Thematic knowledge survives visual crowding and influences object identification

Thematic knowledge survives visual crowding and influences object identification | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Thematic knowledge survives visual crowding and influences object identification Nicolas Slaski, Bilge Sayim, Solène Kalénine This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7442959/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 28 Feb, 2026 Read the published version in Psychological Research → Version 1 posted 9 You are reading this latest preprint version Abstract Object perception is challenged in complex and cluttered visual environments. This phenomenon, known as visual crowding, occurs when object identification is impaired because a target is surrounded by other objects. However, visual environments are also meaningful, and objects are often arranged in coherent and predictable ways. In particular, thematic relations linking objects that play complementary roles in events (for example, hammer–nail) have been shown to facilitate conceptual processing, especially for manipulable objects. However, it remains unclear whether such semantic relations can also influence perceptual processes. This study investigated whether thematic information can be processed under visual crowding and facilitate object identification in clutter. Participants were briefly presented with pairs of object images in peripheral vision and asked to perform an object identification task. The object pairs were either thematically related or unrelated, and correctly positioned for being used together or not. Additionally, objects were presented either in isolation or crowded by meaningless flankers. Experiment 1 demonstrated that objects in thematically related pairs were identified more accurately than those in unrelated pairs, even under crowding. Experiment 2 showed that while the thematic benefit may occur at the decisional stage when objects are presented in isolation, thematic grouping may be at play in crowding conditions when perceiving objects. These findings suggest that semantic knowledge involving the use of manipulable objects can survive visual crowding and influence object recognition. Our findings align with an embodied perspective of perception, suggesting that action-related semantic knowledge can influence relatively early stages of visual processing. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Public Significance This study shows that meaningful relations between objects that are typically used together, such as the relation between hammer and nail, can help people identify individual objects when perceptual conditions are difficult, in particular when objects appear in visually cluttered environments. This suggests that higher-level knowledge derived from interactions with objects may influence lower-level perceptual processes. Findings highlight the importance of considering semantic context to understand the perception of objects in everyday, naturalistic environments. Introduction Our everyday visual environments are rich and usually involve many objects. Objects that appear within the same visual context are often semantically linked. For example, a kitchen environment usually includes specific furniture, food, appliances and utensils. These objects are typically used together when preparing meals or eating (e.g., spatula on a pan, a spoon on a bowl). Temporal, spatial, causal and functional relations between objects that play complementary roles in the same event (e.g., pen and paper in writing) have been extensively studied and have been referred to as ‘thematic relations’ (Estes et al., 2011 ). The central goal of the present study is to better understand the influence of thematic relations on the perception of (objects in) complex visual environments. A large body of evidence now indicates that thematic relations are an important organizing principle of conceptual knowledge. For example, in a series of experiments Lin & Murphy, ( 2001 ) demonstrated that thematic relations are meaningful and used by adults in a wide range of categorization tasks, similar to other relations, such as taxonomic relations (categorization based on shared features). In one of their experiments, participants performed a classical forced-choice category task. Participants were presented with a triad of images, consisting of a target (e.g. pencil), an image thematically related to the target (e.g., eraser), and an image taxonomically related to the target (e.g., pen). They were asked to indicate which image “goes best with the target to form a category”. “Category” was defined as “a set of things or people that share some commonalities – be it genetic makeup, functions, purposes, physical and perceptual characteristics, or behavioral predispositions.” Despite the instructions emphasizing taxonomic categorization, results showed that the majority (66%) of participants chose predominantly thematic alternatives. This study demonstrated that thematic relations play a crucial role in semantic organization. Further studies demonstrated that thematic relations are particularly salient and for manipulable artifact objects at both the behavioral and brain levels (see below, Kalénine et al., 2012 ; Kalénine & Bonthoux, 2008 ; Wamain et al., 2014 ). In one study, participants had to identify taxonomic and thematic relations in a force-choiced categorization task where a semantically-related picture and an unrelated picture were presented below a target picture (Kalénine et al., 2009 ). Depending on the condition, the related picture was thematically or taxonomically related to the target. The triads could be artefacts or natural objects and manipulable or non-manipulable. Despite strict control of the degree of semantic relatedness between conditions, behavioral results showed that thematic relations were more rapidly identified for artefacts (e.g., bed-person asleep) than natural targets (e.g. squirrel-hazelnut), especially for manipulable artefacts (e.g., bowl-toast). On the contrary, taxonomic relations were more rapidly identified for natural than artefact targets. Thus, taxonomic conceptual processing relies on perceptual similarities and is facilitated for natural objects that are defined by their perceptual features. On the other hand, thematic conceptual processing relies on functional relations and is facilitated for artefacts, especially when manipulable as they are defined by their functional features (Kalénine et al., 2009 ). This study demonstrates that taxonomic and thematic relations are based on different sensory-motor processes. In particular, thematic relation processing is closely related to action knowledge about artefacts. The relevance of thematic relations for structuring knowledge about manipulable artefacts may be related to their proximity with the representations of actions. These representations are acquired through experience of object use. Neuroimaging results (Kalénine et al., 2009 ) showed that the posterior temporo-parietal cortex, a region typically associated with action tool knowledge (i.e. information about tool use) was selectively activated during the processing of thematic relations for manipulable artefacts. This region has been further shown critical for both identification of thematic relations for artefacts and recognition of object use gestures in people with brain lesions (Kalénine & Buxbaum, 2016 ; Tsagkaridis et al., 2014 ). Consistent with the involvement of action-related processes in the identification of thematic relations between objects, the existence of a “paired-affordance effect” has been demonstrated in the literature on visual attention and object perception in both healthy individuals (Borghi et al., 2012 ; Roux-Sibilon et al., 2018 ; Xu et al., 2015 ; Yoon et al., 2010 ) and brain-damaged patients (Riddoch et al., 2003 , 2006 ). This effect corresponds to better performance in correctly judging whether object pairs are used together when objects are correctly rather than incorrectly positioned for action. A correct position for action involves spatially arranging the objects in a way that is congruent with their common use (e.g. corkscrew positioned above wine bottle) and/or orienting the active object of the pair (i.e. the tool) towards dominant hand. Such a facilitatory effect has also been found when detecting (Riddoch et al., 2003 ) and identifying objects (Roberts & Humphreys, 2011 ). In a study of Robert and Humphreys (2011), pairs of objects were presented centrally to right-handed participants for 50 ms. Pairs consisted of two types of objects, one active (e.g. corkscrew; light match) and one passive (e.g. wine bottle; candles). Right-handed participants had to verbally identify the two objects. The two objects could be thematically related (wine bottle and corkscrew) or non-related (wine bottle and light match). The active object (corkscrew or light match) could be presented on the right side or on the left side of the pair. The active object could also be correctly positioned for use on the passive object (i.e., functionally oriented towards the passive object). Results showed that both passive and active objects were more often correctly identified when they were thematically related than unrelated. Importantly, results also showed more correct identifications when objects were correctly positioned compared to incorrectly positioned (for right-handed participants). Taken together, findings indicate that thematic relations and action position of both objects play a role in identifying objects in pairs. In this line of research, objects are usually presented in simple viewing conditions. Object pairs are shown in isolation in central vision. Similarly, the role of thematic relations in object processing has been mostly studied with objects (object pairs or triads) presented in isolation. In naturalistic settings, however, objects frequently appear surrounded by other objects in complex, often cluttered scenes and rapidly changing visual environments. One of the main limits of visual perception in cluttered environments is crowding: When a target is surrounded by other stimuli (flankers), target identification is worse compared to presentation of the target in isolation (Herzog et al., 2015 ; Levi, 2008 ). Crowding has been considered as a fundamental limitation of our peripheral vision, and extensive research has been conducted into the principles governing crowding using comparatively simple stimuli such as Verniers (Malania et al., 2007 ; Manassi et al., 2012 ; Sayim et al., 2008 , 2010 ; Westheimer & Hauske, 1975 ), Gabor patches (Greenwood et al., 2010 ; Parkes et al., 2001 ; Saarela et al., 2009 ; Wilkinson et al., 1997 ) and letters (Liu & Arditi, 2001 ; Sayim et al., 2014 ; Song et al., 2014 ). These studies have identified several low- and mid-level factors that influence crowding. For instance, crowding is strongly affected by the spacing between the target and the surrounding flankers: The closer the flankers are to the target, the stronger the crowding effect (Bouma, 1970 ; Levi et al., 2002 ; Pelli, 2008 ; Pelli et al., 2004 ; but (Melnik et al., 2018 , 2020 , below). Similarly, the greater the similarity of the target and the flankers, the stronger the crowding effect (but see Rummens & Sayim, 2019 , 2021 , 2022 ). For instance, Kooi et al., ( 1994 ) presented participants with peripheral letters Ts, where one T served as the target and was surrounded by four other Ts. The stimuli varied in contrast polarity (e.g., black target and white flankers). Results indicated better performance when the target and flankers were dissimilar. Subsequent studies replicated these findings and extended them to other features such as size, orientation, spatial frequency, color, and motion (see Whitney & Levi, 2011 ) for a review). Historically, these data were interpreted as interference from low-level property similarity between flankers and target with stronger interference when they are similar and weaker interference when they are dissimilar. Generally, segregating the target from the flankers becomes more difficult when they share common low-level properties. However, recent studies found inverse patterns and suggested that mechanisms on the level of perceptual organization, such as grouping, also play a key role in crowding. For example, Melnik et al. ( 2018 , 2020 ) showed that reducing distance between the flankers and the target, thereby maximizing strong target-grouping, can yield emergent features that in turn reduced crowding effects. Orientation discrimination of a chevron target stimulus improved when closely spaced flankers allowed the formation of a diamond configuration (“<>”) compared to an alternative configuration. These results show that crowding is modulated not only by target-flanker similarity but also by the emergent features of particular configurations. Despite the failure to identify the target, some information of the target appears to be still processed, influencing subsequent behavior. This has been shown for several lower-level visual features such as orientation (He et al., 1996 ), motion (Aghdaee, 2005 ), or position (Whitney, 2005 ). Importantly, a few studies demonstrated that higher-level semantic features of the target, such as facial expressions (Kouider et al., 2011 ), biological motion (Ikeda et al., 2013 ) and semantic information of Chinese words (Chien et al., 2022 ; Yeh et al., 2012 ; Zhou et al., 2016 ) may survive crowding. For example, in Yeh et al. ( 2012 ), naïve Chinese participants were presented with Chinese target words in the periphery, under both crowded and isolated conditions. Subsequently, a single Chinese character appeared centrally that could be semantically related to the target or not, and participants had to judge whether the character was a word or non-word (lexical decision task). Participants then performed a second task where they had to identify the crowded word seen earlier. Results showed that despite the inability to identify crowded words, crowded stimuli elicited semantic priming effects in the lexical decision task: participants were better when the words were semantically related than unrelated. In the present study, we investigated whether thematic relations between objects can still be processed when everyday manipulable objects are presented under crowded conditions. Given that thematically related objects tend to co-occur in visual environments and are often used together, we hypothesized that thematic relations between objects could survive crowding and help object identification, possibly reflecting semantic grouping. In addition, we also hypothesized that the spatial configuration of thematically related objects may influence object identification. Specifically, we expected better performance when related objects were positioned in a way that affords their direct use, as compared to when they were positioned incorrectly. EXPERIMENT 1: Processing of thematic relation under visual crowding. The aim of Experiment 1 was to assess whether effects of thematic relations between objects are preserved under visual crowding. To this aim, we presented pairs of objects that were either thematically related or unrelated pairs. Additionally, objects were either positioned in a way that is compatible with their co-actor from the participant’s perspective or not. Finally, objects were presented either in isolation or crowded by meaningless flankers. We measured the performance of identification of objects in these different conditions. Methodology Participants Sixty-five right-handed participants (Edinburgh Handedness Inventory, Oldfield, 1971) (mean age: 20.2, sd: 2.6) with normal or corrected-to normal vision took part in the study. All participants were students at the University of Lille. They gave informed consent and participated in exchange for course credits. The protocol was approved by the Ethical Committee of the University of Lille (reference: 2022-605-S106). Stimuli and procedure Stimuli A total of 12 sets (Appendix 1) of 3D-grayscaled (8-bit) images of manipulable manufactured objects selected from Roux-Sibilon et al. (2018) were used as stimuli. Image sizes ranged from 0.6° to 2.3° in width and from 1.1° to 2.3° in height. Each set involved an 'active' object usually held and manipulated by the dominant hand, (e.g., a pen), a 'passive' object usually held but not manipulated by the non-dominant hand (e.g., a notebook), and an 'unrelated' object not semantically related and used in co-action with either the active or the passive object (e.g., a hat). The 3 objects of each set were arranged in pairs to manipulate the relation between objects: objects were either thematically related (T) or unrelated (UN). The T condition was created with the active and passive objects, which together form a typical object-action pair (e.g. pen and notebook for writing). The UN condition was created by replacing either the active or passive object with the unrelated object (i.e., active + unrelated and passive + unrelated). Additionally, within each pair, objects were presented on half of the trials on the right side and in the other half on the left side of the pair. For the T condition, when the passive object was on the left side and active object on the right side, the objects were correctly positioned for action (A+), when swapped the objects were incorrectly positioned for action (A-). (Figure 1). We controlled our stimuli with objective and subjective measures of visual similarity and semantic relatedness between objects. For each pair, images were compared using the Feature Similarity algorithm (FSim; Zhang et al., 2011), which compares two images and provides a similarity value (from 0 to 1) based on their low-level visual features (i.e., phase congruency and gradient magnitude). FSim values were averaged between T and UN pairs. No significant differences were found between the two conditions (T = 0.657, UN = 0.634, Welch’s t-test = 0.6671, p = 0.512). Subjective ratings were collected in an additional sample of 21 participants. For each pair, participants answered two questions sequentially. They were first asked to rate the extent to which the two objects were visually similar (considering overall shape, details, colors, etc.), and second, the extent to which the two objects were semantically linked. Participants responded on a 7-point scale from not at all (1) to very (7) visually similar/semantically linked. As expected, participants judged T pairs (mean = 6.83) to be more semantically related than UN pairs (mean = 1.75, KS = 1, p0.124). Nonetheless, FSim measures and subjective ratings of visual similarity were entered as a control variable in the statistical model (see Data Analysis and Results section.). Finally, pairs were presented in isolation (isolated condition) or surrounded by six flankers (crowded condition). The flankers were of three distinct grayscale object designs (Figure 2). Crowding is usually strong when the target and flankers are highly similar (Whitney & Levi, 2011 for a review, but see Rummens & Sayim, 2021, 2022). Therefore, we have chosen object-like flankers similar to the targets to ensure sufficient crowding. Note that object-like flankers were meaningless and identical across all pairs for two reasons. First, we wanted to avoid potential variations in perceptual or semantic similarity between targets and flankers. Finally, we wanted to prevent participants from being misled by recognizable objects during identification. In crowded conditions, the flankers were randomly positioned directly to the outer edges of the object images (Figure 2). Procedure The experiment has been conducted using Psychopy software (v2022.1.3) (Peirce et al., 2019). We used a 15.6-inch screen (1920x1080, 60Hz). Stimuli appeared at 6.2 degrees eccentricity either above or below a fixation cross (0.8 degrees) in the center of the screen. The two images of each pair were separated by 0.2 degrees (edge-to-edge of the image rectangles). Participants were positioned approximately 57 cm from the monitor. Prior to the main task, participants performed a familiarization task where stimuli were presented individually in the center of the screen with their corresponding names displayed below. The familiarization task aimed to avoid ambiguity about individual object identification prior to the experiment. Then participants performed a practice of the main task to familiarize themselves with the procedure. The procedure of the practice task was the same as in the main task, except that it used different stimuli (16 trials) and provided written feedback of the correctness of their responses after each trial. After the practice, participants performed the main task. During the task, participants were presented with pairs of objects under different conditions and were asked to identify one object from each pair. Each trial proceeded as follows: a fixation cross was displayed for 1000 ms, followed by the presentation of a pair of objects for 150 ms. After a 250 ms interval, the written name of an object appeared in the center. The object name could correspond to one of the objects of the pair presented previously or not (Figure 3.). Participants were instructed to determine as quickly and accurately as possible whether the word corresponded to one of the images presented in that trial or not (“yes” or “no”). Participants responded by pressing ‘p’ or ‘l’ on the keyboard with their right forefinger and middle finger respectively. The key assignments were counterbalanced between participants. For T pairs, the target object appeared on both sides of the pair, once on the left and once on the right. Conversely, in half of the UN trials, the target object appeared on one side of the pair, either left or right, with the assignment counterbalanced across sets (i.e., left object for active/unrelated pairs and right object for passive/unrelated pairs, or vice versa). This ensured an equal number of yes and no trials in the T and UN conditions. In total, participants performed 768 trials: 12 (sets) * 2 (Pair Relatedness conditions) * 2 (Position for Action conditions) * 2 (Crowding conditions) * 2 (Target side) * 2 (Target Presence) * 2 (Visual field conditions). Trials were presented in random order for each participant. A break was proposed every 96 trials. Response, accuracy, and response times (RT) in milliseconds were recorded. Results Data processing We verified that the performance of each participant exceeded the chance level in the isolated condition (i.e. above 54,2%, p<0.05). All participants performed above chance level. Next, participants with accuracy scores more than 2.5 standard deviations above or below the group mean in the isolated or crowded conditions were excluded (Berger & Kiefer, 2021). One participant fell below the upper threshold in the isolated condition and was subsequently removed from the analysis, leaving 64 participants. The same procedure was applied at the set level to assess whether performance from a given set deviated excessively from