The world is perceived as being hierarchically organized and includes global perceptions comprising local details (D’Souza et al., 2016); a human being is able to process information at both global and local levels. This capacity, which is denominated central coherence, is implied in everyday activities such as categorization, inspection of details in our surroundings, perception of the structure of a scene and analysis of information (Nayar et al., 2017). Central coherence is, therefore, the normal tendency to process and give meaning, in a global manner, to incoming information in its context (Noens and van Berckelaer-Onnes, 2008). It is thanks to central coherence that when we receive a message, we prioritize comprehension of meaning and not just the literal form (Crespo-Eguílaz et al., 2012).

Attwood (Attwood, 2007) provides a metaphor that may be useful to understand the nature of weak central coherence: “imagine rolling up a sheet of paper to form a tube and with one eye closed bring it up to other open eye, as if it were a telescope, and look at the world through it: you see the details but do not perceive the context.” A person with difficulty in central coherence (that is, a preference for analytical processing rather than global processing) has specific difficulties in simultaneously processing information perceived and in giving coherent, integral meaning to that information. The cognitive style of such a person is, therefore, characterized by a tendency to process details (Navon, 1977; Schooler, 2002; Förster and Dannenberg, 2010; Crespo-Eguílaz and Narbona, 2011) This style of processing is slower and more demanding from the cognitive point of view (Nayar et al., 2017). An ability to carry out tasks requiring central coherence rapidly is of fundamental importance in learning and in social behavior (Crespo-Eguílaz et al., 2012). Therefore, weakness in central coherence entails difficulties in contextual comprehension of social situations and in adaption to these situations (López and Leekam, 2007). Difficulties in central coherence are also described by (Lamb and Robertson, 1989) as local bias and by (Vermeulen, 2015), in his studies on people with autism spectrum disorders (ASD), as context blindness. To date, dysfunction in central coherence has principally been studied in ASD (López and Leekam, 2007), although it has also been described in other disorders, such as, Down’s syndrome, Williams syndrome (D’Souza et al., 2016) non-verbal learning disorder (NVLD) (Gillberg, 2003, 2009; Crespo-Eguílaz and Narbona, 2011); and thus weak central coherence is apparently not specific to ASD.

There were no significant differences in performance between medicated and non-medicated participants. This was the case for all dimensions of the CCT and for all tasks. The finding applies to the two clinical groups in which there were patients under pharmacological treatment: the NVLD+ADHD group and the ADHD group (Table 4).

There were no significant differences between clinical groups or the control group in performance in the Verbal detail dimension.

With regard to the Simultaneity dimension, the performance levels of the NVLD+ADHD group and the SCD group were significantly poorer than that of the control group (Figure 1); the effect size is high (Table 5). Performance levels of the ADHD and ASD1 groups were typical (Figure 1). There were significant differences in this respect between the NVLD+ADHD group and the ADHD and ASD1 groups (Figure 1; Table 5).

The performance of NVLD+ADHD and ASD1 groups in the Inference dimension was significantly lower than that of the control group, while the performance of the ADHD group was typical. The effect size of the differences was high (Table 5). The performance of the SCD group was also apparently lower, but this difference was not statistically significant (Figure 1).

In the Context dimension, the mean performance levels of the NVLD+ADHD group and the ASD1 group were significantly lower than that of the control group (Figure 1 and Table 5). Mean performance of subjects in ADHD and SCD groups was adequate. The effect size was high for all comparisons.

Finally, for the CCT as a whole, mean performance levels of the NVLD+ADHD group and of the SCD group were significantly lower (with large effect size) than that of the control group, while mean performance levels of the ADHD and ASD1 groups were average (Figure 1 and Table 5). Overall mean performance of the NVLD+ADHD group was significantly lower than that of the ADHD group (Table 5).

