4. Quantifying non-associative learning

Habituation and sensitization are two well-known forms of non-associative learning, which change the response probability of an animal towards a stimulus following repeated exposure to the same stimulus (habituation) or exposure to a strong yet different stimulus (sensitization). Habituation is a response decrement to a monotonously repeated stimulus (Braun and Bicker, 1992), while sensitization implies that a strong or particularly salient stimulus enhances the response of the bee to a test stimulus (Menzel et al., 1991).

A honey bee extends its proboscis when its antennae are stimulated with a sucrose solution whose concentration exceeds its sucrose response threshold (see section 2.2.). When low-concentrated sucrose solutions are repeatedly applied to the antennae of a bee with a short interval between two stimulations ("inter-trial interval"), the proboscis extension response habituates (Braun and Bicker, 1992; Scheiner, 2004). The bee no longer shows proboscis extension when its antennae are stimulated with this sucrose stimulus.

To perform a habituation experiment, adapt the following standard protocol to your needs:

1. Harness the bee (see section 2.4).

2. Test for each bee whether it shows the PER to antennal stimulation with 1 % sucrose.

3. Stimulate both antennae of each bee repeatedly with 1 % sucrose (for example 30 times), leaving a break of 3 seconds in between the stimulations ("inter-trial interval").

Take care not to stimulate the extended proboscis with sucrose solution.

4. For each individual, count the number of trials that the bee needs to habituate the PER, i.e. count how often the bee shows the PER before it stops extending its proboscis.

Alternatively, calculate a habituation score (i.e. the sum of PERs during the 30 habituation trials). You may also calculate response frequencies over all trials across groups.

5. Compare scores between two groups using T Tests or non-parametric equivalents, if necessary. More than two groups can be compared using analysis of variance (ANOVA) or non-parametric equivalents with respective post-hoc tests. Response frequencies can be compared using Chi square tests or multiple-measurement ANOVA (for further details see the BEEBOOK paper on statistical methods (Pirk et al., 2013)).

One disadvantage of this method is that it is difficult to interpret why a bee has stopped responding. Because the researcher does not know whether a bee has become exhausted of extending its proboscis, whether it really habituated or whether it may show peripheral adaptation, a dishabituation test is necessary at the end of the experiment. Here, a high-concentrated sucrose solution (30 % or 50 % for example) is applied to the antennae. If the bee shows dishabituation, it responds again both to the high-concentrated sucrose solution and to a subsequent stimulation with the low sucrose concentration used for habituation. Therefore, you need to add these steps to your protocol:

5. Apply a 30 % sucrose solution to both antennae 3 seconds after your final habituation trial.

6. Three seconds later, test whether the bee shows the PER to the 1 % sucrose solution used in the habituation session.

7. Only compare habituation data of bees which responded in the tests for dishabituation.

The course of habituation is determined by individual gustatory responsiveness and the sucrose concentration used as habituating stimulus. Bees with low gustatory responsiveness and satiated bees habituate faster than bees that are more responsive to gustatory stimuli (Braun and Bicker, 1992; Scheiner et al., 2004). High-concentrated sucrose concentrations lead to slower habituation than low sucrose concentrations (Scheiner, 2004).

This paradigm is very simple and allows a high throughput of bees in a short time. For that reason, habituation tests have been used repeatedly to study effects of pesticides or viruses on honey bee behaviour. Guez et al. (2001) showed, for example, how the insecticide imidacloprid affected habituation of the proboscis extension response. Iqbal and Müller (2007) demonstrated that DWV infection has no effect on habituation in honey bees and Kralj et al. (2007) analysed whether infection with Varroa mites affects habituation.

Sensitization can be easily tested in honey bees using an odorant and a sucrose stimulus. Most bees do not show spontaneous proboscis extension when their antennae are stimulated with an odorant such as carnation, citral or geraniol. When their antennae are briefly stimulated with a high-concentrated sucrose solution (30 % or 50 % for example) and with an odorant immediately afterwards, the bees become sensitized to the odorant (Menzel et al., 1993; Hammer et al., 1994). Now they display proboscis extension in response to antennal stimulation with the odorant which previously did not evoke a visible response. Note that sensitization to an odour is different from classical olfactory conditioning (see section 4.), because: (1) It does not require repeated pairings of odour and sucrose solution; (2) It only occurs within up to two minutes after olfactory stimulation.

Satiated bees can also be sensitized to gustatory stimuli which do not elicit a PER such as low sucrose concentrations (1 %) or water. Normally, satiated bees rarely extend their probosces when their antennae are stimulated with these solutions. But when their antennae are briefly stimulated with a high sucrose concentration and subsequently with a low sucrose concentration, bees can become sensitized to the low sucrose concentration and now respond to it. To analyse different groups, compare the response frequencies to the sensitized odour or low sucrose stimulus using Chi square tests  (for further details see the BEEBOOK paper on statistical methods (Pirk et al., 2013)).

Similar to habituation, the degree of sensitization depends on the gustatory responsiveness of the bee and the strength (i.e. the sucrose concentration) of the sensitizing stimulus (Menzel et al., 1993; Scheiner 2004). Sensitization to odours or low sucrose concentrations is a very effective and high-throughput assay, which allows the experimenter to draw conclusions on the non-associative learning abilities of a bee or a population of bees and which includes different stimulus modalities (i.e. gustatory and olfactory).