Figures 3 and 4 illustrate the global dissemination of research outputs in this domain. The United States leads with 1,028 relevant studies, demonstrating both the earliest research initiation (centrality index = 0.12) and sustained dominance in the field. This is followed by Canada (254), Australia (209), the United Kingdom (195), and China (76). Figure 5 and Table 1 present the top 10 institutions by publication volume in autism family therapy research, with the University of California System ranking first at 132 publications. Harvard University (100), the University of Toronto (91), the State University System of Florida (79), and the University of London (73) follow closely. Notably, the majority of these top institutions are U.S.-based academic powerhouses, with only a small number from other national education systems. This underscores the dominance of American universities in autism family therapy research, attributed to their internationally leading research platforms and academic excellence. It should be noted that Harvard University and Harvard Medical School are listed separately in the dataset but are administratively affiliated within the same institutional network.

A total of 873 authors contributed to the body of research on family therapy for autism ( Figure 6 ). The network visualization uses nodes to represent individual researchers and edges to denote collaborative relationships, with node size proportional to publication output and edge thickness reflecting collaboration intensity. Color gradients from warm to cool tones encode temporal proximity, with warmer hues indicating more recent contributions. Table 2 lists the top 10 most prolific authors, led by Baranek, Grace T. (19 publications), followed by Cordier, Reinie (13) and Storch, Eric A. (13). According to Price Law, the minimum number of articles by core authors (M) = 0.749 × √nmax (nmax is the number of articles by the most prolific authors), 15 M = 0.749 × 4.3 ≈ 3.22, rounded to an integer of 3, which means that the core authors have ≥4 publications. This identified 207 core authors responsible for 1,007 articles, accounting for 53.2% (1,007/1,891) of total publications—exceeding the 50% threshold for significant core contribution. The dense network structure observed in Figure 6 indicates robust collaborative relationships among researchers in this field.

Figure 7 displays the keyword co-occurrence network derived from the literature using the Log-Likelihood Ratio (LLR) algorithm, with clusters labeled #0 through #10. The visualization reveals 11 distinct thematic clusters, as visualized in Figure 7 . The average silhouette value (S = 0.7833) exceeds the 0.5 threshold, indicating strong cluster validity and coherence. The compact distribution of clusters in the mapping suggests high inter-cluster relatedness and thematic convergence among research topics. Table 3 presents weighted keywords within each cluster, highlighting their relative importance.

Integrating these cluster findings, early intervention emerges as a central research theme in autism family therapy. Subthemes include cognitive function training, attention deficit intervention, and family-based vocational rehabilitation, which collectively constitute the primary focus areas within this domain.

Keywords serve as highly condensed representations of scientific research features, reflecting disciplinary hotspots and inter-theme connections. Keywords were extracted using CiteSpace, which also constructed a visual mapping based on keyword co-occurrence relationships. In Figure 8 , each node represents a keyword, with node size proportional to keyword frequency. Edges connect related keywords, with line thickness indicating co-occurrence strength. Collectively, the visualization depicts thematic associations and their relative prominence. Notably, Figure 8 reveals two primary research trajectories. The first focuses on early intervention within home environments, including cognitive training, behavioral modification for children/adolescents with autism, and family-centered diagnostic practices. The second trajectory highlights emerging trends in quality of life enhancement for autistic families, communication skill development in children, and vocational rehabilitation programs for adult patients. These findings indicate an evolving research landscape balancing traditional clinical interventions with broader societal support systems.

