Psychological well-being is a complex and multi-faceted aspect of human health, with mental disorders exerting a profound and lasting impact on individuals. Recognizing and addressing these disorders effectively is of paramount importance. To this end, the choice of appropriate assessment tools is the initial step in identifying and understanding these problems. One such tool frequently employed by researchers is the questionnaire, tailored to specific methodologies and target groups (1).

Social functioning, a cornerstone of an individual’s overall well-being and daily life, encompasses the capacity to form and sustain relationships, adapt to social contexts, and communicate effectively (2, 3). Understanding the intricate domain of social functioning is indispensable. Effective adult functioning hinges on the ability to interpret others’ behaviors and respond appropriately, guided by widely accepted norms of social interaction (4). An individual’s social functioning is defined by their interactions with their surroundings and their capacity to carry out their role within such contexts (5). Social dysfunction, encompassing a broad spectrum of maladaptive behaviors and impaired interactions, has become a pressing concern in contemporary society. It hampers the functioning of individuals, communities, and societies, leading to a range of adverse consequences (6). This underscores the importance of understanding and addressing social dysfunction as it not only affects individual well-being but also has far-reaching implications for the broader social fabric.

The psychosis spectrum presents three vital components of social ability: social cognition, social interaction, and social functioning. These components are intertwined, with competency in one skill often fostering another. Social cognition, in particular, stands as a linchpin of successful interactions, enabling and reinforcing effective social functioning (7). Understanding of the factors that contribute to social dysfunction is crucial for policymakers, clinicians, and researchers seeking to address these challenges effectively. Research on social dysfunction has revealed its multifaceted nature, influenced by various biological, psychological, and sociocultural factors. Previous studies have highlighted the role of cognitive functioning in shaping social behavior and interactions. Cognitive deficits have been linked to impaired social adaptability and difficulties in understanding and interpreting social cues (8).

Emotional states, particularly anxiety and depression, have also garnered attention in the context of social dysfunction. Individuals experiencing anxiety may withdraw from social interactions due to fear or discomfort, leading to social dysfunction (9). Similarly, depression can contribute to social withdrawal and impaired social functioning, as individuals may struggle with motivation and engagement in social activities. Recognizing the influence of anxiety and depression in predicting social dysfunction is essential for devising interventions that address emotional well-being. A linkage between the presence of social dysfunction in depression and anxiety has long been noted (10, 11).

Major depressive disorder (MDD) is a prevalent psychiatric condition characterized by persistent sadness and a loss of interest or pleasure in daily activities (12). Notably, MDD often leads to profound and pervasive impairments in social functioning, defined as the capacity to fulfill standard social roles (13). Importantly, depression frequently results in the loss or disruption of significant social relationships. Another critical clinical feature of depression is the presence of somatic symptoms which can be interpreted as the clinical manifestation of psychological issues (14, 15). According to research, women who exhibit a greater extent of physical symptoms are more prone to being diagnosed with depression (16). Although women are known to report higher somatic symptoms, less is known regarding within-group disparities between women of varying education and age.

Depression is one of the most prevalent mental disorders in the 21st century, and it is related to low socioeconomic status (17), poor physical health (18), sleep problems (19), and a drastic suicide risk (20). Annually, an estimated 703,000 individuals commit suicide throughout the world. For every suicide, there are about 20 others who attempt suicide and several more who have serious suicidal ideation (21). Suicide can be prompted by a variety of risk factors such as poor socioeconomic status, financial hardship, and a low educational level (22, 23). It has been observed in older people that prior adverse experiences influence the management of future crises (24), while others argue that these can lead to depression (25), while still, others contend that older people, in particular, should be considered for suicide prevention programs owing to the numerous risk factors that relate to them (26). The gender paradox in suicide is long-established, with women having more suicidal thoughts and attempts than males (27, 28) with a prevalence of suicidal ideation in women. However, the intra-group variation in the prevalence of suicidal ideation in women has not been widely investigated. Furthermore, these social deficits are linked to an array of negative outcomes, including increased mortality rates (29, 30). This intricate web of connections between suicide, social dysfunction, and depression highlights the urgent need for comprehensive approaches to address these interrelated issues. By understanding these relationships, we can develop more effective strategies for suicide prevention and mental health support, especially among at-risk populations.

