Leveraging machine learning for bibliometric analysis of emerging fields

Psychotherapy researchers have emphasized the importance of a paradigm shift toward translational psychological treatment. However, the publication landscape on this topic is uncharted. This makes it difficult to assess the state of translational psychotherapy research. Hence, we developed a framework that leverages machine learning to find relevant studies. Based on this, we bibliometrically analyzed 682 publications to provide insights for the integration and development of the research field. Specifically, citation and network analyses were performed to examine the strengths and weaknesses of the research field. Six main findings emerged: 1. Machine learning proved valuable in finding eligible publications and generating an open dataset. 2. Translation comes primarily from physiological psychology/neuroscience and experimental psychology, with a focus on fear and anxiety. 3. Translational research is characterized by international collaborations. 4. It has an impact within and beyond academia. 5. The lack of standardized terminology might threaten scientific progress. 6. There is thematic fragmentation in the field. To foster a paradigm shift towards translational psychological treatment, a consistent terminology would greatly facilitate its development and dissemination.

The field of translational psychological treatment has significant impact, but its full potential is yet to be realized. To better understand potential barriers, we conduct a comprehensive literature review. Based on Ehring’s et al. (2022) proposed framework for translational research in clinical psychology (Steps 1 - 6) and we categorize the previously mapped out dataset to identify the exact translational stage of every publication. Translational research is characterized by fluid transitions or overlaps between its various stages, making it challenging to neatly distinguish between them. In such cases, we begin by determining the direction of translation: when the knowledge transfer is from basic science to implementation then assign the highest step, and when the movement is towards basic principles assign the lowest one. Additionally, we categorize the dataset into three distinctive stages: “pre-clinical trials, clinical trials and implementation” and subsequently, we compare it to the categorization based on Ehring’s model. This high resolution categorization allows us to specify the object of translation, the targeted disorder, the proposed intervention and the direction of translation, shedding light on possible barriers and enabling their resolution (facilitating their elimination). Additionally, we analyze temporal delay in translation to identify both hiccups and untapped potential for further advancement in translational psychological treatment research.