ORIGINAL RESEARCH published: 25 July 2022 doi: 10.3389/fpsyg.2022.857630 Figure-Disembedding Is Inferior in Non-autistic Compared to Autistic Individuals but Can Be Improved by Training Christine M. Falter-Wagner 1*, Carola Bloch 1, 2, Marta Robles 1, 3, Lea Horch 2, Kai Vogeley 2,4 and Alexandra Livia Georgescu 5 1 Department of Psychiatry and Psychotherapy, Medical Faculty, LMU Munich, Munich, Germany, 2 Department of Psychiatry and Psychotherapy, University Hospital of Cologne, Cologne, Germany, 3 Department of Clinical and Health Psychology, Autonomous University of Barcelona, Barcelona, Spain, 4 Institute of Neuroscience and Medicine, Cognitive Neuroscience (INM-3), Research Centre Jülich, Jülich, Germany, 5 Department of Psychology, Institute of Psychiatry, Psychology and Neuroscience, King’s College London, London, United Kingdom Edited by: Timothy L. Hubbard, Background: Figure-disembedding is one of the most discussed visuo-cognitive Arizona State University, functions, in which individuals with Autism Spectrum Disorder (ASD) have been reported United States to outperform non-autistic individuals. A local processing bias has been assumed to Reviewed by: Massimiliano Conson, underlie such superior performance patterns. The aim of the current study is to investigate University of Campania Luigi whether processing preferences can be modified by procedural priming. Vanvitelli, Italy Natalia Zaretskaya, Method: The current study used a procedural priming task (Navon figures) to induce University of Graz, Austria more local or global processing in 25 autistic and 21 typically developing (TD) control Johannes Hönekopp, Northumbria University, participants, using hierarchical figures preceding the figure-disembedding task. United Kingdom Results: Participants with ASD outperformed non-autistic individuals in the unprimed baseline *Correspondence: Christine M. Falter-Wagner task version. The performance was not modulated by priming in either direction (toward a christine.falter@cantab.net local or global processing style) in both groups. However, the performance of TD control participants was improved by training to the same level as that observed in the ASD group. Specialty section: This article was submitted to Conclusion: Figure-disembedding performance in ASD is superior to that in TD control Cognition, participants and robust against procedural priming or training. In contrast, performance a section of the journal Frontiers in Psychology in the TD control group can be improved up to the level of the ASD group. Any studies Received: 18 January 2022 reporting superiority in individuals with ASD in figure-disembedding should consider Accepted: 09 June 2022 training effects when evaluating group differences. Published: 25 July 2022 Citation: Keywords: Autism Spectrum Disorder, local processing, priming, figure-disembedding, Gestalt Falter-Wagner CM, Bloch C, Robles M, Horch L, Vogeley K and Georgescu AL (2022) Figure- Disembedding Is Inferior in INTRODUCTION Non-autistic Compared to Autistic Individuals but Can Be Improved by In the past decades, autism researchers have begun to investigate strengths in the cognitive Training. profile of individuals with Autism Spectrum Disorder (ASD). This may have been inspired by Front. Psychol. 13:857630. the seminal work by Shah and Frith (1983, 1993), who showed superior performance in doi: 10.3389/fpsyg.2022.857630 autistic compared to typically developing (TD) children on the block design sub-components Frontiers in Psychology | www.frontiersin.org 1 July 2022 | Volume 13 | Article 857630 Falter-Wagner et al. Figure-Disembedding in Autistic and Non-autistic Individuals of the Wechsler intelligence scales (Wechsler, 1981) and the with ASD can process globally if explicitly instructed to do Children’s Embedded-Figures Task (Witkin et al., 1971). The so, e.g., in a selective attention task (Plaisted et al., 1999; authors noted that a key characteristic of both tasks was that Mottron et al., 2003). the detail-focused and “local” performance benefit in ASD Priming is a traditional technique to modify individual might have been based on a disruption of Gestalt grouping. differences in perception despite identical stimulus material. As a result, the salience of parts (local processing) over wholes It is used to analyze the influence of situational variables on (global processing) might increase as distraction from the perception, cognition, feelings and behavior (Bargh and overall Gestalt is reduced. This could have caused an advantage Chartrand, 2000). Indeed, a whole series of studies has for the autistic over the TD group (e.g., Falter et al., 2008), demonstrated the effect of priming on the visuo-spatial perception which is in accordance with the “Weak-Central-Coherence” of TD participants (e.g., Lin et al., 2008). Many researchers theory (WCC, Frith, 1989) predicting superior local processing are of the opinion that situational factors in daily life trigger (Happé and Frith, 2006) and originally also reduced global similar priming effects and therefore have a significant influence processing (Frith, 1989). The assumption that autistic persons on our perception (Kühnen et al., 2001; Kühnen and Oyserman, have a preference for local processing has indeed been widely 2002; Lin et al., 2008; Oyserman and Lee, 2008; Oyserman supported (Mottron et al., 1999; Behrmann et al., 2006; Wang et al., 2009). However, to date, research on ASD and priming et al., 2007). Supporting this view, the results of a meta-analysis is sparse. Thus, the aim of the current study was to test, that included 35 studies comparing among others performance whether a processing preference can be overcome through on Block Design and Figure-Disembedding found that autistic priming in ASD as in TD. participants are, on average, superior at tasks that require Some studies have already employed priming tasks in autistic detail-focused processing compared to control persons (Muth people. Semantic priming makes use of a semantic relationship et al., 2014). On the other hand, the second (original) assumption between the content being primed and the task that follows. of the WCC theory, concerning the corresponding global It has been used in some studies to explore the spontaneous processing impairments, has been challenged (Brosnan et al., use of a contextual, i.e., global, processing mechanism in autistic 2004; Wang et al., 2007; Scherf et al., 2008; Koldewyn et al., and TD individuals (Kamio and Toichi, 2000; Toichi and Kamio, 2013; Hadad and Ziv, 2015; Booth and Happé, 2018). Indeed, 2001; López and Leekam, 2003; Hala et al., 2007; Kamio et al., some studies report that autistic individuals can successfully 2007). Toichi and Kamio (2001) examined the ability of autistic re-organize the integral structure of visual information. For individuals to use semantic connections to complete words. example, researchers have found superior performance of autistic The semantic priming was performed by presenting a semantically individuals on tasks such as conjunctive visual search, where related word to the target word. Autistic and TD individuals a target needs to be identified among distractors (Plaisted et al., showed comparable abilities in completing target words. As 1998). Moreover, Dakin and Frith’s (2005) review article provides in the TD group, the autistic individuals performed worse if evidence only for superior local processing in persons with there was no semantic connection between the priming word ASD, but not for impaired global processing. In response to and the target word. Thus, all participants, irrespective of this, alternative accounts of the WCC have been proposed, diagnosis, showed no differences in receptivity to a semantic such as the Enhanced Perceptual Functioning account (EPF, priming (Toichi and Kamio, 2001). Interestingly, Kamio and Mottron and Burack, 2001; Mottron et al., 2006) and the Toichi (2000) differentiated in their study between two different Reduced Generalization Account (RG, Plaisted, 2001). According forms of semantic priming: they used either related words or to EPF, the local preference of autistic individuals is not a images as references to the target word to be completed. While consequence of a global deficit but merely of superior perceptual control persons showed a comparable improvement after both processes. It further suggests that local preference can priming methods, autistic individuals were significantly better be overcome by global requirements, and that the ability to when image priming was used compared to word priming. process globally is not restricted. Moreover, the “low-level” Kamio et al. (2007) examined autistic individuals without speech perceptual processes are characterized by an atypical autonomy development delay in a similar paradigm. Using semantically with reduced influence by “high-level” processes (Mottron and close words, a priming was performed, in which participants Burack, 2001; Mottron et al., 2006). The RG account assumes were asked to decide whether a target word is an existing that autistic individuals tend to differentiate more, but generalize word or not. While control persons performed better after less strongly between stimuli, leading to a more detail-focused priming, no priming effect was observed in the autistic processing style. In light of the evidence, the original WCC participants. This contrasts with the previously mentioned theory was honed to “detail-focused cognitive style” (Happé studies of the same team, rendering the literature on semantic and Frith, 2006). priming in autism inconclusive. In summary, the preference for a local processing mode While semantic priming lends itself to language studies, (if not otherwise instructed) by autistic individuals stands in the present study set out to investigate procedural priming, contrast to that by TD individuals, who tend to show a global which assumes that the use of a cognitive procedure on a processing preference (if not otherwise instructed). This raises given task can heighten the likelihood that the same procedure the intriguing question of whether the respective default being used on a subsequent task. To test the modifiability of processing style might be influenced dynamically under certain a default processing style we made use of a figure-disembedding task designs. For instance, it has been shown that participants task, in which autistic individuals have previously been shown Frontiers in Psychology | www.frontiersin.org 2 July 2022 | Volume 13 | Article 857630 Falter-Wagner et al. Figure-Disembedding in Autistic and Non-autistic Individuals to have a superior performance to TD individuals (Falter et al., data sets (16 males, 9 females) and 21 TD data sets (8 males, 2008). This task requires participants to detect a smaller shape 13 females). All participants were between 24 and 56 years as a constituent part of a more complex figure. For procedural old. The autistic individuals were older (M = 47.48, SD = 6.56) priming preceding the figure-disembedding task, we used the than TD individuals (M = 40.24, SD = 7.58). However, the groups Navon task (Navon, 1977) requiring either local or global were matched by verbal IQ, performance IQ, and full IQ (see processing, as already used for priming purposes before (Förster Table 1 for descriptive statistics and matching). et al., 2008). We therefore investigated whether procedural Participants were recruited through the Autism Outpatient priming of global processing (henceforth: global priming) might Unit of the University Hospital Cologne. Study inclusion criteria deteriorate performance on a figure-disembedding task in autistic for the TD group were lack of a neurological or psychiatric individuals compared to their baseline performance, which has diagnosis, no use of psychotropic medication, normal or been found to show a local processing tendency in previous corrected-to-normal vision and age between 18 and 60 years. perceptual studies (e.g., Falter and Bailey, 2011; Falter, 2012; Study inclusion criteria for the ASD group were a diagnosis Falter et al., 2013). We also investigated whether procedural of autistic disorder (F84.0) and Asperger’s Syndrome (F84.5) priming of local processing (henceforth: local priming) might according to ICD-10 [World Health Organization (WHO), improve performance in TD individuals compared to their 1993], with an at least average Full Scale IQ (FIQ > 85). Further, baseline performance. The potential of dynamically changing participants were included if they had no comorbid neurological processing styles on figure-disembedding would show the extent or psychiatric diagnosis, no use of psychotropic medication, to which the alleged processing differences in ASD compared and if they had normal or corrected-to-normal vision and age to TD are modifiable by context and task design. between 18 and 60 years. Participants took part in an unrelated First, in concordance with findings by Muth et al. (2014), study on alexithymia and emotion recognition on the same we hypothesized (H1) that the performance on the figure- testing day (unpublished data). The study was approved by disembedding task would be superior in autistic compared to the ethics committee of the Medical Faculty of the University TD individuals in the unprimed condition (i.e., baseline). of Cologne. Participants provided consent prior to initiating Second, given the idea of a general default tendency