the others, but no outliers were detected. Regarding response times (RT), we excluded trials with incorrect responses (25.3%). A global trimming procedure was applied to remove aberrant responses. RT inferior to 200 ms and RT superior to five times the median (4171 ms) were excluded (0.38% trials). Then, we removed RTs superior or inferior to 2.5 standard deviations from the mean RT of each participant in each condition (Participant x Target Presence x Crowding x Pair Relatedness x Position for Action). This global trimming procedure excluded 3.24% of trials. Data analysis Logistic and linear mixed-effect regression models were used to analyze accuracy and response times, respectively. We hypothesized that thematic relations between objects would facilitate the correct identification of objects that were actually present in the display. Thus, analysis of identification performance was restricted to trials on which the target was present. In our mixed-effect regression models, fixed effects corresponded to the factors of interest, including Crowding (isolated, crowded), Pair Relatedness (T, UN pairs), Position for Action (correct, incorrect) and their interactions. Subjective ratings of visual similarity and FSim values were entered as additional fixed effects to control for visual similarity differences between the two objects in the pairs. Contrasts were manually coded as -0.5 and +0.5. The maximal random effect structures reaching model convergence were selected following the guidelines of Barr et al., (2013) and Bates et al., (2015). Random structure simplification was conducted using an iterative procedure based on Principal Component Analysis with the rePCA function from the lme4 package. Final models included random intercepts and slopes for both participants and targets. Participants were further nested in task order and targets were nested in item sets (see syntax in Appendix 3). For linear mixed-effect models, Westfall’s d measures were computed as an alternative to Cohen’s d (Brysbaert & Stevens, 2018; Judd et al., 2017; Westfall et al., 2014). Westfall’s d for a given contrast is obtained by dividing the estimate for that contrast (difference of estimated means) by the square root of the sum of the variance of all random effects and residuals of the model. Overall, we expected a main effect of Crowding, with worse identification performance in the Crowded compared to Isolated conditions. Furthermore, we wanted to evaluate whether Pair Relatedness would be modulated by crowding through the Crowding x Pair Relatedness interaction. In particular, we wanted to test whether an advantage for T pairs would occur in both the isolated and crowded conditions. Additionally, in line with studies on the paired-object affordance effect (Yoon & Humphreys, 2010), Position for Action of thematic pairs was expected to impact performance, with correct position for action facilitating target identification in thematic pairs in both the isolated and crowded conditions. Accuracy Descriptive statistics for the different conditions are presented in Table 1. First, logistic mixed-effect models showed a significant effect of crowding on correct responses with 4.76 times worse target identification performance in the crowded compared to isolated condition (Odds ratio (OR) = 4.76, CI 95% = 3.57 – 6.25, p<0.001) (Figure 4). There was also a significant main effect of Pair Relatedness with 1.85 times better target identification performance in the thematically related compared to unrelated condition (OR = 1.85, CI 95% = 1.55 – 2.2, p<0.001). Importantly, there was a significant interaction between Crowding and Pair Relatedness (OR = 1.67, CI 95% = 1.429 – 1.961, p<0.001). The more accurate identification for T pairs than UN pairs was stronger in isolated than crowded conditions. While crowded UN pairs were identified at chance level (t = 0.478, CI 95% = 0.471 – 0.546, p = 0.634), crowded T pairs were identified above chance (t = 3.311, CI 95% = 0.529 – 0.619, p = 0.001). Finally, Position for Action did not influence correct responses, neither as a main effect (OR=1.05, CI 95% = [0.94 – 1.16], p>0.42) nor in interaction with Crowding (OR = 0.08, CI 95% = 0.89 – 1.31, p>0.42). Table 1 Means and Standard Errors in Accuracy (proportion of correct responses) and Response Times (RTs) for correct responses in the different conditions of Experiment 1. Note that in thematically unrelated pairs, the target object differed across the Position for Action conditions Accuracy RTs Condition M SE M SE Thematically Related Isolated Correctly positioned 0.863 0.010 799 12 Incorrectly positioned 0.868 0.009 807 12 Crowded Correctly positioned 0.574 0.012 974 17 Incorrectly positioned 0.574 0.012 964 13 Thematically Unrelated Isolated Correctly positioned 0.716 0.011 870 13 Incorrectly positioned 0.743 0.010 833 11 Crowded Correctly positioned 0.499 0.010 1037 18 Incorrectly positioned 0.519 0.010 972 15 Response times Considering the percentage of correct responses in isolated and crowded conditions, response times for correct responses (RTs) were mainly investigated to verify the absence of speed-accuracy trade-off in participants’ responses. There was a main effect of Crowding, with slower RTs in the crowded than isolated condition (estimate = 0.15, CI 95% = 0.12 – 0.19, p<0.001, Westfall’s d = 0.41) and a main effect of Pair Relatedness, with faster RTs for T pairs than UN pairs (estimate = -0.08, CI 95% = -0.11 – -0.05, Westfall’s d = 0.21). There was no interaction between Pair Relatedness and Crowding (estimate = 0.02,CI 95% = -0.01–0.04 , p>0.15, Westfall’s d = 0.04). Again, Position for Action did not influence RTs, neither as a main effect (estimate = 0.003, CI 95% = -0.01 – 0.02, p>0.77, Westfall’s d = 0) nor in interaction with Crowding (estimate = 0.002, CI 95% = -0.02 – 0.03, p>0.89, Westfall’s d = 0). Discussion of experiment 1 In Experiment 1, participants were presented with pairs of objects that could be thematically related or unrelated. Experiment 1 has demonstrated that target objects (e.g., pen) were identified more accurately when they appeared in thematically-related (e.g., pen – notebook) than semantically unrelated object pairs (e.g., pen – hat). This thematic advantage was also present in crowded conditions, albeit to a lesser extent. We found no evidence that performance was modulated as a function of the position for action of related pairs, either in isolation or under visual crowding. While the thematic advantage observed is unequivocal, its locus remains to be clarified. In Experiment 1, the target noun (i.e., “pen”) was systematically related to the other, non-target object of the pair (i.e. notebook) in thematically related pairs (e.g., pen – notebook). Therefore, the advantage of the thematic effect could originate from two different stages. During the visual presentation of object pairs (e.g. pen-notebook), thematic pairs could benefit from perceptual grouping and be better perceived (Auckland et al., 2007; Green & Hummel, 2006; Nah & Geng, 2022; Roberts & Humphreys, 2011). Alternatively, after the visual presentation of the objects and during the presentation of the target nouns (e.g., when the word “pen” is presented), processing of the thematic relation between the target noun (e.g., word “pen”) and one of the perceived objects (i.e., the non-target object, (e.g., “notebook”) could bias the decision towards target presence, even without the thematic relation being perceived in the visual object pair (Biederman et al., 1982; Hollingworth, 1998; Hollingworth & Henderson, 1999). In other words, participants may have identified only the passive object (e.g., notebook) and inferred based on their semantic knowledge that the thematically related object (e.g., pen) was likely to be presented. This guessing strategy could explain in part or in full the observed benefit of thematic relation on object identification. Indeed, our design did not include trials where the target noun (e.g., “pen”) was related to the non-target object (e.g., notebook) but was actually absent. In either case, results suggest that relatively high level semantic information of manipulable objects survived crowding and improved performance. Experiment 2 was then designed to investigate whether the observed thematic advantage on object identification originated from the thematic relation between objects during pair presentation and/or from a decisional strategy based on semantic information of single objects during the target noun presentation. Experiment 2: Locus of the thematic advantage Experiment 2 aimed to further investigate the locus of the thematic advantage observed in Experiment 1. We manipulated the relationship between the target noun and the non-target object to evaluate its influence on participants’ decisions. Critically, we designed “no” trials in which the target was absent but target nouns could still be thematically related to one object of the pair. This design allowed us to disentangle perceptual and decisional contributions to the thematic advantage by computing measures of sensitivity and bias. A decisional strategy would be expected to increase bias toward ‘yes’ responses and reduce sensitivity when the target noun was thematically related –compared to unrelated– to one of the objects of the pair. In contrast, a perceptual contribution to the thematic advantage should not affect either bias or sensitivity based on the relation between the target noun and the non-target object. Experiment 2 was conducted online and aimed at replicating the findings from Experiment 1. Methodology Participants Right-handed participants native speakers of English were recruited via Prolific. Given the methodological similarities between Experiments 1 and 2, a post hoc power analysis of Experiment 1 was conducted to determine the appropriate sample size for Experiment 2. Utilizing a simulation-based approach (Kumle et al., 2021), we computed the post hoc power to detect the observed interaction between Pair Relatedness and Crowding in the mixed logistic regression model used in Experiment 1. The analysis indicated a power of 98% with the actual 64 participants. Considering the high power of Experiment 1, we aimed to include an equivalent number of participants in Experiment 2. Initially, 75 participants were recruited. Exclusion criteria led to the removal of 10 participants: one due to a screen refresh rate below 30 Hz, five for responding in less than 150 ms, seven for a mean accuracy at chance level in the isolated condition, and one for having an accuracy in the isolated condition exceeding 2.5 standard deviations above the mean of the group. Consequently, the final sample comprised 65 participants (28 female, 34 male, 1 unknown, mean age: 25 years, SD: 3.2). Stimuli Images of objects were the same as in Experiment 1 with the addition of five new sets ( Appendix 2) to improve the reliability of the statistical analyses. The main differences from Experiment 1 were the following. In Experiment 2, UN pairs were created with the passive object (e.g., notebook) and the unrelated object (e.g., screwdriver). Additionally, among each set, the identification was limited to the unrelated object (e.g. screwdriver) and the active object (e.g., pen), while the passive object (e.g., book) was always present but remained task-irrelevant as it was never the target for identification (Figure 5). The object to detect appeared in 50% of the trials. Critically, the passive object could either be thematically related or unrelated to the target noun. In the Target Related (TR) condition, the target noun (e.g., pen) was always thematically related to the passive object (e.g., book) regardless of whether the target was present or not. In contrast, in the Target Unrelated (TU) condition, the target noun (e.g., screwdriver) was never thematically related to the passive object regardless of whether the target was present or not. With this design, we can quantify the influence of the relation between the target noun and the non-target object on object detection performance. As in experiment 1, we collected two types of subjective ratings on a 7-point scale on two additional samples of participants on the visual and semantic properties of the object pairs. One sample of 15 participants rated the extent to which the two objects were semantically related. As expected, T pairs (mean = 6.69) were rated to be more semantically linked than UN pairs (mean = 1.59, STATS). Another sample of 22 participants rated the extent to which the two objects were visually similar. T and UN pairs were rated to be equally visually similar (mean = 1.96). Procedure The experiment was conducted using Psychopy software (v2023.2.2) and was hosted online on Pavlovia (Peirce et al., 2019). To normalize stimuli size presentations, we used a screen calibration: participants adjusted their screen resolution by matching a virtual standardized credit card with their own one. To control the distance between participants and screen, we used a blind spot calibration method: participants were asked to close one eye, fixate a fixation cross and adjust their position until a red dot in their visual periphery was no longer visible. The correct distance was set to be around 60 cm away from the screen. The experimental procedure was identical to that of Experiment 1, except for three minor modifications. First, to ensure proper attention to the familiarization phase, a second phase was added. In this additional phase, participants were required to discriminate between the objects they had previously seen and visually similar exemplars of the same objects. Feedback was provided after each trial. Second, the number of trials in the practice phase was doubled (32 trials) to improve task familiarization in the online setting. Third, in the main task, each trial ended either upon the participant’s response or after 6 seconds in the absence of a response. Data analysis To evaluate the replication of the results from Experiment 1, we conducted the same analysis of accuracy and correct response times using logistic and linear mixed-effect regression respectively, following the same procedure (Appendix 4). In addition, for the purpose of this second Experiment, we computed a measure of sensitivity and bias (Signal Detection Theory) using a logistic mixed-effect regression model (Zloteanu & Vuorre, 2024). The purpose of this analysis was to estimate the influence of Target Relatedness on discrimination performance and bias. Specifically, we wanted to evaluate whether participants would erroneously identify the target ‘pen’ when they were presented with the unrelated object pair 'screwdriver-notebook’ because of the presence of the notebook. This model estimated the probability to answer “yes” and included fixed effects for Target Presence (i.e., whether the object target was present or absent), Target Relatedness, Crowding and Position for Action. In this framework, fixed effects that interact with Target Presence represent perceptual sensitivity (analogous to d’), whereas main effects and interactions that do not involve Target Presence represent systematic response biases (analogous to decision criterion). Compared to traditional signal detection theory analyses, this modeling approach offers the advantage of accounting for variability across participants and items by incorporating random effects. Results Data processing Initially, we ensured that the performance of each participant exceeded chance level on isolated trials (i.e. above 55%, p<0.05). Seven participants were at chance in the isolated condition and were thus excluded. Next, participants with accuracy scores above or below 2.5 standard deviations from the group mean in the crowding conditions were excluded. One participant was above the upper threshold in the isolated and crowded conditions and was subsequently removed from the analysis. This procedure was also applied at the set level, and no outliers were detected. Finally, two participants were excluded for responding in less than 200 ms on at least three consecutive trials, repeated seven or more times throughout the experiment. The final sample included 65 participants. Regarding response times (RT), we excluded incorrect response trials (37.2%). A global trimming procedure was applied to remove aberrant responses, RTs shorter than 200 ms & RTs longer than five times the median (3946 ms) were then excluded (1.2% trials removed). Then, we removed RTs superior or inferior to 2.5 standard deviations from the mean RT of each participant in each condition (participant x Target Presence x Crowding x Pair Relatedness x Position for Action). This global trimming procedure excluded 2.9% of trials. Accuracy Descriptive statistics of the different conditions are presented in Table 2. Logistic mixed-effect models showed a significant effect of crowding on correct responses with 4.17 times worse target identification performance in the crowded compared to isolated conditions (Odds Ratio (OR) = 4.17, 95% CI = 3.12–5.55, p < 0.001) (Figure 6). There was also a significant main effect of Pair Relatedness (OR = 1.66, CI = 1.16 – 2.38, p=0.005). More importantly, there was a significant interaction between Crowding and Pair Relatedness (OR = 1.41, 95% CI = 1.22–1.64, p < 0.001). The higher accuracy for T pairs than UN pairs was stronger in isolated than crowded conditions. We also found an interaction between Position for Action and Crowding (OR = 1.27, 95% CI = 1.10–1.46, p = 0.001), indicating that the negative effect of crowding was weaker when objects were correctly positioned for action compared to when they were incorrectly positioned. Table 2 Means and Standard Errors in Accuracy and correct Response Times (RTs) according to Pair Relatedness, Position for Action and Crowding in Experiment 2 Accuracy RTs Condition M SE M SE Thematically Related Isolated Correctly positioned 0.809 0.014 746 11 Incorrectly positioned 0.825 0.016 744 10 Crowded Correctly positioned 0.541 0.014 834 13 Incorrectly positioned 0.516 0.015 857 18 Thematically Unrelated Isolated Correctly positioned 0.701 0.014 793 12 Incorrectly positioned 0.718 0.013 792 11 Crowded Correctly positioned 0.471 0.014 887 20 Incorrectly positioned 0.440 0.015 889 14 Responses times Given the high accuracy in both the isolated and crowded conditions, RTs were mainly examined to verify the absence of speed-accuracy trade-off in participants' performance. Results revealed a significant main effect of Crowding, with slower RTs in the crowded condition compared to the isolated condition (estimate = 0.10, 95% CI = 0.06–0.14, p < 0.001, Westfall’s d = 0.24). Additionally, a significant main effect of Pair Relatedness was observed, with faster RTs for thematically related (T) pairs compared to unrelated (UN) pairs (estimate = -0.06, 95% CI = -0.10 to -0.01, p = 0.012, Westfall’s d = 0.14). In contrast, Position for Action had no significant effect on RTs, either as a main effect ( p > 0.41) or in interaction with Crowding ( p > 0.06), Pair Relatedness, or both ( p > 0.88). Sensitivity The overall sensitivity of the task is represented by the main effect of Target Presence (OR = 3.90, 95% CI = 2.72–5.52, p < .001) (Figure 6). Sensitivity was reduced by crowding, as indicated by the significant interaction between Target Presence and Crowding (OR = 0.24, 95% CI = 0.22–0.27, p < 0.001), with sensitivity being 4.17 times greater in the isolated than in the crowded condition. Importantly, we found an interaction between Target Presence and Target Relatedness (OR = 1.20, 95% CI = 1.09–1.32, p < 0.001) which evaluates the sensitivity to Target Relatedness. When there was no relation between the target noun and the passive, non-target object, sensitivity was 1.19 higher than when there was a thematic relation. Interestingly, this effect was significant for the isolated conditions (OR = 1.23, 95% CI = 1.07–1.42, p0.06). We also found an interaction between Target Presence and Position for Action (OR = 1.25, 95% CI = 1.03–1.52, p<0.024), suggesting that the crowding effect was 1.25 times weaker when objects were correctly positioned for action, compared to when they were incorrectly positioned for action. Bias The overall bias of the task is represented by the Intercept (OR = 0.95, 95% CI = 0.79- 1.11, p>0.6). We found a main effect of Crowding (OR = 1.92, 95% CI = 1.61-2.27, p<0.001), participants were 1.92 times less likely to respond “yes” in the crowded compared to the isolated condition. We found a main effect of Target Relatedness (OR = 1.84, 95% CI = 1.39- 2.44), participants were 1.84 more likely to respond “yes” when the target was related than unrelated. We also found an interaction between Crowding and Target Relatedness (OR = 1.44, 95% CI = 1.28 -1.56, p<0.001), the effect of Target Relatedness was 1.44 times weaker in the crowded than isolated conditions. In addition, the impact of Target Relatedness on the bias was present both in isolated (OR = 2.19, 95% CI= 1.64 – 2.91, p<0.001) and crowded conditions (OR = 1.55, 95% CI = 1.16- 2.06, p<0.003). Finally, we found an interaction between Position for Action and Crowding (OR = 1.14, 95% CI = 1.03-1.25, p<0.01) suggesting that the impact of crowding on bias was weaker for objects correctly than incorrectly positioned for action. Discussion of Experiment 2 Experiment 2 replicated the finding that objects in thematically related pairs were identified more accurately than those in unrelated pairs, in both isolated and crowded conditions. This demonstrates the robustness of the effect, as we obtained an effect of similar size in an online setting with a new set of stimuli. In addition, Experiment 2 investigated whether the origin of this effect occurs during the presentation of object pairs and/or during the presentation of the target nouns. We found a systematic bias toward “yes” decisions when the target noun was thematically related to the nontarget object in both isolated and crowded trials, although the effect was weaker under crowding. In addition, target discrimination was poorer when the target was thematically related, but only for isolated pairs: it was more difficult to detect a “pen” compared to a “screwdriver” when there was a book among the object pair. These results will be further addressed in the General Discussion. Contrary to Experiment 1, we found a significant interaction between visual crowding and position for action on identification performance. Specifically, the effect of crowding on accuracy and sensitivity was reduced when objects were correctly positioned for action compared to incorrectly positioned for action. Crucially, this benefit was independent from the presence of a semantic relationship between objects. This indicates that it is the action-related configuration itself, rather than semantic knowledge, that interacts with visual crowding and improves identification. This finding is consistent with prior studies showing that objects correctly positioned to be grasped and manipulated in a plausible interaction benefit from perceptual grouping (Auckland et al., 