We investigate the intersubject signal variability of near-infrared spectroscopy (NIRS), which is commonly used for noninvasive measurement of the product of the optical path length and the concentration change in oxygenated hemoglobin (DeltaC'oxy) and deoxygenated hemoglobin (DeltaC'deoxy) and their sum (DeltaC'total) related to human cortical activation. We do this by measuring sensorimotor cortex activation in 31 healthy adults using 24-measurement-position near-infrared (NIR) topography. A finger-tapping task is used to activate the sensorimotor cortex, and significant changes in the hemisphere contralateral to the tapping hand are assessed as being due to the activation. Of the possible patterns of signal changes, 90% include a positive DeltaC'oxy, 76% included a negative DeltaC'deoxy, and 73% included a positive DeltaC'total. The DeltaC'deoxy and DeltaC'total are less consistent because of a large intersubject variability in DeltaC'deoxy; in some cases there is a positive DeltaC'deoxy. In the cases with no positive DeltaC'oxy in the contralateral hemisphere, there are cases of other possible changes for either or both hemispheres and no cases of no change in any hemoglobin species in either hemisphere. These results suggest that NIR topography is useful for observing brain activity in most cases, although intersubject signal variability still needs to be resolved.