Parallel, non-convergent, interactions between separate cortical loci underlie perceptual unity: implications for a new view of object recognition

Any physical device, including computers, when comparing A to B, must send the information to point C. Explanations of brain processing take such a convergence for granted thus generating models relying on increasingly converging hierarchical streams. Such models, however, consistently fail to explain many perceptual phenomena. To see whether the brain, at times, can compare (integrate, process) events that take place at different loci without sending the information to a common target, I performed experiments in three modalities, somato-sensory, auditory, and visual, where 2 different loci at the primary cortex were stimulated.  Subjects were able to integrate inputs in time and space affecting small separate cortical loci. The ability to correlate activity between loci was independent of cortical distance up to 2-4 cm. Given the organization of sensory cortex where localized responses in primary cortex do not interact while convergence in downstream areas results in loss of individual stimulus identity and in decreasing selectivity to elementary stimuli, those results cannot be explained by conventional convergence models. We must thus assume a non-converging mechanism whereby two (or more) activated cortical loci can be integrated without sending information via axons into another downstream integrating site. Once we allow for such a non-converging mechanism, many perceptual phenomena can be viewed differently. Object recognition and representation is such a phenomenon that, I suggest, does not result from hierarchical convergence of cells with ever-increasing feature selectivity but rather from parallel interactions between various visual and non-visual areas.  If my hypothesis of the brain ability to relate activity taking place at separate loci without using convergence-by-wires is correct, it implies that the brain can use heretofore unconsidered (unknown?) parallel processing and that conventional models, including computer programs, would not be able to capture many brain processes.