At early stages of visual processing in the cortex, neurons view the visual environment through a window – or receptive field - which affords only a restricted view of the world. Receptive fields are located at every position in the visual field and tuned into different ranges along multiple visual dimensions. For example, the same neuron can signal the tilt, size, color, motion direction and absolute depth of an object at one location in the visual field. The output of this coding scheme is a piecemeal representation of the visual scene which cannot be used to unambiguously guide sensorimotor behavior. Subsequent stages of cortical visual processing almost certainly serve to reconstitute early visual signals into a coherent representation of a visual scene, most likely by striking a delicate, and possibly flexible, balance between the integration and segmentation of early visual signals. Yet we have still not figured out the computational principles governing this fundamentally important combinatorial process.
The proposed Research Topic provides an ideal opportunity for researchers working independently across the scientific spectrum of this problem to draw together their work within a common forum. The Research topic will be implicitly subdivided into different sub-themes, dealing with rapidly evolving aspects of this field. The first sub-theme will build the foundation for the Research Topic, providing reviews of the outstanding theoretical and computational issues surrounding the integration of different visual attributes. In the second and third sub-themes, we call for contributions from psychophysicists and physiologists who attempt to empirically establish the nature of the neural computations governing the combinatorial process. In the final subtheme, we will take a broader perspective and consider physiologically-inspired computational models of integration.
At early stages of visual processing in the cortex, neurons view the visual environment through a window – or receptive field - which affords only a restricted view of the world. Receptive fields are located at every position in the visual field and tuned into different ranges along multiple visual dimensions. For example, the same neuron can signal the tilt, size, color, motion direction and absolute depth of an object at one location in the visual field. The output of this coding scheme is a piecemeal representation of the visual scene which cannot be used to unambiguously guide sensorimotor behavior. Subsequent stages of cortical visual processing almost certainly serve to reconstitute early visual signals into a coherent representation of a visual scene, most likely by striking a delicate, and possibly flexible, balance between the integration and segmentation of early visual signals. Yet we have still not figured out the computational principles governing this fundamentally important combinatorial process.
The proposed Research Topic provides an ideal opportunity for researchers working independently across the scientific spectrum of this problem to draw together their work within a common forum. The Research topic will be implicitly subdivided into different sub-themes, dealing with rapidly evolving aspects of this field. The first sub-theme will build the foundation for the Research Topic, providing reviews of the outstanding theoretical and computational issues surrounding the integration of different visual attributes. In the second and third sub-themes, we call for contributions from psychophysicists and physiologists who attempt to empirically establish the nature of the neural computations governing the combinatorial process. In the final subtheme, we will take a broader perspective and consider physiologically-inspired computational models of integration.
