My lab performs lots of memory experiments: both Classical (Pavlovian) Conditioning experiments where a cue is presented at the same time as a reward, and Operant (Instrumental) Conditioning experiments where the flies need to perform a task in order to receive a reward. These memory experiments involve responses to cues, in order to best understand memory retrieval and provide a quantitative response to how rewarding flies found the reward we are studying. The most common cues used in memory experiments are odors.
An impressive amount of information is known about how odors are encoded in Drosophila, especially the anatomy of the larval and adult mushroom body system (Eichler et al, 2017; Li et al, 2020). Odors are encoded through the antennal-lobe (Castro et al, 2023; Lazar et al, 2023)->lateral horn (Das Chakraborty & Sache, 2021) and mushroom body (Modi et al, 2020)->mushroom body output neuron circuit (Aso & Rubin, 2024). This is one of the main reasons why odors are often used as cues for memory experiments. There are a plethora of tools to understand the mechanisms through which odors are encoded in the fly brain.
When designing a memory experiment, your focus is typically the response to a cue (also called a conditioned stimulus) that changes as a consequence of a punishment or reward (also called an unconditioned stimulus). This means to maximize your understanding of how the reward/punishment influences learned response, you need an odor cue that doesn’t have a strong innate preference. The problem is of course, that odors always seen to have an innate valence. That is, flies will demonstrate preference or aversion to an odor, depending on the concentration and situational context (Siju et al, 2020; Eschbach et al, 2021; Das Chakraborty et al, 2022; Wu et al, 2022). So basically, there is no optimal neutral odor for learning and memory experiments and preference for an odor can be highly variable and depend on prior odor experience, behavioral arena type, intensity of odor and length of odor exposure.
A lot of experiments use 3-octanol (OCT) and 4-methylcyclohexanol (MCH), for both shock-odor and sugar-odor memories. I’m not entirely sure why the field ended up using these odors so widely but I suspect it’s largely historical and can be traced back to the original Quinn et al, 1974 countercurrent memory study. Masek & Heisenberg (2008) did a careful characterization of how odor concentration impacts memory, and used 3-octanol (OCT), benzaldehyde (BAL), isoamylacetate (IAA) and amylacetate (AM).
My lab studies reward, and so when I first characterized a new reward assay for alcohol (Kaun et al, 2011), I tried a number of different odors paired with alcohol before finding ones that provided the highest aversive and appetitive memory for alcohol. I never published this characterization, but basically I found memory (aversive or appetitive, Nunez et al, 2018) was poor with OCT and MCH, and higher with more ethologically relevant odors (acetic acid, isoamyl alcohol, ethyl acetate and isoamyl acetate). I used a variety of sources to find the odors I experimented with for characterizing alcohol memory and found one of the most helpful sources to be an older paper looking at attraction to a panel of chemicals using a simple olfactory trap (Kikuchi, 1973) and comparing known fly-responsive odors to those found in fermenting fruit (for example in fermenting grapes Loscos et al, 2007). You might be surprised at how much information there is out there from the wine and food industry on food and fermentation-related odors.
So what odors should you use to maximize response? My recommendation is to start with odors that are really well characterized (like OCT and MCH) and see if you can get the type of memory/behavior you are interested in. If not, then think about the assay you are developing and the desired outcome, then dive into the literature to find odors with responses that are flexible (ie preference can go up or down). If you are using an odor that isn’t super well characterized, make sure you run some control tests looking at preference for that odor in different concentrations and contexts (for example if flies are hungry vs satiated) before running all the memory tests.
Make sure if you are designing an odor-memory assay that you design a ‘balanced’ assay. By this I mean that you need to make sure the response you are seeing isn’t just because flies have an innate preference for that odor. Maybe you look at preference before versus after pairing the odor with a reward or punishment. Although in this case you need to worry about whether the preference for that odor changed just because of repetitive odor exposures or because it was paired with a reward/punishment, so you need to run an odor-only control alongside it to test for this.
Alternatively, you can design a 'reciprocal' assay with two different odors and have one group of flies getting exposed to an unpaired odor (Odor 1) and then a paired odor (Odor 2) and another group of flies in which the unpaired odor is Odor 2 and the paired odor is Odor 1. If this was an odor-sugar memory experiment, you would expect the first group of flies to go towards odor 2 and the second group of flies toward odor 1. You calculate a preference index (PI) for each group (Paired-Unpaired)/Total, then average these two PI’s for a conditioned preference index (CPI) or learning index (LI). This way you control for innate odor preferences and have a ‘balanced’ assay. This is the typical way memory studies are designed in flies, but not in rodent models. I think it’s because you need a higher sample size to do these types of balanced experiments and that gets expensive when working with mice or rats.
Finally - a couple last notes of warning:
Make sure to pay special attention the isomer and purity of the odor chemical you are ordering for experiments because these can change the smell.
Make sure to store your odors in glass jars because some of these chemicals erode plastic
Make sure to store your odors in the dark because light can cause chemical changes
Make sure each person makes their own working stock because you don’t want everyone in the lab using the same bottle: higher chances of contamination
I recommend buying the smaller bottles, as opposed to the bigger bottle that can last the lifetime of the lab because the odor quality can change over time.