In total CCT score and in all dimensions except Verbal detail, the NVLD+ADHD group performed significantly worse than the control group. The performance of the ADHD group, however, was similar to that of the control group in all dimensions. Furthermore, the NVLD+ADHD and ADHD groups differed in overall CCT (Table 6). As the participants in both groups have attention deficit, we infer that the clear difficulties in central coherence in subjects with NVLD+ADHD is not to be explained by attention deficit but rather as being characteristic of NVLD.

NVLD+ADHD and ASD1 groups had significantly lower mean performance than the control group in the Inference and Context dimensions (Table 6). The two groups differ in mean performance in Simultaneity: children with NVLD+ADHD performing worse than those with ASD1.

CCT performance profiles for NVLD+ADHD and SCD groups were similar. In both groups, mean performance in Simultaneity was significantly lower than that of the control group. However, the performance deficit of the NVLD+ADHD group was more serious than that of the SCD group because the symptomatology was more pronounced: the mean performance of the NVLD+ADHD group was significantly lower than that of the control group in most dimensions of the CCT (Table 6).

The results of our study indicate that deficit in central coherence is not an effect exclusive to ASD. While various studies have found a reduction in global processing in ASD (Rinehart et al., 2000; Pellicano et al., 2005; Seernani et al., 2020) or that people with ASD perform significantly better in tests of local preference than people with typical development (Shah and Frith, 1983; Plaisted et al., 1998; O’Riordan et al., 2001; Pellicano et al., 2006; Gambra, 2020), there are also numerous studies that contradict such findings (Brian and Bryson, 1996) and that demonstrate that ability for global processing in people with ASD is intact (Bertone et al., 2003; Wang et al., 2004; Johnson et al., 2010).

Mottron and Belleville (Mottron and Belleville, 1993) demonstrate that people with autism process information, at both the local and the global level, as well as control subjects, but that in ASD the local interferes with the global when stimuli are incongruous. Subsequent studies confirm these findings (Jolliffe and Baron-Cohen, 1997; Rinehart et al., 2001). These somewhat contradictory findings have also been obtained in studies using more than one evaluation tool (Edgin and Pennington, 2005). Syriopoulou Delli et al. (2016) suggest that disparity in results might be best explained by considering that the style of information processing in autism is personal rather than a distinct characteristic of ASD.

In our research, the general performance of the ASD1 group was lower than that of the control group, but this difference was not always significant (depending on the dimension of the CCT); this is in agreement with the results of Syriopoulou Delli et al. (2016) who found that in a global preference task the scores of children with typical development were apparently higher than the scores of children with ASD, but the difference was not statistically significant. With regard to central coherence abilities, schoolchildren with ASD1 had average performance in simultaneity tasks (which were principally visual tasks) and average-low performance in tasks involving inference, verbal detail and understanding in context. Relative to the control group, differences in performance at tasks involving inference and context were significant. (The same differences were found in the NVLD+ADHD group, but children with ASD1 performed better than children with NVLD+ADHD in simultaneity tasks.)

In contrast to our findings, Loth (Loth, 2003) concluded that only about 35% of children with autism show weak central coherence in different tasks, while 48% have mixed styles of processing, with good performance in conceptual tasks. In our study, however, 90% of the schoolchildren in the ASD1 group had difficulty with one or more abilities related to central coherence. The difference in results can be explained in various ways. First, ASD is by nature heterogeneous and symptoms vary according to the level of ASD (Santangelo and Folstein, 1999; Tager-Flusberg and Joseph, 2003); in our study, all participants with ASD were level 1. Second, within ASD there can be subgroups defined by performance in terms of central coherence (Tager-Flusberg and Joseph, 2003). Third, as already discussed, the different tasks used by different studies make different demands on subjects and therefore give different measures of performance. Finally, Mottron et al. (Mottron et al., 2006) suggest that autism is often associated with improved perceptual processing, but this is not evident in all children with ASD.

