US researchers have made a breakthrough in detecting what kind of text subjects are listening to or imagining silently by monitoring their brainwaves. The study, published in Nature Neuroscience, used functional magnetic resonance imaging (fMRI) to identify blood flow changes linked to neuronal impulses. While fMRI imaging is slow compared to other brain imaging techniques, it was able to extract semantics and meaning, a level of detail unseen before. The researchers now want to see how the technology can be applied more widely, including for the development of communication aids for severely paralyzed patients.
The Texas-based team behind the study say they were able to reproduce the content of a silent film that subjects watched and decode previously unknown texts from fMRI data. However, they emphasised that decoding brain activity is dependent on the test subjects’ willingness to cooperate, without which the method does not work. Indeed, the researchers stress that the results cannot be generalised, because the method has only been tested on three subjects. A notable advantage is that fMRI can provide high spatial resolution.
The researchers used stories sourced from the storytelling website The Moth, which they said were chosen because they covered a wide range of topics and were entertaining. However, to preserve accuracy, they discouraged distraction by subjects while the stories were being related. The study also found that individual subjects had to be specifically trained for the monitoring to work. While the technology is still heavily dependent on MRI scanners, the researchers are optimistic that future developments will make it less costly and easier to apply.
