Juan Linde-Domingo

Maintaining multiple items in working memory (WM) requires the allocation of limited resources, which may involve prioritizing individual items over others. Previous work has suggested that attentional prioritization can vary rhythmically between WM samples, but direct empirical evidence for such temporal fluctuations remains scarce. Here, we recorded eye-tracking and EEG to track the temporal dynamics of multi-item processing while participants maintained up to three object orientations in WM. Whereas EEG encoding was only short-lived, miniature gaze patterns reflected the orientations of all three items throughout the entire WM delay and showed systematic load and serial position effects. Contrary to rhythmic sampling accounts, however, we found no evidence that eye movements alternated between the concurrently maintained items. Furthermore, the strength of WM encoding in gaze was not modulated by frontal theta activity in EEG, suggesting that oculomotor encoding was not directly coupled to rhythmic neural dynamics associated with WM and internal attention. Together, our findings indicate that the oculomotor signature of visuospatial WM processing was rather stably shaped by all of the simultaneously maintained items. The results further corroborate that eye-tracking and EEG provide complementary windows into the temporal dynamics of WM.

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.

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'.

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.

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.

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.

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.

Prior work has shown that control brain regions encode upcoming novel instructed actions. Similarly, visual working memory (WM) representations can reflect different priority states. However, it remains unknown whether the priority status of WM-guided novel actions similarly modulates their neural coding format, and how such dynamics unfold over time. We addressed these questions using EEG while human participants performed two consecutive choice reaction tasks. At the start of each trial, participants encoded two novel stimulus-response (S-R) mappings. A cue then indicated whether these mappings would be relevant immediately (“current” condition), later after an intervening task (“prospective”), or after a free delay (“delayed”). Using multivariate pattern analysis, we found that the S-R category was decodable in all conditions during relevant time windows. Critically, pattern similarity analyses revealed that while mere maintenance demands allow for temporary preservation of neural codes (i.e., between current and delayed trials), shielding from interference (i.e., prospective trials) induced significant alterations to the neural code. Further exploration of the representational geometry revealed that priority status gained prominence dynamically over S-R category coding when preparing for such shielding demands. Importantly, some of these changes emerged anticipatorily, prior to target onset. Overall, our results show that, similar to visual WM, the priority of intended actions dynamically and anticipatorily reshapes their neural format. They further reveal how different demands induce content geometries compatible with previously proposed coding schemes, and that such representational changes can be implemented flexibly in time. Significance statement: Successfully translating visuomotor plans into action relies not only on maintaining such plans but also on prioritizing relevant information at the right time. While prior work has shown that working memory representations adapt to prioritize visual content, it remains unknown whether similar mechanisms apply to action-guiding content. Using EEG and multivariate analyses, we show that priority states—whether immediate, delayed, or prospectively relevant—modulate the representational geometry of novel stimulus-response mappings. Notably, the anticipation of interference compared to mere maintenance leads to greater alterations in the neural code during a retention interval. These results demonstrate that neural representations of novel instructed actions are shaped by their priority state and upcoming demands and reveal how goal- relevant information is restructured over time.

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 1,854 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.