toward the study. a local processing style in ASD and global processing style in TD, we predicted (H2) reduced performance after global priming Neuropsychological Tests and in the autistic group and (H3) increased performance after Questionnaires local priming in the TD group, whereas priming modes that Wechsler Adult Intelligence Scale—Fourth Edition already fit with the general tendency should not make a The cognitive abilities were measured with the Wechsler Adult difference in performance in either group. Intelligence Scale—Fourth Edition (WAIS-IV, Petermann, 2012). There were no significant differences between the groups for MATERIALS AND METHODS either performance or verbal IQ (see Table 1). Participants Beck Depression Inventory—II A total of 29 autistic participants and 22 TD adult control The Beck Depression Inventory—II (BDI-II, Beck et al., 1996), participants took part in the study. Technical issues resulted was used to measure the severity of depressive symptomology. in incomplete data sets of four ASD and one TD participant. The reasoning for this is because depression is a common The analyzed samples therefore included 25 evaluable ASD comorbidity in autism (e.g., Hollocks et al., 2019) and it is TABLE 1 | Demographic data. TD (n = 21) ASD (n = 25a) W p-value Rank biserial correlation Mean SD Mean SD Age 40.24 7.77 47.48 6.70 409.00 0.001 0.56 BDI 2.52 2.75 13.00 8.13 479.50 <0.001 0.83 VIQ 110.86 17.00 116.76 17.11 320.50 0.204 0.22 PIQ 103.33 15.88 111.68 19.26 345.00 0.070 0.31 FIQ 108.24 16.8 115.68 18.51 331.00 0.133 0.26 AQ 14.62 4.59 43.16 2.76 525.00 <0.001 1.00 EQ 50.67 13.79 14.38 7.62 5.50 <0.001 −0.98 SQ 22.24 8.16 43.25 14.01 441.00 <0.001 0.75 Mann Whitney U-test was performed, because most variables violated the assumption of normality (determined by Shapiro–Wilk test). Rank biserial correlation is a method of reporting the effect size for the Mann–Whitney U-test. BDI, Beck’s Depression Inventory; VIQ, verbal intelligence quotient; PIQ, performance intelligence quotient; FIQ, full intelligence quotient; VIQ, PIQ, and FIQ were measured using the Wechsler Adult Intelligence Scale. AQ, Autism Quotient; EQ, Empathy Questionnaire; SQ, Systemizing Questionnaire. Standard deviations (SD), scores for both the Autism Spectrum Disorder (ASD) group and the typically developed (TD) group. an = 24 for EQ and SQ. Frontiers in Psychology | www.frontiersin.org 3 July 2022 | Volume 13 | Article 857630 Falter-Wagner et al. Figure-Disembedding in Autistic and Non-autistic Individuals associated with differences in motivation and cognition (Grahek stimuli were shown repeatedly, in randomized order, so that et al., 2019). The participants with ASD had a significantly each participant worked on a total of 72 NFs per condition. higher BDI score (see Table 1). Participants were instructed to identify as quickly but as accurately as possible, which of the two target letters was presented and to press the “F” or “H” keys accordingly. The Autism Spectrum Quotient Test, Systemizing progression was self-paced. The number of correct answers Quotient, and Empathy Quotient and the corresponding reaction times were recorded. It is The Autism Spectrum Quotient Test (AQ, Baron-Cohen et al., important to highlight at this stage that the “baseline” condition 2001), The Systemizing Quotient (SQ, Baron-Cohen et al., 2003) was one where there was no priming and hence no Navon and The Empathy Quotient (EQ, Baron-Cohen and Wheelwright, task to complete at all. It was always the first block to complete. 2004), were used to measure traits associated with ASD. One A computerized version of a figure-disembedding task (i.e., participant with ASD had not filled out the EQ and the SQ. As EFT), was used to measure visual perception before and after expected, the AQ and SQ scores of the ASD participants were priming, which has already been successfully used in several significantly higher and the EQ score of the ASD participants previous studies (Falter et al., 2006, 2008; Brosnan et al., 2012). significantly lower than the control scores (see Table 1). All Each trial consisted of a colored, complex image and two control participants achieved an AQ score below the CUT-OFF shapes below it. The participants were asked to decide as value of ≥32 points. For demographic data see Table 1. quickly but as accurately as possible which of the two shapes was embedded in the complex image. The progression was Design, Task, and Procedure therefore self-paced (see Figure 1B). They were instructed to The study has a mixed 2 × 3 design with a between-subject give their answers by pressing one of two keys: the “F” key factor Group (with two levels: ASD or TD) and a within- for the left form, the “H” key for the right form. In half of subject factor Priming (with three levels: baseline, local or the trials the right shape was embedded, in the other half the global). The Navon Figures (NF) task was used for priming left shape. Auditory and visual feedback was given for 2,000 ms; (Navon, 1977). The dependent variable was the performance two different cues for a correct and a wrong answer were (accuracy and reaction times) on a computerized version of played (auditory) and the embedded shape was marked in the a figure-disembedding task, known as the Embedded Figures complex picture by a border (visual). Task (EFT) to measure the priming effect on visual perception Regarding the overall procedure, after completing the (Falter et al., 2006, 2008). Each of these two tasks (Navon questionnaires and going through the instruction for the main Figures for priming and EFT for measuring visual perception) task, the presentation of the above-mentioned tasks was as follows. are detailed below, after which the general procedure is outlined. Participants worked through two practice trials requiring figure- The priming consisted of the Navon Figures task (Navon, disembedding. The inter-trial interval was always 500 ms. The 1977). NFs consisted of small letters (0.5°–0.5° of visual angle), first block was always unprimed and consisted of six figure- which were grouped to form large letters (2.5°–2.5° of visual disembedding trials to measure a baseline performance against angle). The idea behind them, namely that small letters are which to compare each of the two priming blocks to follow. The arranged to form large letters, offer both local and global second and third blocks were primed with a priming condition features to be processed. The NFs always contained the target using the NF task (local then global or vice-versa, counterbalanced letters “F” or “H” (see Figure 1). In the “local” condition, across participants within each group). The priming condition one of the two target letters