Apple develops brain signal reader

By Alimat Aliyeva

The American company Apple has developed a technology codenamed PARS for analyzing electroencephalograms (EEG) using artificial intelligence and headphones, Azernews reports.

Unlike traditional methods, which require costly manual annotation of data by specialists, the new system uses a self-learning algorithm based on raw, unmarked data. The algorithm learns to predict time intervals between different brainwave segments, enabling the detection of deep structures of neural activity.

The technology has already demonstrated impressive results in tests, surpassing or matching modern methods in three of the four standard EEG datasets. Of particular interest is the use of the EESM17 dataset, which contains information collected using intra-auricular brain activity monitoring systems during sleep. This confirms the ability to record key neurological indicators—including sleep stages and epileptic patterns—through sensors located directly in the ear canal.

Apple’s 2023 patent documentation details the technical implementation: the system places an array of redundant electrodes around the AirPods’ ear pads, where intelligent algorithms select in real time the sensor combinations with the best signal. This approach solves the problem of maintaining constant skin contact, taking into account the unique anatomy of each ear.

Future versions of PARS could potentially do more than track sleep or epileptic events. They might also monitor a user’s emotional state or fatigue levels in real time—directly through their headphones. This opens the door to a highly personalized “smart” audio assistant that adapts to your brain activity throughout the day.