- Homepage
- Programme
- Registration
- Practical Guide
- About Neuronus
Marius Peelen
Donders Institute for Brain, Cognition and Behaviour Radboud University, Nijmegen, The Netherlands
Expectations derived from scene context influence perception. For example, objects presented in their typical context (e.g., a car on a road) are more easily recognized than objects presented in an atypical context. Recent behavioural studies have shown that context-based expectations influence not only semantic judgements, but also how sharply we perceive objects. Furthermore, there is now also evidence for the reverse influence, with objects affecting scene perception. Here, I present results from fMRI and MEG studies investigating the neural basis of such bidirectional interactions between object and scene processing. Results provide evidence for scene-based sharpening of object representations in visual cortex from around 280 ms after stimulus onset, reflecting feedback signals after the initial parallel processing of scenes and objects. This expectation-based modulation was observed even when the stimuli were task-irrelevant and attention was temporally and spatially directed away from the scenes. Interestingly, the reverse influence - with objects sharpening scene representations - was found at the same latency, in line with a common predictive processing mechanism for bidirectional object-scene interactions. These results indicate that objects and scenes, while initially processed in parallel pathways, engage in mutual and facilitatory interactions. These interactions then shape the feedback signals propagated within each pathway, modulating activity in hierarchically lower levels of the visual system, thereby resulting in overall reduced uncertainty and sharpened visual perception.
Natalia Rutkowska1, Maksymilian Bielecki2, Michał Bola3
1Laboratory of Brain Imaging, Nencki Institute of Experimental Biology, PAS, Warsaw, Poland
2Institute of Psychology, SWPS University, Warsaw, Poland
3Centre for Brain Research, Jagiellonian University, Krakow, Poland
The present study aimed to elucidate how perceptual representations of scene elements - specifically, backgrounds and objects - develop and interact over time. In the conducted experiment participants (N=31) were presented with grayscale images of real-world scenes, which depicted a natural or a man-made background and a single natural or man-made foreground object, combined in semantically congruent or incongruent ways. Participants performed a speeded classification of backgrounds or objects (in separate blocks) as natural or man-made. We analysed reaction-times (RTs) of a manual response to target stimuli and two measures derived from the drift diffusion model (DDM): boundary separation (a) and drift rate (v). Classification was faster for objects as relative to backgrounds and, in case of man-made targets, for congruent as relative incongruent images. In line, the DDM analysis indicated that less perceptual evidence was needed to categorise objects in comparison to backgrounds; and that accumulation of evidence was faster for congruent as relative to incongruent images, with the effect stronger for man-made targets. Thus, our findings indicate that local objects can be classified before the global background, and provide insights into how semantic incongruence slows down and hampers visual recognition.
Funding: National Science Center Poland grants (2018/29/B/HS6/02152; 2022/46/E/HS6/00150).
Michał Bola
Centre for Brain Research, Jagiellonian University, Kraków, Poland
In real-world scenes objects do not occur in isolation, but rather in relation to other scene elements. That such relations play a key role in the object recognition process has been shown by studies investigating perception of semantically incongruent objects - defined as objects with a very low probability of occurring in a given context - recognition of which is slower and less accurate in comparison to congruent ones. However, whether semantic relations present in scenes are able to guide spatial attention automatically remains an open research question. To address this question, in the first study (N = 25) we investigated whether semantically incongruent objects automatically capture attention (i.e. cause a semantic “pop-out”) using a combination of behavioural and ERP indicators of attention shifts. In the second study (N = 46), which was conducted and published in the registered report format, we investigated whether semantically incongruent objects automatically hold (or engage) attention for a longer time than congruent ones. Therefore, results of both studies inform us to what extent semantic relations present in natural scenes are able to automatically guide attentional selection and scene exploration.
Funding: National Science Centre Poland grant (2022/46/E/HS6/00150).
Marek A. Pedziwiatr
School of Biological and Behavioural Sciences, Queen Mary University of London, London, UK
Abstract: The visual input that the eyes usually receive contains temporally continuous information about unfolding events. Therefore, humans can accumulate knowledge about their current environment. Typical studies on scene perception, however, involve presenting multiple images that are unrelated to each other and, thereby, render this accumulation unnecessary. Our study, instead, facilitated it and explored its effects. Specifically, we investigated how recently-accumulated prior knowledge affects gaze behaviour. In our preregistered study, participants viewed sequences of static film frames from films directed by Alfred Hitchcock. Each sequence contained several ‘context frames’ followed by a ‘critical frame’. The context frames showed either events from which the situation depicted in the critical frame naturally followed or events unrelated to this situation. Therefore, identical critical frames were viewed by participants possessing prior knowledge that was either relevant or irrelevant to their content. When that knowledge was relevant, participants' gaze behaviour was less exploratory, as revealed by two complementary analyses: one based on traditional metrics of oculomotor behaviour and one based on modelling individual gaze traces using hidden Markov models. Therefore, our results demonstrate that recently gained prior knowledge impacts oculomotor exploration.
Funding: Funded by a Leverhulme Trust grant (RPG-2020-024).
Diana Kollenda1,2, Elaheh Akbarifathkouhi1,2, Maximilian D. Broda1,2, Benjamin de Haas1,2
1Experimental Psychology, Justus Liebig University, Giessen, German
2Center for Mind, Brain and Behavior (CMBB), Marburg and Giessen, Germany
Recent research has demonstrated systematic differences in the gaze patterns of observers when freely viewing complex scenes, with prominent differences in terms of text and social salience as well as the extent of visual exploration (de Haas et al., 2019). These findings give rise to further questions: Are individual gaze patterns associated with different impressions of identical scenes and what neural correlates underlie individual biases in visual attention and exploration?
I will present recent work showing that similarities in observers' gaze patterns serve as predictors of similarities in subsequent scene descriptions. In addition, a direct relationship was found between an observer's systematic visual biases and subsequent references to corresponding object categories, such as text and people. Furthermore, I will present preliminary results of a neuroimaging study investigating the relationship between, individual salience biases, visual exploration, and the functional layout of visual cortices in the individual brain.
Taken together, I will present data on the underlying mechanisms of our individual perception of the world.
Funding: This work was supported by European Research Council Starting Grant 852885 INDIVISUAL; BdH was further supported by Deutsche Forschungsgemeinschaft (DFG,German Research Foundation) Project Nos. 222641018-SFB/TRR 135 TP C9 as well as "The Adaptive Mind", funded by the Excellence Program of the Hessian Ministry of Higher Education, Science, Research and Art.