corresponded to the small letters, (local or global) remained the same within a block. The second and in the “global” condition one of the two target letters block started with 48 Navon priming trials (either local or global, corresponded to the large letters. There were 14 different stimuli depending on the condition) followed by six figure-disembedding for the local and 14 for the global priming condition. These trials. Before each figure-disembedding trial, an additional four A B C FIGURE 1 | Sample stimuli or Navon Figures (NF) for priming. Instruction: “Does the stimulus contain an F or an H?”; (A) global priming, (B) local priming, and (C) presents an example stimulus of the Figure-Disembedding Task with a complex figure at the top (with shape inserted in red for illustration purposes) and two answer options at the bottom (the left shape would be the correct answer option). Frontiers in Psychology | www.frontiersin.org 4 July 2022 | Volume 13 | Article 857630 Falter-Wagner et al. Figure-Disembedding in Autistic and Non-autistic Individuals TABLE 2 | Reaction time (ms) and accuracy by participant group. TD ASD Mean SD Min Max Mean SD Min Max ACC baseline 3.43 1.40 1.00 6.00 4.16 0.75 3.00 5.00 ACC global 4.29 1.35 2.00 6.00 4.16 1.28 2.00 6.00 ACC local 4.00 1.18 1.00 6.00 4.04 1.27 2.00 6.00 RT baseline 12042.86 11474.65 2447.00 51050.00 14473.08 12221.67 2167.00 58569.00 RT global 10358.52 7162.06 2363.00 35832.00 12593.44 9343.76 2705.00 41895.00 RT local 12376.76 8436.30 3079.00 37127.00 13056.12 10100.69 2866.00 45205.00 Reaction times (RT) in milliseconds and accuracy scores (ACC) for baseline, globally primed (global) and locally primed (local) conditions. Standard deviations (SD), minimum (Min) and maximum (Max) scores for both the Autism Spectrum Disorder (ASD) group and the typically developed (TD) group. priming trials were presented in order to prevent the priming with ASD compared to TD individuals using directed Mann effect from fading. The third block had the same structure as Whitney U-test due to singular model fit: [W(44) = 348, p = 0.025, the second block, except that the alternative priming condition r = 0.290].There was no group difference concerning RT (we (local or global, respectively) was presented preceding the figure- performed the analysis only for the RTs in correct trials) in disembedding trials. The allocation of figure-disembedding trials the baseline condition [W(44) = 308, p = 0.162] Thus, to the local, global or unprimed condition was randomized for we concentrated on accuracy data in the following analyses each individual experimental run. and are reporting the non-significant results of the analyses The total duration of the study, including the questionnaires for the RTs in the Supplementary Materials. was about 60 min. At the end participants were debriefed and paid for their participation. To keep the distance between eye Priming Effect and screen constant at 59 cm a chin and forehead support Testing for H2, a global priming effect, first, we fitted models was used. The stimulus presentation and response recording for data from baseline and global conditions only. We included in both tasks were performed by the software package Presentation fixed effects for Group (ASD; TD), Condition (baseline; global), (version 12.2; Neurobehavioral Systems, Inc.).1 and their interaction. We found a trend of a main effect of Condition [χ2(1) = 3.19, p = 0.074], but in the non-expected Statistical Analysis direction, with greater accuracies in the global condition As data was collected as repeated measures within subjects, compared to baseline. There was no significant main effect and considering that the variance in individuals and items is for Group [χ2(1) = 1.05, p = 0.305], but a trend of an interaction clustered, mixed effects models have been fitted in R (R Core [χ2(1) = 3.22, p = 0.073], indicating that global priming rather Team, 2019). These are a robust alternative to Analyses of improved accuracies in the TD group. Variance (Baayen et al., 2008; Jaeger, 2008; Singmann and Testing for H3, a local priming effect, we fitted models for Kellen, 2019). To account for the nested data structure, random data from baseline and local conditions only. We included intercepts have been modeled for variance clustered in subjects fixed effects for Group (ASD; TD), Condition (baseline; local), and stimuli. For accuracy (i.e., binary response variable) and their interaction. Results showed no main effects of Condition generalized linear mixed effects models have been fitted using [χ2(1) = 1.05, p = 0.305], Group [χ2(1) = 2.05, p = 0.152], and their the lme4 package in R (Bates et al., 2014). Model diagnostics interaction [χ2(1) = 1.61, p = 0.204]. have been visually inspected for all models. Likelihood ratio Given a trend to better performance in the global priming tests were calculated to test if a predictor in question significantly condition compared to the baseline condition, despite of the explained variance in the response variable against an alpha expected detrimental effect of priming toward a global processing level of 0.05. Important R packages that were used during style, we had to assume a training effect, which was analyzed mixed models analysis were afex (Singmann et al., 2015) and in the following. performance (Lüdecke et al., 2021). Training Effect We ran a model with accuracy predicted by Trialnumber, RESULTS Group, and their interaction and found a significant effect of Trialnumber [χ2(1) = 6.96, p = 0.008], indicating improvement Descriptive statistics can be found in Table 2. in performance over time. Group was non-significant [χ2(1) = 0.63, p = 0.427], but the interaction showed that there was a trend Baseline Condition indicating a larger training effect for the TD group [χ2(1) = 3.51, To test H1, investigating a group difference in the baseline p = 0.061]. Figure 2 depicts accuracies in percentage per trial condition, we found significantly superior accuracy of participants numbers in groups as well as priming condition orders. The positive slopes show an increase in accuracies over time, which www.neurobs.com/ 1 was especially pronounced in the TD group. The training effect Frontiers in Psychology | www.frontiersin.org 5 July 2022 | Volume 13 | Article 857630 Falter-Wagner et al. Figure-Disembedding in Autistic and Non-autistic Individuals FIGURE 2 | Average accuracies per group (ASD left panel; TD right panel) for each trial (1–18) depending on priming condition order in the Embedded Figure Task (EFT). One point depicts the percentage of accurate responses per trial number in groups (ASD left panel; TD right panel) and condition orders (global first in blue; local first in red). Linear regression