Intracortically distributed neurovascular coupling relationships within and between human somatosensory cortices

The coupling of the brain's neural activity to its blood supply, termed neurovascular coupling, and its mechanisms are a fundamental feature of brain physiology that have been under discussion for more than a century and are still not fully understood. Recent studies have identified a linear relationship between measures of hemodynamic change and neuronal activity in rats (Mathiesen and others 1998; Ngai and others 1999) and primates (Mathiesen and others 1998; Heeger and others 2000; Rees and others 2000; Logothetis and others 2001) and between functional magnetic resonance imaging (fMRI) blood oxygen level–dependent (BOLD) responses in humans and primate neuronal activity (Mathiesen and others 1998; Heeger and others 2000; Rees and others 2000; Logothetis and others 2001). Some nonlinearity in this relationship has also been identified in relation to animal experiments (Ances and others 2000; Jones and others 2004; Hewson-Stoate and others 2005). The relationship is less well characterized in humans, but modern neuroimaging methods combined with more traditional electrophysiological techniques now allow for a definition of the neurovascular relationship in normal human subjects. A further physiological question relates to whether the brain increases blood flow to functionally active areas at the expense of other nonfunctioning areas. Neuroimaging has thrown some light on this: some studies have identified decreases in blood flow responses in functionally related (but not adjacent) cortical areas in both sensory (Drevets and others 1995; Peyron and others 1999) and motor areas (Allison and others 2000) during activation. However, whether these findings are related to decreased neuronal activity causing reductions in hemodynamic change (i.e., an underlying negative neurovascular coupling mechanism), or are a purely vascular phenomenon, remains unresolved. The experiments presented here investigate the direction and linearity of the neurovascular coupling relationship in normal human subjects. We compared changes in cerebral blood flow (CBF) (using blocked design fMRI BOLD) and scalp electrophysiology (using somatosensory evoked potentials [SEPs]) in parallel experiments with changes in the intensity of a median nerve electrical stimulus. By weighting the fMRI images according to the evoked potential amplitudes at corresponding intensities, we sought to test the hypothesis that these 2 data sets covaried with each other, either positively or negatively, in a linear or nonlinear manner. We thus identified whether these relationships might vary within the somatosensory cortex itself or between hemispheres