Assessing spinosad effect on honey bee olfactory conditioning using a microcontroller-based device

Honey bee health is adversely affected by numerous biotic and abiotic stressors, such as the extensive use of pesticides, which reduce survival and, even at sublethal doses, impair fitness and behaviour, including learning ability. The proboscis extension reflex (PER) is a widely used method to investigate associative learning in different insect species, such as Apis mellifera ligustica, when exposed to biotic and abiotic stresses. PER protocols require careful insect handling and accurate timing of stimulus application for effective conditioning and precise learning behaviour assessment. Despite the considerable amount of literature on this topic, information on automatising PER protocols for evaluating a large number of insects is limited. In this study, under laboratory conditions we evaluated whether exposure to different lethal concentrations (LC_0.2 and LC₂₀) of the natural bioinsecticide spinosad affects olfactory learning in groups of twenty worker bees within the same experimental cycle. For this purpose, we developed and validated an automated, microcontroller-based open-source device designed to standardise key steps of the PER conditioning protocol. Spinosad is a neurotoxic molecule that is known to induce mortality in honey bees, as well as behavioural impairment on locomotion and orientation at low doses. Surprisingly, possible effects on associative learning evaluated through PER are unknown. The sequential steps of the PER protocol were mechanised using a rotating carousel, where the conditioned stimulus was automatically released. The operator manually applied the unconditioned stimulus to the antennae, provided the rewarding solution, and recorded whether each insect exhibited PER behaviour. Bee responses were instantly transmitted to a laptop using a basic serial data transfer protocol and revealed decreased PER response (about 25%) for both spinosad concentrations at 20 min after the last conditioning trial. After small adaptations and validation, we expect that the device can be adopted both in laboratory and field conditions to investigate PER response in other bees, e.g., wild pollinators and bumblebees. This work aligns with current attempts to develop novel tools for monitoring biopesticide effects on pollinator health in changing environments.