Optogenetic olfactory behavior depends on illumination characteristics

Optogenetics has become an important tool for the study of behavior, enabling neuroscientists to infer causations by examining behavior after activating genetically circumscribed neurons with light. Light-induced neural activity is affected by illumination parameters used in experiments, such as intensity, duration, and frequency. Here, we hypothesized that the intensity of light and the presence of oscillations in illumination would alter optogenetically induced olfactory behaviours. To test this, we activated olfactory receptor neurons (ORNs) in Drosophila by using either static or pulsed light stimuli across a range of light intensities. The various regimes elicited distinct behavioral valence responses (attraction, aversion, indifference) from several ORN types. Our results demonstrate the importance of both frequency and intensity for interpreting optogenetic behavioral experiments accurately; successfully generalizing optogenetic results requires the use of more than a single illumination regime.


Fly care and strains
Flies were prepared as previously described (Mohammad et al., 2017) . Briefly, flies were kept in the dark on a regular fly medium (Temasek Life Sciences Laboratories, 2018) , and transferred on 0.5 mM all-trans-retinal (Sigma-Aldrich, USA) supplemented food two days prior to optogenetic experiments. Cantonized w 1118 flies were used as wild type.

Optogenetic behavior apparatus
An optogenetic apparatus was designed somewhat similar to one previously described (Mohammad et al., 2017) . A

Triggering valence with optogenetic activation
Artificial activation of single classes of  Flies are loaded in a darkened arena and given a choice between red light and dark.
B. An example trace of a fly that spends most of its time in the dark zones after encountering red light (blue arrows). Red boxes indicate when the fly is walking through red illumination.
Artificial activation of Gr21a neurons is sufficient to elicit avoidance (Suh et al., 2007) . Here, we tested Gr21a avoidance by optogenetically activating these neurons with either static or pulsed light. While continuous photoactivation was aversive only at the lowest light intensity (4.65 μW/mm 2 , Fig 2A), pulsed light triggered aversion at all three intensities (Fig 2B). This dramatic behavioral difference between static-and pulsed-light activations indicates that the presence or absence of oscillations can alter the behavioral effect of optogenetic actuation ( Fig 2C). μW/mm 2 , Fig 3A), whereas pulsed light did not trigger any behavioral response (Fig 3B). The discrepancy between the two types of light stimuli is largest at the highest light intensity with a ΔwTSALE of -0.22 [95CI -0.35, -0.09] (Fig 3C).

D.
Static-light stimulation of the Or85c neurons does not elicit any valence response.

E.
Pulsed-light stimulation of the Or85c neurons is aversive at the two lower intensities.

Discussion
Optogenetic temporal structure influences

olfactory-related behavior
Vinegar and CO 2 are important cues for Drosophila to find food and sense stress (Jones et al., 2007;Semmelhack & Wang, 2009;Suh et al., 2004) . but not static light (Fig 4A-F). Interestingly, all of the olfactory responses induced by Or59c and Or85c neurons occured at the lower light intensities (Fig 4A-B-E), and did not increase as the light intensity increased.
Of the four ORNs tested in the present study, static-light stimulation performed better in eliciting olfactory responses than the pulsed light for two ORNs (Or92a,  Pulver et al., 2009) , and growing evidence now shows that light parameters also change optogenetically induced behaviours.

Should optogenetic activation mimic natural activity?
The extent to which optogenetic activation needs to mimic natural neuronal activity to recreate natural behaviors is not clear (Malyshev, Goz, LoTurco, & Volgushev, 2015;Miesenböck, 2009 inform generalizable conclusions about a neural system, a single optogenetic regime will often be inadequate. As there is no one-size-fits-all guide for choosing light parameters, using multiple intensities and frequencies will be prudent when characterizing systems with optogenetic activation.

Conclusions
Optogenetics has been extensively used to study behavior; however, the extent to which illumination protocols influence experiment results has not been addressed thoroughly. In the present study, we investigated the hypothesis that different light frequencies and intensities would induce distinct olfactory behaviours. Our results support this hypothesis: same ORNs drove different valence responses to static-or pulsed-light activations; notably, none of the ORNs triggered contrasting valence responses (attraction vs. avoidance), but rather inconsistent results (valent vs. non-valent). We propose that manifold illumination regimes must be utilized to obtain generalizable results in behavioral studies.