Although depression is the most prevalent mental disorder, as many as 20 million people were found to have anxiety disorders worldwide.¹ It appears that elderly people have lower prevalence rates of diagnosed anxiety disorders than younger adults (31). However, researchers have discovered a wide range in the prevalence of anxiety disorders during a 12-month period among adults aged 55 and above, ranging from 7 to 12% (32–34). In addition, the prevalence of disorders differs among cultures and living conditions, as well as depending on study design and evaluation methods (35–37). The symptoms of anxiety are more prevalent in low socioeconomic groups (38). It is well recognized that there are differences in mental health across conventional SES variables including income, educational achievement, and other measures of social rank (39). Prins et al. hint of under researched processes in social epidemiology by demonstrating that the effects of class relations on anxiety go beyond those of SES (40). Hence, there are multiple factors related to SES that should be taken into account while investigating anxiety.

In addition to emotional and cognitive factors, sociodemographic characteristics such as education and age have been explored in relation to social dysfunction. Education plays a significant role in social integration and cohesion, with higher educational attainment often associated with improved social adaptability and engagement (41). Age can also influence social functioning, as individuals at different life stages may face distinct social challenges and opportunities (42). Acknowledging the impact of education and age on social dysfunction can inform targeted interventions tailored to specific age groups. Social dysfunction is a multifaceted phenomenon that affects various aspects of patients’ lives, including social interactions, everyday activities, and employment status (43). Vividly, social dysfunction is a multidimensional phenotype that is impacted by a range of socio-demographic factors, such as poor socioeconomic status, as well as psychological disorders/dysfunctions such as depression and anxiety. People’s social roles have been found to change with age and can be influenced by psychopathology (44). Social dysfunction is correlated with unemployment, disability and being single (45, 46).

Understanding the complexity of social dysfunction requires considering various factors beyond just emotional and cognitive aspects. Sociodemographic characteristics like education and age have emerged as significant contributors to the dynamics of social dysfunction. Research has shown that education plays a pivotal role in promoting social integration and cohesion, with higher educational attainment often linked to enhanced social adaptability and engagement (41). Additionally, age can exert a profound influence on social functioning, as individuals at different life stages encounter diverse social challenges and opportunities (47). Social dysfunction, a multifaceted phenomenon impacting various facets of patients’ lives, encompasses social interactions, everyday activities, and employment status (32). This multidimensional phenotype is further shaped by a range of socio-demographic factors, including poor socioeconomic status, as well as psychological disorders and dysfunctions like depression and anxiety. Moreover, people’s social roles undergo transformation with age and can be influenced by psychopathology (44). Notably, social dysfunction exhibits strong correlations with unemployment, disability, and being single, as evidenced by studies such as those conducted by Brown et al. (45) and Rizvi et al. (46).

Machine learning, and specifically Multilayer Perceptron (MLP) modeling, has shown promise in capturing complex relationships and nonlinear dependencies among variables. In social science research, MLP models have been used to predict various social and behavioral outcomes. For instance, Zheng et al. (48) employed an MLP model to predict mental health outcomes, while Lu et al. (49) used MLPs to identify risk factors for social problems. These studies highlight the potential of MLP modeling in social science research, particularly in predicting complex outcomes such as social dysfunction. In recent years, advances in data science and machine learning have opened new avenues for studying complex social phenomena. Among the innovative methodologies, MLP modeling has emerged as a powerful tool for predictive analytics in various domains, including social science research. This study aims to utilize an MLP model to explore the impact of covariates and hidden layers on predicting social dysfunction, offering valuable insights for targeted interventions and policy formulation. The incorporation of covariates into MLP modeling offers an opportunity to gain insights into the impact of various factors on social dysfunction. Covariates encompass independent variables that may influence an individual’s behavior and interactions within society. In the context of social dysfunction, covariates could include demographic information, cognitive functioning, emotional states, and socio-economic status, among others. Understanding the relative importance of these covariates can shed light on the underlying mechanisms driving social dysfunction.

Research studies have demonstrated the utility of MLPs in various social science domains, including predicting mental health outcomes (48) and identifying risk factors for social problems (49). MLPs offer the advantage of uncovering hidden patterns and nonlinear associations, allowing researchers to explore complex interactions among covariates and the dependent variable. By utilizing this modeling approach, researchers can potentially develop more accurate and nuanced models for social dysfunction prediction. However, while existing literature has explored the individual impacts of cognitive functioning, emotional states, education, and age on social dysfunction, limited research has investigated their combined influence within an MLP modeling framework. This study seeks to address this gap by employing an MLP model that incorporates hidden layers to capture intricate relationships among these covariates. By doing so, this research aims to provide a nuanced understanding of social dysfunction prediction and identify the most influential factors for targeted interventions and policy development.