lines with confidence bands fitted per group and condition order combination. Black dashed lines illustrate the block limits: Trials 1–6 were baseline trials, which were followed by two blocks each with 6 trials that included global or local priming in a counterbalanced order. pooled for both priming condition orders is best visualized did group [χ2(1) = 0.06, p = 0.812], and the interaction of BDI in Figure 3. and Group [χ2(1) = 0.11, p = 0.735]. Covariates We tested for potential effects of FIQ, AQ, BDI, and age on DISCUSSION performance. We found a significant effect of FIQ [χ2(1) = 15.96, p < 0.001], and a significant interaction effect between FIQ and The support for H1, namely superior accuracy in participants Group [χ2(1) = 4.52, p = 0.033] showing a stronger increase in EFT with ASD in the unprimed baseline condition corroborates previous accuracies with increasing FIQ in the TD group. In order to findings (e.g., Shah and Frith, 1983; De Jonge et al., 2006; Pellicano analyze whether the training effect might be affected by the et al., 2006; Brosnan et al., 2012; Muth et al., 2014). However, influence of FIQ we have run a full model including FIQ, other studies also show equal performance (e.g., Rodgers, 2000; Trialnumber, as well as Group as fixed factors with their interaction Bölte et al., 2007; Kaland et al., 2007; Chen et al., 2008; White terms and found that the two-way interaction of FIQ and and Saldaña, 2011) as summarized in a recent systematic review Trialnumber was non-significant [χ2(1) = 0.30, p = 0.584] as well (Horlin et al., 2016). Considering this discrepancy in findings, as the three-way interaction with Group * FIQ * Trialnumber the results of the current study that control participants’ performance [χ2(1) = 0.31, p = 0.579]. The main effects for Trialnumber can be improved by training is highly relevant and any findings [χ2(1) = 7.51, p = 0.006] and FIQ [χ2(1) = 16.23, p < 0.001] remained of superior or equal performance in figure-disembedding needs significant, as well as the interaction terms Trialnumber * Group to be evaluated on the basis of potential differential impact of [χ2(1) = 4.01, p = 0.045] and FIQ * Group [χ2(1) = 4.82, p = 0.028]. training on group averages. Additionally, in a model containing Age, Group, and their Against our expectations (H2), we did not find a deterioration interaction, we found a main effect of Age [χ2(1) = 5.60, p = 0.018] of performance after global or local priming in both groups and a trend of Group [χ2(1) = 3.66, p = 0.056]. A non-significant (H3). A performance improvement trend in the TD group interaction term showed that the effect of age was similar across irrespective of priming points to a training effect, which was both groups [χ2(1) = 0.04, p = 0.840]. The effect of age did not indeed confirmed in a multilevel model analysis. This was not cause the superior performance of the ASD group in the baseline visible in the RT data, possibly due to the fact that our condition, but if anything would have made the effect more instruction to participants was “as quickly but as accurately conservative (see Figure 4). In a model containing BDI, Group, as possible” is highlighting the importance of accuracy.” Notably, and their interaction, we found that BDI did not explain variance neither a performance improvement nor decrease was observable in accuracies above chance level [χ2(1) = 0.07, p = 0.795], neither in the ASD group, which renders the ASD group immune Frontiers in Psychology | www.frontiersin.org 6 July 2022 | Volume 13 | Article 857630 Falter-Wagner et al. Figure-Disembedding in Autistic and Non-autistic Individuals FIGURE 3 | Average accuracies per group (ASD left panel; TD right panel) for each trial (1–18) in the Embedded Figure Task (EFT). One point depicts the percentage of accurate responses in the Embedded Figures Task per trial number in groups (ASD left panel; TD right panel). Linear regression lines with confidence bands fitted per group. Black dashed lines illustrate the block limits: Trials 1–6 were baseline trials, which were followed by two blocks each with 6 trials that included global or local priming in a counterbalanced order. FIGURE 4 | Correlations of age and average accuracies in Embedded Figure Tasks (EFT) in ASD group (blue) and TD group (red). One point represents the age and mean accuracy over all Embedded Figure Task (EFT) trials of one participant. Linear regression lines with confidence bands depict negative slopes in both groups. Frontiers in Psychology | www.frontiersin.org 7 July 2022 | Volume 13 | Article 857630 Falter-Wagner et al. Figure-Disembedding in Autistic and Non-autistic Individuals against either priming or training effects in this task. It is had to search at the local or global level to respond. For example, possible that either ASD performance is very stable and robust Tan and colleagues (Tan et al., 2017) did not use target letters, or that ASD individuals perform at ceiling und we therefore but merely instructed participants to attend to either the large cannot observe a training effect. In the case of the first possibility, or the small letters (see also Macrae and Lewis, 2002; Förster this would be in concert with the literature showing that ASD and Higgins, 2005). In yet another version (Gao et al., 2011), individuals tend to take context less into account (i.e., WCC two Navon stimuli were presented to participants simultaneously or EPF theories mentioned in the introduction). in the left and right visual field, respectively, and participants Our results are in accordance with previous studies that were instructed to match the two displays, focusing either on did not find any priming effect in ASD employing semantic the large letters, ignoring the small letters (global level), or on priming (Kamio et al., 2007) but contrast others that did show the small letters, ignoring the large letters (local level). Future semantic priming effects (Kamio and Toichi, 2000). In addition, research should attempt to replicate our findings using other the type of priming used in the current study, has been used procedural priming tasks, in particular Navon variations. in previous research (Förster et al., 2008) and was effectively In conclusion, figure-disembedding seems to be a function a sequential priming task, for which it has been shown that of visuo-spatial cognition that shows solid high performance target-level information (i.e., global or local priming in hierarchical in individuals with ASD without modifiability through priming stimuli) mediate priming effects in subsequent trials (Kim et al., or training. In contrast, performance in control participants 1999). Thus, not just the priming type but also the priming was