  |
 |
      |
Stimulus information contaminates psychophysical tests of independent neural coding A fundamental hypothesis in neuroscience is that the brain works in a largely modular fashion; in other words, different parts of the brain operate on different tasks. For example, there are distinct areas of the brain for the different senses. This modularity is also assumed within the field of vision, as many models of visual processing presume that visual information is analyzed into separable and independent neural codes, or features. One common test of independent features in vision research can be described as a ‘summation’ study, in which performance is measured as the number of available features in a detection, discrimination, or visual search task is increased. Improvement in performance with increasing available features is attributed typically to the summation (or combination) of information across independent features. We have used the optimal observer (ideal observer) to assess the information content of the stimuli in these tasks, and find that, in many instances, increasing the number of available features also increases the stimulus information. Thus, a result that appears to suggest independent features in fact may be due to the information content in the stimuli, irrespective of the intended features to be examined. We are currently assessing the influence of stimulus information on the interpretation of independent coding of features in these summation studies (Shimozaki et al., 2000; Shiimozaki et al., 2001) Publications: Shimozaki, S.S., Abbey, C.K., Eckstein, M.P., Ideal observer analysis of visual search : is performance improvement from feature to disjunction tasks explained by stimulus information? Investigative Ophthalmology and Visual Science (ARVO meeting abstracts, May 2000), 41(4), S423.
|
 |
| | viu home | ucsb psychology | contact us | |