Date of Award

Spring 2026

Language

English

Embargo Period

4-1-2026

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

College/School/Department

Department of Psychology

Program

Behavioral Neuroscience

First Advisor

James Stellar

Committee Members

James Stellar, Andrew Poulos, Damian Zuloaga

Keywords

environmental enrichment, meta-analysis, translational neuroscience, cognitive neuroscience, cognitive psychology

Subject Categories

Animal Experimentation and Research | Behavioral Neurobiology | Cognitive Neuroscience

Abstract

Environmental enrichment --providing animals or people with stimulating, complex surroundings --consistently improves cognitive performance in laboratory rodents. But when similar interventions are tested in humans, the benefits are much smaller. Understanding why matters: if researchers overestimate how well rodent findings will translate, it can mislead expectations for human cognitive interventions.

This dissertation presents the first cross-species meta-analysis comparing environmental enrichment effects on cognition between rodent and human studies. Drawing on 121 studies (71 rodent, 50 human) contributing 478 effect sizes, the analysis used multilevel models with variance correction to account for the statistical dependence created when studies report multiple outcomes.

Results confirmed a substantial cross-species gap. Rodent studies produced medium-to- large effects (g = 0.44, 95% CI [0.25, 0.62]), while human studies showed small effects (g = 0.16, 95% CI [0.11, 0.22]), a 2.6-fold difference. Publication bias analyses were mixed: a correction for potentially missing studies reduced the human effect to g = 0.05 (no longer statistically significant), though formal asymmetry tests did not confirm bias. The resource mismatch hypothesis --that deprived laboratory conditions artificially inflate rodent effects relative to the enriched baseline of typical human life --was not supported; the Species x Enrichment Dose interaction was not significant, and within-species dose-response relationships were weak or absent. No specific intervention type --neither particular rodent enrichment categories nor human cognitive training approaches --was associated with larger cognitive effects than any other. Analyses of cognitive domains (e.g., spatial memory, executive function) showed no significant variation in effect sizes. Contrary to the common concern that enrichment increases individual differences, performance variability was consistently reduced in both species, though non-significantly after correction.

These findings provide a quantified foundation for translational neuroscience. The 2.6-fold gap offers the first concrete benchmark for discounting rodent effect sizes when planning human interventions, while the null moderator findings point toward factorial designs that independently vary enrichment components as a necessary next step.

Comments

Data and code for the dissertation can be found in the following GitHub repository: https://github.com/gregfitzgerald/GSF-dissertation-meta-analysis

License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Download

Share

COinS