The Laser Project

While this project's premise is simple -- using a laser to control the temperature an animal is experiencing -- the implementation is deceivingly complex, and adding heat is much simpler than removing it...

Cooling stage

Caenorhabditis elegans

Despite having only 302 neurons (0.00000035% of yours) C. elegans exhibit complex sensation and behavior, including state changes, directed movement (positive and negative), and importantly, robust thermotaxis. They are able to remember a conditioned temperature and then consistently correctly make environmentally-informed reorientations to efficiently navigate towards that temperature again. This holds whether or not that conditioned temperature was warm or cold! However, since the lasers can only heat, (C. elegans aren't that small!) the real challenge is cooling quickly enough to act on behavioral timescales.

The Flavell Lab focuses on C. elegans research, including my other main project.


Cooling

This is not only important if you want to do experiments below ambient temperature; importantly, cooling rate is proportional to the temperature differential between the heated area and the environment regardless of their absolute values. Since we need to act in the seconds it takes for a worm to swing its head (less than a fifth of a mm) and sample the environment, we need a large differential to allow fast cooling. Unfortunately, heating the worm that much above ambient just isn't feasible. C. elegans are most comfortable around room temp, and have severe heat shock responses, so the cooling stage (above) is a necessary solution to:

  • Expanded experimental temperature range
  • Increased temperature change responsiveness

The system I designed completely solves that temperature change asymmetry (heating rate is bounded by the maximum laser power, but even turning it completely off, cooling is capped by how cold the material around it is) and allows for negative (movement towards colder temperature) thermotaxis experiments.

With the latest version of our cooling system, it can cool down to ~12°C, but is operated at a steady 15°C. From there, if the laser heats to 22° and is then turned off, it only takes 4 seconds to cool to 19°. This is more that sufficient as C. elegans perform best in gradients of 0.5-1.5 C/cm and take a few seconds to move a millimeter. They are usually kept at 20°C, so this is perfect for naturalist experiments.

Importantly, the design achieves this while fulfilling the many requirements imposed by our imaging and heating systems:

To achieve these, the system involves a reference thermistor, sitting out of the way of the worm, as well as a pump, which transports a specifically chosen heat transfer fluid across the cooling device. That device has been optimized to prioritize even cooling (i.e. that the inflow side and center are not significantly colder than the outflow and edges, as tends to happen). The coolant, motor, and device geometry have each gone through many iterations; even the materials of the slide have been optimized, using sapphire for its superior thermal conductivity over glass.

This setup also fulfills two practical requirements from the experimental side:

This last point is important, as the cooling setup creates its own ambient temperature gradient which needs to be accounted for.

A heatmap of the cooling system's gradient

Sampling of the gradient on multiple days shows a remarkably consistent gradient. This simple linear model is accurate to ~0.1°C for the average sample.



However, how have I gotten these measurements? How can we be sure we're actually capturing temperatures experienced by a mm-long nematode?

Measurement

A quick online search will show that small thermistors are quite numerous and affordable. Problem solved!

Unfortunately, thermistors have significantly different IR-absorbing properties than the agar gel our nematodes live on (or the worms themselves). A solid block of plastic or ceramic is going to overshoot the experienced temperature. Likewise, setting a thermistor just outside of the laser radius is going to undershoot (something confirmed after an unfortunate amount of experimenting and debugging). Thus, we need another approach: thermosensitive dyes.

Using these dyes allows us to minimally change the absorbance of our medium, get highly precise temperatures, and even select the specific z-plane we're measuring temperature in, not just xy location. Although getting the dyes working precisely was rather difficult, now that we have them working accurately and reproducibly, we can get much more accurate thermal measurements. These measurements can then be backed up by experimentation. Given C. elegans well-characterized heat shock response, we can demonstrate our precise control over their perceived temperature. Additional gradient and thermogenetic experiments can reinforce this claim if needed.

The Future

Once our environmental control is established, we will be able to perform previously impossible experiments to tackle biologically, neurologically, and computationally interesting questions! Some of the things we'd be able to test are:

Be the Worm!


Laser Power