Abstract
Purpose. There is psychophysical support for a phase model of visual processing involving four channels, each optimally sensitive to one of the following four phase relations; + cosine, − cosine, +sine and − sine. Neurophysiology suggests either an even distribution of neuronal phase responses or else a sine/cosine phase dichotomy that is dependent on spatial bandwidth. We investigated whether there were labelled lines for phase; this can be thought of as testing a strong version of the four channel phase model in which each of the above phases can be discriminated at threshold. Methods. Our stimulus comprised Gaussian weighted (space and time) patches of either Gabor or edge/bar stimuli for which we measure simultaneously detection and phase identification to determine if phase identification can be accomplished at detection threshold (hence labelled lines). We used two bandwidths of Gabor and varied both absolute and relative phase. Results. Subjects could not reliably discriminate at threshold Gabors of even symmetry from Gabors of odd symmetry nor could they discriminate bar from edge stimuli. Conclusion. While there may be labelled lines for polarity there are not labelled lines for bars vs edges. The smallest discriminable step we can reliably make across the phase spectrum at threshold is 180°.
RFH is funded by CIHR (#108-18) and NSERC grant #0GP0046528. FK is funded by NSERC grant #OGP01217130.