2007; Green & Hummel, 2006; Roberts & Humphreys, 2011). For example, Robert & Humphreys (2011) found that briefly presented object pairs (i.e., a corkscrew and a wine bottle) were identified more accurately when they were correctly than incorrectly positioned to be grasped and manipulated, even for unrelated pairs (e.g. a match and a wine bottle). Likewise, studies of patients with visual extinction have shown better recognition of two objects when they are arranged in a plausible action configuration (Riddoch et al., 2003, 2006). These results have been interpreted as evidence that affording spatial relations between objects promotes perceptual grouping. Riddoch et al. (2006) proposed a two‐stage model in which an initial bottom‐up attention stage is sensitive to action‐relevant features: when two objects are correctly positioned to afford a plausible action, attention is automatically spread across both items, making them more likely to be processed and identified. But why do we find this effect only in interaction with crowding? One explanation may lie in the way visual crowding is modulated by the perceptual organization of the target and the flankers. When grouping cues segregate the target from the flankers, crowding is usually reduced (Herzog et al., 2015; but see Rummens & Sayim, 2019, 2022). For instance, a Vernier target flanked by two collinear lines produces strong crowding, but if those lines are perceptually grouped into a rectangle, crowding on the vernier is substantially reduced (Sayim et al., 2010; Manassi et al., 2012). Such effects have been shown with low-level features (e.g., length and color of Vernier stimuli, Malania et al., 2007; Sayim et al., 2008; Manassi et al., 2012) but also higher-level features (e.g., (Mooney) faces, biological motion, (Farzin et al., 2009; Ikeda et al., 2013; Louie et al., 2007). However, the latter authors interpreted these effects not strictly as perceptual grouping effects, but rather as resulting from greater similarity between the target and flankers. In our study, correctly positioned objects may have appeared to be perceptually grouped and thus more segregated from flankers. Grouping of several items has also been shown to improve performance (Rummens and Sayim, 2021, 2022; Sayim et al., 2014). In the present study, such grouping likely diminishes the effect of crowding and improves identification performance. To our knowledge, this is the first study to demonstrate that manipulable objects can be grouped based on action relation spatial configuration, leading to a reduction in visual crowding. However, it remains unclear why this action-related effect occurred only in Experiment 2 and regardless of pair relatedness. One explanation is that in Experiment 1, identification involved passive, active and unrelated objects whereas in Experiment 2, only the active and unrelated objects were task-relevant. In Roberts & Humphreys (2011) the benefit of perceptual grouping was mainly driven by the active object. Thus, in Experiment 2, participants may have focused their attention on the active object, making position for action more influential in facilitating perceptual grouping. General Discussion In this study, we wanted to evaluate the processing of high-level semantic knowledge associated with everyday objects under visual crowding. Pairs of everyday objects were briefly presented in isolation or among meaningless object flankers in the visual periphery. The ability to identify one of the objects was evaluated in different conditions. In the isolated condition, target objects (e.g., pen) were identified more accurately when they appeared in thematically-related (e.g., pen – notebook) than semantically unrelated object pairs (e.g., pen – hat). In the crowded condition, target identification performance for unrelated pairs was at chance level. Critically, we demonstrated a significant performance improvement in target identification when objects were thematically related. Position for action (e.g., pen positioned on the right versus left) did not further modulate this advantage, neither for isolated nor for crowded pairs, although it reduced the deleterious effect of crowding independently from thematic relations in Experiment 2. In the crowding condition, objects in thematic pairs were better identified than objects in unrelated pairs. Yet thematic and unrelated pairs are inherently composed of different objects with different low-level visual properties (overall shape, visual details, etc.). One could argue that low-level visual differences between objects across conditions could account for the advantage of thematic pairs. As the task involved single target identification, participants had to discriminate between the two objects presented. Thus, a lower degree of visual similarity between objects in thematic pairs may have facilitated identification in this condition. To control for this possibility, measures of visual similarity between objects were computed and added to the statistical models. FSim (Zhang et al., 2011) provides an estimation of visual similarity based on the physical properties of the images (Phase Congruency and Gradient Magnitude) and was used as an objective measure of low-level visual similarity. In addition, subjective ratings were collected and used as a complementary estimation of visual similarity between the two images in each pair. The effect of thematic relations on target identification in the crowded conditions was observed even after taking into account these two measures. Thus, while there might be differences in low-level similarity between objects in our sample of stimuli, they do not fully explain the advantage of identifying objects in thematic pairs in crowding conditions. It seems that some aspects about the objects forming a thematic pair survives crowding and facilitates their subsequent identification. But what specific stage do thematic relations impact object identification? One possibility is that thematic relations, through the processing of semantic information about single objects during the perception of the pair, influence responses during decision. Experiment 2 supports this interpretation. Results showed that participants were biased to respond “yes” when the target-noun (e.g., pen) was thematically related to the nontarget object (e.g., the book), even when the target object was absent. In other words, participants may have used the semantic context (i.e., the book) to predict the presence of a thematically congruent object (i.e. the pen). This phenomenon is likely to have been amplified by the fact that related and unrelated objects were relatively similar in shape. This interpretation aligns with the perspective of predictive processing , postulating that expectations can influence visual processing. According to these accounts, context-driven expectations formed through real-world experience can improve perception under ambiguous conditions, such that objects in semantically congruent environments are identified more rapidly and accurately than those in incongruent environments (Bar, 2004; Biederman et al., 1982; Kaiser et al., 2014). However, prior research on object perception in scenes suggests that semantic congruency does not systematically enhance object identification, and that its effects may depend on the specific methodological approach. In a prior study, Hollingworth and Henderson (1998) presented participants with a visual scene followed by a label (i.e., target noun) and asked them to decide whether the labeled object had been present or not. In their first experiment, they found higher identification accuracy for objects semantically congruent with the scene, suggesting a facilitative effect of semantic context on object identification. However, subsequent experiments (Experiments 2 and 3) controlled for semantic congruence in trials where the labeled object was absent. These experiments revealed an increase in false alarms for semantically congruent labels compared to incongruent ones. When computing the sensitivity, incongruent objects were actually better identified than congruent ones. These findings highlight the necessity of distinguishing between perceptual facilitation and decisional biases. The authors gave a functional isolation explanation of their effect, suggesting that context information can orient decisions, but after the perceptual processing stage. In our study, it is therefore likely that the semantic information associated with the nontarget object influenced participants’ decisions during target-noun presentation, as reflected by the observed response bias both in isolation and in crowding conditions. Importantly, our results indicate that semantic knowledge about manipulable objects, such as their functional use with other objects, can get through the bottleneck of visual crowding and influence decision making. Additionally, in line with findings from the broader semantic literature, our results suggest that semantic information from non-target objects is extracted automatically (Biederman, 1972; Biederman et al., 1982; Oliva & Schyns, 1997; Rousselet et al., 2005). However, we cannot rule out that the thematic link between object pairs influenced processing at a perceptual level (i.e., during perception of the pairs). In a follow-up to the work by Hollingworth and Henderson (1998), Auckland, Cave, and Donnelly (2007) further examined the influence of semantic context on object identification. They conducted a six-alternative forced-choice (6-AFC) experiment to isolate perceptual benefits of semantic context from response biases. On each trial, participants briefly saw a target object centrally presented and surrounded by four distractor objects that were either semantically related (e.g., playing cards surrounded by dice) or unrelated (e.g., playing cards surrounded by fruits). Then, participants selected the target from six written words, chosen to dissociate from perceptual and semantic errors. Results showed higher identification for targets embedded with related than unrelated targets. They found semantic errors suggesting that decisional bias can explain a part of the semantic benefits. Importantly, after correcting for this bias, the facilitation remained. This finding suggests that context facilitation can occur not only with natural scenes but also with arrays of objects. It also indicates that semantic influence on object recognition can occur at both decisional and perceptual stages. Although our study was not originally designed to dissociate perceptual enhancement from decisional bias, we can be fairly confident that both of these processes are involved. Indeed, in Experiment 2, we observed a decisional bias toward thematically related targets when pairs were crowded, but sensitivity remained unaffected. Thus, the improved identification for thematically related pairs might be explained at least by a perceptual enhancement arising from the thematic relation between objects, particularly when they were crowded. Indeed, perceiving crowded objects requires greater attentional resources than perceiving isolated ones. Crowding might thereby promote the processing of thematic relations between objects and increase the likelihood that such relations enhance visual perception. The exact mechanism underlying the benefits of thematic relation on visual processing remains unclear. Thematically related objects tend to co-occur in the same event or scenario. Thus, thematic related objects are more often seen together than unrelated objects. This difference may lead to more familiar visual configurations for related than unrelated object pairs. Thematically related objects would therefore group stronger due to their predictable co-occurrence in the world (Kaiser et al., 2014; Nah & Geng, 2022). Several studies using meaningless stimuli have shown that perceptual grouping among the flankers –and ungrouping from the target– is sufficient to reduce crowding effects (Livne & Sagi, 2007; Manassi et al., 2012, 2013; Sayim et al., 2008, 2010). In Experiment 2, such a mechanism seems to be involved for correctly compared to incorrectly positioned objects (see Discussion of Experiment 2), but whether it also contributes to the advantage observed for thematically related over unrelated objects remains an open question. Future research should investigate whether the observed thematic benefits extend to other types of semantic relations, such as taxonomic ones. While thematic relations are based on complementary roles in common events or actions (e.g., pen–notebook), taxonomic relations group objects based on shared categorical features (e.g., pen–pencil) (Mirman et al., 2017). Exploring whether taxonomic relations also facilitate object recognition under crowding would help determine whether the observed effects on visual perception are specific to action representations or based on higher-order semantic representations. Declarations Competing Interests: Authors have no competing interests. Ethical standards : All human studies have been approved by the Ethical Committee of the University of Lille (reference: 2022-605-S106) and have therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The manuscript does not contain clinical studies or patient data. Funding: This work received support from the French National Research Agency (ANR-23-CE28-0015). The first author benefitted from a PhD fellowship from the University of Lille. Open practice statement: Data, materials and analysis codes are available at: https://doi.org/10.57745/EVVBYS Acknowledgement This work was supported by a PhD fellowship from the University of Lille awarded to the first author, and by a grant from the French National Research Agency (ANR) awarded to Solène Kalénine. The authors thank Laurent Ott for his assistance with the programming of the online experiment, and Dominique Knutsen for her help with word translations. Author contributions • Conceptualization: N.Slaski, S.Kalénine, B.Sayim ; Methodology: N.Slaski, S.Kalénine, B.Sayim; Formal analysis and investigation: N.Slaski ; Writing - original draft preparation: N. Slaski; Writing - review and editing: N.Slaski, S.Kalénine, B.Sayim; Funding acquisition: S. Kalénine; Supervision: S.Kalénine, B.Sayim References Aghdaee, S. M. (2005). Adaptation to spiral motion in crowding condition. Perception , 34 (2), 155-162. https://doi.org/10.1068/p5298 Auckland, M. E., Cave, K. R., & Donnelly, N. (2007). Nontarget objects can influence perceptual processes during object recognition. Psychonomic Bulletin & Review , 14 (2), 332-337. https://doi.org/10.3758/BF03194073 Bar, M. (2004). Visual objects in context. Nature Reviews Neuroscience , 5 (8), 617-629. https://doi.org/10.1038/nrn1476 Barr, D. J., Levy, R., Scheepers, C., & Tily, H. J. (2013). Random effects structure for confirmatory hypothesis testing : Keep it maximal. Journal of Memory and Language , 68 (3), 255-278. https://doi.org/10.1016/j.jml.2012.11.001 Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software , 67 , 1-48. https://doi.org/10.18637/jss.v067.i01 Berger, A., & Kiefer, M. (2021). Comparison of Different Response Time Outlier Exclusion Methods : A Simulation Study. Frontiers in Psychology , 12 . https://doi.org/10.3389/fpsyg.2021.675558 Biederman, I. (1972). Perceiving Real-World Scenes. Science , 177 (4043), 77-80. https://doi.org/10.1126/science.177.4043.77 Biederman, I., Mezzanotte, R. J., & Rabinowitz, J. C. (1982). Scene perception : Detecting and judging objects undergoing relational violations. Cognitive Psychology , 14 (2), 143-177. https://doi.org/10.1016/0010-0285(82)90007-X Borghi, A. M., Flumini, A., Natraj, N., & Wheaton, L. A. (2012). One hand, two objects : Emergence of affordance in contexts. Brain and Cognition , 80 (1), 64-73. https://doi.org/10.1016/j.bandc.2012.04.007 Bouma, H. (1970). Interaction Effects in Parafoveal Letter Recognition. Nature , 226 (5241), 177-178. https://doi.org/10.1038/226177a0 Brysbaert, M., & Stevens, M. (2018). Power Analysis and Effect Size in Mixed Effects Models : A Tutorial. Journal of Cognition , 1 (1), 9. https://doi.org/10.5334/joc.10 Chien, S.-E., Yang, Y.-H., Ono, Y., & Yeh, S.-L. (2022). Theta activity in semantic priming under visual crowding as revealed by magnetoencephalography. Neuroscience Research , 185 , 29-39. https://doi.org/10.1016/j.neures.2022.09.004 Estes, Z., Golonka, S., & Jones, L. L. (2011). Chapter eight - Thematic Thinking : The Apprehension and Consequences of Thematic Relations. In B. H. Ross (Éd.), Psychology of Learning and Motivation (Vol. 54, p. 249-294). Academic Press. https://doi.org/10.1016/B978-0-12-385527-5.00008-5 Farzin, F., Rivera, S. M., & Whitney, D. (2009). Holistic crowding of Mooney faces. Journal of Vision , 9 (6), 18. https://doi.org/10.1167/9.6.18 Green, C., & Hummel, J. E. (2006). Familiar interacting object pairs are perceptually grouped. Journal of Experimental Psychology: Human Perception and Performance , 32 (5), 1107-1119. https://doi.org/10.1037/0096-1523.32.5.1107 Greenwood, J. A., Bex, P. J., & Dakin, S. C. (2010). Crowding Changes Appearance. Current Biology , 20 (6), 496-501. https://doi.org/10.1016/j.cub.2010.01.023 He, S., Cavanagh, P., & Intriligator, J. (1996). Attentional resolution and the locus of visual awareness. Nature , 383 (6598), 334-337. https://doi.org/10.1038/383334a0 Herzog, M. H., Sayim, B., Chicherov, V., & Manassi, M. (2015). Crowding, grouping, and object recognition : A matter of appearance. Journal of Vision , 15 (6), 5. https://doi.org/10.1167/15.6.5 Hollingworth, A. (1998). Does consistent scene context facilitate object perception? Journal of Experimental Psychology: General , 127 (4), 398-415. https://doi.org/10.1037/0096-3445.127.4.398 Hollingworth, A., & Henderson, J. M. (1999). Object identification is isolated from scene semantic constraint : Evidence from object type and token discrimination. Acta Psychologica , 102 (2), 319-343. https://doi.org/10.1016/S0001-6918(98)00053-5 Ikeda, H., Watanabe, K., & Cavanagh, P. (2013). Crowding of biological motion stimuli. Journal of Vision , 13 (4), 20. https://doi.org/10.1167/13.4.20 Judd, C. M., Westfall, J., & Kenny, D. A. (2017). Experiments with More Than One Random Factor : Designs, Analytic Models, and Statistical Power. Annual Review of Psychology , 68 (Volume 68, 2017), 601-625. https://doi.org/10.1146/annurev-psych-122414-033702 Kaiser, D., Stein, T., & Peelen, M. V. (2014). Object grouping based on real-world regularities facilitates perception by reducing competitive interactions in visual cortex. Proceedings of the National Academy of Sciences , 111 (30), 11217-11222. https://doi.org/10.1073/pnas.1400559111 Kalénine, S., & Bonthoux, F. (2008). Object manipulability affects children’s and adults’ conceptual processing. Psychonomic Bulletin & Review , 15 (3), 667-672. https://doi.org/10.3758/PBR.15.3.667 Kalénine, S., & Buxbaum, L. J. (2016). Thematic knowledge, artifact concepts, and the left posterior temporal lobe : Where action and object semantics converge. Cortex , 82 , 164-178. https://doi.org/10.1016/j.cortex.2016.06.008 Kalénine, S., Mirman, D., Middleton, E. L., & Buxbaum, L. J. (2012). Temporal dynamics of activation of thematic and functional knowledge during conceptual processing of manipulable artifacts. Journal of Experimental Psychology: Learning, Memory, and Cognition , 38 (5), 1274-1295. https://doi.org/10.1037/a0027626 Kalénine, S., Peyrin, C., Pichat, C., Segebarth, C., Bonthoux, F., & Baciu, M. (2009). The sensory-motor specificity of taxonomic and thematic conceptual relations : A behavioral and fMRI study. NeuroImage , 44 (3), 1152-1162. https://doi.org/10.1016/j.neuroimage.2008.09.043 Kooi, F. L., Toet, A., Tripathy, S. P., & Levi, D. M. (1994). The effect of similarity and duration on spatial interaction in peripheral vision. Spatial Vision , 8 (2), 255-279. https://doi.org/10.1163/156856894x00350 Kouider, S., Berthet, V., & Faivre, N. (2011). Preference is biased by crowded facial expressions. Psychological Science , 22 (2), 184-189. https://doi.org/10.1177/0956797610396226 Kumle, L., Võ, M. L.-H., & Draschkow, D. (2021). Estimating power in (generalized) linear mixed models : An open introduction and tutorial in R. Behavior Research Methods , 53 (6), 2528-2543. https://doi.org/10.3758/s13428-021-01546-0 Levi, D. M. (2008). Crowding—An essential bottleneck for object recognition : A mini-review. Vision Research , 48 (5), 635-654. https://doi.org/10.1016/j.visres.2007.12.009 Levi, D. M., Hariharan, S., & Klein, S. A. (2002). Suppressive and facilitatory spatial interactions in peripheral vision : Peripheral crowding is neither size invariant nor simple contrast masking. Journal of Vision , 2 (2), 3. https://doi.org/10.1167/2.2.3 Lin, E. L., & Murphy, G. L. (2001). Thematic relations in adults’ concepts. Journal of Experimental Psychology: General , 130 (1), 3-28. https://doi.org/10.1037/0096-3445.130.1.3 Liu, L., & Arditi, A. (2001). How Crowding Affects Letter Confusion. Optometry and Vision Science , 78 (1), 50. Livne, T., & Sagi, D. (2007). Configuration influence on crowding. Journal of Vision , 7 (2), 4. https://doi.org/10.1167/7.2.4 Louie, E. G., Bressler, D. W., & Whitney, D. (2007). Holistic crowding : Selective interference between configural representations of faces in crowded scenes. Journal of Vision , 7 (2), 24. https://doi.org/10.1167/7.2.24 Malania, M., Herzog, M. H., & Westheimer, G. (2007). Grouping of contextual elements that affect vernier thresholds. Journal of Vision , 7 (2), 1. https://doi.org/10.1167/7.2.1 Manassi, M., Sayim, B., & Herzog, M. H. (2012). Grouping, pooling, and when bigger is better in visual crowding. Journal of Vision , 12 (10), 13. https://doi.org/10.1167/12.10.13 Manassi, M., Sayim, B., & Herzog, M. H. (2013). When crowding of crowding leads to uncrowding. Journal of Vision , 13 (13), 10. https://doi.org/10.1167/13.13.10 Melnik, N., Coates, D. R., & Sayim, B. (2018). Emergent features in the crowding zone : When target–flanker grouping surmounts crowding. Journal of Vision , 18 (9), 19. https://doi.org/10.1167/18.9.19 Melnik, N., Coates, D. R., & Sayim, B. (2020). Emergent features break the rules of crowding. Scientific Reports , 10 (1), 406. https://doi.org/10.1038/s41598-019-57277-y Mirman, D., Landrigan, J.