Four experiments tested the generality of Hummel's (2001) dual-route model of human object recognition with regard to object-selective attention. According to the model, objects are represented and processed in two different formats-analytic and holistic-that differ in their attentional demands and their invariance properties. A sequential-matching paradigm was used that employed compound stimuli consisting of a picture of an everyday object and a superimposed horizontal or oblique line. On each trial, a reference stimulus was followed by two lateral stimuli, a target and a distractor. Observers either attended (matched) the line (thus ignoring the object) or attended/matched the object (thus ignoring the line) within the compound stimulus. For the picture stimulus, visual similarity between target and reference was manipulated by using either an identical object image, its mirror-reflection, a split image, or an inverted version in the reference. The task was always to semantically match the reference to the target. The results showed distinct facilitation effects, in terms of a reduced response latency relative to a same-category-different-exemplar baseline, which were significantly larger for attended than ignored objects. Furthermore, the facilitation was reduced in the mirror and split relative to the identical condition, and for the inverted relative to the mirror condition. The effects of attention and manipulation were strictly additive-in line with the predictions of the dual-route model. Implications for theories of object recognition are discussed. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
Jüttner et al. (Wed,) studied this question.