Apple could be thinking about designing, in the future, AirPods capable of reading certain signals from the brain thanks to advanced sensors and artificial intelligence models.
Apple research presents a method that allows an AI model to learn patterns of brain electrical activity without the need for previously annotated data.
This approach, called PARS (PAirwise Relative Shift), is based on self-learning. Instead of relying on labeled data indicating sleep stages or epilepsy episodes, PARS trains the model to predict the temporal distance between different segments of EEG signals.
This allows the system to understand the overall structure of brain activity, improving accuracy in tasks such as detecting sleep or identifying neurological abnormalities.
In the tests carried out, the models trained with PARS outperformed or matched the performance of previous methods on multiple data sets. A key part of the study used ear EEG which records brain activity from the ear rather than the scalp. This reduces the visibility of the electrodes and improves user comfort.
Apple had already applied for a patent for portable devices in 2023 capable of measuring biological signals from the ear. The patent describes how multiple electrodes distributed in the headphones could capture the best quality signals using artificial intelligence, optimizing the measurement and combining the data to generate a single signal.
Although the current study does not mention AirPods and is not directly linked to the 2023 patent, It does give an idea of where Cupertino engineers could head to explore advances that could be applied to future AirPods or similar devices..
For now, it’s just research and experimentation. However, if the results continue to be promising, it would not be unreasonable to think that we could see some product of this type in the next decade.