Noise contributed by the probabilistic spiking times of neurons has an important and advantageous role in brain function. We go beyond the deterministic noiseless description of the dynamics of cortical networks and show how the properties of the system are influenced by the spiking noise. We review here recent results that show the direct link between brain activity and psychophysically quantified behaviors during a somatosensory detection task. We focus on the following remarkable observation in this somatosensory task: when a near-threshold stimulus is presented, a sensory percept may or may not be produced. These perceptual judgments are believed to be determined by the fluctuation in activity of early sensory cortices. We show, however, that the behavioral outcomes associated with near-threshold stimuli depend on the neuronal fluctuations of more central areas to early somatosensory cortices. Furthermore, we show that the behavioral correlate of perceptual detection is given by a noise-driven transition in a multistable neurodynamical system. Thus, neuronal fluctuations can be an advantage for brain processing because they lead to probabilistic behavior in decision making in this and other sensory tasks.