Neural dynamics of motion grouping: from aperture ambiguity to object speed and direction Export

@article{Chey1997Neural, abstract = {A neural network model of visual motion perception and speed discriminationis developed to simulate data concerning the conditions under which componentsof moving stimuli cohere or not into a global direction of motion, as in barberpoleand plaid patterns (both type 1 and type 2). The model also simulates howthe perceived speed of lines moving in a prescribed direction depends on theirorientation, length, duration, and contrast. Motion direction and speed bothemerge as part of an interactive motion grouping or segmentation process.The model proposes a solution to the global aperture problem by showing howinformation from feature tracking points, namely, locations from which unambiguousmotion directions can be computed, can propagate to ambiguous motion directionpoints and capture the motion signals there. The model does this without computingintersections of constraints or parallel Fourier and non-Fourier pathways.Instead, the model uses orientationally unselective cell responses to activatedirectionally tuned transient cells. These transient cells, in turn, activatespatially short-range filters and competitive mechanisms over multiple spatialscales to generate speed-tuned and directionally tuned cells. Spatially long-rangefilters and top–down feedback from grouping cells are then used to trackmotion of featural points and to select and propagate correct motion directionsto ambiguous motion points. Top–down grouping can also prime the systemto attend a particular motion direction. The model hereby links low-levelautomatic motion processing with attention-based motion processing. Homologsof model mechanisms have been used in models of other brain systems to simulatedata about visual grouping, figure–ground separation, and speech perception.Earlier versions of the model have simulated data about short-range and long-rangeapparent motion, second-order motion, and the effects of parvocellular andmagnocellular lateral geniculate nucleus lesions on motion perception.}, author = {Chey, Jonathan and Grossberg, Stephen and Mingolla, Ennio}, citeulike-article-id = {2846021}, citeulike-linkout-0 = {http://www.opticsinfobase.org/abstract.cfm?id=1970}, day = {1}, journal = {J. Opt. Soc. Am. A}, keywords = {aperture, apparent, cns, grouping, model, motion, terminator}, month = {October}, number = {10}, pages = {2570--2594}, posted-at = {2008-05-30 00:30:53}, priority = {3}, publisher = {OSA}, title = {Neural dynamics of motion grouping: from aperture ambiguity to object speed and direction}, url = {http://www.opticsinfobase.org/abstract.cfm?id=1970}, volume = {14}, year = {1997} }

TY - JOUR ID - Chey1997Neural L3 - citeulike-article-id:2846021 N2 - A neural network model of visual motion perception and speed discriminationis developed to simulate data concerning the conditions under which componentsof moving stimuli cohere or not into a global direction of motion, as in barberpoleand plaid patterns (both type 1 and type 2). The model also simulates howthe perceived speed of lines moving in a prescribed direction depends on theirorientation, length, duration, and contrast. Motion direction and speed bothemerge as part of an interactive motion grouping or segmentation process.The model proposes a solution to the global aperture problem by showing howinformation from feature tracking points, namely, locations from which unambiguousmotion directions can be computed, can propagate to ambiguous motion directionpoints and capture the motion signals there. The model does this without computingintersections of constraints or parallel Fourier and non-Fourier pathways.Instead, the model uses orientationally unselective cell responses to activatedirectionally tuned transient cells. These transient cells, in turn, activatespatially short-range filters and competitive mechanisms over multiple spatialscales to generate speed-tuned and directionally tuned cells. Spatially long-rangefilters and top–down feedback from grouping cells are then used to trackmotion of featural points and to select and propagate correct motion directionsto ambiguous motion points. Top–down grouping can also prime the systemto attend a particular motion direction. The model hereby links low-levelautomatic motion processing with attention-based motion processing. Homologsof model mechanisms have been used in models of other brain systems to simulatedata about visual grouping, figure–ground separation, and speech perception.Earlier versions of the model have simulated data about short-range and long-rangeapparent motion, second-order motion, and the effects of parvocellular andmagnocellular lateral geniculate nucleus lesions on motion perception. IS - 10 JF - J. Opt. Soc. Am. A EP - 2594 TI - Neural dynamics of motion grouping: from aperture ambiguity to object speed and direction PB - OSA VL - 14 SP - 2570 KW - aperture KW - apparent KW - cns KW - grouping KW - model KW - motion KW - terminator AU - Chey, Jonathan AU - Grossberg, Stephen AU - Mingolla, Ennio PY - 1997/10/01/ UR - http://www.opticsinfobase.org/abstract.cfm?id=1970 ER -