The negative BOLD response and its behavioral correlates

Abstract

Event Abstract Back to Event The negative BOLD response and its behavioral correlates Introduction The negative BOLD response (NBR) describes a phenomenon seen in functional magnetic resonance imaging in which a stimulated cortical region that responds with a positive blood oxygenation-level dependent (BOLD) signal is flanked by an unstimulated region of cortex exhibiting a negative response (Shmuel, Yacoub et al. 2002). It is most easily seen in visual cortex where well-ordered retinotopic maps make it easy to identify stimulated and unstimulated locations. The NBR is primarily a neural effect: reductions in the BOLD signal correlate with reductions in both LFP activity and firing rates (Shmuel, Augath et al. 2006). Using event-related fMRI, we measured the amplitude of the NBR as a function of ongoing background activity in order to test three potential models of response control: response gain control, contrast gain control and a purely subtractive mechanism. We also measured human psychophysical performance using a stimulus very similar to that in our fMRI experiments in order to identify behavioral correlates of the signal changes that we observe. Methods General Our stimulus consisted of a central disk D (diameter 2 degrees) containing a grating of contrast Cd and a surrounding annulus A (diameter 5 degrees) containing a grating of contrast Ca. These components were presented on a uniform mean gray field with a small gap between them. Cd and Ca could be set independently. In our event-related fMRI experiments, the difference between conditions Cd=0% (no center) and Cd=90% (high-contrast center) gave the magnitude of the NBR. We measured the NBR as a function of Ca (0%, 5%, 20% and 45% contrast). In our psychophysical experiments, we used a staircase procedure to measure threshold-versus-contrast curves for both D and A in the presence and absence of high-contrast, spatially-remote maskers (A and D respectively). fMRI methods N subs=5, B0=3T, EPI TR=2s, resolution 2x2x2mm. All data from independently-localized regions in V1 defined on flattened, retinotopically-mapped cortex. Behavioral methods N subs=5, thresholds estimated using QUEST 2 interval forced choice procedure. Results Our fMRI results show that the NBR is best modeled as a multiplicative gain control mechanism with the strongest effects occurring at low background contrast. In addition, the NBR exhibits a strong spatial asymmetry: We measure a strong NBR in the periphery but not in the fovea. Our psychophysical data are in agreement with these results: We measure only weak suppressive effects of a high-contrast annulus on the foveal target but a significant amount of suppression of the surround annulus due to the high-contrast center. Conclusion The NBR is a manifestation of a multiplicative gain control mechanism that acts to suppress neural activity in the periphery when a high-contrast stimulus is present in the fovea.