About City
  1. Rector and President
  2. Schools and departments
  3. Governance and leadership
  4. Education
  5. Working at City
  6. History of City
  7. More about City
  8. Campus map
  1. Health Sciences
  2. Research
  3. Doctoral Studentships
About City

Computation of positional statistics by the human visual system


1st supervisor: Prof Joshua Solomon

2nd supervisors: Prof Michael Morgan

Research centre

Applied Vision Research

Project description

Although visual acuity is surprisingly good and visual memory is surprisingly capacious, there are limits. Under various constraints, observers must sacrifice acuity for capacity. Of these constraints, one of the best studied is viewing eccentricity. For example, outside the centre of the visual field, the addition of just two vertical lines can cause the just-noticeable tilt of a third line to double. This dramatic loss of acuity (‘crowding’) is thought to reflect an obligatory statistical analysis by the visual system. Instead of reporting the target’s tilt, observers report the mean of all three lines. Previous research suggests that normal observers perform relatively well when detecting irregularities in dot arrays, but we have yet to address the more general question of how (and how well) multiple positions can be encoded within the visual system.

This project will adopt the same systematic approach previously used in our laboratory for exploring the encoding multiple orientations. Instead of lines with various orientations, we will use regularly spaced dots extending from fixation to some maximum eccentricity (10 deg, say; see Figure). Discriminations between different mean azimuths and between different variances in azimuth will be analysed in the same way as we have analysed discriminations between different mean orientations and different orientation variances. Of particular interest is whether there is any crowding of position at all. Nearby dots with identical azimuths may have no effect at all on the acuity for a precued target with a slightly different azimuth.

Figure 3. Positional variance discrimination. This is a pretty easy trial. The standard deviation of azimuths in (a) is 2°. In (b) it is 14°.

If you would like to have an informal discussion please contact J.A.Solomon@city.ac.uk.