improved to the level observed in ASD by training. These procedure used in the current study was markedly different results imply that any findings of superior performance in from the semantic priming tasks described above. Yet, our figure-disembedding tasks in ASD should take the potentially results are also in line with previous experiments showing a differential influence of training into account and carefully lack of configural priming in individuals with ASD (Behrmann examine whether the initially found superiority effect holds et al., 2006) with a sequential and non-semantic approach. throughout the experimental trials performed or might lead We found a significant influence of IQ on performance and to equal performance between groups. improvement by training in the TD group only. Arguably, the higher the cognitive abilities the faster improvements by learning occur. Interestingly though performance in the ASD group DATA AVAILABILITY STATEMENT was stable throughout. A further significant influence was found by age with decreasing performance the older the participants The raw data supporting the conclusions of this article will were. The effect of age could not have influence our findings be made available by the authors, without undue reservation. of superior baseline figure-disembedding in ASD, because our ASD group was somewhat older than the TD control group. Thus, our results could if anything be regarded as conservative ETHICS STATEMENT and potentially decreasing the group difference. The findings of this study have to be interpreted in light The studies involving human participants were reviewed and of some limitations. First, the number of trials per condition approved by Ethics Committee of the Medical Faculty of the in the figure-disembedding task was limited. However, increasing University of Cologne. The patients/participants provided their the trial number would potentially cause fatigue in participants. written informed consent to participate in this study. Future research should attempt to replicate the findings with larger trial numbers and possibly a between-subject design regarding the priming factor. Second, we provided feedback on AUTHOR CONTRIBUTIONS the disembedding task, which might have contributed to the training effect. However, the feedback was unique to each trial CF and AG conceptualization; LH data curation; CB, CF, AG, and not strategy based, therefore we do not believe that knowing LH and MR formal analysis; KV funding acquisition; CF, AG, whether or not one was correct or incorrect, would have affected LH investigation; CF and AG methodology; CF, AG, and KV their performance. At most, it could have (de)motivated project administration; KV resources; CF, AG, and KV participants making errors over time. Third, we did not perform supervision; CF, LH, and MR original draft; CF, AG, MR, CB a sample size estimation, due to the fact that there were no and KV review & editing. All authors contributed to the article previous effects to base this on and it is possible that we had and approved the submitted version. too little power to detect priming effects in this study. Future research should attempt to replicate the procedure with larger samples. Further, the study sample were adults with ASD without FUNDING intellectual disabilities limiting generalizability within the spectrum given that up to 50% of autistic disorders have an intellectual CB, MR, and CF were supported by the DFG (German Research disability (Charman et al., 2011). Finally, the use of NF for Council; grant numbers: FA 876/3–1 and 876/5–1). KV was procedural priming can be implemented in different ways. In supported by the DFG (CRC 1252) and by the EC, Horizon the current study, we used a previously used approach (Förster 2020 Framework Programme, FET Proactive (Project et al., 2008). In other versions of priming with NF participants VIRTUALTIMES; Grant agreement ID: 824128). Frontiers in Psychology | www.frontiersin.org 8 July 2022 | Volume 13 | Article 857630 Falter-Wagner et al. Figure-Disembedding in Autistic and Non-autistic Individuals ACKNOWLEDGMENTS SUPPLEMENTARY MATERIAL The authors would like to thank Mathis Jording for help with The Supplementary Material for this article can be found online programming the task and all participants, without whom this at: https://www.frontiersin.org/articles/10.3389/fpsyg.2022.857630/ study would not have been possible. full#supplementary-material REFERENCES Falter, C. M., Arroyo, M., and Davis, G. J. (2006). Testosterone: activation or organization of spatial cognition? Biol. Psychol. 73, 132–140. doi: 10.1016/j. Baayen, R. H., Davidson, D. J., and Bates, D. M. (2008). Mixed-effects modeling biopsycho.2006.01.011 with crossed random effects for subjects and items. J. Mem. Lang. 59, Falter, C. M., and Bailey, A. J. (2011). Perception of mirror symmetry in Autism 390–412. doi: 10.1016/j.jml.2007.12.005 Spectrum Disorders. Autism 16, 622–626. doi: 10.1177/1362361311407353 Bargh, J. A., and Chartrand, T. L. (2000). “The mind in the middle: a practical Falter, C. M., Braeutigam, S., Nathan, R., Carrington, S., and Bailey, A. J. guide to priming and automaticity research,” in Handbook of Research Methods (2013). Enhanced access to early visual processing of perceptual simultaneity in Social and Personality Psychology. eds. H. Reis and C. Judd (New York, NY, in autism spectrum disorders. J. Autism Dev. Disord. 43, 1857–1866. doi: US: Cambridge University Press), 253–285. 10.1007/s10803-012-1735-1 Baron-Cohen, S., Richler, J., Bisarya, D., Gurunathan, N., and Wheelwright, S. Falter, C. M., Plaisted, K. C., and Davis, G. (2008). Visuo-spatial processing (2003). The systemizing quotient: an investigation of adults with Asperger in autism - testing the predictions of extreme male brain theory. J. Autism syndrome or high-functioning autism, and normal sex differences. Philosoph. Dev. Disord. 38, 507–515. doi: 10.1007/s10803-007-0419-8 Trans. Royal Soc. London B Biol. Sci. 358, 361–374. doi: 10.1098/rstb. Förster, J., and Higgins, E. T. (2005). How global versus local perception fits 2002.1206 regulatory focus. Psychol. Sci. 16, 631–636. doi: 10.1111/j.1467-9280.2005.01586.x Baron-Cohen, S., and Wheelwright, S. (2004). The empathy quotient: an Förster, J., Liberman, N., and Kuschel, S. (2008). The effect of global versus investigation of adults with Asperger syndrome or high functioning autism, local processing styles on assimilation versus contrast in social judgment. and Normal sex differences. J. Autism Dev. Disord. 34, 163–175. doi: 10.1023/B: J. Pers. Soc. Psychol. 94, 579–599. doi: 10.1037/0022-3514.94.4.579 JADD.0000022607.19833.00 Frith, U. (1989). Autism: Explaining the Enigma. 