Machine Learning is an incredible technology and it is taking the world towards a whole new direction in leaps and bounds. It has found usage in education as well as healthcare industries – and almost every day, new technologies are coming out. German scientists have now come up with a system which can identify what song is someone listening to by mapping their brain activity.
In an experiment conducted by German scientists, six volunteers were played 40 pieces of music ranging across different genres such as classical, rock, pop, jazz, etc. The MRI machine mapped the neural signature of each of those songs on their brains. A computer was simultaneously mapping the patterns that the brain was showing while listening to these songs.
The system took into account the effect on the brain as well as the sound that was being played. Factors such as the tonality, rhythm, timbre and dynamics were considered. Researchers who tested the accuracy of the system provided it two options at first, and then ten options to understand what the subjects were listening to. The accuracy rate when there were two options was 85% and it dropped to 74% when there were ten options.
This is an elementary level of research that is being carried out in this field. With time and more research, this technology can potentially create a possibility of communicating without the need for a defined ‘language’ that needs to be spoken or written. This research helps scientists understand neural functioning better.
This is a major step forward towards the research of auditory imagination and inner voice experiments. A practical use case of this technology could possibly be of communicating to patients suffering through the locked-in syndrome. With further research, this technology can further enhance the brain-computer interface. While it may sound trivial as of now, this can possibly be the foundation of a major breakthrough in the way we know communication to be!