Research on unconscious perception has faced persistent methodological challenges since the late 19th century. This review discusses key difficulties in demonstrating perception without awareness, examining how objective versus subjective measures of awareness produce different awareness thresholds and lead to alternative experimental approaches. New methodologies include regression-based Bayesian modeling, sensitivity vs. awareness curves from General Recognition Theory, the liminal-prime paradigm, and two-interval forced choice designs. The authors emphasize the need for brain-based approaches and highlight promising studies in this area, along with the role of individual differences and frameworks such as predictive coding and active inference. These advances aim to address challenges in demonstrating cognition without awareness.
In consciousness research, subjective measures (reports of sensory experience) and objective measures (performance on detection or discrimination tasks) are both used to assess visual awareness. While subjective measures sometimes lag behind objective ones, many studies show they converge. Five aspects relevant to achieving convergence are identified, and future work should test the boundary conditions under which convergence or divergence occurs.
Objective performance measures and subjective ratings of visual awareness yield similar thresholds for detecting and discriminating visual stimuli, supporting the validity of both approaches. In a temporal two-alternative forced choice task, participants' accuracy in detecting stimulus presence and discriminating letter case was compared with their subjective ratings using the Perceptual Awareness Scale (PAS) at high and low contrast. Thresholds for objective and subjective measures showed a comparable pattern across tasks and contrast levels, suggesting that both methods provide converging information about the feature-content of a percept and are likely veridical measures of awareness.
Asking participants to rate their awareness of a briefly flashed, masked word during a semantic priming experiment changes the underlying cognitive processes. In two groups performing a lexical decision task, masked primes were shown for 20, 40, or 60 milliseconds. One group also rated prime visibility trial-by-trial using the Perceptual Awareness Scale; the other did not. Reaction-time and drift-diffusion modeling showed general priming effects on reaction time and drift rate only in the group without the rating task. In the rating group, residual priming effects appeared only on trials where participants reported seeing the prime, and affected non-decisional processing time. The concurrent rating task interferes with semantic priming, likely because it diverts attention to identifying the masked prime.