Low-Cost 3D-Printed Mazes with Open-Source ML Tracking for Mouse Behavior

Behavioral neuroscience research often requires substantial financial investment in specialized equipment and software, creating barriers for new investigators and limiting the flexibility of established laboratories. This study explores how 3D printing and machine learning can be combined to reduce startup and operational costs while maintaining research quality. Using 3D printing, we designed and manufactured a mouse T-maze and elevated plus maze to assess cognition and anxiety-like behaviors in male mice. These custom-built mazes demonstrated comparable efficacy with commercial alternatives while offering greater affordability, scalability, and customization. To complement the hardware, we integrated machine learning for automated tracking and analysis of mouse behavior, achieving accuracy equivalent to commercial solutions or experienced human scoring at significantly reduced cost. By combining 3D printing with machine learning, our approach significantly lowers financial barriers for new investigators and enables established research groups to allocate resources more effectively. This approach not only expands research possibilities for established labs but also lowers the barrier to entry for early-career scientists and institutions with limited funding.