BOLD Signal

Why does the BOLD signal increase during activation? It seems like it should decrease since more oxygen is being used up.  
  
It seems paradoxical, but the BOLD signal increases in response to cortical activation.
Don't feel too bad, because this phenomenon was also puzzling to early fMRI investigators as well.
What we now know happens is that in response to brain activation, regional arteriolar dilatation with increased capillary filling occurs so more fresh (oxygenated) blood is supplied than is required for the brain's immediate metabolic needs.
For over a century neurophysiologists mistakenly believed that cerebral hemodynamics were determined by the brain's short-term metabolic requirements. This notion was dispelled in the mid 1980's by by Fox and Raichle who demonstrated using O15-PET a physiological “uncoupling” between brain blood flow and oxygen demand.
During cerebral activation regional cerebral blood flow (CBF) does increase, but the cerebral metabolic rate of oxygen consumption (CMRO2) is not commensurately elevated. CBF and CMRO2 are thus said to be "uncoupled". Paradoxically, more fresh (oxygenated) blood has been supplied to that region of brain than is required for its immediate metabolic needs.
This "overshoot" of oxygenated blood means that the relative concentration of deoxyhemoglobin in activated areas will decrease. The T2/T2* shortening effects of dexoyhemoglobin will be diminished and the BOLD signal in activated areas will therefore increase.

Advanced Discussion (show/hide)»

References
     Fox PT, Raichle ME. Focal physiological uncoupling of cerebral blood flow and oxidative metabolism during somatosensory stimulation in human subjects. Proc Natl Acad Sci USA 1986; 83:1140-1144.
     Fox PT. The coupling controversy. Neuroimage 2012; 62:594-601. (Good historical review of how the concepts of neurovascular coupling evolved in the late 20th Century.)
Related Questions
     How is image contrast produced by BOLD fMRI?