Simple and Honest EEG Drawing
Simple and Honest EEG Drawing has been built by Naoto Hieda and its prototype was exhibited at Axon Art Collective 2015 and DIYBio/NeuroTechTO 2015. Unlike brain-computer interface systems that understand what subjects are thinking, our system shows brain activities in a fuzzy, constellation metaphor to make subjects understand the system. To do so, electroencephalography (EEG) which represents how focused the subject is measured, the data is processed by a machine learning algorithm, and the processed “fuzzy” data is output in a dome as graphics, audio and robotic motion.
For real-time EEG measurement, we use a Muse consumer headband. First, during a one-minute training session, a subject is asked to do several activities (close eyes / stare at something / look around / talk) to collect EEG samples (machine learning) with different focus levels, and they are mapped on a 2D plane as dots, using a machine-learning algorithm (t-SNE). The dots are colored according to the timestamp to represent the activity: for example, red dots represent an activity when the subject is relaxed (i.e., eyes closed), green dots are when the subject is focused (i.e., staring), and blue and light blue dots are when the subject is distracted (i.e., looking around or talking). After the training session, real-time data is overlaid on the dots but this time they form lines with each other.
Development: Naoto Hieda