Brain-Computer Interfaces based on functional magnetic resonance imaging (fMRI-BCI) allow volitional control of anatomically specific regions of the brain. We have developed an fMRI-BCI that provides subjects graphically animated, real-time contingent feedback of a circumscribed region of the brain to train them to volitionally control the region. This non-invasive technique could be used to vary the activity of the neural substrates of a region of interest as an independent variable to study its effects on behavior. Furthermore, if the neurobiological basis of a disorder is known in terms of abnormal activity in certain regions of the brain, fMRI-BCI can be targeted to modify activity in those regions with high specificity. However, there are a few critical drawbacks in the existing fMRI-BCI systems that limit their effectiveness and clinical application. The drawbacks originate from the conventional neuroimaging method that seeks to find how a particular perceptual or cognitive state is encoded by measuring brain activity from many thousands of locations repeatedly, but then analyzing each location separately (univariate analysis). We had proposed to implement an online, generalized (to multiple subjects), multivariate, pattern-based fMRI-BCI that can distinguish multiple brain states and provide real-time feedback of the whole network of brain activity in contrast to activity in single regions of interest. We further proposed to test the efficacy of the system on two important applications: 1. Movement restoration after stroke, 2. Lie detection.