Towards real-time detection of cognitive effort in driving: contribution of a cardiac measurement

Abstract

In 2013, attention deficits accounted for 40 to 50 % of injury accidents. Recent studies have succeeded in detecting impaired states of attention, with a view to assisting the driver, and provide a new opportunity to increase road safety. This study focuses on the detection of drivers' cognitive effort and seeks, through the study of heart rate change (HRC), to identify a sensitive indicator of cognitive effort in short time windows. Eighteen young drivers participated in the study and took part in 8 experimental sessions where they performed a passive or active cognitive task (counting) while driving or not. The counting task had two difficulty levels (counting of beeps vs visuospatial skills and number adding). Participants' heart rates were monitored during all tasks. Previous results recorded in laboratory conditions have been replicated during driving: during the first seconds after a cognitive effort, there is a slight deceleration and a sharp acceleration in heart rate. Conversely, in the absence of cognitive effort, simple cardiac deceleration was observed. Our study confirms that it is possible to distinguish HRC in response to a cognitive effort over short time windows by observing the grand mean of evoked cardiac responses at 0.5 s intervals from stimulus onset when averaged over a significant number of episodes. The new opportunities offered with this cognitive effort indicator are discussed. Recent literature data show that the removal of respiratory influence from heart rate is feasible. With such correction, it seems possible to improve the sensitivity of HRC, and HR acceleration should be observed without averaging the HRC over many trials. If this proves effective, using an algorithm to detect cognitive effort in real time, future assistance devices could warn drivers or overcome their mistakes when they no longer control driving activity because of a cognitive effort.

Respiration and Heart Rate Modulation Due to Competing Cognitive Tasks While Driving

“…In order to minimize the influence of the driving scenario position and driving style, the considered mean values take into consideration only 6 s [following the parameters employed in Pépin et al (2017)] after every stimulus presentation and averaged for all the task duration.…”

Do distinct mind wandering differently disrupt drivers? Interpretation of physiological and behavioral pattern with a data triangulation method

“…This switchover could have a cost in terms of attentional resources. Yet, a recent study has showed the feasibility of detecting the cognitive effort generating by a cognitive demand of a task through heart rate measurement (Pepin et al, 2017). It appears that the specific pattern found here after participants self-reported their MW episodes could correspond to a cognitive effort pattern.…”

“…Heart Rate (HR) values were computed using equations provided by Roy, (2015) and illustrated by Pepin et al, (2017). From the cardiac signal, several time-domain indices bringing information about heart rate variability have been calculated from IBI such as the Root Mean Square Successive Differences of NN intervals (RMSSD) and proportion of the number of interval differences of successive NN intervals greater than 50 ms (pNN50).…”

“…It appears that the specific pattern found here after participants self-reported their MW episodes could correspond to a cognitive effort pattern. Time-windows used for this experiment and in Pepin et al, (2017) are similar. As previously explained, it seems likely that this increase in HR is responsible for the statistical significance highlighted here.…”

Variation of performance, of self-reported alertness and effort as a function of low doses of alcohol and of driving experience

“…This suggests that task demand refers to the intrinsic features of the task while task load describes the subjective impact on the operator [6] and refers to the perceived relationship between the amount of mental processing capability or resources and the amount of resources required by the task. Therefore, measuring workload requires taking into account factors characterizing the task and factors describing how the driver engages in and allocates his/her resources for the processing of the task [47]. In the resource perspective, the level of the driver skill, e.g.…”