Abstract
Background: Elucidating the neural circuits underlying the BOLD signal measured with fMRI is an important goal in human neuroscience. One approach is to correlate the amplitude of the BOLD signal with the field potential power in different temporal frequency bands. However, distinct frequency bands do not always correspond to different neural circuits. In visual cortex, field potential power in the gamma band consists of at least two large, distinct signals arising from distinct neural circuits: an oscillatory (narrowband) component and an asynchronous (broadband) component. We propose that the oscillatory component measures neural synchrony, which is unrelated to BOLD amplitude, whereas the broadband component more closely measures the level of neural activity, which influences BOLD. Methods: Electrocorticography (ECoG) responses to gratings and noise patterns were previously measured in two human subjects in V1-V3. Here we measured fMRI responses to the same stimuli. ECoG responses were separated into broadband and narrowband components; the two components were separately correlated with the fMRI response in corresponding cortical regions. Results: In V1, the ECoG broadband amplitude was positively and strongly correlated with BOLD across stimuli. In contrast, narrowband gamma amplitude was unrelated to BOLD. In V2/V3, BOLD was again positively correlated with broadband ECoG and unrelated to narrowband gamma; in V2/V3, the degree of alpha band suppression was also predictive of BOLD. Conclusion: In visual cortex, the asynchronous broadband component of the field potential is most predictive of the BOLD signal. Narrowband gamma oscillations appear as large signals in the voltage measurement but are uncorrelated with fMRI. The BOLD signal is not caused by, and does not correlate with, a particular frequency band in the LFP. Rather, it is a function of the level of neural activity, which influences the LFP power in many frequency bands, and is best indexed by broadband field potentials.
Meeting abstract presented at VSS 2015