In this context, the present study utilizes an MLP model to predict social dysfunction by incorporating a set of covariates, including education, cognitive functioning, anxiety, depression, and age. By analyzing the importance of these covariates in the context of social dysfunction prediction, this research aims to contribute to a deeper understanding of the multifaceted nature of social dysfunction and provide valuable insights for designing targeted interventions and evidence-based policies. In this study we utilized network analysis which assumes that symptoms of disorders trigger one another (50). In the field of psychopathology, a network approach enables the assessment of how specific behaviors or symptoms are interconnected with several other behaviors or symptoms (51). Implementing network analysis allows for investigating how symptoms of disorders link and become mutually reinforcing by analyzing symptoms as a network (52).

In light of the vulnerability experienced by women in low socioeconomic status and the prevalence of employment inequalities, this study was undertaken to address the imperative need for enhanced research and initiatives aimed at safeguarding and enhancing women’s mental well-being. The primary objective of this study was to assess disparities in Cognitive functioning, social dysfunction, Anxiety, depression Education levels and age within this demographic, shedding light on the pressing mental health concerns faced by women in low socioeconomic groups.

The primary objective of this study is to assess and analyze the disparities in cognitive functioning, social dysfunction, anxiety, depression, education levels, and age within a specific demographic of women in low socioeconomic groups. By employing Multilayer Perceptron (MLP) modeling and incorporating covariates, we aim to gain a deeper understanding of the multifaceted nature of social dysfunction, identify the most influential factors affecting it, and provide valuable insights for designing targeted interventions and evidence-based policies to enhance the mental well-being of women in low socioeconomic status.

The results of our study utilizing the Multilayer Perceptron (MLP) model demonstrated strong predictive performance in assessing social dysfunction. The demographic characteristics of our study sample are summarized in Table 2, indicating key features of the participants, including their educational levels and age categories. A total of 780 participants were included in the analysis (Table 3).

The Multilayer Perceptron (MLP) model demonstrated promising performance in predicting social dysfunction based on the input factors and covariates. The model was trained and tested on a sample of data, with the training phase yielding a Sum of Squares Error of 45.444 and an 11.1% rate of incorrect predictions. During the testing phase, the model achieved a Sum of Squares Error of 24.565 and an 11.9% rate of incorrect predictions, indicating its ability to generalize well to unseen data.

The classification performance of the model was evaluated using the testing sample. The results revealed a high overall accuracy of 88.9%. The model exhibited a remarkable ability to correctly predict cases of high social dysfunction, with a sensitivity of 97.7%. However, its performance in identifying cases of low social dysfunction was relatively lower, with a sensitivity of 34.2%. These findings suggest that the MLP model is particularly adept at identifying individuals with high social dysfunction, but further improvements may be necessary to enhance its predictive accuracy for low social dysfunction cases.

An essential aspect of our study was assessing the relative importance of each covariate in predicting social dysfunction. We quantified the importance of cognitive functioning (HYPO), which emerged as the most influential factor, with an importance score of 0.269. Anxiety (ANX) closely followed, with a score of 0.282, indicating its strong predictive value. Depression (DEP) exhibited high predictive importance as well, with a score of 0.274. In comparison, education (EDU) and age played relatively smaller roles, with importance scores of 0.107 and 0.068, respectively. These importance scores provide a comparative measure within the model, indicating the relative impact of each covariate on social dysfunction prediction. The prominence of cognitive functioning and emotional states underscores the imperative for targeted interventions and support in these domains (Table 4).

Finally, the study assessed the normalized importance of each covariate, offering insights into their relative contributions to the model’s accuracy. Although the contribution of depression (DEP) is higher (97.2%) than that of cognitive functioning (HYPO; 95.2%), it’s important to understand that the normalized importance score does not directly translate to absolute influence. The 100% contribution from anxiety (ANX) underscores its critical role in predicting social dysfunction, but this does not necessarily mean it is inherently more influential than other variables. The normalized importance scores for education (EDU) and age, at 37.7 and 24.1%, respectively, indicate that these factors have a comparatively lower impact on the model’s predictions when compared to other variables (Table 2).

