Abstract
BACKGROUND: Under some conditions, neural responses are highly synchronized across large regions of visual cortex. This synchronized activity is thought to be important for many functions. However, inferring synchrony measured at the scalp (EEG or MEG) can be challenging. This is because measured response levels are not only affected by synchrony, but also by neuronal amplitude and geometry of cortical sources and these variables cannot easily be disentangled. METHODS: We simulated neural signals in V1–V3, with either the same phase (synchronous) or random phases (asynchronous) across cortex. We then combined these with a forward model of MEG sensor responses (axial gradiometers). We compared these model predictions to measured MEG data from individuals (N=11) viewing a full-field high contrast-reversing stimulus. MEG data were separated into two components: time-locked to the stimulus (steady-state visual evoked fields, ‘SSVEFs’) and asynchronous with the stimulus (broadband response). RESULTS: First, SSVEF and broadband responses from the same datasets showed distinct spatial topographies. SSVEF responses were lateralized into two groups of posterior sensors, whereas broadband responses were generally confined to one centralized group. Second, we found that these different spatial topographies could be explained by differences in synchrony rather than source locations according to our forward model. The model predicts that synchronous V1–V3 responses result in two lateralized sensor patches, similar to SSVEFs, due to signal cancellation of sources with opposite facing dipoles. Importantly, using identical V1–V3 sources, the model predicts that sources with random phases result in a spatial topography approximately matching broadband responses (a single, central group of sensors). CONCLUSION: Identical cortical sources can result in very different patterns of sensor activity, depending on the degree of synchrony. In turn, the topography of sensor responses, combined with a forward model, can be used to make inferences about the underlying synchrony of neural responses.
Acknowledgement: NIH Brain Initiative R01 MH111417-01