Manipulating decision processes in the human scalar timing system.

Abstract

Two experiments attempted to manipulate the decision processes used in a temporal generalisation task with humans. In Experiment 1, payoffs (points awarded or deducted) were used to try to alter behaviour when the standard duration was 400ms, and the comparison durations ranged from 100 to 700ms in 100ms steps. Two conditions which either encouraged or discouraged the subject to identify a comparison duration as the standard were compared with a neutral condition. Encouraging identifications of the standard increased the proportion of identifications of the standard, whereas the discouraging manipulation had more ambiguous effects. Using the "modified Church and Gibbon" model, it was shown that the effect of the encourage manipulation was an increase in the response threshold, consistent with the information-processing version of scalar timing theory. A second experiment compared encourage and discourage manipulations with a more difficult discrimination (comparison durations spaced in 50ms steps around the 400ms standard), and with more distinct payoff differences for the different conditions. Behavioural effects were much more marked in Experiment 2, with the encourage condition producing more identifications of a comparison duration as the standard for all comparison durations except the shortest, compared with the discourage condition. Modelling showed that the main theoretical difference between the two conditions was in a change in the response threshold, in a manner consistent with the scalar timing model.