-F., & Britt, A. E. (2017). Taxonomic and thematic semantic systems. Psychological Bulletin , 143 (5), 499-520. https://doi.org/10.1037/bul0000092 Nah, J. C., & Geng, J. J. (2022). Thematic object pairs produce stronger and faster grouping than taxonomic pairs. Journal of Experimental Psychology: Human Perception and Performance , 48 (12), 1325-1335. https://doi.org/10.1037/xhp0001031 Oldfield, R. C. (1971). The assessment and analysis of handedness : The Edinburgh inventory. Neuropsychologia , 9 (1), 97-113. https://doi.org/10.1016/0028-3932(71)90067-4 Oliva, A., & Schyns, P. G. (1997). Coarse Blobs or Fine Edges? Evidence That Information Diagnosticity Changes the Perception of Complex Visual Stimuli. Cognitive Psychology , 34 (1), 72-107. https://doi.org/10.1006/cogp.1997.0667 Parkes, L., Lund, J., Angelucci, A., Solomon, J. A., & Morgan, M. (2001). Compulsory averaging of crowded orientation signals in human vision. Nature Neuroscience , 4 (7), 739-744. https://doi.org/10.1038/89532 Peirce, J., Gray, J. R., Simpson, S., MacAskill, M., Höchenberger, R., Sogo, H., Kastman, E., & Lindeløv, J. K. (2019). PsychoPy2 : Experiments in behavior made easy. Behavior Research Methods , 51 (1), 195-203. https://doi.org/10.3758/s13428-018-01193-y Pelli, D. G. (2008). Crowding : A cortical constraint on object recognition. Current Opinion in Neurobiology , 18 (4), 445-451. https://doi.org/10.1016/j.conb.2008.09.008 Pelli, D. G., Palomares, M., & Majaj, N. J. (2004). Crowding is unlike ordinary masking : Distinguishing feature integration from detection. Journal of Vision , 4 (12), 12. https://doi.org/10.1167/4.12.12 Riddoch, M. J., Humphreys, G. W., Edwards, S., Baker, T., & Willson, K. (2003). Seeing the action : Neuropsychological evidence for action-based effects on object selection. Nature Neuroscience , 6 (1), 82-89. https://doi.org/10.1038/nn984 Riddoch, M. J., Humphreys ,G. W., Hickman ,M., Clift ,J., Daly ,A., & and Colin, J. (2006). I can see what you are doing : Action familiarity and affordance promote recovery from extinction. Cognitive Neuropsychology , 23 (4), 583-605. https://doi.org/10.1080/02643290500310962 Roberts, K. L., & Humphreys, G. W. (2011). Action relations facilitate the identification of briefly-presented objects. Attention, Perception, & Psychophysics , 73 (2), 597-612. https://doi.org/10.3758/s13414-010-0043-0 Rousselet, G., Joubert, O., & Fabre-Thorpe, M. (2005). How long to get to the “gist” of real-world natural scenes? Visual Cognition , 12 (6), 852-877. https://doi.org/10.1080/13506280444000553 Roux-Sibilon, A., Kalénine, S., Pichat, C., & Peyrin, C. (2018). Dorsal and ventral stream contribution to the paired-object affordance effect. Neuropsychologia , 112 , 125-134. https://doi.org/10.1016/j.neuropsychologia.2018.03.007 Rummens, K., & Sayim, B. (2019). Disrupting uniformity : Feature contrasts that reduce crowding interfere with peripheral word recognition. Vision Research , 161 , 25-35. https://doi.org/10.1016/j.visres.2019.05.006 Rummens, K., & Sayim, B. (2021). Broad attention uncovers benefits of stimulus uniformity in visual crowding. Scientific Reports , 11 (1), 23976. https://doi.org/10.1038/s41598-021-03258-z Rummens, K., & Sayim, B. (2022). Multidimensional feature interactions in visual crowding : When configural cues eliminate the polarity advantage. Journal of Vision , 22 (6), 2. https://doi.org/10.1167/jov.22.6.2 Saarela, T. P., Sayim, B., Westheimer, G., & Herzog, M. H. (2009). Global stimulus configuration modulates crowding. Journal of Vision , 9 (2), 5. https://doi.org/10.1167/9.2.5 Sayim, B., Greenwood, J. A., & Cavanagh, P. (2014). Foveal target repetitions reduce crowding. Journal of Vision , 14 (6), 4. https://doi.org/10.1167/14.6.4 Sayim, B., Westheimer, G., & Herzog, M. H. (2008). Contrast polarity, chromaticity, and stereoscopic depth modulate contextual interactions in vernier acuity. Journal of Vision , 8 (8), 12. https://doi.org/10.1167/8.8.12 Sayim, B., Westheimer, G., & Herzog, M. H. (2010). Gestalt Factors Modulate Basic Spatial Vision. Psychological Science , 21 (5), 641-644. https://doi.org/10.1177/0956797610368811 Song, S., Levi, D. M., & Pelli, D. G. (2014). A double dissociation of the acuity and crowding limits to letter identification, and the promise of improved visual screening. Journal of Vision , 14 (5), 3. https://doi.org/10.1167/14.5.3 Tsagkaridis, K., Watson, C. E., Jax, S. A., & Buxbaum, L. J. (2014). The role of action representations in thematic object relations. Frontiers in Human Neuroscience , 8 . https://doi.org/10.3389/fnhum.2014.00140 Wamain, Y., Pluciennicka, E., & Kalénine, S. (2014). Temporal dynamics of action perception : Differences on ERP evoked by object-related and non-object-related actions. Neuropsychologia , 63 , 249-258. https://doi.org/10.1016/j.neuropsychologia.2014.08.034 Westfall, J., Kenny, D. A., & Judd, C. M. (2014). Statistical power and optimal design in experiments in which samples of participants respond to samples of stimuli. Journal of Experimental Psychology: General , 143 (5), 2020-2045. https://doi.org/10.1037/xge0000014 Westheimer, G., & Hauske, G. (1975). Temporal and spatial interference with vernier acuity. Vision Research , 15 (10), 1137-1141. https://doi.org/10.1016/0042-6989(75)90012-7 Whitney, D. (2005). Motion distorts perceived position without awareness of motion. Current Biology , 15 (9), R324-R326. https://doi.org/10.1016/j.cub.2005.04.043 Whitney, D., & Levi, D. M. (2011). Visual crowding : A fundamental limit on conscious perception and object recognition. Trends in Cognitive Sciences , 15 (4), 160-168. https://doi.org/10.1016/j.tics.2011.02.005 Wilkinson, F., Wilson, H. R., & Ellemberg, D. (1997). Lateral interactions in peripherally viewed texture arrays. JOSA A , 14 (9), 2057-2068. https://doi.org/10.1364/JOSAA.14.002057 Xu, S., Humphreys, G. W., & Heinke, D. (2015). Implied actions between paired objects lead to affordance selection by inhibition. Journal of Experimental Psychology: Human Perception and Performance , 41 (4), 1021-1036. https://doi.org/10.1037/xhp0000059 Yeh, S.-L., He, S., & Cavanagh, P. (2012). Semantic Priming From Crowded Words. Psychological Science , 23 (6), 608-616. https://doi.org/10.1177/0956797611434746 Yoon, E. Y., Humphreys, G. W., & Riddoch, M. J. (2010). The paired-object affordance effect. Journal of Experimental Psychology: Human Perception and Performance , 36 (4), 812-824. https://doi.org/10.1037/a0017175 Zhang, L., Zhang, L., Mou, X., & Zhang, D. (2011). FSIM : A Feature Similarity Index for Image Quality Assessment. IEEE Transactions on Image Processing , 20 (8), 2378-2386. IEEE Transactions on Image Processing. https://doi.org/10.1109/TIP.2011.2109730 Zhou, J., Lee, C.-L., Li, K.-A., Tien, Y.-H., & Yeh, S.-L. (2016). Does Temporal Integration Occur for Unrecognizable Words in Visual Crowding? PLOS ONE , 11 (2), e0149355. https://doi.org/10.1371/journal.pone.0149355 Zloteanu, M., & Vuorre, M. (2024). A Tutorial for Deception Detection Analysis or : How I Learned to Stop Aggregating Veracity Judgments and Embraced Signal Detection Theory Mixed Models. Journal of Nonverbal Behavior , 48 (1), 161-185. https://doi.org/10.1007/s10919-024-00456-x Additional Declarations No competing interests reported. Supplementary Files Appendix.docx Cite Share Download PDF Status: Published Journal Publication published 28 Feb, 2026 Read the published version in Psychological Research → Version 1 posted Editorial decision: Revision requested 28 Sep, 2025 Reviews received at journal 28 Sep, 2025 Reviews received at journal 19 Sep, 2025 Reviewers agreed at journal 31 Aug, 2025 Reviewers agreed at journal 30 Aug, 2025 Reviewers invited by journal 29 Aug, 2025 Editor assigned by journal 29 Aug, 2025 Submission checks completed at journal 24 Aug, 2025 First submitted to journal 23 Aug, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7442959","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":508064003,"identity":"ea990af9-68fc-4468-a997-3b5f53f36e3f","order_by":0,"name":"Nicolas Slaski","email":"","orcid":"","institution":"Univ. Lille, CNRS, UMR 9193 - SCALab - Sciences Cognitives et Sciences Affectives","correspondingAuthor":false,"prefix":"","firstName":"Nicolas","middleName":"","lastName":"Slaski","suffix":""},{"id":508064004,"identity":"1535a1b7-2bb1-4dcd-9889-81d727ddc19c","order_by":1,"name":"Bilge Sayim","email":"","orcid":"","institution":"Univ. Lille, CNRS, UMR 9193 - SCALab - Sciences Cognitives et Sciences Affectives","correspondingAuthor":false,"prefix":"","firstName":"Bilge","middleName":"","lastName":"Sayim","suffix":""},{"id":508064005,"identity":"dc31e5f5-0ae7-4ab9-9dd3-386c1a7779c5","order_by":2,"name":"Solène Kalénine","email":"data:image/png;base64,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","orcid":"","institution":"Univ. Lille, CNRS, UMR 9193 - SCALab - Sciences Cognitives et Sciences Affectives","correspondingAuthor":true,"prefix":"","firstName":"Solène","middleName":"","lastName":"Kalénine","suffix":""}],"badges":[],"createdAt":"2025-08-23 19:08:14","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7442959/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7442959/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00426-026-02247-z","type":"published","date":"2026-02-28T15:58:39+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":90898513,"identity":"06680e8f-e35a-40e9-814b-7a2a3575442a","added_by":"auto","created_at":"2025-09-09 11:53:55","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":329038,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eExamples of stimuli used. Object pairs could be thematically related (T condition, first row) or unrelated (UN, two last rows). When thematically related (first row), objects could be correctly positioned to be grasped and used together, with the active object on the side of the dominant hand (Correctly positioned for action, left) or not (Incorrectly positioned for action, right)\u003c/em\u003e\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-7442959/v1/42e247d0f44551e290d4cf32.png"},{"id":90898563,"identity":"acbb2fba-d523-4d42-a829-fc2cfb61b017","added_by":"auto","created_at":"2025-09-09 11:54:01","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":336998,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eExamples of a pair crowded with meaningless flankers positioned edge-to-edge with the object images\u003c/em\u003e\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-7442959/v1/f17b0c406fc23d5be7d83256.png"},{"id":90898544,"identity":"242e33f5-3761-43cb-868b-62c302edc42f","added_by":"auto","created_at":"2025-09-09 11:53:57","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":231551,"visible":true,"origin":"","legend":"\u003cp\u003eProcedure of the experiment. A fixation cross was displayed for 1000 ms, followed by the presentation of a pair of objects for 150 ms. After a 250 ms interval, the written name of an object appeared in the center. The object name could correspond to one of the objects of the pair presented previously or not\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-7442959/v1/cf728905a74eb09dd9f9fa64.png"},{"id":90898565,"identity":"8b05e770-f1e9-4156-93ad-6b58b9b75cdb","added_by":"auto","created_at":"2025-09-09 11:54:01","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":438112,"visible":true,"origin":"","legend":"\u003cp\u003eAccuracy on ‘yes’ trials separated for pairs relations and crowding on Experiment 1. Error-bars represent within error-types and each dot represents mean accuracy of participants.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-7442959/v1/b4c160f5052c0f8b20d6c309.png"},{"id":90898545,"identity":"97725af9-9466-4638-ae3f-13b454da507f","added_by":"auto","created_at":"2025-09-09 11:53:57","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":243986,"visible":true,"origin":"","legend":"\u003cp\u003eExamples of each condition in a set of Experiment 2. Importantly, the relation between the target noun and the nontarget object (passive object) was manipulated. The target could be thematically related (first column) or unrelated (last column) to the nontarget object. The target could be presented in the pairs (first row) or not (last row). Finally, objects could be positioned adequately for grasp and use together (Correctly positioned for action) or not (Incorrectly positioned for action)\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-7442959/v1/516fd700276551a6ed7180b1.png"},{"id":90898553,"identity":"ad2c86d6-aca1-4e4e-81a4-db839ddaf482","added_by":"auto","created_at":"2025-09-09 11:53:59","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":446115,"visible":true,"origin":"","legend":"\u003cp\u003eAccuracy on ‘yes’ trials separated for pairs relations and crowding on Experiment 2. Error-bars represent within error-types and each dot represents mean accuracy of participants\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-7442959/v1/a425bb1f32551c8e9e80e4a9.png"},{"id":90899627,"identity":"b32b1171-8561-47d4-891a-96773c9e5b65","added_by":"auto","created_at":"2025-09-09 12:01:56","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":434558,"visible":true,"origin":"","legend":"\u003cp\u003eSensitivity (hit rate minus false alarm rate) separated for pairs relations and crowding on Experiment 2. Error-bars represent within error-types and each dot represents mean accuracy of participants\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-7442959/v1/58bb5de064a29349f8946e1a.png"},{"id":103765813,"identity":"feca14ee-6446-45f1-ac98-b062ebe11528","added_by":"auto","created_at":"2026-03-02 16:09:12","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3334601,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7442959/v1/14626a76-850e-442e-aa44-f76195934b5e.pdf"},{"id":90898508,"identity":"02eb48b6-573d-435d-a0cf-f4922a6415e8","added_by":"auto","created_at":"2025-09-09 11:53:52","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":17876,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-7442959/v1/d1a02493b0f8a87d8ef691eb.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Thematic knowledge survives visual crowding and influences object identification","fulltext":[{"header":"Public Significance","content":"\u003cp\u003eThis study shows that meaningful relations between objects that are typically used together, such as the relation between hammer and nail, can help people identify individual objects when perceptual conditions are difficult, in particular when objects appear in visually cluttered environments. This suggests that higher-level knowledge derived from interactions with objects may influence lower-level perceptual processes. Findings highlight the importance of considering semantic context to understand the perception of objects in everyday, naturalistic environments.\u003c/p\u003e"},{"header":"Introduction","content":"\u003cp\u003eOur everyday visual environments are rich and usually involve many objects. Objects that appear within the same visual context are often semantically linked. For example, a kitchen environment usually includes specific furniture, food, appliances and utensils. These objects are typically used together when preparing meals or eating (e.g., spatula on a pan, a spoon on a bowl). Temporal, spatial, causal and functional relations between objects that play complementary roles in the same event (e.g., pen and paper in writing) have been extensively studied and have been referred to as \u0026lsquo;thematic relations\u0026rsquo; (Estes et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). The central goal of the present study is to better understand the influence of thematic relations on the perception of (objects in) complex visual environments.\u003c/p\u003e\u003cp\u003eA large body of evidence now indicates that thematic relations are an important organizing principle of conceptual knowledge. For example, in a series of experiments Lin \u0026amp; Murphy, (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2001\u003c/span\u003e) demonstrated that thematic relations are meaningful and used by adults in a wide range of categorization tasks, similar to other relations, such as taxonomic relations (categorization based on shared features). In one of their experiments, participants performed a classical forced-choice category task. Participants were presented with a triad of images, consisting of a target (e.g. pencil), an image thematically related to the target (e.g., eraser), and an image taxonomically related to the target (e.g., pen). They were asked to indicate which image \u0026ldquo;goes best with the target to form a category\u0026rdquo;. \u0026ldquo;Category\u0026rdquo; was defined as \u0026ldquo;a set of things or people that share some commonalities \u0026ndash; be it genetic makeup, functions, purposes, physical and perceptual characteristics, or behavioral predispositions.\u0026rdquo; Despite the instructions emphasizing taxonomic categorization, results showed that the majority (66%) of participants chose predominantly thematic alternatives. This study demonstrated that thematic relations play a crucial role in semantic organization. Further studies demonstrated that thematic relations are particularly salient and for manipulable artifact objects at both the behavioral and brain levels (see below, Kal\u0026eacute;nine et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Kal\u0026eacute;nine \u0026amp; Bonthoux, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Wamain et al., \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). In one study, participants had to identify taxonomic and thematic relations in a force-choiced categorization task where a semantically-related picture and an unrelated picture were presented below a target picture (Kal\u0026eacute;nine et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). Depending on the condition, the related picture was thematically or taxonomically related to the target. The triads could be artefacts or natural objects and manipulable or non-manipulable. Despite strict control of the degree of semantic relatedness between conditions, behavioral results showed that thematic relations were more rapidly identified for artefacts (e.g., bed-person asleep) than natural targets (e.g. squirrel-hazelnut), especially for manipulable artefacts (e.g., bowl-toast). On the contrary, taxonomic relations were more rapidly identified for natural than artefact targets. Thus, taxonomic conceptual processing relies on perceptual similarities and is facilitated for natural objects that are defined by their perceptual features. On the other hand, thematic conceptual processing relies on functional relations and is facilitated for artefacts, especially when manipulable as they are defined by their functional features (Kal\u0026eacute;nine et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). This study demonstrates that taxonomic and thematic relations are based on different sensory-motor processes. In particular, thematic relation processing is closely related to action knowledge about artefacts. The relevance of thematic relations for structuring knowledge about manipulable artefacts may be related to their proximity with the representations of actions. These representations are acquired through experience of object use. Neuroimaging results (Kal\u0026eacute;nine et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) showed that the posterior temporo-parietal cortex, a region typically associated with action tool knowledge (i.e. information about tool use) was selectively activated during the processing of thematic relations for manipulable artefacts. This region has been further shown critical for both identification of thematic relations for artefacts and recognition of object use gestures in people with brain lesions (Kal\u0026eacute;nine \u0026amp; Buxbaum, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Tsagkaridis et al., \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Consistent with the involvement of action-related processes in the identification of thematic relations between objects, the existence of a \u0026ldquo;paired-affordance effect\u0026rdquo; has been demonstrated in the literature on visual attention and object perception in both healthy individuals (Borghi et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Roux-Sibilon et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Xu et al., \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Yoon et al., \u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e2010\u003c/span\u003e) and brain-damaged patients (Riddoch et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2003\u003c/span\u003e, \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). This effect corresponds to better performance in correctly judging whether object pairs are used together when objects are correctly rather than incorrectly positioned for action. A correct position for action involves spatially arranging the objects in a way that is congruent with their common use (e.g. corkscrew positioned above wine bottle) and/or orienting the active object of the pair (i.e. the tool) towards dominant hand. Such a facilitatory effect has also been found when detecting (Riddoch et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2003\u003c/span\u003e) and identifying objects (Roberts \u0026amp; Humphreys, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). In a study of Robert and Humphreys (2011), pairs of objects were presented centrally to right-handed participants for 50 ms. Pairs consisted of two types of objects, one active (e.g. corkscrew; light match) and one passive (e.g. wine bottle; candles). Right-handed participants had to verbally identify the two objects. The two objects could be thematically related (wine bottle and corkscrew) or non-related (wine bottle and light match). The active object (corkscrew or light match) could be presented on the right side or on the left side of the pair. The active object could also be correctly positioned for use on the passive object (i.e., functionally oriented towards the passive object). Results showed that both passive and active objects were more often correctly identified when they were thematically related than unrelated. Importantly, results also showed more correct identifications when objects were correctly positioned compared to incorrectly positioned (for right-handed participants). Taken together, findings indicate that thematic relations and action position of both objects play a role in identifying objects in pairs.