Decoding what one likes or dislikes from single-trial fNIRS measurements

Recent functional neuroimaging studies have shown the possibility of decoding human mental states from their brain activity using noninvasive neuroimaging techniques. In this study, we applied multivariate pattern classification, in conjunction with a short interval of functional near-infrared spectroscopy measurements of the anterior frontal cortex, to decode whether a human likes or dislikes a presented visual object; an ability that is quite beneficial for a number of clinical and technological applications. A variety of objects comprising sceneries, cars, foods, and animals were used as the stimuli. The results showed the possibility of predicting subjective preference from a short interval of functional near-infrared spectroscopy measurements of the anterior frontal regions. In addition, the pattern localization results showed the neuroscientific validity of the constructed classifier.