Tissue Heating During Optogenetic Experiments Using Laser and LEDs

During a recent LinkedIn conversation regarding light sources for Optogenetics a concern was raised about LEDs causing heating of the tissue. The concern was based on information in a paper “Optogenetics in Neural Systems” published by O Yizhar et al., Neuron, 2011.

A careful reading of this paper shows that the concern is relevant only to small LEDs implanted in the tissue, and not regarding fiber coupled LEDs or lasers. The paper states “Possible uses of LEDs include both direct implantation of small LEDs in or on tissue (with heating concerns requiring careful control as noted above), or permanently mounted to optical fiber waveguides carried on the subject"

But there are some interesting questions about heating concerns when using fiber coupled lasers and LEDs.

1. At what power level does heating become an issue?
2. Is there a difference between low NA fiber implants coupled to a laser and LEDs with high NA fibers regarding heating?
3. Do either lasers or LED have an advantage in terms of successful experiments or heating concerns?

The engineers at Prizmatix have supplied me with the following input;

Any light energy at some level will produce heat. At higher levels the heat can be a problem. People using powerful lasers can be close to the safety limits.

Just for example, in order to evaluate the limits we can use International Laser Safety Standard IEC-60825-1: Maximal Permissible Exposure (MPE) for Skin to laser radiation (Table 8 at IEC-60825-1) for 400-700 nm for time of 10 to 1000 sec is 2000W/m^2 which is 2mW/mm^2. In order to get the MPE power this number need to be multiplied by appropriate aperture (Table 7 at IEC-60825-1) which is 3.5 mm diameter for skin. This makes the  power is 2 [mW/mm^2] x 9.6 [mm^2] =~ 20mW

Now, this is assuming that the safety level for skin can be translated to in-vivo behavioral experiments. This may not be so. My guess is that brain tissue would have a lower limit. But 20mW matches what I have heard from researchers as possibly being too much. Here is a quote from an email I received from one of our customers about a similar topic.

>20 mW at cannula sounds a little extreme to me- is it for rodents? We typically see good behavioral responses with 1 mW and at those high power intensities; I would be worried about tissue damage. We have also seen a behavioral effect with the Arch light that we purchased recently- that is likely also around 1 mW

Regarding the difference between low NA and high NA fibers, some researchers have told me that lasers are better for experiment that target axons, since the lower NA fiber shines the light more directly at the target axon.

About this our engineers say:

This may be correct in cases where the target is just ~100 micron in front of the implant tip. This is not the case 99% of the time since it is nearly impossible to implant with such fine resolution.

Practically the target cells sits in some tissue structure with a volume of at least few mm^3, and the implant tip is inserted as close as possible to this structure so the light from tip will reach the structure with enough power per volume to activate the opsins. The characteristic of the opsin to respond to specific wavelength enables selectivity of activation of specific structure. The reason why people ask for high power is that they are afraid the total amount of photons that will reach the structure will not suffice to activate opsin. After 100 or 200 micron the NA of the fiber or coherence effects have no relevance to power density due to scattering of the light through the tissue.

People using laser normally tend to use all available power. In most cases it is not so simple to control laser power during pulsed use, so in practice people using laser use higher power at implant tip than LED users, who have precise control of power during each pulse. Therefore lasers are more prone than LEDs to create a problem of heating of tissue.

I would posit that the amount of light needed will decrease as the accuracy of the fiber placement increases.

When specifically asked if low NA or High NA fibers will deliver more light to the target cells, Prizmatix engineers replied:

Light penetration in tissue is affected by the two major parameters: the scattering coefficient and absorption coefficient of the tissue. Generally speaking the result is an exponential decay in the intensity of light with the depth. See some explanations in the first part of Steve Jacques’s paper: http://omlc.ogi.edu/~jacquess/papers/Jacques_PMB2013_review_opticalProperties.pdf .

In a nutshell, the brain is a turbid media and thus the light scatters immediately when injected into the tissue, regardless of the angle of injection, and a “cloud” of photons will be formed around the fiber’s tip. It’s a “random-walk” effect, thus lower NA will not help you guide the photons better.

The absorption is highly dependent in the wavelength and in the components composing the tissue bulk, so coherence or non-coherence has nothing to do in this aspect. Yet, coherent light will result in speckles (thus non stable illumination at certain point) while non-coherent light won’t.

The power of the light source is related to the number of photons injected to the tissue at a given time. When more photons are injected to the tissue, the probability that a random-walking photon will survive to reach targets with enough energy to carry out its Optogenetic mission, before it is being absorbed, is higher.

So –there we have it. To sum up -

Q.  At what power level does heating become a issue? 

A.  Best to use the lowest power that activates as possible At 20mW there is for sure reason to worry about heat damage to the tissue. 

Q.  Is there a difference regarding heating when using a laser with low NA fiber or LED with high NA fiber? 

A. No, not when used with a fiber 

Q. Do either lasers or LED have an advantage in terms of successful experiments or heating concerns? 

A. Since LED power during pulsed operation can be precisely controlled, LEDs should be less problematic. Regarding hitting the target cell, for most experiments with the cell body as the target, there is no advantage to using lower NA fibers. For experiments using small targets, such as axons, when fiber is placed with extreme precision, a low NA fiber may have an advantage.

Thanks to Eli Pewzner and Assaf Deutsch for their input.

More information on lasers vs. LEDs can be seen in our white paper “Light Sources for Optogenetics Experiments”.

It would be great to hear your comments on this interesting topic. Please use the comments feature to contribute to this discussion.