Keyword emergence analysis focuses on algorithmically deriving research hotspots reflecting the emergence or continued attention over a period of time, reflecting the changes and trends of hotspots in the research field. The γ value was adjusted according to the co-occurrence of keywords, respectively, to obtain the keyword emergence analysis, with the blue line indicating the time interval and the red line indicating the time period in which the keyword emergence was found. “Children,” “autism spectrum disorder,” and “adolescents” are the earliest (1994) terms to appear. “Pervasive developmental disorder” is the earliest (1994) term to appear. “Pervasive developmental disorder” maintains a high intensity of occurrence from 1999 to 2013, with an intensity of 10.72, the highest intensity of emergence, suggesting that most children with autism are accompanied by developmental disorders, and that the US Diagnostic and Statistical Manual of Mental Disorders Fourth Edition (DSM-IV) includes autism as one of the disorders of pervasive developmental disorders. “Mental retardation” is the most frequently occurring keyword; most children with autism have varying degrees of mental retardation, and restoring the intelligence of children with autism has always been a central issue in rehabilitation. “Early intervention” and “patterns” are the most popular keywords in this field in the past 10 years, which indicates that the research on family therapy for autism is increasingly focusing on early intervention for children with autism at home as well as on intervention models ( Figure 9 ).

In the keyword timeline view, the x-axis is the year in which the keyword appears in the cluster, and the y-axis is the cluster in which each keyword is located, which further shows the occurrence, end, and time trend of each cluster, which can reflect the importance of a specific cluster and the period of its distribution. According to the picture, it can be seen that from 1996 to 2010, autism research focused on children’s communication ability and early intervention; from 2010, the research hotspots gradually shifted to parent training, family therapy system, and parental mental health; from 2020 to 2024, the research hotspots shifted to the importance of family therapy for the autistic family intervention model, the emotional management ability of children with autism, comprehensive family wellbeing index, and caregiver psychology (see Figure 10 ).

The superimposed journal overlay map ( Figure 11 ) visualizes citation relationships between source and target journals, revealing interdisciplinary knowledge diffusion patterns in autism family therapy research. Cited journal clusters spanning economics, psychology, education, rehabilitation sciences, and genetics form the foundational knowledge base for understanding psycho-educational health dimensions in autism populations. For example, the Journal of Autism and Developmental Disorders (psychology), Autism Research (genetics), and Rehabilitation Psychology (rehabilitation sciences) serve as core knowledge hubs in this domain.

A newly emerged cluster linking Environmental Toxicology and Clinical Pharmacology highlights an evolving subfield focused on exploring autism’s environmental etiology and developing targeted pharmacotherapeutic interventions. This aligns with recent advancements examining interactions between genetic vulnerabilities, toxicant exposures, and pharmacological responses in autistic individuals.

The overlay map demonstrates robust translational research networks integrating three key dimensions: (1) psycho-educational models such as applied behavior analysis for skill development; (2) economic evaluations assessing cost-effectiveness of family-based interventions; and (3) pharmacogenetic studies investigating personalized medication responses. These integrations underscore the value of interdisciplinary collaboration in addressing autism’s multifactorial nature and improving rehabilitative outcomes.

The co-cited literature helps us to understand the research themes and development context of the field, and these co-cited literature forms the knowledge base of the study. After excluding some of the literature that is not in the field of this study, the 10 most cited of these are shown in Table 4 . The most frequently cited articles were those published by Hayes SA, Pickles A, and Karst J, respectively. These articles focused on stress in families of children with ASD, parent-mediated social interventions for children with autism, and family intervention models for autism.

Since the introduction of family interventions for autism, substantial evidence has demonstrated their positive effects not only on the symptoms of children with autism but also on various other aspects of family wellbeing. Research has shown that parenting a child with autism imposes significant stress on parents, negatively affecting their mental health (25) and potentially influencing the development of young children with autism (26).

In the absence of specialist knowledge, group-based parent training has been identified as a cost-effective approach for providing education and support to parents. This method has proven effective for children with intellectual and physical disabilities (27). Family-Focused Psychoeducational Therapy (FFPT) is another intervention designed to reduce family stress and negative emotional responses, enhance psychosocial functioning, and strengthen social and family support systems during the child’s developmental process (28). A study by Zhou et al. (29) applied FFPT to families of children with autism, demonstrating that FFPT improved parental self-efficacy and psychological wellbeing, enhanced communication skills, and reduced maladaptive emotional responses.

