There is a need to develop brain-computer interface (BCI) systems that are able to remain stable in varying psychophysiological conditions (such as during times of mental fatigue) that can disrupt user capability. Attention resources are finite and how a gradual decrease in attention across time – the so-called time-on-task (TOT) effect – can influence neural signals is yet to be identified. This present study investigated in fifteen male volunteers (i) the time course of cortico-spinal excitability through single-pulse transcranial magnetic stimulation; and (ii) the changes in activity of pre-motor, primary motor, prefrontal cortex and right parietal areas by means of near-infrared spectroscopy (NIRS), during a sustained attention reaction task (RT) of 30 min duration. Applicability of using NIRS-measured cortical activity in order to classify subject attentional state during the sustained attention task was further tested in a sample of 7 subjects.