Ongoing & Completed
To uncover the mysteries of human behavior, AC-DC carries out trailblazing naturalistic studies. The study domains vary, but the majority are either driving or knowledge work studies, revealing how people interact with the two most widely used machines, that is, cars and computers.

Banner Image: Visualization of acceleration and the associated arousal of accelarousal-prone driver S07, shortly after her entry onto a highway. LEFT: The hyphenated blue circle on the map indicates the position associated with the figure's snapshots. MIDDLE: Visual images captured from a dash camera and a facial camera show the surrounding environment and the driver's facial expressions, respectively. RIGHT: Thermal facial image, where the red rectangle outlines the tracked region of interest (ROI). The preponderance of black dots on the zoomed in thermal ROI (see inset) suggests strong transient activation of perspiration pores, a phenomenon associated with the onset of hyperarousal.

Accelarousal may be a widespread phobia and a hidden commuter stressor, research finds

As the COVID 19 vaccination rates increase and the economy is opening up, levels of commuting are slowly picking up. Commuters’ biggest nightmare was always traffic congestion. Recent research, however, found that some commuters have a lot more to worry about.

Prof. Ioannis Pavlidis and his group from the University of Houston’s Computational Physiology Lab, in collaboration with the Texas A&M Transportation Institute, conducted a naturalistic study aiming to understand if there is any underlying grouping in the way drivers react to commonplace acceleration, speed, and steering events. Such events occur multiple times in the normal course of a daily commute, e.g., stop and go on a red light or entering the on ramp to merge onto a highway. The team recruited young but experienced drivers and asked them to drive a Toyota Sienna through a 19 km itinerary in College Station, TX. The driving, which lasted between 25 and 35 minutes, was always done in daytime, under fair weather and light traffic conditions. These ideal conditions were imposed to ensure the elimination of confounding factors.

Throughout the itinerary, the drivers’ physiological stress levels were measured every second by quantifying their perinasal perspiration via thermal imaging, and thus amassing thousands of measurements. At the same time, the vehicle’s computer was recording the acceleration, speed, brake force, and steering angle signals. Using Machine Learning (ML) techniques on these data, the researchers established that about half of the participants have consistently exhibited overarousal (i.e., elevated stress) in commonplace acceleration events, while the other half have not. The researchers named the former group accelaroused drivers, while the latter group non-accelaroused drivers. The differences between the two groups were significant, with accelaroused drivers exhibiting nearly 50% stronger stress reactions to mundane accelerations than non-accelaroused drivers. Moreover, psychometric measurements, taken through standardized questionnaires at the end of the drive, showed that accelaroused drivers felt more overloaded than non-accelaroused drivers. This was a clear indication that accelarousal was taking a toll on drivers.

“We were surprised to find such a clear dichotomy into drivers’ physiological reactions to these low intensity events”, Pavlidis said. “Apparently, drivers are not consciously aware of them”, Pavlidis added. “However, the difference in the physiological responses is highly significant”, Pavlidis emphasized. “This has all the characteristics of a long-term stressor, with all the health implications that this may entail”.

The advent of ubiquitous sensing and big data, makes the detection of such pernicious stressors possible. “Thanks to our research, we now have an understanding of accelarousal – a phobia that was hidden in plain sight”, Tung Huynh – a research assistant in the team, said. “Importantly, we have methods to measure and detect it. “The researchers believe that next generation vehicles will track the stress condition of drivers in correlation with driving variables. For instance, delivery drivers, which is an expanding class in the gig economy, are exposed to stop and go events all the time. Hence, delivery drivers who, unbeknownst to them, suffer from accelarousal, will have a way to detect this condition and account for its long-tern stress effects. “This is not much different than tracking radiation exposure for pilots”, Pavlidis said. “Our technology opens the way for such preventive health systems in cars”.

For the moment, and in the absence of such high-tech systems in current cars, watch out for signs that you have not paid much attention thus far. For example, do you feel that driving wears you down more than others in your circle? “This may be a telltale sign of accelarousal”, Pavlidis concluded.

Paper Presentation

INVESTIGATORS: Ioannis Pavlidis - Mike Manser

FUNDING: Toyota Class Action Settlement Safety Research and Education Program


INSTITUTIONS: Computational Physiology Laboratory, University of Houston, USA

                         Texas A&M Transportation Institute, Texas A&M University, USA

PAPER: To read the paper, go to: Huynh T. et al., CHI 2021

DATA: To access the dataset associated with the paper, go to:

GITHUB: To get the R code, go to: