Search

The implementation or change of information processing routines, known as cognitive control, is traditionally believed to be closely linked to consciousness. It seems that we exert control over our behavior if we know the reasons for, and consequences of, doing so. Recent research suggests, however, that several behavioral phenomena that have been construed as instances of cognitive control can be prompted by events of which actors are not aware. Here we give a brief review of this research, discuss possible reasons for inconsistencies in the empirical evidence, and suggest some lines of future research. Specifically, we suggest to differentiate cognitive control evoked either because of explicit or because of implicit control cues. While the former type of control seems to work outside of awareness, the latter type of control seems to be restricted to consciously registered events that call for control.

Using an explicit task cuing paradigm, we tested whether masked cues can trigger task-set activation, which would suggest that unconsciously presented stimuli can impact cognitive control processes. Based on a critical assessment of previous findings on the priming of task-set activation, we present two experiments with a new method to approach this subject. Instead of using a prime, we varied the visibility of the cue. These cues either directly signaled particular tasks in Experiment 1, or certain task transitions (i.e., task repetitions or switches) in Experiment 2. While both masked task and transition cues affected task choice, only task cues affected the speed of task performance. This observation suggests that task-specific stimulus-response rules can be activated only by masked cues that are uniquely associated with a particular task. Taken together, these results demonstrate that unconsciously presented stimuli have the power to activate corresponding task sets.

The impact of masked stimulation on cognitive control processes is investigated with much interest. In many cases, masked stimulation suffices to initiate and employ control processes. Shifts of attention either happen in the external environment or internally, for example, in working memory. In the former, even masked cues (i.e., cues that are presented for a period too short to allow strategic use) were shown efficient for shifting attention to particular locations in pre-cue paradigms. Internal attention shifting can be investigated using retro-cues: long after encoding, a valid cue indicates the location to-be-tested via change detection, and this improves performance (retro-cue effect). In the present experiment, participants performed in both a pre- and a retro-cue task with masked and normally presented cues. While the masked cues benefitted performance in the pre-cue task, they did not in the retro-cue task. These results inform about limits of masked stimulation.

We used a new methodological approach to investigate whether top-down influences like expertise determine the extent of unconscious processing. This approach does not rely on preexisting differences between experts and novices, but instructs essentially the same task in a way that either addresses a domain of expertise or not. Participants either were instructed to perform a lexical decision task (expert task) or to respond to a combination of single features of word and non-word stimuli (novel task). The stimuli and importantly also the mapping of responses to those stimuli, however, were exactly the same in both groups. We analyzed congruency effects of masked primes depending on the instructed task. Participants performing the expert task responded faster and less error prone when the prime was response congruent rather than incongruent. This effect was significantly reduced in the novel task, and even reversed when excluding identical prime-target pairs. This indicates that the primes in the novel task had an effect on a perceptual level, but were not able to impact on response activation. Overall, these results demonstrate an expertise-based top-down modulation of unconscious processing that cannot be explained by confounds that are otherwise inherent in comparisons between novices and experts.

Humans adapt to context-specific frequencies of response conflicts. Typically, the impact of conflict-inducing information is reduced in contexts with high compared to low frequency of conflict. We investigated how such context-specific conflict adaptation depends on awareness and timing of conflict-eliciting stimuli and conflict-signaling contexts. In a priming paradigm, we varied the visibility of the prime and whether the context is a feature of either prime or target. Concretely, the context was represented by the format of either prime (Experiment 1) or target (Experiment 2), which means that primes or targets of a particular format were associated with a high or low probability of conflict (i.e., prime-target incongruency). In both experiments, we found a context-specific modulation of congruency effects, both with masked and visible primes. To control for mechanisms of event learning in Experiments 3 and 4, context-specific conflict frequency was realized by inducing trials, while stimuli in test trials were associated with equal conflict frequency. We again found a context-specific congruency modulation when the prime represented the context, most interestingly also with masked primes within test trials. When the target represented the context, however, such a modulation occurred with visible primes, but not with masked primes. These results provide a compelling case for the unconscious exertion of a very flexible form of cognitive control. Context-specific conflict adaptation processes can basically operate independently of both conflict awareness and context awareness, but they depend on close temporal proximity of context and conflict information.

Controlling response parameters like the speed and accuracy of responses allows us to adjust our behavior according to particular situational task demands. We investigated whether exertion of cognitive control over speed–accuracy settings is not exclusively based on conscious representations, but can also be elicited by stimuli that are not consciously represented. Participants were instructed to point and click on a target, with a cue signaling before each response whether to prioritize accuracy of the response over speed, or vice versa. In half of the trials, the cue was masked to prevent a conscious representation of the cue. With visible cues, response patterns showed typical speed–accuracy tradeoffs, with faster and less accurate responses after speed cues, and slower but more accurate responses after accuracy cues. Crucially, this was found with masked cues as well. Our results are in line with recent findings on the relation of consciousness and cognitive control processes like task-set activation and response inhibition: masked cues are able to impact on cognitive control processes.

Anticipating where an event will occur enables us to instantaneously respond to events that occur at the expected location. Here we investigated if such spatial anticipations can be triggered by symbolic information that participants cannot consciously see. In two experiments involving a Posner cueing task and a visual search task, a central cue informed participants about the likely location of the next target stimulus. In half of the trials, this cue was rendered invisible by pattern masking. In both experiments, visible cues led to cueing effects, that is, faster responses after valid compared to invalid cues. Importantly, even masked cues caused cueing effects, though to a lesser extent. Additionally, we analyzed effects on attention that persist from one trial to the subsequent trial. We found that spatial anticipations are able to interfere with newly formed spatial anticipations and influence orienting of attention in the subsequent trial. When the preceding cue was visible, the corresponding spatial anticipation persisted to an extent that prevented a noticeable effect of masked cues. The effects of visible cues were likewise modulated by previous spatial anticipations, but were strong enough to also exert an impact on attention themselves. Altogether, the results suggest that spatial anticipations can be formed on the basis of unconscious stimuli, but that interfering influences like still active spatial anticipations can suppress this effect.

Arrow cues and other overlearned spatial symbols automatically orient attention according to their spatial meaning. This renders them similar to exogenous cues that occur at stimulus location. Exogenous cues trigger shifts of attention even when they are presented subliminally. Here, we investigate to what extent the mechanisms underlying the orienting of attention by exogenous cues and by arrow cues are comparable by analyzing the effects of visible and masked arrow cues on attention. In Experiment 1, we presented arrow cues with overall 50% validity. Visible cues, but not masked cues, lead to shifts of attention. In Experiment 2, the arrow cues had an overall validity of 80%. Now both visible and masked arrows lead to shifts of attention. This is in line with findings that subliminal exogenous cues capture attention only in a top-down contingent manner, that is, when the cues fit the observer's intentions.