Additionally, research suggests that family interventions can improve parent–child relationships. Parents who actively engage in their child’s therapy are better able to interpret and respond to their child’s communicative expressions (30). Increased parental involvement in autism interventions is associated with reduced parenting challenges and improved family dynamics. However, there is a lack of large-scale RCTs to fully validate these findings. Studies have also suggested that parent-mediated interventions delivered via telemedicine provide a cost-effective alternative to traditional in-person therapy (31, 32).

A wide range of family interventions for ASD are currently available, and therapists should tailor their selection to address the unique symptoms of each child. A review of the existing ASD literature suggests that family interventions hold significant promise for the future. Family-based interventions have demonstrated positive effects on children, including improvements in core symptoms. However, they are also associated with increased parental stress and varying impacts on family wellbeing, as well as reduced financial burden. Despite these advantages, aspects such as attention, eye gaze, and learning ability in children with ASD have not been extensively validated in the literature. Thus, further research is needed, and multidisciplinary collaboration is essential to maximize the benefits of family-based interventions.

Artificial intelligence (AI) encompasses core technologies such as data discovery and learning, human–computer interaction, and knowledge and data intelligence processing (59). AI has shown potential in improving core ASD symptoms, but the vast range of available technologies has led to a lack of standardized treatment protocols. Linda et al. employed a computer-based social skills intervention, FaceSayTM, with children with ASD, demonstrating improvements in eye gaze, joint attention, and facial recognition skills (60). However, whether such technologies can be effectively implemented in home settings remains unclear. This highlights AI’s potential in ASD interventions and suggests that future research should focus on designing AI-based interventions specifically for home use, addressing both core ASD symptoms and broader aspects of family functioning.

Currently, family intervention approaches primarily involve either direct or indirect therapist-guided interventions delivered through online video conferencing or telephone consultations. Future research could explore software-based interventions that provide structured guidance for caregivers through interactive tasks. While most studies have focused on improving core ASD symptoms, limited research has examined the broader impact of family interventions on overall family functioning.

Caregiver stress plays a pivotal role in the developmental outcomes of children with ASD, as parental mental health is closely linked to child progress. Family interventions have been shown to positively influence children’s verbal abilities, reduce repetitive and stereotyped behaviors, and improve daily living skills. However, in terms of social interaction, family-based interventions face limitations, as they often lack opportunities for children to engage with peers of the same age. This may result in slightly poorer social skills compared to children receiving institution-based rehabilitation. Future research should explore strategies to simulate peer social interactions within family-based intervention models.

Over the past three decades, substantial progress has been made in family-based autism interventions, supported by a growing body of high-quality research. The use of visual analytics can help identify general trends and research hotspots in this field. Moving forward, key areas of research will likely include the effects of family-based early interventions on the social competence of children with ASD, parental stress, associated psychological challenges, and the development of innovative family intervention models.

Most studies on family interventions originate from Western countries, such as the United States and Canada, where individualistic cultural values may shape intervention approaches differently from those in collectivist societies, such as China. In Western settings, parent-mediated interventions often emphasize child-centered goals, whereas Asian families may prioritize family harmony and parental authority. Future research should investigate how cultural factors influence intervention effectiveness and develop culturally adaptive intervention models.

Additionally, many of the included studies lack rigorous methodological designs, limiting the generalizability of their findings. Future research should prioritize well-controlled experimental studies, including RCTs, to ensure robust and reproducible results.

This study is subject to a number of limitations. Firstly, only relevant literature from the WOS core dataset was included in this study for inclusion in the analyses, which may produce some bias in the results. Second, although CiteSpace is a professional bibliometric tool that provides objective analyses, there may be some subjective bias as different researchers may have different views and interpretations of the same content. Finally, the quality of the literature varies and global results on the overall structure of the research field may be more limited.