Carlos González-García

Voluntary and involuntary external attention can differentially modulate perceptual and semantic information, but previous working memory (WM) studies found no corresponding dissociation. Across four experiments (n=140), we tested whether increasing voluntary retro-cue reliability and making involuntary cues fully non-predictive would reveal this dissociation, and whether late visual masking promotes retrospective attention by protecting representations from interference. Peripheral and central retro-cues were used, and performance was analyzed using hierarchical drift-diffusion modeling. No differential prioritization of competing perceptual and semantic WM contents emerged. Beyond this shared pattern, the results revealed distinct influences of cue format, cue reliability, and visual interference on voluntary and involuntary attention. First, increasing central cue reliability enabled voluntary benefits in representational quality that depended on protection from visual interference. Second, peripheral involuntary effects persisted when cue predictiveness was eliminated, whereas evidence for central cues remained ambiguous. Third, removing masking generally did not reduce voluntary retro-cueing effects. Fourth, masking reduced peripheral involuntary benefits in representational quality but enabled retrieval-related benefits, increasing the difference between voluntary and involuntary attention. Finally, more frequent self-reported voluntary cue use was associated with larger attentional effects in the voluntary task, while effects persisted among participants reporting no intentional cue use in the involuntary task. These findings suggest that voluntary and involuntary retrospective attention differ in their sensitivity to cue format, reliability, and visual interference while exerting similar relative effects on competing perceptual and semantic WM representations.

Memory faces a fundamental challenge. It has to become flexible enough to be accessed by novel cues without losing the specific information that defines each individual episode. Whether memory transformation solves this problem by sacrificing details or by adaptively expanding retrieval while preserving them is unclear. We tested this by manipulating the match between a memory and its retrieval cue. Participants retrieved object associations after 6, 24 and 48 hours using either the original encoded image or novel exemplars that were more typical or more distinctive of the same object category. Retrieval initially benefited from the originally encoded image, but this advantage was reduced across delay. By 48 hours, a never-seen, more typical exemplar triggered retrieval as effectively as the original one, while the same convergence was not evident for less-typical exemplars. Crucially, such cue expansion was not accompanied by a loss of perceptual specificity. Participants retained access to the originally encoded exemplar across delays, while errors increasingly favoured more typical alternatives. At the item level, stronger episodic memories predicted better access to perceptual detail, but did not predict the tendency of errors for category-like information, which instead increased with delay. This dissociation suggests that memory generalisation and the continued accessibility of visual detail reflect partly separable transformations rather than opposite ends of a single continuum.

Cognitive operations require recently encountered information to remain available beyond the moment of sensory input. However, how such transient representations are accessed, and how they differ from sensory processing and long-term memory, remains unclear. Here, we combine hierarchical drift-diffusion modelling with behavioural manipulations to dissociate the decision processes underlying feature prioritisation in visual working memory. We first reanalysed a previously collected working memory dataset to characterise semantic and perceptual judgements at the level of latent decision processes. Semantic judgements were associated with reduced non-decision time across conditions, indicating faster access to task-relevant information, while advantages in evidence accumulation emerged selectively under higher cognitive demands. Two further experiments manipulated attentional prioritisation using retro-cues and dissociated the effects of interference from mere maintenance. Across manipulations, semantic prioritisation was selectively expressed in pre-accumulation processes and was amplified when representations fell outside the focus of attention or had to be maintained under interference. Together, these results suggest that semantic representations remain more readily accessible than perceptual details when working memory representations fall outside the focus of attention, consistent with a shift towards more abstract, long-term memory-like formats under conditions of limited attentional support.

Perceptual expectations facilitate perception of expected information by shaping neural coding. However, how the cognitive demands of different tasks may modulate this effect remains unclear. Here, we investigated whether and how expectations derived from sex/gender stereotypes impact behavioral and neural face processing as a function of task goals. Fifty-one participants categorized either the emotional expression (happy vs. angry) or the sex (male vs. female) of faces while mouse trajectories and electroencephalography (EEG) were recorded in separate sessions. Expectations were manipulated through the congruency between facial sex and emotional expression (congruent: angry male, happy female; incongruent: happy male, angry female). Behavioral analyses revealed that incongruent faces elicited larger trajectory deviations, longer movement dynamics and slower reaction times than congruent faces. Representational Similarity Analyses (RSA) showed that face congruency modulated coding geometries of both mouse trajectories and EEG, whereas multivariate classifiers revealed progressive impact of congruency from early stages of face neural coding. Importantly, all measurements revealed stronger congruency effects during emotion than during sex judgments, indicating that the influence of expectations is flexibly tailored to task demands. Furthermore, congruent faces exhibited enhanced decoding performance and reached peak neural discriminability earlier than incongruent ones. Finally, EEG–mouse fusion analyses revealed shared representational dynamics primarily explained by task demands. Together, these findings indicate that expectations facilitate face processing in a task-dependent manner and suggest that predictive facilitation operates, rather than by increasing the rate of evidence accumulation, by accelerating the emergence of neural coding patterns. Significance Statement: Perception is shaped by the combination of sensory input and prior expectations, including socially acquired sex/gender face stereotypes. Learning how these expectations influence perception and how they are flexibly adjusted to current behavioral goals is essential for predictive coding models. Combining EEG and mouse tracking, we show that emotion judgments are more susceptible to expectations than sex categorization, and that expectations accelerate the emergence of neural coding patterns rather than simply increasing the rate of evidence accumulation. These findings characterize the neural mechanism through which predictive processing adapts to task demands and demonstrate that the influence of social stereotypes on face perception is dynamically regulated rather than automatically imposed, providing new insights into when social biases are most likely to shape human perception.

