Beyond the semantic level, the 1998 digital priming experiment was the first in a long series of demonstrations that elementary decisions can be biased non-consciously. In this experiment, we showed that even the motor cortex receives non-conscious influences from subliminal cues. For example, when the subject is about to answer that a target number is greater than 5, the presentation of a non-congruent subliminal cue (a number smaller than 5) results in detectable activation of the motor cortex that controls the hand opposite the correct answer. These effects can be explained within the framework of mathematical models of threshold random walk decision making, detailed in the 2008 lecture. The nervous system seems to make such decisions by accumulating the evidence provided by the target in favor of each possible response. Subliminal priming and the response time modulations it induces can be explained by a partial accumulation of the evidence provided by the primer.
This theoretical view has been strongly reinforced by quantitative data from Vorberg et al (2003). These authors manipulated the amount of evidence provided by the primer by varying the primer-mask time interval. In their experiment, conscious perception varied non-linearly with this temporal parameter. However, the subliminal priming effect, as measured by mean response time, increased in a strictly linear fashion. This result, together with the concomitant increase in the priming-induced error rate, could be captured quantitatively by an evidence accumulation model.
