The current study addressed modality effects in a web-based Concealed Information Test (CIT) by asking participants to encode, and later conceal, crime-related details. Items were encoded and tested verbally or pictorially. A pilot (N = 73) and a preregistered study (N = 158) showed a robust interaction between encoding and testing modality: Items that were encoded and tested in the same modality were associated with better detection. Moreover, recognition of verbally encoded items could not be detected in a pictorial test. Our findings support the existence of a modality-congruency effect when subjects try to conceal their knowledge. In applied scenarios, the modality of test items should be matched to the modality in which crime-related details were encoded. Furthermore, a pictorial CIT might protect informed innocents if leakage happened verbally.

The adaptive gain theory (AGT) posits that activity in the locus coeruleus (LC) is linked to two behavioral modes: exploitation, characterized by focused attention on a single task; and exploration, characterized by a lack of focused attention and frequent switching between tasks. Furthermore, pupil size correlates with LC activity, such that large pupils indicate increased LC firing, and by extension also exploration behavior. Most evidence for this correlation in humans comes from complex behavior in game-like tasks. However, predictions of the AGT naturally extend to a very basic form of behavior: eye movements. To test this, we used a visual-search task. Participants searched for a target among many distractors, while we measured their pupil diameter and eye movements. The display was divided into four randomly generated regions of different colors. Although these regions were irrelevant to the task, participants were sensitive to their boundaries, and dwelled within regions for longer than expected by chance. Crucially, pupil size increased before eye movements that carried gaze from one region to another. We propose that eye movements that stay within regions (or objects) correspond to exploitation behavior, whereas eye movements that switch between regions (or objects) correspond to exploration behavior.

Previous findings suggest a negative correlation between slow oscillations (SO) in posterior brain regions and dreaming. Here we use a precise closed-loop auditory stimulation (CLAS) procedure to causally test whether slow oscillations suppress dreaming. Our results show that boosting posterior SO during NREM sleep decreases the likelihood of dreaming as compared to no SO boosting. This study provides the first causal evidence for the neural correlates of dreaming.

The picture superiority effect is particularly relevant in the context of memory detection. In the current study, participants encoded crime-related details and concealed them in a Concealed Information Test (CIT). Items were encoded and tested verbally or pictorially. Both the pilot study (N=73) and the preregistered study (N=158) showed evidence for a picture superiority effect: Pictorially encoded items were easier to detect. We also found an interaction between encoding and testing: Items that were encoded and tested in the same modality were associated with better detection. Lastly, recognition of verbally encoded items could not be detected in a pictorial test. Our findings support the existence of picture superiority also when subjects conceal their knowledge. In applied scenarios, the modality of test items should be matched to the modality in which crime-related details were encoded. Furthermore, a pictorial CIT might fail to detect concealed knowledge if crime-related details were encoded verbally.