6th Edn. ed. B. Blackwell Baron-Cohen, S., Wheelwright, S., Skinner, R., Martin, J., and Clubley, E. (2001). (Oxford: Basil Blackwell). The autism-Spectrum quotient (AQ): evidence from Asperger syndrome/ Gao, Z., Flevaris, A. V., Robertson, L. C., and Bentin, S. (2011). Priming global high-functioning autism, Malesand females, scientists and mathematicians. and local processing of composite faces: revisiting the processing-bias effect J. Autism Dev. Disord. 31, 5–17. doi: 10.1023/A:1005653411471 on face perception. Atten. Percept. Psychophys. 73, 1477–1486. Bates, D., Mächler, M., Bolker, B., and Walker, S. (2014). Fitting linear mixed- Grahek, I., Shenhav, A., Musslick, S., Krebs, R. M., and Koster, E. H. (2019). effects models using lme4. arXiv 406.5823 [Preprint]. Motivation and cognitive control in depression. Neurosci. Biobehav. Rev. Beck, A. T., Steer, R. A., and Brown, G. K. (1996). Manual for the Beck 102, 371–381. doi: 10.1016/j.neubiorev.2019.04.011 Depression Inventory-II. San Antonio, TX: Psychological Corporation. Hadad, B.-S., and Ziv, Y. (2015). Strong Bias towards analytic perception in Behrmann, M., Avidan, G., Leonard, G. L., Kimchi, R., Luna, B., Humphreys, K., ASD does not necessarily come at the Price of impaired integration skills. et al. (2006). Configural processing in autism and its relationship to face J. Autism Dev. Disord. 45, 1499–1512. doi: 10.1007/s10803-014-2293-5 processing. Neuropsychologia 44, 110–129. doi: 10.1016/J.NEURO Hala, S., Pexman, P. M., and Glenwright, M. (2007). Priming the meaning of PSYCHOLOGIA.2005.04.002 homographs in typically developing children and children with autism. J. Bölte, S., Holtmann, M., Poustka, F., Scheurich, A., and Schmidt, L. (2007). Autism Dev. Disord. 37, 329–340. doi: 10.1007/s10803-006-0162-6 Gestalt perception and local-global processing in high-functioning autism. Happé, F., and Frith, U. (2006). The weak coherence account: detail-focused J. Autism Dev. Disord. 37, 1493–1504. doi: 10.1007/s10803-006-0231-x cognitive style in autism spectrum disorders. J. Autism. Dev. Disord. 36, Booth, R. D. L., and Happé, F. G. E. (2018). Evidence of reduced global 5–25. doi: 10.1007/s10803-005-0039-0 processing in autism Spectrum disorder. J. Autism Dev. Disord. 48, 1397–1408. Hollocks, M. J., Lerh, J. W., Magiati, I., Meiser-Stedman, R., and Brugha, T. S. doi: 10.1007/s10803-016-2724-6 (2019). Anxiety and depression in adults with autism spectrum disorder: Brosnan, M. J., Gwilliam, L. R., and Walker, I. (2012). Brief report: the relationship a systematic review and meta-analysis. Psychol. Med. 49, 559–572. doi: between visual acuity, the embedded figures test and systemizing in autism 10.1017/S0033291718002283 spectrum disorders. J. Autism Dev. Disord. 42, 2491–2497. doi: 10.1007/ Horlin, C., Black, M., Falkmer, M., and Falkmer, T. (2016). Proficiency of s10803-012-1505-0 individuals with autism spectrum disorder at disembedding figures: a systematic Brosnan, M. J., Scott, F. J., Fox, S., and Pye, J. (2004). Gestalt processing in review. Dev. Neurorehabil. 19, 54–63. doi: 10.3109/17518423.2014.888102 autism: failure to process perceptual relationships and the implications for Jaeger, T. F. (2008). Categorical data analysis: away from ANOVAs (transformation contextual understanding. J. Child Psychol. Psychiatry 45, 459–469. doi: or not) and towards logit mixed models. J. Mem. Lang. 59, 434–446. doi: 10.1111/j.1469-7610.2004.00237.x 10.1016/j.jml.2007.11.007 Charman, T., Pickles, A., Simonoff, E., Chandler, S., Loucas, T., and Baird, G. Kaland, N., Mortensen, E. L., and Smith, L. (2007). Disembedding performance (2011). IQ in children with autism spectrum disorders: data from the special in children and adolescents with Asperger syndrome or high-functioning needs and autism project (SNAP). Psychol. Med. 41, 619–627. doi: 10.1017/ autism. Autism 11, 81–92. doi: 10.1177/1362361307070988 S0033291710000991 Kamio, Y., Robins, D., Kelley, E., Swainson, B., and Fein, D. (2007). Atypical Chen, F., Lemonnier, E., Lazartigues, A., and Planche, P. (2008). “Non-superior lexical/semantic processing in high-functioning autism Spectrum disorders disembedding performance in children with high-functioning autism and without early language delay. J. Autism Dev. Disord. 37, 1116–1122. doi: its cognitive style account,” in Research in Autism Spectrum Disorders. ed. 10.1007/s10803-006-0254-3 F. Chen (Lausanne, Switzerland: Elsevier Science). Kamio, Y., and Toichi, M. (2000). Dual access to semantics in autism: is pictorial Dakin, S., and Frith, U. (2005). Vagaries of Visual Perception in Autism. Neuron access superior to verbal access? J. Child Psychol. Psychiatry 41, 859–867. 48, 497–507. doi: 10.1016/j.neuron.2005.10.018 doi: 10.1111/1469-7610.00673 De Jonge, M. V., Kemner, C., and van Engeland, H. (2006). Superior Disembedding Kim, N., Ivry, R. B., and Robertson, L. C. (1999). Sequential priming in performance of high-functioning individuals with autism Spectrum disorders hierarchically organized figures: effects of target level and target resolution. and their parents: the need for subtle measures. J. Autism Dev. Disord. 36, J. Exp. Psychol. Hum. Percept. Perform. 25, 715–729. 677–683. doi: 10.1007/s10803-006-0113-2 Koldewyn, K., Jiang, Y. V., Weigelt, S., and Kanwisher, N. (2013). Global/local Falter, C. M. (2012). Impaired and superior mirror symmetry perception in processing in autism: not a disability, but a disinclination. J. Autism Dev. autism. Autism 17, 117–118. doi: 10.1177/1362361312438618 Disord. 43, 2329–2340. doi: 10.1007/s10803-013-1777-z Frontiers in Psychology | www.frontiersin.org 9 July 2022 | Volume 13 | Article 857630 Falter-Wagner et al. Figure-Disembedding in Autistic and Non-autistic Individuals Kühnen, U., Hannover, B., and Schubert, B. (2001). The semantic–procedural in a selective attention task. J. Child Psychol. Psychiatry, 40, 733–742. doi: interface model of the self: The role of self-knowledge for context-dependent 10.1111/1469-7610.00489 versus context-independent modes of thinking. J. Pers. Soc. Psychol. 80:397. R Core Team (2019). R: A Language and Environment for Statistical Computing. Kühnen, U., and Oyserman, D. (2002). Thinking about the self influences R Foundation for Statistical Computing, Vienna, Austria. thinking in general: Cognitive consequences of salient self-concept. J. Exp. Rodgers, J. (2000). Visual perception and Asperger syndrome: central coherence Soc. Psychol. 38, 492–499. deficit or Hierarchization deficit?: a pilot study. Autism 4, 321–329. doi: Lin, Z., Lin, Y., and Han, S. (2008). Self-construal priming modulates visual 10.1177/1362361300004003007 activity underlying global/local perception. Biol. Psychol. 77, 93–97. doi: Scherf, K. S., Luna, B., Kimchi, R., Minshew, N., and Behrmann, M. (2008). 