Decoding Answers to Four-Choice Questions Using Functional near Infrared Spectroscopy

This paper presents a functional near infrared (NIR) spectroscopy-based paradigm that can be used to decode answers to four-choice questions. Ten healthy subjects were asked to perform one of the four different brain activities, that is, right-hand motor imagery (RMI), left-hand motor imagery (LMI), mental arithmetic (MA) and mental counting (MC), to answer the given four-choice questions. In selecting the A, B, C or D choices, the subjects were asked to perform RMI, LMI, MA or MC, respectively. Signals from the primary motor and prefrontal cortices were acquired simultaneously using a continuous-wave functional NIR spectroscopy system. The four activities were classified using multiclass linear discriminant analysis to decode the answers to an average accuracy of 73.3% across the 10 subjects. The results demonstrate the potential of functional NIR spectroscopy to decode answers to four-choice questions using four different intentionally generated brain activities as control signals.