Aimed to assess, in the light of current attentional theories, the nature of the attentional deficit in a group of severely traumatically head-injured subjects, relative to a group of orthopaedic rehabilitation patients, and to establish which neuropsychological measures best reflected the deficit. Three separate studies were conducted in order to meet these aims. The first study focused on selective attention; the second, on vigilance or sustained attention; the third, on the Supervisory Attentional System. Results provided no evidence for the presence of deficits of focused attention, sustained attention, or supervisory attentional control, but ample evidence for the presence of a deficit in speed of information processing. Those neuropsychological measures shown to be the best measures of this deficit included the Symbol Digit Modalities Test, simple and choice reaction-time tasks, colour naming and word reading scores on the Stroop, and the Paced Auditory Serial Addition Test.