Sudden perceptual learning refers to the abrupt disambiguation of an initially ambiguous stimulus. Despite the limited encoding time, these brief perceptual experiences transform into enduring memory engrams. Still, a comprehensive framework of why these memories are so long-lasting and their neural underpinnings, in particular the role of the hippocampus, is currently lacking. Here, we build on the apparent connection of sudden perceptual learning with long-term memory to outline how a representational perspective may help to overcome current limitations, similar to other areas of memory research. In short, we claim that sudden perceptual learning triggers initial prediction errors across the hierarchy, including strong higher-level (semantic) ones. Once a solution is found, the formerly ambiguous image can be connected to previous schemas and a disambiguation cascade of prediction error minimization is triggered, starting with semantic prediction errors. As a consequence, these events lead to low-dimensional representations that show signs of early semanticization, thereby favouring semantics over perceptual details. In sum, we propose that multiple memory traces are simultaneously encoded during sudden perceptual learning, and while low-dimensional representations engage neocortical structures, high-dimensional, episodic representations require the hippocampus. This perspective offers a promising framework to advance research on sudden perceptual learning and related phenomena. This article is part of the theme issue 'The role of hippocampal predictions in cognition: bridging perception and memory'.

While attention is known to prioritize stimuli held temporally in our working memory (WM), it remains poorly understood how different features (and primarily, spatial vs. non-spatial) involuntarily select complex WM contents to guide action. Building on the Binding and Retrieval in Action Control (BRAC) framework and the Theory of Event Coding (TEC), we hypothesized that any encoded feature should be capable of reactivating an Stimulus-Response (S-R) event file in WM. Across six experiments (N=199), divided into two Experimental Series—in which we systematically manipulated timing, perceptual salience, and attentional engagement—we pre-registered and tested the hypothesis that non-predictive, color-based retro-cues could involuntarily select and prioritize complex WM events associated with S-R bindings. Contrary to our expectations, we found no significant involuntary (facilitatory) effect of color retro-cues and observed moderate meta-analytic evidence against such an effect. This comprehensive lack of an effect stands in stark contrast to meta-analytic evidence from previous studies using the exact same design but with spatial retro-cues, which demonstrated a small but robust effect. This provides compelling evidence that not all retro-cues can interact with WM representations with the same efficacy, challenging the assumption that all features hold equal status within an event representation. Instead, our findings suggest a critical bias in involuntary WM selection: color-based retro-cues fail to involuntarily select complex WM contents containing integrated S-R associations, while spatial retro-cues do. This distinction implies that the functional relevance of a perceptual feature to action planning is the critical determinant for its ability to trigger the prioritization of complex internal representations, upholding the essential link between action and memory in adaptive behavior.

On the basis of verbal instructions, humans can accomplish novel and diverse demands on the very first try. This complex phenomenon recruits structured brain activity across the frontoparietal Multiple Demand Network (MDN), which is thought to encode upcoming task parameters and guide behavior. However, it is uncertain how novel instructions are translated into efficient neural task representations. To address this, we collected functional magnetic resonance imaging (fMRI) data while participants followed a rich set of novel verbal instructions. These varied along three core dimensions: the overarching task demand (to select or to integrate stimuli information), the relevant target category (animate or inanimate items), and the visual feature that participants responded to (color or shape). We used multivariate pattern analysis (MVPA) to examine whether and how each of these dimensions was reflected in MDN activity. We contrasted two alternative representational geometries that may underpin novel task coding: low-dimensional spaces based on abstract and generalizable representations, and high-dimensional architectures hosting context-unique, conjunctive neural codes. Our results showed that anticipatory MDN activity was sensitive to the content of instructions. While selection vs. integration task demands were broadly encoded across the MDN, coding of the relevant categories and features was restricted to lateral MDN regions, namely, the intraparietal sulcus and the inferior frontal junction. Critically, the representational spaces across the MDN displayed a mixture of geometrical motifs, partially supporting our two alternative hypotheses. On the one hand, Cross-Condition Generalization Performance revealed the presence of abstract and transferable neural codes for task demand information. On the other hand, Shattering Dimensionality showed complex, high dimensional coding spaces across the MDN, structured around both task-informative and non-informative axes. Despite this, no evidence of conjunctive neural codes was observed. Overall, these findings highlight that novel instructed behavior may recruit both abstraction and high dimensionality to promote generalization while maximizing the expressivity of MDN coding spaces. More broadly, they emphasize the importance of considering encoding geometry for a computational understanding of cognitive control processes.