The results of the Multilayer Perceptron (MLP) model for social dysfunction prediction are promising, demonstrating its capacity to capture complex relationships and dependencies among covariates. The model showcased high overall accuracy and strong sensitivity in identifying individuals with high social dysfunction. Cognitive functioning (HYPO), depression (DEP) and anxiety (ANX) emerged as the most influential predictors, highlighting their significance in predicting social dysfunction. These findings offer valuable insights for developing targeted interventions and evidence-based policies to address social dysfunction and promote societal well-being (Figures 1, 2).

To provide a more comprehensive understanding of the implications of our findings, it’s essential to discuss how the model’s performance aligns with our research objectives. Our study aimed to predict social dysfunction among a specific demographic of women in low socioeconomic status, and the results show the MLP model’s promising ability to achieve this goal. The model’s high overall accuracy and sensitivity are encouraging, as they indicate that it excels in identifying individuals with high social dysfunction. These findings hold significant promise for practical applications and policy development. By recognizing that cognitive functioning (HYPO), anxiety (ANX), and depression (DEP) are the most influential factors in predicting social dysfunction, we can now tailor interventions and policies to address these critical areas. For instance, targeted support programs for improving cognitive functioning, mental health, and coping with anxiety may be developed to aid individuals experiencing social dysfunction. Policymakers can use these insights to prioritize resources and design evidence-based initiatives. While our MLP model excels in identifying high social dysfunction cases, it’s important to acknowledge that further improvements are needed to enhance its predictive accuracy for low social dysfunction cases. Understanding the challenges and limitations of the model’s performance can guide future research and model refinement efforts. Future studies may focus on refining the model to better capture nuanced variations in social dysfunction among individuals with relatively lower dysfunction levels.

The findings derived from the Multilayer Perceptron (MLP) model, employed to forecast social dysfunction, provide valuable insights into the intricate nuances of this societal concern. The model’s remarkable performance, characterized by its elevated overall accuracy and sensitivity in identifying instances of high social dysfunction, underscores its efficacy in discerning the intricate interplay among input variables and covariates. Nevertheless, the comparatively lower sensitivity in detecting cases of low social dysfunction highlights the imperative for further exploration and enhancement of the model’s predictive capabilities in this particular domain. One of the remarkable attributes of artificial neural networks is their adaptability, achieved through the fine-tuning of model factors via weight adjustments, allowing for specific function training, as elucidated by Hagan and Dennutu (64). Furthermore, the inherent parallel processing capabilities of artificial neural networks translate into swift classification speeds, as articulated by the same authors. Additionally, they help mitigate the oversimplification of complex relationships, a notable advantage. These merits underscore the preference for employing this method over traditional statistical techniques such as logistic regression for predictive purposes.

Our study findings, echoing the insights of Farahani et al. (65), underscore the significant role of artificial neural networks in making precise predictions based on various risk factors. Farahani et al.’s (65) research, which achieved an impressive 90.5% accuracy in classifying high and low-risk individuals in predicting suicide tendencies, highlights the potential of these networks in modeling complex, non-linear relationships. In essence, both of these research work underline the potential of artificial neural networks in unraveling complex social phenomena and stress the importance of addressing mental health concerns in interventions for healthier and thriving communities.

The finding that cognitive functioning (HYPO) emerged as an influential predictor of social dysfunction aligns with existing literature on the role of cognitive processes in shaping human behavior and interactions (66). Cognitive functioning encompasses a broad range of mental processes, such as memory, attention, and problem-solving, which play a pivotal role in social functioning and adaptability. The strong impact of cognitive functioning on social dysfunction underscores the significance of cognitive assessments and interventions in mitigating social challenges.

Anxiety and depression were also identified as crucial predictors of social dysfunction. These findings are consistent with research that highlights the association between mental health issues and impaired social interactions (67). Individuals experiencing anxiety or depression may struggle with social interactions, leading to withdrawal and social dysfunction. The MLP model’s ability to capture the influence of these emotional states emphasizes the importance of addressing mental health concerns in interventions aimed at promoting healthy social functioning (45).

