Semester

Summer

Date of Graduation

2022

Document Type

Dissertation

Degree Type

PhD

College

Eberly College of Arts and Sciences

Department

Psychology

Committee Chair

Melissa Blank

Committee Co-Chair

Cole Vonder Haar

Committee Member

Cole Vonder Haar

Committee Member

Christa Lilly

Committee Member

Michael Young

Abstract

Preclinical behavioral neuroscience often uses choice paradigms to capture psychiatric symptoms. In particular, the subfield of operant research produces nested datasets with many discrete choices in a session. The standard analytic practice is to aggregate choice into a continuous variable and analyze using ANOVA or linear regression. However, choice data often have multiple interdependent outcomes of interest, violating an assumption of general linear models. The aim of the current study was to quantify the accuracy of linear mixed-effects regression (LMER) for analyzing data from a 4-choice operant task called the Rodent Gambling Task (RGT), which measures decision-making in the context of various manipulations (e.g., brain injury). Prior analysis of RGT data from intact rats (Sham; n = 58) and brain-injured rats (TBI; n = 51) revealed five distinct decision-making phenotypes for this task. To generate datasets for parametric analysis, trial-level data was simulated using a Monte Carlo approach recapitulating those phenotypes. Population parameters were defined from existing data, and repeated sampling was conducted to generate 1000 datasets for four sample sizes (n = 6, 10, 14, 20) and four effect sizes (f = 0.0, 0.3, 0.4 and 0.5). Two LMER models were performed to compare TBI versus Sham across datasets: a full LMER where choice of all four outcomes was analyzed simultaneously, and a control LMER where choice of a single outcome was analyzed. The full LMER exceeded 75% false positives across all sample sizes, and the control LMER was underpowered to detect expected effects. These results suggest analyzing trial-level data in a mixed effects logistic regression will be necessary to accurately analyze RGT data. More broadly, these types of errors must be remedied to improve translation to clinical research.

Share

COinS