Ambiguous Figures and Attention Affecting Appearance

Previously I posted about our use of a morphing ambiguous figures battery. The images start from one un-ambiguous image, e.g. an umbrella and pass via a series of small change through a point of ambiguity to another un-ambiguous image, e.g. a bat. We have series that go from inanimate to inanimate; animate to animate; and the crosses. The normative study has been available as a manuscript copy for awhile, but the "pretty" online versions and pdf copies of the manuscripts are now available from Behavior Research Methods here.

Although I am growing increasingly strident about the field's loose use of the term attention, I remain fascinated by the phenomena that commonly fall under the rubric. In particular, I have been much intrigued and much puzzled by the observation that attention can affect appearance as originally claimed by Carrasco and colleages in Nature Neuroscience in 2004. Basically I didn't want to believe the result, because it went against my intuition that if attention is something, then it ought to be good for something, and that something, as far as I could figure out, ought to be improving (or maybe prioritizing) our perceptions. The consequence seemed to me to be that attention should make perception better, and yet the Carrasco et al results found that attention made perception demonstrably less accurate. I eventually couldn't resist trying this for myself, and while I did replicate the effect on contrast reports with one method (choosing the more contrasted), I failed to with another (contrast matching). Certainly in my mind there is no doubt that the behavioral observation of Carrasco and colleagues (since extended to many other domains) is real, but I don't think it is as simple as a change in appearance, but rather a first glimmer that our sense of a single object, a single appearance is illusory, and that this can be teased apart with multiple assessments, via different procedures, for a single percept. My own thinking on what these data mean is still in flux, and I look forward to hearing from others. If you want to know what I am thinking as of this moment you can read the paper on-line here or download the pdf here.