COVID-19 Information

COVID-19 Information

The University of Texas at Dallas is closely monitoring the outbreak of COVID-19 (coronavirus).

Covid-19 Updates

Comet Calendar

Bayesian Modeling of Response Time and Accuracy for a Visual Mental Rotation Task - Psychology Lecture Series

Tuesday, October 18 at 2:30pm to 3:30pm

JO 4.306

Psychology Lecture Series - In-Person and Virtual Event

Title - Bayesian Modeling of Response Time and Accuracy for a Visual Mental Rotation Task

Speaker: Ritesh Kumar Malaiya
PhD Student, UT Dallas

Drift Diffusion models are designed to explain the mental processes that we may use to make a decision. Such a mental process accumulates evidence as soon as we see a stimulus and continues until a decision is made. Diffusion based models have been widely studied to explain the response time and accuracy in terms of the participant’s ability (drift parameter) and response cautiousness (boundary parameter). The Q-diffusion model used in current study, is an extension of the diffusion model and incorporates stimulus-specific parameters as well. A prior distribution specifies the researcher’s expectations about possible parameter values. In Bayesian inference, the prior distribution is used with the probability model and the data to obtain a posterior distribution which specifies parameter value likelihoods. Evaluating robustness of posterior statistic (e.g., mean) against various choices of prior is important for understanding whether or not two totally different choices of parameter priors will make identical predictions. Current study empirically investigated, given a mental rotation response dataset, the robustness of Bayesian inference of the Q-diffusion drift parameter for both strong and weakly informed prior distribution. The study found that Bayesian inference of the Q-diffusion model is robust against minor variations in the strongly informed priors. However, the weakly informed priors where not able to converge to the assumed true posterior distribution. This study emphasizes the importance of prior sensitivity analysis along with other model checks such as model convergence, when utilizing Bayesian inference for parameter estimation.

This talk is in-person in JO 4.306 and also is a virtual event. At 2:30pm on October 18, join the talk on MS Teams.
Conference ID: 249 923 855 045
Passcode: aRCGUd

Persons with disabilities may submit a request for accommodations to participate in this event at UT Dallas' ADA website. You may also call (972) 883-5331 for assistance or send an email to All requests should be received no later than 10 business days prior to the event.

Event Type

Lectures & Workshops

Target Audience

Undergraduate Students, Faculty & Staff, General Public, Graduate Students


Research, Science & Technology


bayesian, Response Time, Decisions

Behavioral and Brain Sciences
Contact Information
Lena McGee
Google Calendar iCal Outlook