\u003c/p\u003e\u003cp\u003eIn this line of research, objects are usually presented in simple viewing conditions. Object pairs are shown in isolation in central vision. Similarly, the role of thematic relations in object processing has been mostly studied with objects (object pairs or triads) presented in isolation. In naturalistic settings, however, objects frequently appear surrounded by other objects in complex, often cluttered scenes and rapidly changing visual environments. One of the main limits of visual perception in cluttered environments is crowding: When a target is surrounded by other stimuli (flankers), target identification is worse compared to presentation of the target in isolation (Herzog et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Levi, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). Crowding has been considered as a fundamental limitation of our peripheral vision, and extensive research has been conducted into the principles governing crowding using comparatively simple stimuli such as Verniers (Malania et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Manassi et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Sayim et al., \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2008\u003c/span\u003e, \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Westheimer \u0026amp; Hauske, \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e1975\u003c/span\u003e), Gabor patches (Greenwood et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Parkes et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2001\u003c/span\u003e; Saarela et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Wilkinson et al., \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e1997\u003c/span\u003e) and letters (Liu \u0026amp; Arditi, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2001\u003c/span\u003e; Sayim et al., \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Song et al., \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). These studies have identified several low- and mid-level factors that influence crowding. For instance, crowding is strongly affected by the spacing between the target and the surrounding flankers: The closer the flankers are to the target, the stronger the crowding effect (Bouma, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e1970\u003c/span\u003e; Levi et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Pelli, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Pelli et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2004\u003c/span\u003e; but (Melnik et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, below). Similarly, the greater the similarity of the target and the flankers, the stronger the crowding effect (but see Rummens \u0026amp; Sayim, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2021\u003c/span\u003e, \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). For instance, Kooi et al., (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e1994\u003c/span\u003e) presented participants with peripheral letters Ts, where one T served as the target and was surrounded by four other Ts. The stimuli varied in contrast polarity (e.g., black target and white flankers). Results indicated better performance when the target and flankers were dissimilar. Subsequent studies replicated these findings and extended them to other features such as size, orientation, spatial frequency, color, and motion (see Whitney \u0026amp; Levi, \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2011\u003c/span\u003e) for a review). Historically, these data were interpreted as interference from low-level property similarity between flankers and target with stronger interference when they are similar and weaker interference when they are dissimilar. Generally, segregating the target from the flankers becomes more difficult when they share common low-level properties. However, recent studies found inverse patterns and suggested that mechanisms on the level of perceptual organization, such as grouping, also play a key role in crowding. For example, Melnik et al. (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) showed that reducing distance between the flankers and the target, thereby maximizing strong target-grouping, can yield emergent features that in turn reduced crowding effects. Orientation discrimination of a chevron target stimulus improved when closely spaced flankers allowed the formation of a diamond configuration (\u0026ldquo;\u0026lt;\u0026gt;\u0026rdquo;) compared to an alternative configuration. These results show that crowding is modulated not only by target-flanker similarity but also by the emergent features of particular configurations.\u003c/p\u003e\u003cp\u003eDespite the failure to identify the target, some information of the target appears to be still processed, influencing subsequent behavior. This has been shown for several lower-level visual features such as orientation (He et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1996\u003c/span\u003e), motion (Aghdaee, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), or position (Whitney, \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). Importantly, a few studies demonstrated that higher-level semantic features of the target, such as facial expressions (Kouider et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), biological motion (Ikeda et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) and semantic information of Chinese words (Chien et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Yeh et al., \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Zhou et al., \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) may survive crowding. For example, in Yeh et al. (\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), na\u0026iuml;ve Chinese participants were presented with Chinese target words in the periphery, under both crowded and isolated conditions. Subsequently, a single Chinese character appeared centrally that could be semantically related to the target or not, and participants had to judge whether the character was a word or non-word (lexical decision task). Participants then performed a second task where they had to identify the crowded word seen earlier. Results showed that despite the inability to identify crowded words, crowded stimuli elicited semantic priming effects in the lexical decision task: participants were better when the words were semantically related than unrelated.\u003c/p\u003e\u003cp\u003eIn the present study, we investigated whether thematic relations between objects can still be processed when everyday manipulable objects are presented under crowded conditions. Given that thematically related objects tend to co-occur in visual environments and are often used together, we hypothesized that thematic relations between objects could survive crowding and help object identification, possibly reflecting semantic grouping. In addition, we also hypothesized that the spatial configuration of thematically related objects may influence object identification. Specifically, we expected better performance when related objects were positioned in a way that affords their direct use, as compared to when they were positioned incorrectly.\u003c/p\u003e"},{"header":"EXPERIMENT 1: Processing of thematic relation under visual crowding.","content":"\u003cp\u003eThe aim of Experiment 1 was to assess whether effects of thematic relations between objects are preserved under visual crowding. To this aim, we presented pairs of objects that were either thematically related or unrelated pairs. Additionally, objects were either positioned in a way that is compatible with their co-actor from the participant\u0026rsquo;s perspective or not. Finally, objects were presented either in isolation or crowded by meaningless flankers. We measured the performance of identification of objects in these different conditions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethodology\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eParticipants\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSixty-five right-handed participants (Edinburgh Handedness Inventory,\u0026nbsp;Oldfield, 1971) (mean age: 20.2, sd: 2.6) with normal or corrected-to normal vision took part in the study. All participants were students at the University of Lille. They gave informed consent and participated in exchange for course credits. The protocol was approved by the Ethical Committee of the University of Lille (reference: 2022-605-S106).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStimuli and procedure\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStimuli\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 12 sets (Appendix 1) of 3D-grayscaled (8-bit) images of manipulable manufactured objects selected from Roux-Sibilon et al. (2018) were used as stimuli. Image sizes ranged from 0.6\u0026deg; to 2.3\u0026deg; in width and from 1.1\u0026deg; to 2.3\u0026deg; in height. Each set involved an \u0026apos;active\u0026apos; object usually held and manipulated by the dominant hand, (e.g., a pen), a \u0026apos;passive\u0026apos; object usually held but not manipulated by the non-dominant hand (e.g., a notebook), and an \u0026apos;unrelated\u0026apos; object not semantically related and used in co-action with either the active or the passive object (e.g., a hat). The 3 objects of each set were arranged in pairs to manipulate the relation between objects: objects were either thematically related (T) or unrelated (UN). The T condition was created with the active and passive objects, which together form a typical object-action pair (e.g. pen and notebook for writing). The UN condition was created by replacing either the active or passive object with the unrelated object (i.e., active + unrelated and passive + unrelated). Additionally, within each pair, objects were presented on half of the trials on the right side and in the other half on the left side of the pair. For the T condition, when the passive object was on the left side and active object on the right side, the objects were correctly positioned for action (A+), when swapped the objects were incorrectly positioned for action (A-). (Figure 1). We controlled our stimuli with objective and subjective measures of visual similarity and semantic relatedness between objects. For each pair, images were compared using the Feature Similarity algorithm (FSim; Zhang et al., 2011), which compares two images and provides a similarity value (from 0 to 1) based on their low-level visual features (i.e., phase congruency and gradient magnitude). FSim values were averaged between T and UN pairs. No significant differences were found between the two conditions (T = 0.657, UN = 0.634, Welch\u0026rsquo;s t-test = 0.6671, p = 0.512). Subjective ratings were collected in an additional sample of 21 participants. For each pair, participants answered two questions sequentially. They were first asked to rate the extent to which the two objects were visually similar (considering overall shape, details, colors, etc.), and second, the extent to which the two objects were semantically linked. Participants responded on a 7-point scale from not at all (1) to very (7) visually similar/semantically linked. As expected, participants judged T pairs (mean = 6.83) to be more semantically related than UN pairs (mean = 1.75, KS = 1, p\u0026lt;0.001). Concerning visual similarity, we found no differences between T pairs (mean = 1.85) and UN pairs (mean = 2.83, KS = 0.42, p\u0026gt;0.124). Nonetheless, FSim measures and subjective ratings of visual similarity were entered as a control variable in the statistical model (see Data Analysis and Results section.). Finally, pairs were presented in isolation (isolated condition) or surrounded by six flankers (crowded condition). The flankers were of three distinct grayscale object designs (Figure 2). Crowding is usually strong when the target and flankers are highly similar (Whitney \u0026amp; Levi, 2011 for a review, but see Rummens \u0026amp; Sayim, 2021, 2022). Therefore, we have chosen object-like flankers similar to the targets to ensure sufficient crowding. Note that object-like flankers were meaningless and identical across all pairs for two reasons. First, we wanted to avoid potential variations in perceptual or semantic similarity between targets and flankers. Finally, we wanted to prevent participants from being misled by recognizable objects during identification. In crowded conditions, the flankers were randomly positioned directly to the outer edges of the object images (Figure 2).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eProcedure\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe experiment has been conducted using Psychopy software (v2022.1.3) (Peirce et al., 2019). We used a 15.6-inch screen (1920x1080, 60Hz). Stimuli appeared at 6.2 degrees eccentricity either above or below a fixation cross (0.8 degrees) in the center of the screen. The two images of each pair were separated by 0.2 degrees (edge-to-edge of the image rectangles). Participants were positioned approximately 57 cm from the monitor. Prior to the main task, participants performed a familiarization task where stimuli were presented individually in the center of the screen with their corresponding names displayed below. The familiarization task aimed to avoid ambiguity about individual object identification prior to the experiment. Then participants performed a practice of the main task to familiarize themselves with the procedure. The procedure of the practice task was the same as in the main task, except that it used different stimuli (16 trials) and provided written feedback of the correctness of their responses after each trial. After the practice, participants performed the main task. During the task, participants were presented with pairs of objects under different conditions and were asked to identify one object from each pair. Each trial proceeded as follows: a fixation cross was displayed for 1000 ms, followed by the presentation of a pair of objects for 150 ms. After a 250 ms interval, the written name of an object appeared in the center. The object name could correspond to one of the objects of the pair presented previously or not (Figure 3.). Participants were instructed to determine as quickly and accurately as possible whether the word corresponded to one of the images presented in that trial or not (\u0026ldquo;yes\u0026rdquo; or \u0026ldquo;no\u0026rdquo;). Participants responded by pressing \u0026lsquo;p\u0026rsquo; or \u0026lsquo;l\u0026rsquo; on the keyboard with their right forefinger and middle finger respectively. The key assignments were counterbalanced between participants. For T pairs, the target object appeared on both sides of the pair, once on the left and once on the right. Conversely, in half of the UN trials, the target object appeared on one side of the pair, either left or right, with the assignment counterbalanced across sets (i.e., left object for active/unrelated pairs and right object for passive/unrelated pairs, or vice versa). This ensured an equal number of yes and no trials in the T and UN conditions. In total, participants performed 768 trials: 12 (sets) * 2 (Pair Relatedness conditions) * 2 (Position for Action conditions) * 2 (Crowding conditions) * 2 (Target side) * 2 (Target Presence) * 2 (Visual field conditions). Trials were presented in random order for each participant. A break was proposed every 96 trials. Response, accuracy, and response times (RT) in milliseconds were recorded.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData processing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe verified that the performance of each participant exceeded the chance level in the isolated condition (i.e. above 54,2%, p\u0026lt;0.05). All participants performed above chance level. Next, participants with accuracy scores more than 2.5 standard deviations above or below the group mean in the isolated or crowded conditions were excluded (Berger \u0026amp; Kiefer, 2021). One participant fell below the upper threshold in the isolated condition and was subsequently removed from the analysis, leaving 64 participants. The same procedure was applied at the set level to assess whether performance from a given set deviated excessively from the others, but no outliers were detected.\u003c/p\u003e\n\u003cp\u003eRegarding response times (RT), we excluded trials with incorrect responses (25.3%). A global trimming procedure was applied to remove aberrant responses. RT inferior to 200 ms and RT superior to five times the median (4171 ms) were excluded (0.38% trials). Then, we removed RTs superior or inferior to 2.5 standard deviations from the mean RT of each participant in each condition (Participant x Target Presence x Crowding x Pair Relatedness x Position for Action). This global trimming procedure excluded 3.24% of trials.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLogistic and linear mixed-effect regression models were used to analyze accuracy and response times, respectively. We hypothesized that thematic relations between objects would facilitate the correct identification of objects that were actually present in the display. Thus, analysis of identification performance was restricted to trials on which the target was present. In our mixed-effect regression models, fixed effects corresponded to the factors of interest, including Crowding (isolated, crowded), Pair Relatedness (T, UN pairs), Position for Action (correct, incorrect) and their interactions. Subjective ratings of visual similarity and FSim values were entered as additional fixed effects to control for visual similarity differences between the two objects in the pairs. Contrasts were manually coded as -0.5 and +0.5. The maximal random effect structures reaching model convergence were selected following the guidelines of Barr et al., (2013) and Bates et al., (2015). Random structure simplification was conducted using an iterative procedure based on Principal Component Analysis with the \u003cem\u003erePCA\u003c/em\u003e function from the \u003cem\u003elme4\u003c/em\u003e package. Final models included random intercepts and slopes for both participants and targets. Participants were further nested in task order and targets were nested in item sets (see syntax in Appendix 3). For linear mixed-effect models, Westfall\u0026rsquo;s d measures were computed as an alternative to Cohen\u0026rsquo;s d (Brysbaert \u0026amp; Stevens, 2018; Judd et al., 2017; Westfall et al., 2014). Westfall\u0026rsquo;s d for a given contrast is obtained by dividing the estimate for that contrast (difference of estimated means) by the square root of the sum of the variance of all random effects and residuals of the model.\u003c/p\u003e\n\u003cp\u003eOverall, we expected a main effect of Crowding, with worse identification performance in the Crowded compared to Isolated conditions. Furthermore, we wanted to evaluate whether\u0026nbsp;Pair Relatedness would be modulated by crowding through the Crowding x Pair Relatedness interaction. In particular, we wanted to test whether an advantage for T pairs would occur in both the isolated and crowded conditions. Additionally, in line with studies on the paired-object affordance effect (Yoon \u0026amp; Humphreys, 2010), Position for Action of thematic pairs was expected to impact performance, with correct position for action facilitating target identification in thematic pairs in both the isolated and crowded conditions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAccuracy\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDescriptive statistics for the different conditions are presented in Table 1. First, logistic mixed-effect models showed a significant effect of crowding on correct responses with 4.76 times worse target identification performance in the crowded compared to isolated condition (Odds ratio (OR) = 4.76, CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= 3.57 \u0026ndash; 6.25, p\u0026lt;0.001) (Figure 4). There was also a significant main effect of\u0026nbsp;Pair Relatedness with 1.85 times better target identification performance in the thematically related compared to unrelated condition\u0026nbsp;(OR = 1.85, CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= 1.55 \u0026ndash; 2.2, p\u0026lt;0.001).\u0026nbsp;Importantly, there was a significant interaction between Crowding and\u0026nbsp;Pair Relatedness\u0026nbsp;(OR = 1.67, CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= 1.429 \u0026ndash; 1.961, p\u0026lt;0.001). The more accurate identification for T pairs than UN pairs was stronger in isolated than crowded conditions. While crowded UN pairs were identified at chance level (t = 0.478, CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= 0.471 \u0026ndash; 0.546, p = 0.634), crowded T pairs were identified above chance (t = 3.311, CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= 0.529 \u0026ndash; 0.619, p = 0.001). Finally, Position for Action did not influence correct responses, neither as a main effect (OR=1.05, CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= [0.94 \u0026ndash; 1.16], p\u0026gt;0.42) nor in interaction with Crowding (OR = 0.08, CI\u003csub\u003e95% \u0026nbsp;\u003c/sub\u003e= 0.89 \u0026ndash; 1.31, p\u0026gt;0.42).