The Network Model of Mentalization, Social Vulnerability, and the Self in Autism: A Comparison With Neurotypical Adults.
Szilárd Holka, Dániel Sörnyei, Ágota Vass and 2 others
The psychological vulnerability network in autistic adults is less tightly connected than in neurotypical adults.
Anxiety showed a trend toward greater centrality in the autistic network, and mentalization was central in both. Cross-sectional, self-report data. Patterns, not causes.
Key findings
1The autistic group's network had significantly lower global strength than the neurotypical group's (Δ = 0.773, p = 0.003), meaning the psychological constructs—mentalization, attachment, anxiety, self-concept, social support—were less tightly interconnected in autistic adults. The overall network structure, however, did not differ significantly between groups (p = 0.469).
2Autistic traits (AQ) were significantly more central in the neurotypical network than in the autistic network (strength: Δ = −0.459, p = 0.027; expected influence: Δ = −0.431, p = 0.047). Trait anxiety showed a trend toward greater expected influence in the autistic network (Δ = 0.265, p = 0.096) but did not reach significance. Mentalization was the most central node in both networks.
3The autistic group scored significantly higher than the matched neurotypical group on mentalization difficulties (d = 0.83), insecure attachment, trait anxiety, psychological inflexibility, and self-related measures, and lower on perceived social support (d = −0.48), with medium to large effect sizes across the board.
Still to come
How it was doneWhat they found
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
Cross-sectional design: the associations describe how variables co-occur, not which causes which. The authors explicitly state no causal inferences can be drawn. All measures were self-report. The authors note that individuals with greater mentalizing difficulties may overestimate their own capacities, and that perceived social support is shaped by subjective interpretation and social desirability. Convenience sampling: participants were recruited through media, social media, and outpatient units. The authors acknowledge that individuals with higher psychological vulnerability may have been less likely to complete the survey, potentially limiting the range of symptom severity represented. 48.4% of the autistic group (75 of 156) had at least one additional psychiatric diagnosis (most commonly mood disorder or anxiety), so the 'autistic' network reflects a clinically complex population rather than a pure autism signal. The schizophrenia group (n = 62) was excluded from all main analyses because its network failed stability criteria, so no three-way comparison is possible. The anxiety centrality finding was not a priori hypothesized and did not reach significance (p = 0.096); the authors label it exploratory and hypothesis-generating. Networks were estimated at the between-person level, reflecting covariation across individuals rather than dynamic within-person processes over time.
Declared interests
No conflicts of interest declared. Funded by the Hungarian National Research, Development and Innovation Office (two grants) and the University Excellence Fund of Eötvös Loránd University. No commercial sponsor.
The easy way to misread this
Do not read the anxiety centrality finding as evidence that anxiety drives or causes the social difficulties in autism. It was a trend (p = 0.096), not a significant result, and the authors state it was not a priori hypothesized. The study is cross-sectional and entirely self-report, so no causal direction can be inferred from any of the network edges. Additionally, nearly half the autistic participants had a co-occurring psychiatric diagnosis, so the 'autistic network' is not a clean autism signal, and the lower global connectivity may partly reflect the smaller, more heterogeneous sample rather than a specific autism-related disconnection.
Summarised by AI from the full paper, without a clinician reviewing it. Check it against the source before it changes what you do. Read it on PubMed →