10.1016/j.biopsycho.2007.08.002 Missing the big picture: impaired development of global shape processing López, B., and Leekam, S. R. (2003). Do children with autism fail to process in autism. Autism Res. 1, 114–129. doi: 10.1002/AUR.17 information in context? J. Child Psychol. Psychiatry 44, 285–300. doi: Shah, A., and Frith, U. (1983). An islet of ability in autistic children: a research 10.1111/1469-7610.00121 note. J. Child Psychol. Psychiatry 24, 613–620. doi: 10.1111/j.1469-7610.1983. Lüdecke, D., Ben-Shachar, M. S., Patil, I., Waggoner, P., and Makowski, D. tb00137.x (2021). Performance: an R package for assessment, comparison and testing Shah, A., and Frith, U. (1993). Why do autistic individuals show superior of statistical models. J. Open Source Softw. 6:3139. doi: 10.21105/joss.03139 performance on the block design task? J. Child Psychol. Psychiatry 34, Macrae, C. N., and Lewis, H. L. (2002). Do I know you? Processing orientation 1351–1364. doi: 10.1111/j.1469-7610.1993.tb02095.x and face recognition. Psychol. Sci. 13, 194–196. doi: 10.1111/1467-9280.00436 Singmann, H., Bolker, B., Westfall, J., Aust, F., and Ben-Shachar, M. S. (2015). Mottron, L., Belleville, S., and Ménard, E. (1999). Local bias in autistic subjects Afex: Analysis of Factorial Experiments. R Package Version 0.13–145. as evidenced by graphic tasks: perceptual hierarchization or working memory Singmann, H., and Kellen, D. (2019). “An introduction to mixed models for deficit? J. Child Psychol. Psychiatry 40, 743–755. doi: 10.1111/1469-7610.00490 experimental psychology.” in New Methods in Cognitive Psychology. eds. Mottron, L., and Burack, J. A. (2001). “Enhanced perceptual functioning in the D. H. Spieler and F. Schumacher (New York, NY: Routledge), 4–31. development of autism,” in The Development of Autism: Perspectives From Theory Tan, E. W. S., Lim, S. W. H., and Manalo, E. (2017). Global-local processing and Research. eds. J. A. Burack, T. Charman, N. Yirmiya and P. R. Zelazo impacts academic risk taking. Q. J. Exp. Psychol. 70, 2434–2444. doi: (Mahwah, NJ: Erlbaum), 131–148. 10.1080/17470218.2016.1240815 Mottron, L., Burack, J. A., Iarocci, G., Belleville, S., and Enns, J. T. (2003). Toichi, M., and Kamio, Y. (2001). Verbal association for simple common words Locally oriented perception with intact global processing among adolescents in high-functioning autism. J. Autism Dev. Disord. 31, 483–490. doi: with high-functioning autism: evidence from multiple paradigms. J. Child 10.1023/A:1012216925216 Psychol. Psychiatry Allied Discip. 44, 904–913. doi: 10.1111/1469-7610.00174 Wang, L., Mottron, L., Peng, D., Berthiaume, C., and Dawson, M. (2007). Mottron, L., Dawson, M., Soulières, I., Hubert, B., and Burack, J. (2006). Local bias and local-to-global interference without global deficit: a robust Enhanced perceptual functioning in autism: an update, and eight principles finding in autism under various conditions of attention, exposure time, and of autistic perception. J. Autism Dev. Disord. 36, 27–43. doi: 10.1007/ visual angle. Cogn. Neuropsychol. 24, 550–574. doi: 10.1080/13546800701417096 S10803-005-0040-7 Wechsler, D. (1981). The psychometric tradition: developing the wechsler adult Muth, A., Hönekopp, J., and Falter, C. M. (2014). Visuo-spatial performance intelligence scale. Contemp. Educ. Psychol. 6, 82–85. doi: 10.1016/0361- in autism: a Meta-analysis. J. Autism Dev. Disord. 44, 3245–3263. doi: 10.1007/ 476X(81)90035-7 s10803-014-2188-5 White, S. J., and Saldaña, D. (2011). Performance of children with autism on Navon, D. (1977). Forest before trees: The precedence of global features in the embedded figures test: a closer look at a popular task. J. Autism Dev. visual perception. Cogn. Psychol. 9, 353–383. doi: 10.1016/0010-0285(77)90012-3 Disord. 41, 1565–1572. doi: 10.1007/s10803-011-1182-4 Oyserman, D., and Lee, S. W. (2008). Does culture influence what and how Witkin, H. A., Oltman, P. K., Raskin, E., and Karp, S. A. (1971). A Manual we think? Effects of priming individualism and collectivism. Psychol. Bull. for the Group Embedded Figures Test. Palo Alto, CA: Consulting 134:311. Psychologists Press. Oyserman, D., Sorensen, N., Reber, R., and Chen, S. (2009). Connecting and World Health Organization (WHO) (1993). The ICD-10 Classification of mental separating mind-sets: culture as situated cognition. J. Pers. Soc. Psychol. and Behavioural Disorders. World Health Organization. 97:217. Pellicano, E., Maybery, M., Durkin, K., and Maley, A. (2006). Multiple Conflict of Interest: The authors declare that the research was conducted in cognitive capabilities/deficits in children with an autism spectrum the absence of any commercial or financial relationships that could be construed disorder: “weak” central coherence and its relationship to theory of mind as a potential conflict of interest. and executive control. Dev. Psychopathol. 18, 77–98. doi: 10.1017/ S0954579406060056 Publisher’s Note: All claims expressed in this article are solely those of the Petermann, F. (Hrsg.) (2012). Wechsler Adult Intelligence Scale - Fourth Edition. authors and do not necessarily represent those of their affiliated organizations, Deutschsprachige Adaptation der WAIS-IV von D. Wechsler. Frankfurt/M.: or those of the publisher, the editors and the reviewers. Any product that may Pearson Assessment & Information GmbH. be evaluated in this article, or claim that may be made by its manufacturer, is Plaisted, K. C. (2001). “Reduced generalization in autism: an alternative to not guaranteed or endorsed by the publisher. weak central coherence,” in The Development of Autism: Perspectives From Theory and Research. eds. J. A. Burack, T. Charman, N. Yirmiya and P. R. Copyright © 2022 Falter-Wagner, Bloch, Robles, Horch, Vogeley and Georgescu. Zelazo (Mahwah, NJ: Erlbaum), 149–169. This is an open-access article distributed under the terms of the Creative Commons Plaisted, K. C., O’Riordan, M., and Baron-Cohen, S. (1998). Enhanced visual Attribution License (CC BY). The use, distribution or reproduction in other forums search for a conjunctive target in autism: a research note. J. Child Psychol. is permitted, provided the original author(s) and the copyright owner(s) are credited Psychiatry 39, 777–783. doi: 10.1111/1469-7610.00376 and that the original publication in this journal is cited, in accordance with accepted Plaisted, K. C., Swettenham, J., and Rees, L. (1999). Children with autism academic practice. No use, distribution or reproduction is permitted which does not show local precedence in a divided attention task and global precedence comply with these terms. Frontiers in Psychology | www.frontiersin.org 10 July 2022 | Volume 13 | Article 857630