Our study found that children with NVLD had difficulties with central coherence, confirming that this difficulty is a characteristic in the cognitive profile of NVLD. This finding is in line with different studies (Crespo-Eguílaz and Narbona, 2009; Magallón, 2011). Other authors (Chow and Skuy, 1999; Schoemaker et al., 2001; Molenaar-Klumper, 2002; Viñuela, 2007; Mammarella and Pazzaglia, 2010; Mammarella et al., 2019) do not refer explicitly to the construct of central coherence but nonetheless affirm that children with NVLD have difficulty with global perception and with analysing, organizing and synthesizing information.

General performance of the NVLD+ADHD group in central coherence tasks was lower than that of the control group (p < 0.001). However, performance was average for abilities related to verbal details. That is, for children with NVLD+ADHD, performance in tasks that involve verbal aspects of central coherence seems to stand out against the background of their overall performance. This concurs with the more general observation that the verbal abilities of children with NVLD+ADHD are better than their manipulative abilities. In our study, performance was average-low for abilities to understand in a simultaneous manner, to make inferences and to understand in a context (all of which were statistically significant differences relative to the control group). These results are coherent with those of Crespo-Eguílaz et al. (2012), who reported that rapid interpretation of a chimerical image - to make sense of which it was necessary to perceive and integrate both parts of the chimera - posed difficulty for 85% of schoolchildren with NVLD+ADHD but for only 5% of controls. Magallón (2011) found that children with NVLD+ADHD had difficulty in noticing incongruities in chimerical images (62.1% carried out this task badly in comparison to 8% of controls) and in visual scenes (about 60% performed poorly at this task). These results are along the same lines (Díaz Lucero et al., 2011), who evaluated the neuropsychological profile of a group of 22 children with DAMP and concluded that they did not understand globally but rather by parts, which made it difficult for them to integrate concepts, to carry out abstraction and, therefore, to adapt correctly to the context. Another research group, (Drumond et al., 2005), also found that children with NVLD had difficulty in tasks that involved construction of a whole from parts. Semrud-Clikeman et al. (2010) reported findings similar to ours for a group of children with NVLD (with difficulties in central coherence) relative to other groups of children with level-1 ASD and ADHD (without difficulties in central coherence).

In our study, the children with NVLD also had ADHD. However, as in the study of Makris et al. (2021), the difficulties in central coherence that we have discussed were not observed with a separate ADHD group. Therefore, these difficulties cannot be explained by attention deficit and appear to be characteristic of NVLD. In contrast to these results, (Cardillo, 2018) found that the central coherence profile of a group of children with ADHD was heterogeneous: the children had difficulty with visuoconstruction abilities when they had to battle with global configurations but performed visual-perception tasks correctly. In this study, the SCD group was characterized by average-low performance for all central coherence abilities studied. Thus, as a group, schoolchildren with a deficit in social communication skills have below-average performance, which indicates that their neuropsychological profile is also characterized by weak central coherence.

The study’s limitations include that all children in the NVLD sample have ADHD. It would have been preferable to identify another pure NVLD group, although this is challenging due to the frequent presence of comorbid disorders. Additionally, the sample size of the ASD group could be larger to draw definitive conclusions. Environmental or socio-economic factors not accounted for in the study could influence central coherence abilities. These factors might limit the generalizability of findings to different populations.

The current study, through the profiles of central coherence for the clinical groups described, is consistent with a deficit in central coherence is not exclusive to autism spectrum disorders. This study evidence that that children with other neurodevelopmental and learning disorders, such as, non-verbal learning disorder and social communication disorder, experience difficulties related with this cognitive function. In addition, it was found that schoolchildren with ADHD did not have difficulty with central coherence. Finally, we establish that the Central Coherence Test provides complementary information that is useful for differential diagnosis between neurodevelopmental disorders involving weak central coherence.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving humans were approved by the Ethics committee of the University of Navarra (Ethics approval number: 2017.004mod1). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin. Written informed consent was obtained from the minor(s)’ legal guardian/next of kin for the publication of any potentially identifiable images or data included in this article.

LG: Writing–original draft, Conceptualization, Methodology, Writing–review and editing. SM: Formal analysis, Writing–original draft, Writing–review and editing. NC-E: Formal analysis, Writing–review and editing.