Scene wheels: Measuring perception and memory of real-world scenes with a continuous stimulus space

The upside of cumulative conceptual interference on exemplar-level mnemonic discrimination

Although long-term visual memory (LTVM) has a remarkable capacity, the fidelity of its episodic representations can be influenced by at least two intertwined interference mechanisms during the encoding of objects belonging to the same category: the capacity to hold similar episodic traces (e.g., different birds) and the conceptual similarity of the encoded traces (e.g., a sparrow shares more features with a robin than with a penguin). The precision of episodic traces can be tested by having participants discriminate lures (unseen objects) from targets (seen objects) representing different exemplars of the same concept (e.g., two visually similar penguins), which generates interference at retrieval that can be solved if efficient pattern separation happened during encoding. The present study examines the impact of within-category encoding interference on the fidelity of mnemonic object representations, by manipulating an index of cumulative conceptual interference that represents the concurrent impact of capacity and similarity. The precision of mnemonic discrimination was further assessed by measuring the impact of visual similarity between targets and lures in a recognition task. Our results show a significant decrement in the correct identification of targets for increasing interference. Correct rejections of lures were also negatively impacted by cumulative interference as well as by the visual similarity with the target. Most interestingly though, mnemonic discrimination for targets presented with a visually similar lure was more difficult when objects were encoded under lower, not higher, interference. These findings counter a simply additive impact of interference on the fidelity of object representations providing a finer-grained, multi-factorial, understanding of interference in LTVM.

Brief category learning distorts perceptual space for complex scenes

The formation of categories is known to distort perceptual space: representations are pushed away from category boundaries and pulled toward categorical prototypes. This phenomenon has been studied with artificially constructed objects, whose feature dimensions are easily defined and manipulated. How such category-induced perceptual distortions arise for complex, real-world scenes, however, remains largely unknown due to the technical challenge of measuring and controlling scene features. We address this question by generating realistic scene images from a high-dimensional continuous space using generative adversarial networks and using the images as stimuli in a novel learning task. Participants learned to categorize the scene images along arbitrary category boundaries and later reconstructed the same scenes from memory. Systematic biases in reconstruction errors closely tracked each participant's subjective category boundaries. These findings suggest that the perception of global scene properties is warped to align with a newly learned category structure after only a brief learning experience.

The Accuracy and Precision of Memory for Natural Scenes: A Walk in the Park

It is challenging to quantify the accuracy and precision of scene memory because it is unclear what 'space' scenes occupy (how can we quantify error when misremembering a natural scene?). To address this, we exploited the ecologically valid, metric space in which scenes occur and are represented: routes. In a delayed estimation task, participants briefly saw a target scene drawn from a video of an outdoor 'route loop', then used a continuous report wheel of the route to pinpoint the scene. Accuracy was high and unbiased, indicating there was no net boundary extension/contraction. Interestingly, precision was higher for routes that were more self-similar (as characterized by the half-life, in meters, of a route's Multiscale Structural Similarity index), consistent with previous work finding a 'similarity advantage' where memory precision is regulated according to task demands. Overall, scenes were remembered to within a few meters of their actual location.

Differential effects of intrinsic properties of natural scenes and interference mechanisms on recognition processes in long-term visual memory

Humans display remarkable long-term visual memory (LTVM) processes. Even though images may be intrinsically memorable, the fidelity of their visual representations, and consequently the likelihood of successfully retrieving them, hinges on their similarity when concurrently held in LTVM. In this debate, it is still unclear whether intrinsic features of images (perceptual and semantic) may be mediated by mechanisms of interference generated at encoding, or during retrieval, and how these factors impinge on recognition processes. In the current study, participants (32) studied a stream of 120 natural scenes from 8 semantic categories, which varied in frequencies (4, 8, 16 or 32 exemplars per category) to generate different levels of category interference, in preparation for a recognition test. Then they were asked to indicate which of two images, presented side by side (i.e. two-alternative forced-choice), they remembered. The two images belonged to the same semantic category but varied in their perceptual similarity (similar or dissimilar). Participants also expressed their confidence (sure/not sure) about their recognition response, enabling us to tap into their metacognitive efficacy (meta-d'). Additionally, we extracted the activation of perceptual and semantic features in images (i.e. their informational richness) through deep neural network modelling and examined their impact on recognition processes. Corroborating previous literature, we found that category interference and perceptual similarity negatively impact recognition processes, as well as response times and metacognitive efficacy. Moreover, images semantically rich were less likely remembered, an effect that trumped a positive memorability boost coming from perceptual information. Critically, we did not observe any significant interaction between intrinsic features of images and interference generated either at encoding or during retrieval. All in all, our study calls for a more integrative understanding of the representational dynamics during encoding and recognition enabling us to form, maintain and access visual information.

Increasing stimulus similarity drives nonmonotonic representational change in hippocampus

Studies of hippocampal learning have obtained seemingly contradictory results, with manipulations that increase coactivation of memories sometimes leading to differentiation of these memories, but sometimes not. These results could potentially be reconciled using the nonmonotonic plasticity hypothesis, which posits that representational change (memories moving apart or together) is a U-shaped function of the coactivation of these memories during learning. Testing this hypothesis requires manipulating coactivation over a wide enough range to reveal the full U-shape. To accomplish this, we used a novel neural network image synthesis procedure to create pairs of stimuli that varied parametrically in their similarity in high-level visual regions that provide input to the hippocampus. Sequences of these pairs were shown to human participants during high-resolution fMRI. As predicted, learning changed the representations of paired images in the dentate gyrus as a U-shaped function of image similarity, with neural differentiation occurring only for moderately similar images.

Generative adversarial networks unlock new methods for cognitive science

Generative adversarial networks (GANs) enable computers to learn complex data distributions and sample from these distributions. When applied to the visual domain, this allows artificial, yet photorealistic images to be synthesized. Their success at this very challenging task triggered an explosion of research within the field of artificial intelligence (AI), yielding various new GAN findings and applications. After explaining the core principles behind GANs and reviewing recent GAN innovations, we illustrate how they can be applied to tackle thorny theoretical and methodological problems in cognitive science. We focus on how GANs can reveal hidden structure in internal representations and how they offer a valuable new compromise in the trade-off between experimental control and ecological validity.