels decrease within a short time during RAGT, but are restored and maintained by intermittent warning sound stimulation.

Statistical modelling of networked human-automation performance using working memory capacity

This study examines the challenging problem of modelling the interaction between individual attentional limitations and decision-making performance in networked human-automation system tasks. Analysis of real experimental data from a task involving networked supervision of multiple unmanned aerial vehicles by human participants shows that both task load and network message quality affect performance, but that these effects are modulated by individual differences in working memory (WM) capacity. These insights were used to assess three statistical approaches for modelling and making predictions with real experimental networked supervisory performance data: classical linear regression, non-parametric Gaussian processes and probabilistic Bayesian networks. It is shown that each of these approaches can help designers of networked human-automated systems cope with various uncertainties in order to accommodate future users by linking expected operating conditions and performance from real experimental data to observable cognitive traits like WM capacity. Practitioner Summary: Working memory (WM) capacity helps account for inter-individual variability in operator performance in networked unmanned aerial vehicle supervisory tasks. This is useful for reliable performance prediction near experimental conditions via linear models; robust statistical prediction beyond experimental conditions via Gaussian process models and probabilistic inference about unknown task conditions/WM capacities via Bayesian network models.

Effects of sleep deprivation on decisional support utilisation

To inform development of decisional support systems for the sleep deprived, this study examined the effect of sleep debt, time pressure and risk on the ability to use a decision aid. A total of 19 participants were tested when well rested and sleep deprived. Participants played computerised forms of Blackjack, which varied a 1- or 4-second response deadline, at two levels of risk, and could be supplied with online advice. Mean bets served as indications of confidence. Although confidence was less when play was fast or higher risk participants did not bet significantly less when sleep deprived, suggesting an impaired calibration of judgement that was supported by evidence of rallying. This failure to adjust confidence was accompanied by slower responses at low risk when sleep deprived. Sleep-deprived participants were less able to use decisional support under time pressure and made more errors without advice and time pressure.

PRACTITIONER SUMMARY

Decisional support is becoming more pervasive. To inform development of decisional support systems to assist the sleep deprived, an experiment considered the use of decisional support as a function of time pressure and risk. Advisory systems require processing and will be less efficacious under time pressure when sleep deprived.

Driving performance impairments due to hypovigilance on monotonous roads

Drivers' ability to react to unpredictable events deteriorates when exposed to highly predictable and uneventful driving tasks. Highway design reduces the driving task mainly to a lane-keeping manoeuvre. Such a task is monotonous, providing little stimulation and this contributes to crashes due to inattention. Research has shown that driver's hypovigilance can be assessed with EEG measurements and that driving performance is impaired during prolonged monotonous driving tasks. This paper aims to show that two dimensions of monotony - namely road design and road side variability - decrease vigilance and impair driving performance. This is the first study correlating hypovigilance and driver performance in varied monotonous conditions, particularly on a short time scale (a few seconds). We induced vigilance decrement as assessed with an EEG during a monotonous driving simulator experiment. Road monotony was varied through both road design and road side variability. The driver's decrease in vigilance occurred due to both road design and road scenery monotony and almost independently of the driver's sensation seeking level. Such impairment was also correlated to observable measurements from the driver, the car and the environment. During periods of hypovigilance, the driving performance impairment affected lane positioning, time to lane crossing, blink frequency, heart rate variability and non-specific electrodermal response rates. This work lays the foundation for the development of an in-vehicle device preventing hypovigilance crashes on monotonous roads.