Skip to content Skip to navigation

Stanford Cars and Brains Project


In a previous blog post, I tried to persuade you that all automated vehicle research projects require a team of neuroscientists. My main argument was fairly high-level; we expect to have a human in the loop, at least under some driving conditions, for the foreseeable future, and that understanding the neurobiological underpinnings of how the humans interact with an automated systems is vital when designing a system, that supports, rather than distracts or confuses the user.

I would (and do!) argue that studying the cognitive processes underlying driver (or passenger) behavior is vital for the goal of making a person’s driving, transportation, or mobility experience safer and more enjoyable.

In my opinion, some of the most interesting examples of areas in which neuroimaging can help in design principles for automated vehicles and future transportation systems are:
· Identifying driving or transportation scenarios that demand increased mental workload from the user, and hence focusing automated and assistive technologies on those scenarios

· Evaluating automated systems, car-driver communication channels, and ADAS in terms of their effect on the user’s cognitive load, to ensure that the system, designed to reduce mental workload, does not induce the opposite effect

· Personalizing or adapting the experience to the specific needs, preferences, or goals of a particular user, or population of users, in a given situation

Today, I would like to focus on the final example, as I firmly believe that there a huge opportunity, as mobility becomes more automated, to tailor the transportation experience to a particular user, and some of the recent results from the Stanford CAB (Cars And Brains) Project are especially exciting and relevant to this idea.

We have found that particular personality traits (assessed by the standard NEO five-factor personality inventory) correlate with different approaches to a standard maneuver (double lane-change) under challenging, high cognitive load conditions. In particular, a subset of drivers appear to rely more on feedback information to successfully complete a driving maneuver, while others tend to favor feed forward control as a strategy. These differences are associated with significantly different scores on the “Big Five” personality traits (openness, conscientiousness, extraversion, agreeableness, and neuroticism).

Intuitively, it makes sense that someone’s fixed personality traits, in addition to their more transient emotional state, affect their driving style, but quantifying it in this way gives rise to the intriguing possibility of being able to predict a person’s preference for the design of an automated transportation system, and their interaction with it, from their personality profile.

But, I hear you cry; where does the neuroscience, which you are so relentlessly championing, come into play? It turns out that our latest research shows that the different personality traits, mentioned above, are also correlated with different patterns of cortical activity during distinct portions of the driving maneuver, further suggesting that we can identify different driver populations, not only by their behavior when interacting with automated systems, but also by the neurocognitive resources they employ while doing so.

The results of research programs such as the Stanford CAB Project contribute to a goal very close to my heart; that automation makes mobility and transportation more accessible, useable, and enjoyable to more people. Whether it is the ageing population, who’s reaction time and visual acuity is declining; populations with traumatic brain injury, such as veterans, or simply the driver who wishes for full automation during a tedious commute, but wants to drive for fun along the coastal road on a sunny weekend, our results add weight to the argument that one-size does not fit all when it comes to mobility, and help inform methods aimed at tailoring the transportation solution to particular users in the future.