My research is interested in understanding in vivo modelling of depression using systematic review and meta-analysis.
Systematic review is an incredibly useful tool to assess the relevant literature and achieve a clear overview of the current literature, which is becoming increasingly difficult with the amount of papers published in scholarly journals increasing exponentially (Bornmann & Mutz, 2014, www). It can also provide better understanding of the laboratory methods used to induce the condition, the range of outcome measures used to assess depressive-like phenotypes, and the variables that might impact on the efficacy of different treatments (de Vries et al., 2011, www).
However, systematic reviews are time consuming and often not produced quickly enough to inform the field before they need to be updated (Tsafnat et al., 2014, www.). Automation techniques such as machine learning and text mining can aid the systematic review process and reduce the workload at various stages of the systematic review process, mainly the screening and data extraction stages (Jonnalagadda et al., 2013, www).
The method of model induction, outcome measures, treatments tested, and study quality in preclinical investigations of depression are of key interest in this project. I aim to use ContentMine to help investigate these key areas by using dictionaries, such as genera and species, to aid with document classification and clustering. Identifying the language used when authors report measures to reduce bias in their experiments, using regular expressions, can improve document tagging and data extraction for study quality measures. Identifying these key features in papers can aid the systematic review process by making grouping easier, for example grouping similar models, outcome measures, or animal species, for follow up analyses. These tools increase the efficiency in which systematic reviews are carried out, shortening the time from search date to publication, and thus allow for more up-to-date reviews to be produced. All measures to automate and streamline the systematic review process can reduce human workload, decrease costs associated with human time and expedite scientific advances.