Masked priming - a window to non-conscious cognition

Lähteenmäki, Hyönä, Koivisto and Nummenmaa (2015) introduced new ideas to tackle the old and notorious question whether it can be reliably shown that stimuli, which participants are not aware of, will be evaluatively or semantically processed. Methodologically, they suggested a trial-by-trial subjective rating task (perceptual awareness scale; PAS) administered directly after each masked priming trial to assess awareness of the prime stimulus more directly. They claimed to have found no priming for subjectively unaware primes in two response priming experiments. Beside the super-ordinate question whether the proposal should in principle be considered a better solution compared to more traditional ones (e.g., a direct prime detection block administered subsequently to the priming block), their experiments can be inherently criticized for deviating from typical masked priming experiments with regard to several details. Therefore, in two experiments we integrated the PAS-rating in a more standard masked priming paradigm. We obtained priming effects even for subjectively unaware primes.

Priming in a lexical decision task elicited by briefly presented masked words is frequently considered as an index of unconscious automatic semantic processing. In order to confirm that the masked prime is unconsciously processed, an explicit psychophysical prime identification test is typically administered after the priming experiment. Recently, it is has been criticized that this classical sequential approach to assess prime identification after the main experiment does not capture trial-wise fluctuations of prime awareness within the priming phase. Based on this criticism, a trial-by-trial prime awareness rating using the Perceptual Awareness Scale (Ramsøy & Overgaard, 2004) within the priming experiment has been introduced. However, it is possible that ratings of the perceptual experience related to the masked prime interferes with concurrent semantic prime processing. The present study therefore compared masked semantic priming effects assessed within the classical sequential procedure, in which prime identification is assessed after the priming experiment, with those obtained in a condition, in which prime awareness is rated trial-wise within the priming experiment. Analysis of mean reaction times as well as drift diffusion modelling showed that priming effects were significantly smaller in the condition with trial-wise awareness ratings compared to the condition without such ratings. This shows that assessing subjective perceptual experience on a trial-by-trial basis is not a neutral tool to assess fluctuations of prime awareness. Instead, the rating procedure heavily interferes with semantic processes underlying masked priming, presumably due to attentional demands associated with concurrent prime identification.

For decades, cathode ray tube (CRT) monitors have been the standard for visual stimulus presentation in experimental psychology. Nowadays, these monitors are replaced by liquid crystal displays (LCDs). Early generation LCDs were, however, often found to yield imprecise and unreliable stimulus presentation (e.g., Elze & Tanner, 2012; Wiens & Öhman, 2007) and the high variation of specific characteristics in current general models (Elze, 2010a) puts into question whether they can be used for experiments which require millisecond precise presentation times. The present study aimed at showing that LCDs can even be used for experiments requiring masked presentation conditions, if the differences in the technical characteristics of LCDs and CRTs are accounted for. Participants were administered a masked number priming task and a subsequent forced-choice prime discrimination task at a CRT as well as a LCD monitor. Using sequential Bayesian testing, we found no evidence for monitor differences. Specifically, we obtained evidence for masked number priming under conditions of zero awareness with both monitors.

The NCE is a reversal in priming effects that can occur when a masked arrow prime is followed by an arrow target at a long stimulus-onset asynchrony (SOA). To counter the explanation that the NCE is actually a positive priming effect elicited by mask features associated with the prime-opposed response, we devise masks that always point in the same direction as the prime, eliminating all antiprime features. We find large positive priming effects for arrow primes without masks and for arrow masks without primes. When a neutral mask is introduced, priming effects turn negative at long SOAs. In the critical case where the mask is an arrow in the same direction as the prime, the prime does not add to the positive priming effect from the mask, but instead strongly diminishes it. This feature-free inhibition is accompanied by a drop in response accuracy 280 ms after the target. No feature-free inhibition is seen when arrows are replaced by color stimuli. We conclude that even though response activation by stimulus features plays a major role in the NCE, there is a strong inhibitory component (though perhaps not in all feature domains) that is not based on visual features.

Milner and Goodale proposed two separable pathways of visual information processing, one that generates conscious experiences of the stimuli (vision-for-perception), and another that uses action-related information from visual signals to control movements (vision-for-action). This distinction is supported by findings from neurological patients and dissociative effects of visual illusions in healthy participants. Backward masking experiments accord with the dual pathway perspective by showing priming effects of visual stimuli independent from perception of the prime. A recent report of suppressive effects of visual forward masks provided new evidence for a dual pathway perspective: When a forward mask precedes a prime that is followed by a backward masking target stimulus, the forward mask is limiting the effects of the prime independent from its effects on conscious perception. In a previous study, we found increased suppressive effects of strong rather than weak forward masks. Here we report a study that examined the time course of the suppressive effects of forward masks. Four experiments revealed that suppression of priming effects increases with the duration of the forward mask, but decreases when the inter-stimulus interval between the forward mask and the prime is increased. Findings suggest that forward masking procedures provide a tool to examine the vison-for-action pathway.