Main content



Loading wiki pages...

Wiki Version:
Title Applications of artificial intelligence technique in predicting tinnitus intervention outcomes Abstract Objectives: Psychological approaches such as cognitive behavioral therapy (CBT) have the most evidence-based. However, the studies so far have examined group effects and there is little knowledge on the predictors of outcomes. Our previous efforts on using traditional statistical methods have not resulted in the identification of any key predictors. In the current study, we have applied Artificial Intelligence (AI) techniques on a dataset of individuals who underwent Internet-based CBT (ICBT) to examine the predictors of intervention outcomes as well as to identify the participant group that are most/least likely to be benefited. Design: 228 individuals with tinnitus who underwent ICBT in three different trials were included in this study. A positive treatment effect was defined as a reduction of 13-points following the intervention on the Tinnitus Functional Index (TFI). Various demographic, tinnitus-related, and comorbidities were used as predictor variables. AI techniques with various decision trees, and random forest were applied to this data to examine the predictors of outcomes and to identify participant groups who are most likely to be benefited. Results: The CART decision tree and gradient boosting models (GB) resulted in the identification of some key factors (e.g., education level, baseline tinnitus severity) that may influence the ICBT outcomes. The predictive models based on CART and GB had a mean AUC of 0.69 (SD:0.001) and 0.68 (SD:0.02), respectively. Though, not strong, these classifiers show moderate discrimination power in distinguishing the participants who tend to yield successful outcomes with the treatment. Characteristics. In fact, three participant groups had shown at least 85% success with the ICBT intervention, including the participants who had an education master’s level or above. Conclusions: We believe that this research may help choose right intervention based on individual variations and help achieve optimal outcomes at individual patient level. Meeting Times Interactive Presentation Session May 3, and additional discussion sessions on May 4 and May 5. Contact Dr. Vinaya Manchaiah, Jo Mayo Endowed Professor of Speech and Hearing Sciences, Lamar University, Texas;
OSF does not support the use of Internet Explorer. For optimal performance, please switch to another browser.
This website relies on cookies to help provide a better user experience. By clicking Accept or continuing to use the site, you agree. For more information, see our Privacy Policy and information on cookie use.

Start managing your projects on the OSF today.

Free and easy to use, the Open Science Framework supports the entire research lifecycle: planning, execution, reporting, archiving, and discovery.