AUTONEWS

Passengers' brain signals may help self-driving cars make safer choices

Cars from companies like Tesla already promise hands-free driving, but recent crashes show that today's self-driving systems can still struggle in risky, fast-changing situations.

Now, researchers say the next safety upgrade may come from an unexpected source: The brains of the people riding inside those cars.

In a new study appearing in Cyborg and Bionic Systems, Chinese researchers tested whether monitoring passengers' brain activity could help self-driving systems make safer decisions in risky situations.

The team used a noninvasive technology called functional Near-Infrared Spectroscopy (fNIRS), which tracks brain activity linked to stress, emotions and risk perception in real time.

"Functional Near-Infrared Spectroscopy (fNIRS), as a non-invasive real-time brain activity monitoring method, can provide cognitive information related to human risk perception and emotional states, and is thus considered a tool that can enhance autonomous driving systems," lead author Xiaofei Zhang, a professor at Tsinghua University in Beijing, said in a news release.

"Our study introduces an intelligent decision-making algorithm based on fNIRS by analyzing passengers' physiological states, aiming to improve the safety and decision-making efficiency of autonomous vehicles when facing risky scenarios," Zhang added.

The researchers built a new system that combines brain data from passengers with an autonomous vehicle's driving software.

When the system detects that passengers are experiencing higher levels of risk or stress, the vehicle automatically shifts to a more cautious driving strategy.

Based on a form of deep reinforcement learning, the algorithm is designed to learn faster and make safer choices by factoring in human reactions.

In tests, the system switched to a more conservative driving mode when passengers appeared uneasy, helping the vehicle respond more carefully in dangerous situations.

The study found that this approach outperformed traditional autonomous driving methods in several areas, including learning speed, overall safety and comfort.

However, the researchers noted some limits. The driving scenarios tested were relatively simple, and participants came from a narrow age range and similar backgrounds. Because of this, the findings may not apply to all real-world driving situations.

"Future research will aim to validate the algorithm in more complex and realistic driving scenarios and further enhance the accuracy and robustness of driving risk assessment by integrating information from vehicle sensors," Zhang said.

Recent research published in December 2025 indicates that self-driving cars can become safer by integrating passengers' brain signals into their decision-making systems. 

How the technology works:

Brain Monitoring: The system uses non-invasive functional Near-Infrared Spectroscopy (fNIRS) to track brain activity related to stress, emotion, and risk perception in real time.

Adaptive Driving: When the vehicle's AI detects that a passenger is experiencing high levels of stress or unease, it automatically shifts to a more cautious driving strategy.

Early Warning Signals: Electroencephalogram (EEG) studies have also shown that passengers' brains emit "early warning signals" in the frontal lobe up to two seconds before a risky event, such as another car cutting into their lane. 

Key benefits and findings:

Faster Learning: The algorithm, which uses deep reinforcement learning, learns to make safer choices faster by factoring in human reactions compared to traditional autonomous driving methods.

Enhanced Safety and Comfort: In tests, these systems outperformed standard autonomous models in learning speed, overall safety, and passenger comfort.

Predictive Accuracy: Recent EEG-driven models have achieved up to 93.9% accuracy in predicting potential risks based on passenger brain signals. 

Current limitations(below)

Simplicity of Tests: Current research has mostly used simple driving scenarios and participants from narrow age ranges.

Real-World Application: Researchers caution that more testing is needed to validate these findings in complex, real-world driving environments before the technology can be widely deployed.

2025 HealthDay. All rights reserved.