Selective stimulation of colour mechanisms: An empirical perspective

J. J. Kulikowski, D. J. Mckeefry, A. G. Robson

    Research output: Contribution to journalArticlepeer-review


    Anatomically distinct parvo and magno visual pathways show considerable functional overlap. However, specific stimulation of the most sensitive colour-opponent parvo-neurones is still possible, provided that colour stimuli are verified for selectivity. The authors have shown that gratings of low contrast, low spatial frequency and of restricted spatial content (6 or less spatial cycles) are optimal stimuli for distinguishing between colour-related (tritan and red/green) from achromatic or partly chromatic responses. This is particularly important when recording global responses, such as visual evoked potentials, VEPs. The crucial point is that at low presentation rates (<2 Hz), colour-related onset VEPs are maximally different from contrast reversal VEPs, thereby reflecting the activity of sustained-type parvo mechanisms. Achromatic onset, offset and reversal VEPs are similar, reflecting mediation by transient-type magno mechanisms. A stringent test of colour-response specificity is to check whether the chromatic reversal VEP has a low-pass temporal tuning curve, since it becomes band-pass when substantial achromatic intrusions are present. Specification of chromatic isoluminant stimuli, e.g along cardinal axes, does not guarantee their colour-selectivity, if chromatic aberration and variable macular pigmentation changes the chromatic content of the retinal image. It is shown here how chromatic stimuli, namely (1) red/green and (2) purple/green (tritanopic) gratings, can be optimized for selective stimulation of the colour system.
    Original languageEnglish
    Pages (from-to)379-402
    Number of pages23
    JournalSpatial vision
    Issue number4
    Publication statusPublished - 1997


    Dive into the research topics of 'Selective stimulation of colour mechanisms: An empirical perspective'. Together they form a unique fingerprint.

    Cite this