Brain-Computer Interface: Annotation Pillows
The foundation of predictive models created via supervised learning is a labeled dataset. To ease annotation during long, continuously collected datasets, I designed and sewed devices to add labels in realtime. The devices are soft and plush, welcoming interaction. Buttons allow indication of instantaneous events or a binary status. Sliders and a rotary knob express continuous conditions. Gentle background lighting reflects current settings or provide feedback from a classifier deployed on the main collection device. This device is small (Raspberry Pi 4/5) but can simultaneously collect EEG (Muse S headband) and ECG (Polar strap) data, and be placed anywhere within Bluetooth range.