Detection of one-, two-, and three-dimensional Markov constraints in visual displays

Abstract

One-, two- and three-dimensional Markov constraints were introduced into visual displays by a common method. To obtain three-dimensional displays, two-dimensional spatially-encoded displays were presented successively in time. Following random initialization, all additional display elements were generated by rule. With non-probabilistic rules, sequences with horizontal (X) constraint, or with vertical (Y) constraint, alone were easily detected. Sequences with temporal (T) constraint alone, and with two-dimensional constraints in XY, XT, and YT, also were detected. Three-dimensional XYT constraints, however, could only be detected at chance level. Discrimination thresholds with probabilistic rules also show the relative superiority of one- over two-dimensional constraints. One- and two-dimensional constraints in T are sensitive to the rate of presentation of successive displays, whether memory is carried by the display or by the eye.