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1\u003c/strong\u003e Means and Standard Errors in Accuracy (proportion of correct responses) and Response Times (RTs) for correct responses in the different conditions of Experiment 1. Note that in thematically unrelated pairs, the target object differed across the Position for Action conditions\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"605\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 17.4422%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 14.3737%;\"\u003e\n \u003cp\u003eAccuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 7.4291%;\"\u003e\n \u003cp\u003eRTs\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 9.3671%;\"\u003e\n \u003cp\u003eCondition\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 8.0751%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e\u003cem\u003eM\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e\u003cem\u003eSE\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e\u003cem\u003eM\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e\u003cem\u003eSE\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 22.2873%;\"\u003e\n \u003cp\u003eThematically Related\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 8.6404%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 16.1502%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 17.4422%;\"\u003e\n \u003cp\u003eIsolated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 14.3737%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 7.4291%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.0188%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 15.4234%;\"\u003e\n \u003cp\u003eCorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e0.863\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e0.010\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e799\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.0188%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 15.4234%;\"\u003e\n \u003cp\u003eIncorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e0.868\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e0.009\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e807\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 17.4422%;\"\u003e\n \u003cp\u003eCrowded\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 14.3737%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 7.4291%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.0188%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 15.4234%;\"\u003e\n \u003cp\u003eCorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e0.574\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e974\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.0188%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 15.4234%;\"\u003e\n \u003cp\u003eIncorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e0.574\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e964\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 22.2873%;\"\u003e\n \u003cp\u003eThematically Unrelated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 8.6404%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 16.1502%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 17.4422%;\"\u003e\n \u003cp\u003eIsolated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.0188%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 15.4234%;\"\u003e\n \u003cp\u003eCorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e0.716\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e870\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.0188%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 15.4234%;\"\u003e\n \u003cp\u003eIncorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e0.743\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e0.010\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e833\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 17.4422%;\"\u003e\n \u003cp\u003eCrowded\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.0188%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 15.4234%;\"\u003e\n \u003cp\u003eCorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e0.499\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e0.010\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e1037\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e18\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.0188%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 15.4234%;\"\u003e\n \u003cp\u003eIncorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 10.0131%;\"\u003e\n \u003cp\u003e0.519\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 4.3606%;\"\u003e\n \u003cp\u003e0.010\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6.5408%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e972\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 3.7145%;\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cbr\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResponse times\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConsidering the percentage of correct responses in isolated and crowded conditions,\u0026nbsp;response times for correct responses (RTs) were mainly investigated to verify the absence of speed-accuracy trade-off in participants\u0026rsquo; responses. There was a main effect of Crowding, with slower RTs in the crowded than isolated condition (estimate = 0.15, CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= 0.12 \u0026ndash; 0.19, p\u0026lt;0.001, Westfall\u0026rsquo;s d = 0.41) and a main effect of Pair Relatedness, with faster RTs for T pairs than UN pairs (estimate = -0.08, CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= -0.11 \u0026ndash; -0.05, Westfall\u0026rsquo;s d = 0.21). There was no interaction between Pair Relatedness and Crowding (estimate = 0.02,CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= -0.01\u0026ndash;0.04 , p\u0026gt;0.15, Westfall\u0026rsquo;s d = 0.04). Again, Position for Action did not influence RTs, neither as a main effect (estimate = 0.003, CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= -0.01 \u0026ndash; 0.02, p\u0026gt;0.77, Westfall\u0026rsquo;s d = 0) nor in interaction with Crowding (estimate = 0.002, CI\u003csub\u003e95%\u0026nbsp;\u003c/sub\u003e= -0.02 \u0026ndash; 0.03, p\u0026gt;0.89, Westfall\u0026rsquo;s d = 0).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDiscussion of experiment 1\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn Experiment 1, participants were presented with pairs of objects that could be thematically related or unrelated. Experiment 1 has demonstrated that target objects (e.g., pen) were identified more accurately when they appeared in thematically-related (e.g., pen \u0026ndash; notebook) than semantically unrelated object pairs (e.g., pen \u0026ndash; hat). This thematic advantage was also present in crowded conditions, albeit to a lesser extent. We found no evidence that performance was modulated as a function of the position for action of related pairs, either in isolation or under visual crowding. While the thematic advantage observed is unequivocal, its locus remains to be clarified. In Experiment 1, the target noun (i.e., \u0026ldquo;pen\u0026rdquo;) was systematically related to the other, non-target object of the pair (i.e. notebook) in thematically related pairs (e.g., pen \u0026ndash; notebook). Therefore, the advantage of the thematic effect could originate from two different stages. During the visual presentation of object pairs (e.g. pen-notebook), thematic pairs could benefit from perceptual grouping and be better perceived (Auckland et al., 2007; Green \u0026amp; Hummel, 2006; Nah \u0026amp; Geng, 2022; Roberts \u0026amp; Humphreys, 2011). Alternatively, after the visual presentation of the objects and during the presentation of the target nouns (e.g., when the word \u0026ldquo;pen\u0026rdquo; is presented), processing of the thematic relation between the target noun (e.g., word \u0026ldquo;pen\u0026rdquo;) and one of the perceived objects (i.e., the non-target object, (e.g., \u0026ldquo;notebook\u0026rdquo;) could bias the decision towards target presence, even without the thematic relation being perceived in the visual object pair (Biederman et al., 1982; Hollingworth, 1998; Hollingworth \u0026amp; Henderson, 1999). In other words, participants may have identified only the passive object (e.g., notebook) and inferred based on their semantic knowledge that the thematically related object (e.g., pen) was likely to be presented. This guessing strategy could explain in part or in full the observed benefit of thematic relation on object identification. Indeed, our design did not include trials where the target noun (e.g., \u0026ldquo;pen\u0026rdquo;) was related to the non-target object (e.g., notebook) but was actually absent. In either case, results suggest that relatively high level semantic information of manipulable objects survived crowding and improved performance. Experiment 2 was then designed to investigate whether the observed thematic advantage on object identification originated from the thematic relation between objects during pair presentation and/or from a decisional strategy based on semantic information of single objects during the target noun presentation.\u003c/p\u003e"},{"header":"Experiment 2: Locus of the thematic advantage","content":"\u003cp\u003eExperiment 2 aimed to further investigate the locus of the thematic advantage observed in Experiment 1. We manipulated the relationship between the target noun and the non-target object to evaluate its influence on participants\u0026rsquo; decisions. Critically, we designed \u0026ldquo;no\u0026rdquo; trials in which the target was absent but target nouns could still be thematically related to one object of the pair. This design allowed us to disentangle perceptual and decisional contributions to the thematic advantage by computing measures of sensitivity and bias. A decisional strategy would be expected to increase bias toward \u0026lsquo;yes\u0026rsquo; responses and reduce sensitivity when the target noun was thematically related \u0026ndash;compared to unrelated\u0026ndash; to one of the objects of the pair. In contrast, a perceptual contribution to the thematic advantage should not affect either bias or sensitivity based on the relation between the target noun and the non-target object. Experiment 2 was conducted online and aimed at replicating the findings from Experiment 1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethodology\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eParticipants\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRight-handed participants native speakers of English were recruited via Prolific. Given the methodological similarities between Experiments 1 and 2, a post hoc power analysis of Experiment 1 was conducted to determine the appropriate sample size for Experiment 2. Utilizing a simulation-based approach (Kumle et al., 2021), we computed the post hoc power to detect the observed interaction between Pair Relatedness and Crowding in the mixed logistic regression model used in Experiment 1. The analysis indicated a power of 98% with the actual 64 participants. Considering the high power of Experiment 1, we aimed to include an equivalent number of participants in Experiment 2. Initially, 75 participants were recruited. Exclusion criteria led to the removal of 10 participants: one due to a screen refresh rate below 30 Hz, five for responding in less than 150 ms, seven for a mean accuracy at chance level in the isolated condition, and one for having an accuracy in the isolated condition exceeding 2.5 standard deviations above the mean of the group. Consequently, the final sample comprised 65 participants (28 female, 34 male, 1 unknown, mean age: 25 years, SD: 3.2).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStimuli\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eImages of objects were the same as in Experiment 1 with the addition of five new sets ( Appendix 2) to improve the reliability of the statistical analyses. The main differences from Experiment 1 were the following. In Experiment 2, UN pairs were created with the passive object (e.g., notebook) and the unrelated object (e.g., screwdriver). Additionally, among each set, the identification was limited to the unrelated object (e.g. screwdriver) and the active object (e.g., pen), while the passive object (e.g., book) was always present but remained task-irrelevant as it was never the target for identification (Figure 5). The object to detect appeared in 50% of the trials. Critically, the passive object could either be thematically related or unrelated to the target noun. In the Target\u003cem\u003e\u0026nbsp;\u003c/em\u003eRelated (TR) condition, the target noun (e.g., pen) was always thematically related to the passive object (e.g., book) regardless of whether the target was present or not. In contrast, in the Target\u003cem\u003e\u0026nbsp;\u003c/em\u003eUnrelated (TU) condition, the target noun (e.g., screwdriver) was never thematically related to the passive object regardless of whether the target was present or not. With this design, we can quantify the influence of the relation between the target noun and the non-target object on object detection performance.\u003cbr\u003eAs in experiment 1, we collected two types of subjective ratings on a 7-point scale on two additional samples of participants on the visual and semantic properties of the object pairs. One sample of 15 participants rated the extent to which the two objects were semantically related. As expected, T pairs (mean = 6.69) were rated to be more semantically linked than UN pairs (mean = 1.59, STATS). Another sample of 22 participants rated the extent to which the two objects were visually similar. T and UN pairs were rated to be equally visually similar (mean = 1.96).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eProcedure\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe experiment\u0026nbsp;was\u0026nbsp;conducted using Psychopy software (v2023.2.2) and was hosted online on Pavlovia\u0026nbsp;(Peirce et al., 2019). To normalize stimuli size presentations, we used a screen calibration: participants adjusted their screen resolution by matching a virtual standardized credit card with their own one. To control the distance between participants and screen, we used a blind spot calibration method: participants were asked to close one eye, fixate a fixation cross and adjust their position until a red dot in their visual periphery\u0026nbsp;was no longer visible. The correct distance was set to be around 60 cm away from the screen. The experimental procedure was identical to that of Experiment 1, except for three minor modifications. First, to ensure proper attention to the familiarization phase, a second phase was added. In this additional phase, participants were required to discriminate between the objects they had previously seen and visually similar\u0026nbsp;exemplars\u0026nbsp;of the same objects. Feedback was provided after each trial. Second, the number of trials in the practice phase was doubled (32 trials) to improve task familiarization in the online setting. Third, in the main task, each trial ended either upon the participant\u0026rsquo;s response or after 6 seconds in the absence of a response.\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u003cstrong\u003eData analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the replication of the results from Experiment 1, we conducted the same analysis of accuracy and correct response times using logistic and linear mixed-effect regression respectively, following the same procedure (Appendix 4). In addition, for the purpose of this second Experiment, we computed a measure of sensitivity and bias (Signal Detection Theory) using a logistic mixed-effect regression model (Zloteanu \u0026amp; Vuorre, 2024). The purpose of this analysis was to estimate the influence of Target Relatedness on discrimination performance and bias. Specifically, we wanted to evaluate whether participants would erroneously identify the target \u0026lsquo;pen\u0026rsquo; when they were presented with the unrelated object pair \u0026apos;screwdriver-notebook\u0026rsquo; because of the presence of the notebook. This model estimated the probability to answer \u0026ldquo;yes\u0026rdquo; and included fixed effects for Target Presence (i.e., whether the object target was present or absent), Target Relatedness, Crowding and Position for Action. In this framework, fixed effects that interact with Target Presence represent perceptual sensitivity (analogous to d\u0026rsquo;), whereas main effects and interactions that do not involve Target Presence represent systematic response biases (analogous to decision criterion). Compared to traditional signal detection theory analyses, this modeling approach offers the advantage of accounting for variability across participants and items by incorporating random effects.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData processing\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInitially, we ensured that the performance of each participant exceeded chance level on isolated trials (i.e. above 55%, p\u0026lt;0.05). Seven participants were at chance in the isolated condition and were thus excluded. Next, participants with accuracy scores above or below \u0026nbsp;2.5 standard deviations from the group mean in the crowding conditions were excluded. One participant was above the upper threshold in the isolated and crowded conditions and was subsequently removed from the analysis. This procedure was also applied at the set level, and no outliers were detected. Finally, two participants were excluded for responding in less than 200 ms on at least three consecutive trials, repeated seven or more times throughout the experiment. The final sample included 65 participants.\u003c/p\u003e\n\u003cp\u003eRegarding response times (RT), we excluded incorrect response trials (37.2%). A global trimming procedure was applied to remove aberrant responses, RTs shorter than 200 ms \u0026amp; RTs longer than five times the median (3946 ms) were then excluded (1.2% trials removed). Then, we removed RTs superior or inferior to 2.5 standard deviations from the mean RT of each participant in each condition (participant x Target Presence x Crowding x Pair Relatedness x Position for Action). This global trimming procedure excluded 2.9% of trials.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAccuracy\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDescriptive statistics of the different conditions are presented in Table 2. Logistic mixed-effect models showed a significant effect of crowding on correct responses with 4.17 times worse target identification performance in the crowded compared to isolated conditions (Odds Ratio (OR) = 4.17, 95% CI = 3.12\u0026ndash;5.55, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001) (Figure 6). There was also a significant main effect of Pair Relatedness (OR = 1.66, CI = 1.16 \u0026ndash; 2.38, p=0.005). More importantly, there was a significant interaction between Crowding and Pair Relatedness (OR = 1.41, 95% CI = 1.22\u0026ndash;1.64, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001). The higher accuracy for T pairs than UN pairs was stronger in isolated than crowded conditions. We also found an interaction between Position for Action and Crowding (OR = 1.27, 95% CI = 1.10\u0026ndash;1.46, \u003cem\u003ep\u003c/em\u003e = 0.001), indicating that the negative effect of crowding was weaker when objects were correctly positioned for action compared to when they were incorrectly positioned.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2\u003c/strong\u003e Means and Standard Errors in Accuracy and correct Response Times (RTs) according to Pair Relatedness, Position for Action and Crowding in Experiment 2\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"594\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 28.5028%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 18.4649%;\"\u003e\n \u003cp\u003eAccuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 20.3238%;\"\u003e\n \u003cp\u003eRTs\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.1619%;\"\u003e\n \u003cp\u003eCondition\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3409%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e\u003cem\u003eM\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e\u003cem\u003eSE\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e\u003cem\u003eM\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e\u003cem\u003eSE\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 31.3531%;\"\u003e\n \u003cp\u003eThematically Related\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 18.4649%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 22.3066%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 28.5028%;\"\u003e\n \u003cp\u003eIsolated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 18.4649%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 22.3066%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.1619%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3409%;\"\u003e\n \u003cp\u003eCorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e0.809\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e0.014\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e746\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.1619%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3409%;\"\u003e\n \u003cp\u003eIncorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e0.825\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e0.016\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e744\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 28.5028%;\"\u003e\n \u003cp\u003eCrowded\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 18.4649%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 22.3066%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.1619%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3409%;\"\u003e\n \u003cp\u003eCorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e0.541\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e0.014\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e834\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.1619%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3409%;\"\u003e\n \u003cp\u003eIncorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e0.516\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e0.015\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e857\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e18\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 31.3531%;\"\u003e\n \u003cp\u003eThematically Unrelated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 18.4649%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 22.3066%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 28.5028%;\"\u003e\n \u003cp\u003eIsolated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.1619%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3409%;\"\u003e\n \u003cp\u003eCorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e0.701\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e0.014\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e793\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.1619%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3409%;\"\u003e\n \u003cp\u003eIncorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e0.718\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e792\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 28.5028%;\"\u003e\n \u003cp\u003eCrowded\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.1619%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3409%;\"\u003e\n \u003cp\u003eCorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e0.471\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e0.014\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e887\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e20\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 2.8503%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.1619%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18.3409%;\"\u003e\n \u003cp\u003eIncorrectly positioned\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7.1877%;\"\u003e\n \u003cp\u003e0.440\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11.4011%;\"\u003e\n \u003cp\u003e0.015\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.2858%;\"\u003e\n \u003cp\u003e889\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9.914%;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResponses times\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGiven the high accuracy in both the isolated and crowded conditions, RTs were mainly examined to verify the absence of speed-accuracy trade-off in participants\u0026apos; performance. Results revealed a significant main effect of Crowding, with slower RTs in the crowded condition compared to the isolated condition (estimate = 0.10, 95% CI = 0.06\u0026ndash;0.14, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001, Westfall\u0026rsquo;s \u003cem\u003ed\u003c/em\u003e = 0.24). Additionally, a significant main effect of Pair Relatedness was observed, with faster RTs for thematically related (T) pairs compared to unrelated (UN) pairs (estimate = -0.06, 95% CI = -0.10 to -0.01, \u003cem\u003ep\u003c/em\u003e = 0.012, Westfall\u0026rsquo;s \u003cem\u003ed\u003c/em\u003e = 0.14). In contrast, Position for Action had no significant effect on RTs, either as a main effect (\u003cem\u003ep\u003c/em\u003e \u0026gt; 0.41) or in interaction with Crowding (\u003cem\u003ep\u003c/em\u003e \u0026gt; 0.06), Pair Relatedness, or both (\u003cem\u003ep\u003c/em\u003e \u0026gt; 0.88).