Education and age have demonstrated comparatively lower predictive significance in relation to social dysfunction when juxtaposed with cognitive functioning, anxiety, and depression. Nevertheless, their roles remain substantively meaningful and warrant careful consideration in comprehensive analyses. Education, for instance, has been consistently correlated with heightened social integration and cohesion, often highlighting that individuals with higher educational attainment tend to exhibit superior social adaptability (41). Similarly, age can exert a notable influence on social functioning, with individuals at various life stages grappling with distinct social challenges (47). Recognizing the pivotal roles of education and age in the context of predicting social dysfunction paves the way for tailoring age-appropriate interventions and support systems.

The relative impact of each covariate on the model’s predictive accuracy is further underscored by the normalized importance scores. Notably, cognitive functioning and depression exhibit substantial normalized importance, signifying their significant contributions to the model’s overall performance. Furthermore, the comprehensive normalized importance attributed to anxiety underscores its pivotal role in precisely predicting social dysfunction. This understanding of the varying importance of covariates enables the prioritization and customization of interventions to effectively target the most influential factors. The current study serves as a compelling testament to the potential of Multilayer Perceptron (MLP) modeling in unraveling intricate social phenomena, such as social dysfunction. The model’s capacity to harness hidden layers and capture nonlinear relationships among covariates offers an invaluable tool for deciphering the multifaceted dynamics that underlie social dysfunction. As MLP modeling continues to evolve, it opens up new avenues for the development of refined predictive models and a nuanced comprehension of social behavior, thereby supporting the formulation of evidence-based policies and the implementation of precisely targeted interventions.

In summation, the discussion of the MLP model’s outcomes underscores the critical significance of cognitive functioning, anxiety, and depression as major predictors of social dysfunction. These findings accentuate the imperative of addressing mental health concerns and cognitive well-being within interventions aimed at fostering healthy social interactions. Moreover, the acknowledgment of the roles played by education and age provides insights for tailoring interventions tailored to specific age groups and educational backgrounds. The application of MLP modeling in social science research presents a promising avenue toward gaining deeper insights into social dysfunction and the formulation of effective strategies to nurture thriving communities.

The present study utilized a Multilayer Perceptron (MLP) model to predict social dysfunction, drawing insights from a comprehensive set of input factors and covariates. The MLP model demonstrated favorable performance in both training and testing phases, highlighting its capability to generalize well to unseen data and make accurate predictions. The high overall accuracy of 88.9% underscores the effectiveness of the MLP model in classifying social dysfunction cases.

An essential aspect of the study was the assessment of covariate importance in predicting social dysfunction. Cognitive functioning (HYPO) emerged as an influential predictor, emphasizing its critical role in shaping social behavior and interactions. Anxiety (ANX) and depression (DEP) also contributed significantly to the model’s accuracy. These findings offer valuable insights into the drivers of social dysfunction and call for targeted interventions addressing cognitive functioning and emotional well-being.

The normalized importance scores provided further clarity on the relative contributions of covariates to the model’s predictive accuracy. Anxiety (ANX), cognitive functioning (HYPO) and depression (DEP) demonstrated the highest relative importance, indicating their significant impact on the model’s performance. Anxiety (ANX), in particular, exhibited a full contribution, indicating its indispensable role in predicting social dysfunction. Education (EDU) and age, though less influential, remain important factors in understanding social dysfunction dynamics.

Overall, the study highlights the potential of MLP modeling in social science research, particularly in predicting social dysfunction. The model’s ability to capture complex relationships and nonlinear dependencies among covariates makes it a valuable tool for unraveling the intricate mechanisms underlying social dysfunctions. The findings presented here contribute to a deeper understanding of social dysfunction’s multifaceted nature and inform evidence-based interventions and policies aimed at addressing these challenges.

While the MLP model showed robust performance, there is room for further research and refinement. Enhancing the model’s sensitivity in identifying low social dysfunction cases could be explored to improve its comprehensive predictive capacity. Additionally, expanding the dataset and incorporating additional relevant covariates may enrich the model’s predictive power, providing a more nuanced understanding of social dysfunction dynamics.

In conclusion, the study underscores the significance of Multilayer Perceptron Modeling in social dysfunction prediction, offering valuable insights into the impact of covariates on this complex societal issue. The results have implications for policymakers, clinicians, and researchers seeking to address social dysfunction and promote healthier and more harmonious communities.

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 Tarbiat Modares University, Tehran, Iran. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

SB: Writing – original draft. ST: Writing – original draft. HF: Methodology, Writing – original draft, Writing – review & editing. PW: Methodology, Writing – review & editing. ER: Writing – original draft.