Purchase this article with an account.
Lora T. Likova, Christopher W. Tyler; Cortical network dynamics of figure/ground categorization. Journal of Vision 2007;7(9):312. doi: 10.1167/7.9.312.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Introduction: FMRI of asynchronous figure/ground structures has revealed a mechanism of top-down suppression from hMT+ to the stimulated background-representation in V1/V2 in the human brain (Likova & Tyler, 2005).
Purpose and Methods: In order to (i) analyze the whole-brain figure/ground network and (ii) understand its neural dynamics, we employed a novel instantaneous-stimulation paradigm: We find that instantaneous refreshes in subregions of random-dot fields create distinct and prolonged percepts of spatial structure. Different transient spatial-configurations were generated by refreshing (i) only the figure region, (ii) only the background, (iii) the whole stimulus field, or (iv) figure then background in asynchronous fashion. FMRI responses were obtained throughout the brain in a GE Signa 3T scanner at 1.5 sec TR. The dynamics of the neural activation was determined by deconvolution of the BOLD signal with the standard estimate of the hemodynamic response function (HRF, defined as the transform from the neural response to the BOLD activation waveform).
Results: Inter-condition comparisons revealed remarkable differences in one or more properties of the response waveform (latency, width, polarity, and number of phases), not only between, but within each individual cortical area. The instantaneous-stimulus paradigm generated a wide variety of BOLD signal waveforms and corresponding neural response estimates throughout a distributed network extending through the occipital, parietal and frontal cortex.
Conclusions: Such distinct response waveforms within a given cortical region evoked by different conditions assure that the differences could be securely attributed to the neural dynamics, not to spatial variations in the HRF. Integrated analysis of the neural response reconstruction throughout the cortex enabled us to quantify the neural dynamics of the involved cortical areas. The activation pattern specific to figure/ground categorization implies a distributed recurrent architecture with functional feedback loops.
This PDF is available to Subscribers Only