A high-throughput opto- & thermogenetic behaviour box
FlyMAD was precise but tested one fly at a time. I designed a behaviour box that tests 18 flies at once with precise, programmable control of light and temperature, and that can be reproduced for under $500.
- Where
- HHMI Janelia Research Campus · Ashburn, VA, USA
- When
- 2015
- Role
- Designer & builder
- Tools
- Laser-cut parts, custom circuit boards, Arduino / Raspberry Pi, machine vision

The problem
Opto- and thermogenetic experiments need tight, reproducible control of light and temperature, and enough animals for real statistical power. FlyMAD delivered precision, but only one fly at a time.
The design
- 18 chambers, one video. Each arena records 18 behavioural bouts in a single video for automated analysis.
- Fully configurable RGB lighting. LEDs under the chambers provide blue backlighting for the cameras that doesn’t trigger red-sensitive optogenetic tools; red light bright enough to activate neurons expressing Chrimson and other red-light-gated channels; and all three channels together for white light to support vision-guided behaviour.
- Closed-loop temperature control. A heated, high-resistance glass floor with feedback control raises and lowers temperature precisely and reproducibly to drive thermogenetic tools such as TrpA1 and shibirets.
- Built in-house. I designed and laser-cut every structural part, with electronics on custom circuit boards driven by Arduino and Raspberry Pi microcontrollers.
- Reproducible. Each rig can be built for less than $500.
Outcome
I built three rigs at Janelia, and students and postdocs went on using them for years after I left. I designed the box so it could be reproduced cheaply, with teaching labs and lower-budget research groups in mind. Wider sharing is an opportunity that remains open.
My contribution
I conceived, designed, fabricated and programmed the system.
Gallery