Visual inputs during natural perception are highly ambiguous: objects are frequently occluded, lighting conditions vary, and object identification depends significantly on prior experiences. However, why do certain images remain unidentifiable while others can be recognized immediately, and what visual features drive subjective clarification? To address these critical questions, we developed a unique dataset of 1854 ambiguous images and collected more than 100,000 ratings (from a total of 947 participants) evaluating their identifiability before and after seeing undistorted versions of the images. Relating the representations of a brain-inspired neural network model in response to our images with human ratings, we show that subjective identification depends largely on the extent to which higher-level visual features from the original images are preserved in their ambiguous counterparts. In line with these results, an image-level regression analysis showed that the subjective identification of ambiguous images was best explained by high-level visual dimensions. Notably, the predominance of higher-level features over lower-level ones softens after participants disambiguate the images, suggesting that the visual system flexibly shifts between top-down guessing to bottom-up matching after disambiguation. Moreover, we found that the process of ambiguity resolution was accompanied by a notable decrease in semantic distance and a greater consistency in object naming among participants. However, the relationship between information gained after disambiguation and subjective identification was non-linear, indicating that acquiring more information does not necessarily enhance subjective clarity. Instead, we observed a U-shaped relationship, suggesting that subjective identification improves when the acquired information either strongly matches or mismatches prior predictions. Collectively, these findings advance our understanding on how we resolve ambiguity and extract meaning from incomplete visual information.

Attention shapes the maintenance and retrieval of information in working memory (WM). In external attention, interactions between attentional voluntariness and representational level have been reported, with voluntary and involuntary mechanisms differentially modulating low- and high-level content. In contrast, a study using central retro-cues in WM has found no such interaction. Because central and peripheral cues recruit attention through partially distinct mechanisms, and peripheral cues more effectively elicit automatic orienting, we tested whether peripheral retro-cues alter how attention modulates low- and high-level WM representations. In a preregistered experiment, 100 participants completed two retro-cueing tasks in which peripheral cues directed either voluntary or involuntary attention to perceptual (gray vs. sepia) and semantic (natural vs. artificial) features of real-world stimuli competing in WM. Reaction times and accuracy were analyzed using hierarchical drift-diffusion modeling to dissociate decisional from non-decisional processes. The results revealed the following pattern. First, attentional voluntariness did not differentially influence perceptual versus semantic WM content. Second, drift rates showed reliable retro-cueing effects for both voluntary and involuntary attention; voluntary attention produced larger, primarily cost-driven effects, whereas involuntary attention yielded smaller, predominantly benefit-driven effects. Third, non-decision times showed larger effects (reflecting both benefits and costs) for voluntary attention and smaller, benefit-only effects for involuntary attention. Fourth, in contrast to previous findings with central retro-cues, involuntary attentional modulation of non-decision times was observed. In conclusion, although peripheral retro-cues more effectively recruit involuntary attention, this enhancement does not differentially prioritize perceptual over semantic WM representations, thereby limiting direct analogies between internal and external attention.

Suddenly finding the solution to a problem after a period of impasse often comes with a feeling of insight. This subjective experience is proposed to arise as a consequence of prediction errors. Accordingly, previous studies have revealed that more incorrect initial predictions result in more intense insights. Crucially however, prominent models of Bayesian inference suggest levels of computationally-defined surprise are not a simple feature of distance between predictions and inputs, but also their precision or certainty. Yet, how these two factors interact to give rise to insight experiences remains unknown. In this pre-registered study, participants were exposed to ambiguous images while they tried to guess the correct label of the image (to derive prediction accuracy) and rated their confidence in that label (for prediction uncertainty). We then measured the intensity of their insight when a solution was given. As predicted, we found that the intensity of insight was a result of both the prediction accuracy and the uncertainty awarded to it. More specifically, when initial predictions were far from the true label, those made with lower confidence induced weaker insights, while the opposite pattern was observed when predictions were closer to the reality. Trial-by-trial estimations of prediction errors from participants’ responses closely mirrored insight ratings. Finally, we analysed data from two additional independent datasets with different modalities and setups and replicated the interaction between prediction accuracy and uncertainty on the intensity of insight. Altogether, these findings suggest that insight experiences are read out from prediction errors and highlight the key role of uncertainty in characterising this relationship.