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe overall sensitivity of the task is represented by the main effect of Target Presence (OR = 3.90, 95% CI = 2.72\u0026ndash;5.52, p \u0026lt; .001) (Figure 6). Sensitivity was reduced by crowding, as indicated by the significant interaction between Target Presence and Crowding (OR = 0.24, 95% CI = 0.22\u0026ndash;0.27, p \u0026lt; 0.001), with sensitivity being 4.17 times greater in the isolated than in the crowded condition. Importantly, we found an interaction between Target Presence and Target Relatedness (OR = 1.20, 95% CI = 1.09\u0026ndash;1.32, p \u0026lt; 0.001) which evaluates the sensitivity to Target Relatedness. When there was no relation between the target noun and the passive, non-target object, sensitivity was 1.19 higher than when there was a thematic relation. Interestingly, this effect was significant for the isolated conditions (OR = 1.23, 95% CI = 1.07\u0026ndash;1.42, p\u0026lt;0.005), but failed to reach significance for the crowded conditions (OR = 1.14, 95% CI = 0.99\u0026ndash;1.3, p\u0026gt;0.06). We also found an interaction between Target Presence and Position for Action (OR = 1.25, 95% CI = 1.03\u0026ndash;1.52, p\u0026lt;0.024), suggesting that the crowding effect was 1.25 times weaker when objects were correctly positioned for action, compared to when they were incorrectly positioned for action.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBias\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe overall bias of the task is represented by the Intercept (OR = 0.95, 95% CI = 0.79- 1.11, p\u0026gt;0.6). We found a main effect of Crowding (OR = 1.92, 95% CI = 1.61-2.27, p\u0026lt;0.001), participants were 1.92 times less likely to respond \u0026ldquo;yes\u0026rdquo; in the crowded compared to the isolated condition. We found a main effect of Target Relatedness (OR = 1.84, 95% CI = 1.39- 2.44), participants were 1.84 more likely to respond \u0026ldquo;yes\u0026rdquo; when the target was related than unrelated. We also found an interaction between Crowding and Target Relatedness (OR = 1.44, 95% CI = 1.28 -1.56, p\u0026lt;0.001), the effect of Target Relatedness was 1.44 times weaker in the crowded than isolated conditions. In addition, the impact of Target Relatedness on the bias was present both in isolated (OR = 2.19, 95% CI= 1.64 \u0026ndash; 2.91, p\u0026lt;0.001) and crowded conditions (OR = 1.55, 95% CI = 1.16- 2.06, p\u0026lt;0.003). Finally, we found an interaction between Position for Action and Crowding (OR = 1.14, 95% CI = 1.03-1.25, p\u0026lt;0.01) suggesting that the impact of crowding on bias was weaker for objects correctly than incorrectly positioned for action.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDiscussion of Experiment 2\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eExperiment 2 replicated the finding that objects in thematically related pairs were identified more accurately than those in unrelated pairs, in both isolated and crowded conditions. This demonstrates the robustness of the effect, as we obtained an effect of similar size in an online setting with a new set of stimuli. In addition, Experiment 2 investigated whether the origin of this effect occurs during the presentation of object pairs and/or during the presentation of the target nouns. We found a systematic bias toward \u0026ldquo;yes\u0026rdquo; decisions when the target noun was thematically related to the nontarget object in both isolated and crowded trials, although the effect was weaker under crowding. In addition, target discrimination was poorer when the target was thematically related, but only for isolated pairs: it was more difficult to detect a \u0026ldquo;pen\u0026rdquo; compared to a \u0026ldquo;screwdriver\u0026rdquo; when there was a book among the object pair. These results will be further addressed in the General Discussion.\u003c/p\u003e\n\u003cp\u003eContrary to Experiment 1, we found a significant interaction between visual crowding and position for action on identification performance. Specifically, the effect of crowding on accuracy and sensitivity was reduced when objects were correctly positioned for action compared to incorrectly positioned for action. Crucially, this benefit was independent from the presence of a semantic relationship between objects. This indicates that it is the action-related configuration itself, rather than semantic knowledge, that interacts with visual crowding and improves identification. This finding is consistent with prior studies showing that objects correctly positioned to be grasped and manipulated in a plausible interaction benefit from perceptual grouping (Auckland et al., 2007; Green \u0026amp; Hummel, 2006; Roberts \u0026amp; Humphreys, 2011). For example, Robert \u0026amp; Humphreys (2011) found that briefly presented object pairs (i.e., a corkscrew and a wine bottle) were identified more accurately when they were correctly than incorrectly positioned to be grasped and manipulated, even for unrelated pairs (e.g. a match and a wine bottle). Likewise, studies of patients with visual extinction have shown better recognition of two objects when they are arranged in a plausible action configuration (Riddoch et al., 2003, 2006). These results have been interpreted as evidence that affording spatial relations between objects promotes perceptual grouping. Riddoch et al. (2006) proposed a two‐stage model in which an initial bottom‐up attention stage is sensitive to action‐relevant features: when two objects are correctly positioned to afford a plausible action, attention is automatically spread across both items, making them more likely to be processed and identified.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBut why do we find this effect only in interaction with crowding? One explanation may lie in the way visual crowding is modulated by the perceptual organization of the target and the flankers. When grouping cues segregate the target from the flankers, crowding is usually reduced (Herzog et al., 2015; but see Rummens \u0026amp; Sayim, 2019, 2022). For instance, a Vernier target flanked by two collinear lines produces strong crowding, but if those lines are perceptually grouped into a rectangle, crowding on the vernier is substantially reduced (Sayim et al., 2010; Manassi et al., 2012). Such effects have been shown with low-level features (e.g., length and color of Vernier stimuli, Malania et al., 2007; Sayim et al., 2008; Manassi et al., 2012) but also higher-level features (e.g., (Mooney) faces, biological motion, (Farzin et al., 2009; Ikeda et al., 2013; Louie et al., 2007). However, the latter authors interpreted these effects not strictly as perceptual grouping effects, but rather as resulting from greater similarity between the target and flankers.\u003c/p\u003e\n\u003cp\u003eIn our study, correctly positioned objects may have appeared to be perceptually grouped and thus more segregated from flankers. Grouping of several items has also been shown to improve performance (Rummens and Sayim, 2021, 2022; Sayim et al., 2014). In the present study, such grouping likely diminishes the effect of crowding and improves identification performance. To our knowledge, this is the first study to demonstrate that manipulable objects can be grouped based on action relation spatial configuration, leading to a reduction in visual crowding. However, it remains unclear why this action-related effect occurred only in Experiment 2 and regardless of pair relatedness. One explanation is that in Experiment 1, identification involved passive, active and unrelated objects whereas in Experiment 2, only the active and unrelated objects were task-relevant. In Roberts \u0026amp; Humphreys (2011) the benefit of perceptual grouping was mainly driven by the active object. Thus, in Experiment 2, participants may have focused their attention on the active object, making position for action more influential in facilitating perceptual grouping.\u003c/p\u003e"},{"header":"General Discussion","content":"\u003cp\u003eIn this study, we wanted to evaluate the processing of high-level semantic knowledge associated with everyday objects under visual crowding. Pairs of everyday objects were briefly presented in isolation or among meaningless object flankers in the visual periphery. The ability to identify one of the objects was evaluated in different conditions. In the isolated condition, target objects (e.g., pen) were identified more accurately when they appeared in thematically-related (e.g., pen \u0026ndash; notebook) than semantically unrelated object pairs (e.g., pen \u0026ndash; hat). In the crowded condition, target identification performance for unrelated pairs was at chance level. Critically, we demonstrated a significant performance improvement in target identification when objects were thematically related. Position for action (e.g., pen positioned on the right versus left) did not further modulate this advantage, neither for isolated nor for crowded pairs, although it reduced the deleterious effect of crowding independently from thematic relations in Experiment 2.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn the crowding condition, objects in thematic pairs were better identified than objects in unrelated pairs. Yet thematic and unrelated pairs are inherently composed of different objects with different low-level visual properties (overall shape, visual details, etc.). One could argue that low-level visual differences between objects across conditions could account for the advantage of thematic pairs. As the task involved single target identification, participants had to discriminate between the two objects presented. Thus, a lower degree of visual similarity between objects in thematic pairs may have facilitated identification in this condition. To control for this possibility, measures of visual similarity between objects were computed and added to the statistical models. FSim (Zhang et al., 2011) provides an estimation of visual similarity based on the physical properties of the images (Phase Congruency and Gradient Magnitude) and was used as an objective measure of low-level visual similarity. In addition, subjective ratings were collected and used as a complementary estimation of visual similarity between the two images in each pair. The effect of thematic relations on target identification in the crowded conditions was observed even after taking into account these two measures. Thus, while there might be differences in low-level similarity between objects in our sample of stimuli, they do not fully explain the advantage of identifying objects in thematic pairs in crowding conditions. It seems that some aspects about the objects forming a thematic pair survives crowding and facilitates their subsequent identification.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBut what specific stage do thematic relations impact object identification? One possibility is that thematic relations, through the processing of semantic information about single objects during the perception of the pair, influence responses during decision. Experiment 2 supports this interpretation. Results showed that participants were biased to respond \u0026ldquo;yes\u0026rdquo; when the target-noun (e.g., pen) was thematically related to the nontarget object (e.g., the book), even when the target object was absent. In other words, participants may have used the semantic context (i.e., the book) to predict the presence of a thematically congruent object (i.e. the pen). This phenomenon is likely to have been amplified by the fact that related and unrelated objects were relatively similar in shape. This interpretation aligns with the perspective of \u003cem\u003epredictive processing\u003c/em\u003e, postulating that expectations can influence visual processing. According to these accounts, context-driven expectations formed through real-world experience can improve perception under ambiguous conditions, such that objects in semantically congruent environments are identified more rapidly and accurately than those in incongruent environments (Bar, 2004; Biederman et al., 1982; Kaiser et al., 2014). However, prior research on object perception in scenes suggests that semantic congruency does not systematically enhance object identification, and that its effects may depend on the specific methodological approach. In a prior study, Hollingworth and Henderson (1998) presented participants with a visual scene followed by a label (i.e., target noun) and asked them to decide whether the labeled object had been present or not. In their first experiment, they found higher identification accuracy for objects semantically congruent with the scene, suggesting a facilitative effect of semantic context on object identification. However, subsequent experiments (Experiments 2 and 3) controlled for semantic congruence in trials where the labeled object was absent. These experiments revealed an increase in false alarms for semantically congruent labels compared to incongruent ones. When computing the sensitivity, incongruent objects were actually better identified than congruent ones. These findings highlight the necessity of distinguishing between perceptual facilitation and decisional biases. The authors gave a \u003cem\u003efunctional isolation\u0026nbsp;\u003c/em\u003eexplanation of their effect, suggesting that context information can orient decisions, but after the perceptual processing stage. In our study, it is therefore likely that the semantic information associated with the nontarget object influenced participants\u0026rsquo; decisions during target-noun presentation, as reflected by the observed response bias both in isolation and in crowding conditions. Importantly, our results indicate that semantic knowledge about manipulable objects, such as their functional use with other objects, can get through the bottleneck of visual crowding and influence decision making. Additionally, in line with findings from the broader semantic literature, our results suggest that semantic information from non-target objects is extracted automatically (Biederman, 1972; Biederman et al., 1982; Oliva \u0026amp; Schyns, 1997; Rousselet et al., 2005).\u003c/p\u003e\n\u003cp\u003eHowever, we cannot rule out that the thematic link between object pairs influenced processing at a perceptual level (i.e., during perception of the pairs). In a follow-up to the work by Hollingworth and Henderson (1998), Auckland, Cave, and Donnelly (2007) further examined the influence of semantic context on object identification. They conducted a six-alternative forced-choice (6-AFC) experiment to isolate perceptual benefits of semantic context from response biases. On each trial, participants briefly saw a target object centrally presented and surrounded by four distractor objects that were either semantically related (e.g., playing cards surrounded by dice) or unrelated (e.g., playing cards surrounded by fruits). Then, participants selected the target from six written words, chosen to dissociate from perceptual and semantic errors. Results showed higher identification for targets embedded with related than unrelated targets. They found semantic errors suggesting that decisional bias can explain a part of the semantic benefits. Importantly, after correcting for this bias, the facilitation remained. This finding suggests that context facilitation can occur not only with natural scenes but also with arrays of objects. It also indicates that semantic influence on object recognition can occur at both decisional and perceptual stages.\u0026nbsp;\u003cbr\u003e\u0026nbsp;Although our study was not originally designed to dissociate perceptual enhancement from decisional bias, we can be fairly confident that both of these processes are involved. Indeed, in Experiment 2, we observed a decisional bias toward thematically related targets when pairs were crowded, but sensitivity remained unaffected. Thus, the improved identification for thematically related pairs might be explained at least by a perceptual enhancement arising from the thematic relation between objects, particularly when they were crowded. Indeed, perceiving crowded objects requires greater attentional resources than perceiving isolated ones. Crowding might thereby promote the processing of thematic relations between objects and increase the likelihood that such relations enhance visual perception.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe exact mechanism underlying the benefits of thematic relation on visual processing remains unclear. Thematically related objects tend to co-occur in the same event or scenario. Thus, thematic related objects are more often seen together than unrelated objects. This difference may lead to more familiar visual configurations for related than unrelated object pairs. Thematically related objects would therefore group stronger due to their predictable co-occurrence in the world (Kaiser et al., 2014; Nah \u0026amp; Geng, 2022). Several studies using meaningless stimuli have shown that perceptual grouping among the flankers \u0026ndash;and ungrouping from the target\u0026ndash; is sufficient to reduce crowding effects (Livne \u0026amp; Sagi, 2007; Manassi et al., 2012, 2013; Sayim et al., 2008, 2010). In Experiment 2, such a mechanism seems to be involved for correctly compared to incorrectly positioned objects (see Discussion of Experiment 2), but whether it also contributes to the advantage observed for thematically related over unrelated objects remains an open question.\u003c/p\u003e\n\u003cp\u003eFuture research should investigate whether the observed thematic benefits extend to other types of semantic relations, such as taxonomic ones. While thematic relations are based on complementary roles in common events or actions (e.g., pen\u0026ndash;notebook), taxonomic relations group objects based on shared categorical features (e.g., pen\u0026ndash;pencil) (Mirman et al., 2017). Exploring whether taxonomic relations also facilitate object recognition under crowding would help determine whether the observed effects on visual perception are specific to action representations or based on higher-order semantic representations.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eCompeting Interests:\u0026nbsp;\u003c/strong\u003eAuthors have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical standards\u003c/strong\u003e: All human studies have been approved by the Ethical Committee of the University of Lille (reference: 2022-605-S106) and have therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.\u003c/p\u003e\n\u003cp\u003eThe manuscript does not contain clinical studies or patient data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003eThis work received support from the French National Research Agency (ANR-23-CE28-0015). The first author benefitted from a PhD fellowship from the University of Lille.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOpen practice statement:\u0026nbsp;\u003c/strong\u003eData, materials and analysis codes are available at: \u0026nbsp;https://doi.org/10.57745/EVVBYS\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by a PhD fellowship from the University of Lille awarded to the first author, and by a grant from the French National Research Agency (ANR) awarded to Sol\u0026egrave;ne Kal\u0026eacute;nine. The authors thank Laurent Ott for his assistance with the programming of the online experiment, and Dominique Knutsen for her help with word translations.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026bull; Conceptualization: N.Slaski, S.Kal\u0026eacute;nine, B.Sayim ; Methodology: N.Slaski, S.Kal\u0026eacute;nine, B.Sayim; Formal analysis and investigation: N.Slaski ; Writing - original draft preparation: N. Slaski; Writing - review and editing: N.Slaski, S.Kal\u0026eacute;nine, B.Sayim; Funding acquisition: S. Kal\u0026eacute;nine; Supervision: S.Kal\u0026eacute;nine, B.Sayim\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAghdaee, S. M. (2005). Adaptation to spiral motion in crowding condition. \u003cem\u003ePerception\u003c/em\u003e, \u003cem\u003e34\u003c/em\u003e(2), 155-162. https://doi.org/10.1068/p5298\u003c/li\u003e\n\u003cli\u003eAuckland, M. E., Cave, K. R., \u0026amp; Donnelly, N. (2007). Nontarget objects can influence perceptual processes during object recognition. \u003cem\u003ePsychonomic Bulletin \u0026amp; Review\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(2), 332-337. https://doi.org/10.3758/BF03194073\u003c/li\u003e\n\u003cli\u003eBar, M. (2004). Visual objects in context. \u003cem\u003eNature Reviews Neuroscience\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(8), 617-629. https://doi.org/10.1038/nrn1476\u003c/li\u003e\n\u003cli\u003eBarr, D. J., Levy, R., Scheepers, C., \u0026amp; Tily, H. J. (2013). Random effects structure for confirmatory hypothesis testing : Keep it maximal. \u003cem\u003eJournal of Memory and Language\u003c/em\u003e, \u003cem\u003e68\u003c/em\u003e(3), 255-278. https://doi.org/10.1016/j.jml.2012.11.001\u003c/li\u003e\n\u003cli\u003eBates, D., M\u0026auml;chler, M., Bolker, B., \u0026amp; Walker, S. (2015). Fitting Linear Mixed-Effects Models Using lme4. \u003cem\u003eJournal of Statistical Software\u003c/em\u003e, \u003cem\u003e67\u003c/em\u003e, 1-48. https://doi.org/10.18637/jss.v067.i01\u003c/li\u003e\n\u003cli\u003eBerger, A., \u0026amp; Kiefer, M. (2021). Comparison of Different Response Time Outlier Exclusion Methods : A Simulation Study. \u003cem\u003eFrontiers in Psychology\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e. https://doi.org/10.3389/fpsyg.2021.675558\u003c/li\u003e\n\u003cli\u003eBiederman, I. (1972). Perceiving Real-World Scenes. \u003cem\u003eScience\u003c/em\u003e, \u003cem\u003e177\u003c/em\u003e(4043), 77-80. https://doi.org/10.1126/science.177.4043.77\u003c/li\u003e\n\u003cli\u003eBiederman, I., Mezzanotte, R. J., \u0026amp; Rabinowitz, J. C. (1982). Scene perception : Detecting and judging objects undergoing relational violations. \u003cem\u003eCognitive Psychology\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(2), 143-177. https://doi.org/10.1016/0010-0285(82)90007-X\u003c/li\u003e\n\u003cli\u003eBorghi, A. M., Flumini, A., Natraj, N., \u0026amp; Wheaton, L. A. (2012). One hand, two objects : Emergence of affordance in contexts. \u003cem\u003eBrain and Cognition\u003c/em\u003e, \u003cem\u003e80\u003c/em\u003e(1), 64-73. https://doi.org/10.1016/j.bandc.2012.04.007\u003c/li\u003e\n\u003cli\u003eBouma, H. (1970). Interaction Effects in Parafoveal Letter Recognition. \u003cem\u003eNature\u003c/em\u003e, \u003cem\u003e226\u003c/em\u003e(5241), 177-178. https://doi.org/10.1038/226177a0\u003c/li\u003e\n\u003cli\u003eBrysbaert, M., \u0026amp; Stevens, M. (2018). Power Analysis and Effect Size in Mixed Effects Models : A Tutorial. \u003cem\u003eJournal of Cognition\u003c/em\u003e, \u003cem\u003e1\u003c/em\u003e(1), 9. https://doi.org/10.5334/joc.10\u003c/li\u003e\n\u003cli\u003eChien, S.-E., Yang, Y.-H., Ono, Y., \u0026amp; Yeh, S.-L. (2022). Theta activity in semantic priming under visual crowding as revealed by magnetoencephalography. \u003cem\u003eNeuroscience Research\u003c/em\u003e, \u003cem\u003e185\u003c/em\u003e, 29-39. https://doi.org/10.1016/j.neures.2022.09.004\u003c/li\u003e\n\u003cli\u003eEstes, Z., Golonka, S., \u0026amp; Jones, L. L. (2011). Chapter eight - Thematic Thinking : The Apprehension and Consequences of Thematic Relations. In B. H. Ross (\u0026Eacute;d.), \u003cem\u003ePsychology of Learning and Motivation\u003c/em\u003e (Vol. 54, p. 249-294). Academic Press. https://doi.org/10.1016/B978-0-12-385527-5.00008-5\u003c/li\u003e\n\u003cli\u003eFarzin, F., Rivera, S. M., \u0026amp; Whitney, D. (2009). Holistic crowding of Mooney faces. