Abstract:
The functional relevance of consciousness has been questioned over the last seven decades. Masked priming is the gold standard for demonstrating that unconscious processing can be performed at a higher cognitive level. To do this, researchers compare two tasks: in an indirect task, participants respond quickly to a target stimulus that is preceded by a masked prime. Here, the responses to the target stimulus are faster in trials with congruent prime-target pairs than in trials with incongruent pairs, which leads to a clear and significant priming effect. In a direct task, participants are then asked to classify the masked prime stimulus. Typically, they perform close-to-chance, which is interpreted as poor conscious perception of the prime. On the other hand, the significant priming effect is interpreted as evidence for good unconscious classification of the prime.
However, this standard reasoning is flawed. A significant effect cannot be taken as evidence for a good classification. Instead of comparing a non-significant close-to-chance performance in the direct task with a significant priming effect in the indirect task, researchers need to use the same metric in both tasks to compare them directly. For this, we propose the sensitivity analysis and introduce the descriptive term indirect task advantage (ITA). An ITA means a higher sensitivity in the indirect task than in the direct task. To infer better unconscious processing, two main steps are necessary: First, the sensitivities for both tasks are directly compared. If the indirect sensitivity is higher than the direct sensitivity (i.e., an ITA is present), the next step is to relate both measures to unconscious or conscious processing. If there is no ITA, there is no better sensitivity of the indirect measure, and thus there is no basis for concluding that unconscious processing is better. Hence, finding an ITA is the methodological basis for building theories of unconscious processing.
In three projects, we investigate whether there is a condition that shows an ITA. First, we replicate a highly influential landmark study on unconscious number priming and apply parametric variations. Second, we develop an approach to measure sensitivity d’ in electroencephalography (EEG). Third, we upscale the direct measure by using confidence ratings as a continuous measure instead of the traditional binary forced choice response.
In nine experiments, we compare the sensitivity in direct and indirect tasks and find no evidence for an ITA. Neither with different levels of complexity of neural processing (i.e., using EEG and RTs), nor with different stimulus sets (numbers, simple lines), nor with different conditions (parametric variations of prime contrast, prime duration, and mask-target SOA), and barely or not at all with methodological improvements (continuous scale in the direct task, i.e., confidence ratings). Our empirical results are highly consistent across experiments and fit well with the literature.
However, we show that there is no evidence for an ITA and thus the methodological basis for the assumption that unconscious processing is performed at a higher cognitive level is missing, casting doubt on current theories of unconscious processing. To summarize, the conclusions of previous studies imply that our brain can unconsciously perform astonishing processing at a higher cognitive level, which would question the role of consciousness. But the sensitivity analysis revealed no evidence for an ITA. Thus, there is no evidence to question the functional relevance of consciousness based on the priming literature.
electroencephalography