The trial-spacing effect in olfactory patterning discriminations in honeybees.

Harnessed bees conditioned to associate odors and sucrose reward learn to discriminate between olfactory mixtures and their odor components in negative (NP: A+, B+, AB-) and positive (PP: A-, B-, AB+) patterning experiments. They thus extend the proboscis to the reinforced (CS+) but not to the non-reinforced (CS-) stimuli. Using the same protocol, we studied whether or not trials, which are spaced in time, are more effective in supporting patterning discrimination than massed trials which succeed fast to each other ('trial-spacing effect'). Training followed a NP (4 A+, 4 B+, 8 AB-) or a PP (4 A-, 4 B-, 8 AB+) schedule, with a 1:1 ratio between CS+ and CS- trials (8 CS+ and 8 CS- trials). ITIs of 1, 3, 5 and 8min were used in both tasks. Increasing ITI resulted in better differentiation between reinforced and non-reinforced CSs in both NP and PP tasks. However, whereas only the longest ITI of 8min allowed discrimination in NP, PP could already be solved with an ITI of 5min. This difference might be due to the fact that NP, but not PP, would require the formation of a unique cue and thus longer processing times. We thus show that the trial-spacing effect, previously demonstrated for single stimulus conditioning, also determines performance in patterning tasks in which three different stimuli (A, B, AB) alternate so that elements have to be discriminated from their compound.