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(6), 18. https://doi.org/10.1167/9.6.18\u003c/li\u003e\n\u003cli\u003eGreen, C., \u0026amp; Hummel, J. E. (2006). Familiar interacting object pairs are perceptually grouped. \u003cem\u003eJournal of Experimental Psychology: Human Perception and Performance\u003c/em\u003e, \u003cem\u003e32\u003c/em\u003e(5), 1107-1119. https://doi.org/10.1037/0096-1523.32.5.1107\u003c/li\u003e\n\u003cli\u003eGreenwood, J. A., Bex, P. J., \u0026amp; Dakin, S. C. (2010). Crowding Changes Appearance. \u003cem\u003eCurrent Biology\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(6), 496-501. https://doi.org/10.1016/j.cub.2010.01.023\u003c/li\u003e\n\u003cli\u003eHe, S., Cavanagh, P., \u0026amp; Intriligator, J. (1996). Attentional resolution and the locus of visual awareness. \u003cem\u003eNature\u003c/em\u003e, \u003cem\u003e383\u003c/em\u003e(6598), 334-337. https://doi.org/10.1038/383334a0\u003c/li\u003e\n\u003cli\u003eHerzog, M. H., Sayim, B., Chicherov, V., \u0026amp; Manassi, M. (2015). Crowding, grouping, and object recognition : A matter of appearance. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(6), 5. https://doi.org/10.1167/15.6.5\u003c/li\u003e\n\u003cli\u003eHollingworth, A. (1998). Does consistent scene context facilitate object perception? \u003cem\u003eJournal of Experimental Psychology: General\u003c/em\u003e, \u003cem\u003e127\u003c/em\u003e(4), 398-415. https://doi.org/10.1037/0096-3445.127.4.398\u003c/li\u003e\n\u003cli\u003eHollingworth, A., \u0026amp; Henderson, J. M. (1999). Object identification is isolated from scene semantic constraint : Evidence from object type and token discrimination. \u003cem\u003eActa Psychologica\u003c/em\u003e, \u003cem\u003e102\u003c/em\u003e(2), 319-343. https://doi.org/10.1016/S0001-6918(98)00053-5\u003c/li\u003e\n\u003cli\u003eIkeda, H., Watanabe, K., \u0026amp; Cavanagh, P. (2013). Crowding of biological motion stimuli. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(4), 20. https://doi.org/10.1167/13.4.20\u003c/li\u003e\n\u003cli\u003eJudd, C. M., Westfall, J., \u0026amp; Kenny, D. A. (2017). Experiments with More Than One Random Factor : Designs, Analytic Models, and Statistical Power. \u003cem\u003eAnnual Review of Psychology\u003c/em\u003e, \u003cem\u003e68\u003c/em\u003e(Volume 68, 2017), 601-625. https://doi.org/10.1146/annurev-psych-122414-033702\u003c/li\u003e\n\u003cli\u003eKaiser, D., Stein, T., \u0026amp; Peelen, M. V. (2014). Object grouping based on real-world regularities facilitates perception by reducing competitive interactions in visual cortex. \u003cem\u003eProceedings of the National Academy of Sciences\u003c/em\u003e, \u003cem\u003e111\u003c/em\u003e(30), 11217-11222. https://doi.org/10.1073/pnas.1400559111\u003c/li\u003e\n\u003cli\u003eKal\u0026eacute;nine, S., \u0026amp; Bonthoux, F. (2008). Object manipulability affects children\u0026rsquo;s and adults\u0026rsquo; conceptual processing. \u003cem\u003ePsychonomic Bulletin \u0026amp; Review\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(3), 667-672. https://doi.org/10.3758/PBR.15.3.667\u003c/li\u003e\n\u003cli\u003eKal\u0026eacute;nine, S., \u0026amp; Buxbaum, L. J. (2016). Thematic knowledge, artifact concepts, and the left posterior temporal lobe : Where action and object semantics converge. \u003cem\u003eCortex\u003c/em\u003e, \u003cem\u003e82\u003c/em\u003e, 164-178. https://doi.org/10.1016/j.cortex.2016.06.008\u003c/li\u003e\n\u003cli\u003eKal\u0026eacute;nine, S., Mirman, D., Middleton, E. L., \u0026amp; Buxbaum, L. J. (2012). Temporal dynamics of activation of thematic and functional knowledge during conceptual processing of manipulable artifacts. \u003cem\u003eJournal of Experimental Psychology: Learning, Memory, and Cognition\u003c/em\u003e, \u003cem\u003e38\u003c/em\u003e(5), 1274-1295. https://doi.org/10.1037/a0027626\u003c/li\u003e\n\u003cli\u003eKal\u0026eacute;nine, S., Peyrin, C., Pichat, C., Segebarth, C., Bonthoux, F., \u0026amp; Baciu, M. (2009). The sensory-motor specificity of taxonomic and thematic conceptual relations : A behavioral and fMRI study. \u003cem\u003eNeuroImage\u003c/em\u003e, \u003cem\u003e44\u003c/em\u003e(3), 1152-1162. https://doi.org/10.1016/j.neuroimage.2008.09.043\u003c/li\u003e\n\u003cli\u003eKooi, F. L., Toet, A., Tripathy, S. P., \u0026amp; Levi, D. M. (1994). The effect of similarity and duration on spatial interaction in peripheral vision. \u003cem\u003eSpatial Vision\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(2), 255-279. https://doi.org/10.1163/156856894x00350\u003c/li\u003e\n\u003cli\u003eKouider, S., Berthet, V., \u0026amp; Faivre, N. (2011). Preference is biased by crowded facial expressions. \u003cem\u003ePsychological Science\u003c/em\u003e, \u003cem\u003e22\u003c/em\u003e(2), 184-189. https://doi.org/10.1177/0956797610396226\u003c/li\u003e\n\u003cli\u003eKumle, L., V\u0026otilde;, M. L.-H., \u0026amp; Draschkow, D. (2021). Estimating power in (generalized) linear mixed models : An open introduction and tutorial in R. \u003cem\u003eBehavior Research Methods\u003c/em\u003e, \u003cem\u003e53\u003c/em\u003e(6), 2528-2543. https://doi.org/10.3758/s13428-021-01546-0\u003c/li\u003e\n\u003cli\u003eLevi, D. M. (2008). Crowding\u0026mdash;An essential bottleneck for object recognition : A mini-review. \u003cem\u003eVision Research\u003c/em\u003e, \u003cem\u003e48\u003c/em\u003e(5), 635-654. https://doi.org/10.1016/j.visres.2007.12.009\u003c/li\u003e\n\u003cli\u003eLevi, D. M., Hariharan, S., \u0026amp; Klein, S. A. (2002). Suppressive and facilitatory spatial interactions in peripheral vision : Peripheral crowding is neither size invariant nor simple contrast masking. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(2), 3. https://doi.org/10.1167/2.2.3\u003c/li\u003e\n\u003cli\u003eLin, E. L., \u0026amp; Murphy, G. L. (2001). Thematic relations in adults\u0026rsquo; concepts. \u003cem\u003eJournal of Experimental Psychology: General\u003c/em\u003e, \u003cem\u003e130\u003c/em\u003e(1), 3-28. https://doi.org/10.1037/0096-3445.130.1.3\u003c/li\u003e\n\u003cli\u003eLiu, L., \u0026amp; Arditi, A. (2001). How Crowding Affects Letter Confusion. \u003cem\u003eOptometry and Vision Science\u003c/em\u003e, \u003cem\u003e78\u003c/em\u003e(1), 50.\u003c/li\u003e\n\u003cli\u003eLivne, T., \u0026amp; Sagi, D. (2007). Configuration influence on crowding. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(2), 4. https://doi.org/10.1167/7.2.4\u003c/li\u003e\n\u003cli\u003eLouie, E. G., Bressler, D. W., \u0026amp; Whitney, D. (2007). Holistic crowding : Selective interference between configural representations of faces in crowded scenes. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(2), 24. https://doi.org/10.1167/7.2.24\u003c/li\u003e\n\u003cli\u003eMalania, M., Herzog, M. H., \u0026amp; Westheimer, G. (2007). Grouping of contextual elements that affect vernier thresholds. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(2), 1. https://doi.org/10.1167/7.2.1\u003c/li\u003e\n\u003cli\u003eManassi, M., Sayim, B., \u0026amp; Herzog, M. H. (2012). Grouping, pooling, and when bigger is better in visual crowding. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e(10), 13. https://doi.org/10.1167/12.10.13\u003c/li\u003e\n\u003cli\u003eManassi, M., Sayim, B., \u0026amp; Herzog, M. H. (2013). When crowding of crowding leads to uncrowding. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(13), 10. https://doi.org/10.1167/13.13.10\u003c/li\u003e\n\u003cli\u003eMelnik, N., Coates, D. R., \u0026amp; Sayim, B. (2018). Emergent features in the crowding zone : When target\u0026ndash;flanker grouping surmounts crowding. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e18\u003c/em\u003e(9), 19. https://doi.org/10.1167/18.9.19\u003c/li\u003e\n\u003cli\u003eMelnik, N., Coates, D. R., \u0026amp; Sayim, B. (2020). Emergent features break the rules of crowding. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(1), 406. https://doi.org/10.1038/s41598-019-57277-y\u003c/li\u003e\n\u003cli\u003eMirman, D., Landrigan, J.-F., \u0026amp; Britt, A. E. (2017). Taxonomic and thematic semantic systems. \u003cem\u003ePsychological Bulletin\u003c/em\u003e, \u003cem\u003e143\u003c/em\u003e(5), 499-520. https://doi.org/10.1037/bul0000092\u003c/li\u003e\n\u003cli\u003eNah, J. C., \u0026amp; Geng, J. J. (2022). Thematic object pairs produce stronger and faster grouping than taxonomic pairs. \u003cem\u003eJournal of Experimental Psychology: Human Perception and Performance\u003c/em\u003e, \u003cem\u003e48\u003c/em\u003e(12), 1325-1335. https://doi.org/10.1037/xhp0001031\u003c/li\u003e\n\u003cli\u003eOldfield, R. C. (1971). The assessment and analysis of handedness : The Edinburgh inventory. \u003cem\u003eNeuropsychologia\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(1), 97-113. https://doi.org/10.1016/0028-3932(71)90067-4\u003c/li\u003e\n\u003cli\u003eOliva, A., \u0026amp; Schyns, P. G. (1997). Coarse Blobs or Fine Edges? Evidence That Information Diagnosticity Changes the Perception of Complex Visual Stimuli. \u003cem\u003eCognitive Psychology\u003c/em\u003e, \u003cem\u003e34\u003c/em\u003e(1), 72-107. https://doi.org/10.1006/cogp.1997.0667\u003c/li\u003e\n\u003cli\u003eParkes, L., Lund, J., Angelucci, A., Solomon, J. A., \u0026amp; Morgan, M. (2001). Compulsory averaging of crowded orientation signals in human vision. \u003cem\u003eNature Neuroscience\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(7), 739-744. https://doi.org/10.1038/89532\u003c/li\u003e\n\u003cli\u003ePeirce, J., Gray, J. R., Simpson, S., MacAskill, M., H\u0026ouml;chenberger, R., Sogo, H., Kastman, E., \u0026amp; Lindel\u0026oslash;v, J. K. (2019). PsychoPy2 : Experiments in behavior made easy. \u003cem\u003eBehavior Research Methods\u003c/em\u003e, \u003cem\u003e51\u003c/em\u003e(1), 195-203. https://doi.org/10.3758/s13428-018-01193-y\u003c/li\u003e\n\u003cli\u003ePelli, D. G. (2008). Crowding : A cortical constraint on object recognition. \u003cem\u003eCurrent Opinion in Neurobiology\u003c/em\u003e, \u003cem\u003e18\u003c/em\u003e(4), 445-451. https://doi.org/10.1016/j.conb.2008.09.008\u003c/li\u003e\n\u003cli\u003ePelli, D. G., Palomares, M., \u0026amp; Majaj, N. J. (2004). Crowding is unlike ordinary masking : Distinguishing feature integration from detection. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(12), 12. https://doi.org/10.1167/4.12.12\u003c/li\u003e\n\u003cli\u003eRiddoch, M. J., Humphreys, G. W., Edwards, S., Baker, T., \u0026amp; Willson, K. (2003). Seeing the action : Neuropsychological evidence for action-based effects on object selection. \u003cem\u003eNature Neuroscience\u003c/em\u003e, \u003cem\u003e6\u003c/em\u003e(1), 82-89. https://doi.org/10.1038/nn984\u003c/li\u003e\n\u003cli\u003eRiddoch, M. J., Humphreys ,G. W., Hickman ,M., Clift ,J., Daly ,A., \u0026amp; and Colin, J. (2006). I can see what you are doing : Action familiarity and affordance promote recovery from extinction. \u003cem\u003eCognitive Neuropsychology\u003c/em\u003e, \u003cem\u003e23\u003c/em\u003e(4), 583-605. https://doi.org/10.1080/02643290500310962\u003c/li\u003e\n\u003cli\u003eRoberts, K. L., \u0026amp; Humphreys, G. W. (2011). Action relations facilitate the identification of briefly-presented objects. \u003cem\u003eAttention, Perception, \u0026amp; Psychophysics\u003c/em\u003e, \u003cem\u003e73\u003c/em\u003e(2), 597-612. https://doi.org/10.3758/s13414-010-0043-0\u003c/li\u003e\n\u003cli\u003eRousselet, G., Joubert, O., \u0026amp; Fabre-Thorpe, M. (2005). How long to get to the \u0026ldquo;gist\u0026rdquo; of real-world natural scenes? \u003cem\u003eVisual Cognition\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e(6), 852-877. https://doi.org/10.1080/13506280444000553\u003c/li\u003e\n\u003cli\u003eRoux-Sibilon, A., Kal\u0026eacute;nine, S., Pichat, C., \u0026amp; Peyrin, C. (2018). Dorsal and ventral stream contribution to the paired-object affordance effect. \u003cem\u003eNeuropsychologia\u003c/em\u003e, \u003cem\u003e112\u003c/em\u003e, 125-134. https://doi.org/10.1016/j.neuropsychologia.2018.03.007\u003c/li\u003e\n\u003cli\u003eRummens, K., \u0026amp; Sayim, B. (2019). Disrupting uniformity : Feature contrasts that reduce crowding interfere with peripheral word recognition. \u003cem\u003eVision Research\u003c/em\u003e, \u003cem\u003e161\u003c/em\u003e, 25-35. https://doi.org/10.1016/j.visres.2019.05.006\u003c/li\u003e\n\u003cli\u003eRummens, K., \u0026amp; Sayim, B. (2021). Broad attention uncovers benefits of stimulus uniformity in visual crowding. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(1), 23976. https://doi.org/10.1038/s41598-021-03258-z\u003c/li\u003e\n\u003cli\u003eRummens, K., \u0026amp; Sayim, B. (2022). Multidimensional feature interactions in visual crowding : When\u0026thinsp; configural\u0026thinsp; cues\u0026thinsp; eliminate the polarity advantage. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e22\u003c/em\u003e(6), 2. https://doi.org/10.1167/jov.22.6.2\u003c/li\u003e\n\u003cli\u003eSaarela, T. P., Sayim, B., Westheimer, G., \u0026amp; Herzog, M. H. (2009). Global stimulus configuration modulates crowding. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(2), 5. https://doi.org/10.1167/9.2.5\u003c/li\u003e\n\u003cli\u003eSayim, B., Greenwood, J. A., \u0026amp; Cavanagh, P. (2014). Foveal target repetitions reduce crowding. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(6), 4. https://doi.org/10.1167/14.6.4\u003c/li\u003e\n\u003cli\u003eSayim, B., Westheimer, G., \u0026amp; Herzog, M. H. (2008). Contrast polarity, chromaticity, and stereoscopic depth modulate contextual interactions in vernier acuity. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(8), 12. https://doi.org/10.1167/8.8.12\u003c/li\u003e\n\u003cli\u003eSayim, B., Westheimer, G., \u0026amp; Herzog, M. H. (2010). Gestalt Factors Modulate Basic Spatial Vision. \u003cem\u003ePsychological Science\u003c/em\u003e, \u003cem\u003e21\u003c/em\u003e(5), 641-644. https://doi.org/10.1177/0956797610368811\u003c/li\u003e\n\u003cli\u003eSong, S., Levi, D. M., \u0026amp; Pelli, D. G. (2014). A double dissociation of the acuity and crowding limits to letter identification, and the promise of improved visual screening. \u003cem\u003eJournal of Vision\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(5), 3. https://doi.org/10.1167/14.5.3\u003c/li\u003e\n\u003cli\u003eTsagkaridis, K., Watson, C. E., Jax, S. A., \u0026amp; Buxbaum, L. J. (2014). The role of action representations in thematic object relations. \u003cem\u003eFrontiers in Human Neuroscience\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e. https://doi.org/10.3389/fnhum.2014.00140\u003c/li\u003e\n\u003cli\u003eWamain, Y., Pluciennicka, E., \u0026amp; Kal\u0026eacute;nine, S. (2014). Temporal dynamics of action perception : Differences on ERP evoked by object-related and non-object-related actions. \u003cem\u003eNeuropsychologia\u003c/em\u003e, \u003cem\u003e63\u003c/em\u003e, 249-258. https://doi.org/10.1016/j.neuropsychologia.2014.08.034\u003c/li\u003e\n\u003cli\u003eWestfall, J., Kenny, D. A., \u0026amp; Judd, C. M. (2014). Statistical power and optimal design in experiments in which samples of participants respond to samples of stimuli. \u003cem\u003eJournal of Experimental Psychology: General\u003c/em\u003e, \u003cem\u003e143\u003c/em\u003e(5), 2020-2045. https://doi.org/10.1037/xge0000014\u003c/li\u003e\n\u003cli\u003eWestheimer, G., \u0026amp; Hauske, G. (1975). Temporal and spatial interference with vernier acuity. \u003cem\u003eVision Research\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(10), 1137-1141. https://doi.org/10.1016/0042-6989(75)90012-7\u003c/li\u003e\n\u003cli\u003eWhitney, D. (2005). Motion distorts perceived position without awareness of motion. \u003cem\u003eCurrent Biology\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(9), R324-R326. https://doi.org/10.1016/j.cub.2005.04.043\u003c/li\u003e\n\u003cli\u003eWhitney, D., \u0026amp; Levi, D. M. (2011). Visual crowding : A fundamental limit on conscious perception and object recognition. \u003cem\u003eTrends in Cognitive Sciences\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(4), 160-168. https://doi.org/10.1016/j.tics.2011.02.005\u003c/li\u003e\n\u003cli\u003eWilkinson, F., Wilson, H. R., \u0026amp; Ellemberg, D. (1997). Lateral interactions in peripherally viewed texture arrays. \u003cem\u003eJOSA A\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(9), 2057-2068. https://doi.org/10.1364/JOSAA.14.002057\u003c/li\u003e\n\u003cli\u003eXu, S., Humphreys, G. W., \u0026amp; Heinke, D. (2015). Implied actions between paired objects lead to affordance selection by inhibition. \u003cem\u003eJournal of Experimental Psychology: Human Perception and Performance\u003c/em\u003e, \u003cem\u003e41\u003c/em\u003e(4), 1021-1036. https://doi.org/10.1037/xhp0000059\u003c/li\u003e\n\u003cli\u003eYeh, S.-L., He, S., \u0026amp; Cavanagh, P. (2012). Semantic Priming From Crowded Words. \u003cem\u003ePsychological Science\u003c/em\u003e, \u003cem\u003e23\u003c/em\u003e(6), 608-616. https://doi.org/10.1177/0956797611434746\u003c/li\u003e\n\u003cli\u003eYoon, E. Y., Humphreys, G. W., \u0026amp; Riddoch, M. J. (2010). The paired-object affordance effect. \u003cem\u003eJournal of Experimental Psychology: Human Perception and Performance\u003c/em\u003e, \u003cem\u003e36\u003c/em\u003e(4), 812-824. https://doi.org/10.1037/a0017175\u003c/li\u003e\n\u003cli\u003eZhang, L., Zhang, L., Mou, X., \u0026amp; Zhang, D. (2011). FSIM : A Feature Similarity Index for Image Quality Assessment. \u003cem\u003eIEEE Transactions on Image Processing\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(8), 2378-2386. IEEE Transactions on Image Processing. https://doi.org/10.1109/TIP.2011.2109730\u003c/li\u003e\n\u003cli\u003eZhou, J., Lee, C.-L., Li, K.-A., Tien, Y.-H., \u0026amp; Yeh, S.-L. (2016). Does Temporal Integration Occur for Unrecognizable Words in Visual Crowding? \u003cem\u003ePLOS ONE\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(2), e0149355. https://doi.org/10.1371/journal.pone.0149355\u003c/li\u003e\n\u003cli\u003eZloteanu, M., \u0026amp; Vuorre, M. (2024). A Tutorial for Deception Detection Analysis or : How I Learned to Stop Aggregating Veracity Judgments and Embraced Signal Detection Theory Mixed Models. \u003cem\u003eJournal of Nonverbal Behavior\u003c/em\u003e, \u003cem\u003e48\u003c/em\u003e(1), 161-185. https://doi.org/10.1007/s10919-024-00456-x\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"psychological-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"prpf","sideBox":"Learn more about [Psychological Research](http://link.springer.com/journal/426)","snPcode":"426","submissionUrl":"https://submission.nature.com/new-submission/426/3","title":"Psychological Research","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7442959/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7442959/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eObject perception is challenged in complex and cluttered visual environments. This phenomenon, known as visual crowding, occurs when object identification is impaired because a target is surrounded by other objects. However, visual environments are also meaningful, and objects are often arranged in coherent and predictable ways. In particular, thematic relations linking objects that play complementary roles in events (for example, hammer\u0026ndash;nail) have been shown to facilitate conceptual processing, especially for manipulable objects. However, it remains unclear whether such semantic relations can also influence perceptual processes. This study investigated whether thematic information can be processed under visual crowding and facilitate object identification in clutter. Participants were briefly presented with pairs of object images in peripheral vision and asked to perform an object identification task. The object pairs were either thematically related or unrelated, and correctly positioned for being used together or not. Additionally, objects were presented either in isolation or crowded by meaningless flankers. Experiment 1 demonstrated that objects in thematically related pairs were identified more accurately than those in unrelated pairs, even under crowding. Experiment 2 showed that while the thematic benefit may occur at the decisional stage when objects are presented in isolation, thematic grouping may be at play in crowding conditions when perceiving objects. These findings suggest that semantic knowledge involving the use of manipulable objects can survive visual crowding and influence object recognition. Our findings align with an embodied perspective of perception, suggesting that action-related semantic knowledge can influence relatively early stages of visual processing.\u003c/p\u003e","manuscriptTitle":"Thematic knowledge survives visual crowding and influences object identification","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-09 11:53:31","doi":"10.21203/rs.3.rs-7442959/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-09-28T13:25:02+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-28T05:13:27+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-19T10:31:06+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"159656560561979584175487064493036114283","date":"2025-08-31T14:02:17+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"50486372877522886669194537354057596149","date":"2025-08-30T07:41:28+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-08-29T11:03:33+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-29T09:04:45+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-25T02:09:53+00:00","index":"","fulltext":""},{"type":"submitted","content":"Psychological Research","date":"2025-08-23T18:58:47+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"psychological-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"prpf","sideBox":"Learn more about [Psychological Research](http://link.springer.com/journal/426)","snPcode":"426","submissionUrl":"https://submission.nature.com/new-submission/426/3","title":"Psychological Research","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"e46131fd-fbf7-4156-992a-4546a517382d","owner":[],"postedDate":"September 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-03-02T16:06:31+00:00","versionOfRecord":{"articleIdentity":"rs-7442959","link":"https://doi.org/10.1007/s00426-026-02247-z","journal":{"identity":"psychological-research","isVorOnly":false,"title":"Psychological Research"},"publishedOn":"2026-02-28 15:58:39","publishedOnDateReadable":"February 28th, 2026"},"versionCreatedAt":"2025-09-09 11:53:31","video":"","vorDoi":"10.1007/s00426-026-02247-z","vorDoiUrl":"https://doi.org/10.1007/s00426-026-02247-z","workflowStages":[]},"version":"v1","identity":"rs-7442959","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7442959","identity":"rs-7442959","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}