Patients with mild traumatic brain injury (MTBI) challenge physicians' skills and test their patience. Their manifold symptomatology is often not supported by objective neurological findings. We sought to compare regional cerebral blood flow (rCBF) between symptomatic subjects with longstanding MTBI and healthy controls, and to examine the correspondence between neuropsychological deficit and rCBF compromise. Twenty-eight clinically symptomatic male subjects with MTBI and twenty matched controls underwent neuropsychological testing and Tc-99m-HMPAO brain SPECT imaging. Neuropsychological test data were used to categorize subjects into sub-groups according to the presumed location of lesions based on their neurobehavioral deficits. Image subtraction comparisons were made between controls, all MTBI subjects and sub-groups. MTBI patients demonstrated regions of hypoperfusion in frontal, pre-frontal and temporal cortices, and sub-cortical structures. Hypoperfusion in 'frontal', 'left posterior' and to a lesser extent 'sub-cortical' sub-groups was concordant with neuropsychological localization. This was not the case for the 'right posterior' group, where no concordance was found. The rCBF is reduced in symptomatic patients with longstanding MTBI and unremarkable structural brain imaging. Although group analysis is appropriate for the generation of statistically significant differences, the clinical application of brain SPECT imaging in MTBI calls for a capability to associate clinical examination, neuropsychological assessment and cerebral perfusion at the individual subject level. Such competence is still to be attained.