Human brain activity can be recorded non-invasively using measures such as changes in focal brain blood flow as well as recording the spontaneously occurring electrical activity of the brain from sensors located on the scalp. Blood flow measurement, performed using functional magnetic resonance imaging (fMRI), allows the active structures in the brain to be readily identified. However, the blood flow response has very poor temporal resolution and cannot give a good sense of the dynamics of brain activity as it unfolds over the millisecond range. Electrical activity methods, such as electroencephalography (EEG), allow brain dynamics to be studied with millisecond accuracy, and allow differentiation between electrical activity that is either phase-locked (evoked) or correlated (induced) with an incoming sensory stimulus, or tied to a decision or motor response. EEG, however, has relatively poor spatial resolution. It is difficult to identify the generators for particular EEG signals, because the inverse problem---estimating 3D current source distributions from noisy 2D measurements---has no unique solution.
By using the strengths of each method to answer specific scientific questions, functions of the brain can be studied. Some examples will be provided in the study of visual stimulation where the human brain appears to have specialized circuitry that processes the face, hand and body actions of others.