Abrupt learning in illusory contour perception

N. Rubin, R. Shapley, K. Nakayama, A. Grossetete

Research output: Contribution to journalArticlepeer-review


Purpose. Preliminary findings indicated that in a task which involves the perception of illusory contours, subjects may exhibit very rapid - often abrupt - improvement, which is specific to the retinal size of the stimulus. We studied the learning effects in this paradigm to explore interactions between high-level and low-level (stimulus specific) factors in perceptual learning. Method. The stimuli were Kanizsa-like squares where the (illusory) edges were slightly curved into 'thin' or 'fat' shapes, which the subjects had to discriminate between (Ringach and Shapley Arvo '94). The stimulus duration was between 70ms and 97ms; it was followed by a blank screen (duration between 56ms and 140ms), and then by a mask. Subjects sat 60 cm away from the screen in the 'near' condition and 150-240 cm away in the 'far' condition. The physical size of the stimulus was fixed - 15 cm. The experiments consisted of three short phases (80 to 180 trials each): a testing phase, where only difficult (small curvature) stimuli were presented; a training phase, where high-curvature stimuli were added to the set; and a re-testing phase, with a set of stimuli identical to that of the first (testing) phase. The timing parameters, as well as the inclusion of feedback, varied between subjects. Results. A total of 33 subjects were tested. Thirteen of them showed significant improvement between testing and training phases, which was retained in the re-testing phase. Six were tested for transfer of the learning from the 'near' to 'far' condition: none showed transfer. Two were tested for 'far' to 'near' transfer: one subject showed significant transfer and the other did not. Five were tested for transfer across reversal of the contrast polarity of the stimulus: two subjects did not transfer, and three did. Conclusions. Early, retinotopically-organized visual areas are clearly involved in performing our task. However, the abrupt improvements that can occur, and the long retinal distances between the inducers (up to 11deg), suggest that the synaptic modifications cannot be confined to these early areas. Our results are more consistent with the view that the learning is the result of an interaction between higher-level processes and early visual cortical areas.

Original languageEnglish (US)
Pages (from-to)S180
JournalInvestigative Ophthalmology and Visual Science
Issue number3
StatePublished - Feb 15 1996

ASJC Scopus subject areas

  • Ophthalmology
  • Sensory Systems
  • Cellular and Molecular Neuroscience


Dive into the research topics of 'Abrupt learning in illusory contour perception'. Together they form a unique fingerprint.

Cite this