Cognition
If there is one theme of research in animal information processing and decision making, it is that insects qualitatively seem to have (almost?) all the cognitive skills we find in humans (see also youtube lectures on intelligence). We contribute by showing that bumble bees assess reliability of different information channels and use them accordingly (visual vs olfactory, social vs personal).
However, we also show that innate predispositions contribute a large amount to what we think of as ‘intelligent’ behavior; and indeed that information is costly, particularly sampling large sets of stimuli or learning large numbers of cues, and thus may also be ignored altogether to increase speed and thus efficiency of foraging.
Speed vs Accuracy
Accuracy, i.e. using information or even learning, trades off with speed; and bees care about the cost of errors.
Chittka, L, Dyer, A, Bock, F, Dornhaus, A 2003 ‘Bees trade off foraging speed for accuracy’, Nature 424: 388 - pdf - some individuals are consistently faster but more error-prone; if errors are costly, all bees distinguish colors more accurately
also see Franks et al. 2003 below for group-level speed vs accuracy
Complex tasks and innovation
Some tasks improve more with learning than others; and tasks that give no rewards until after learning may depend on innate persistence to ever be learnt.
Muth F, Keasar T, Dornhaus A 2015 ‘Trading off short-term costs for long-term gains: how do bumblebees decide to learn morphologically complex flowers?’, Animal Behaviour 101: 191-199 - pdf - bees learned a complex task if they had an innate preference for the flower color; if not, they did not persist enough to be able to overcome initial failure
Leonard AS, Brent J, Papaj DR, Dornhaus A 2013 ‘Floral nectar guide patterns discourage nectar robbing by bumble bees’, PLoS One 8: e55914 - pdf - if you make legitimacy easier, more people stick to it - simple
Barker J, Dornhaus A, Bronstein J, Muth F 2018 ‘Learning about larceny: experience can bias bumble bees to rob nectar’, Behavioral Ecology and Sociobiology 72:68 - pdf - learning may improve efficiency of some tasks more than others; and bees tend to repeat behaviors they know
Meta-Learning
Bumble bees learn both how reliable cues are and how often the environment changes - hierarchically a level ‘up’ from associative learning, i.e. perhaps ‘learning when to learn’.
Dunlap AS, Nielsen ME, Dornhaus A, Papaj DR 2016 ‘Foraging bumble bees weigh the reliability of personal and social information’, Current Biology 26: 1195-1199 - pdf - if information such as flower color is reliable, it is used - otherwise ignored; but if bee-or-no-bee on a flower is reliable, it is the preferred cue, if not, it is actively avoided (against optimality, everything else being equal) - so the reliability of social information is overvalued by bees
Dunlap A, Papaj D, Dornhaus A 2017 ‘Sampling and tracking a changing environment: persistence and reward in the foraging decisions of bumble bees’, Interface Focus (Royal Society Journal) 7: 20160149 - pdf - bees measure environmental stability, i.e. how often the signal-reward pairing changes, and react to faster change by more frequent sampling
complex stimuli
Unsurprisingly, additional informative signal components can help learning or signal robustness; however, signal elements that to not help the distinction task also may set context and thus help recall (and thus improve learning/certainty).
Salient signal components (or those that receivers have innate predisposition to attend to) can overshadow (prevent learning of) other signal components, and this is reward-specific (e.g. only for positive/sugar reward). This can lead to imperfect mimicry being quite successful.
And finally, we stereotype: individuals stick to processing key traits instead of all signal components if signals are too complex/diverse, even at the cost of making errors.
Kikuchi DW, Dornhaus A, Gopeechund V, Sherratt TN 2019 ‘Signal categorization by foraging animals depends on ecological diversity’ ELife e43965 - pdf - using humans as ‘foragers’ (in a digital visual search task), we show that diversity and complexity of signals leads individuals to stereotype, i.e. focus on a single somewhat predictive aspect, ignoring the others (presumably to save time and cognitive investment). Many implications, one is that mimicry may be easier to evolve in complex stimulus settings
Kikuchi DW, Dornhaus A 2018 ‘How cognitive biases select for imperfect mimicry: a study of asymmetry in learning with bumblebees’, Animal Behavior 144: 125-134 - pdf - predispositions of receivers affect interpretation and learning of signals: here, blue overshadows (prevents) learning of other cues when paired with reward, so ‘mimics’ only needed to match the blue not other signal parts; blue did not overshadow on stimuli paired with negative outcomes. Costs of errors surprisingly had no effect.
Leonard AS, Dornhaus A, Papaj DR 2011 ‘Flowers help bees cope with uncertainty: signal detection and the function of complex floral signals’, Journal of Experimental Biology 214: 113-121 - pdf - using a signal detection paradigm, we show that bees are less uncertain about color when scent is also present, even if the scent is not informative; possibly the second signal component (scent) helped recall by setting context
Kulahci IG, Dornhaus A, Papaj D 2008 ‘Multimodal signals enhance decision-making in foraging bumble-bees’ Proceedings of the Royal Society: Biological Sciences 275: 797-802 - pdf - bees learn faster and are more accurate if given visual+olfactory cues of reward; higher accuracy does not come at a cost to speed when a modality is added
Kaczorowski RL, Leonard AS, Dornhaus A, Papaj D 2012 ‘Floral signal complexity as a possible adaptation to environmental variability: a test using nectar foraging bumble bees’, Animal Behaviour 83: 905-913 - pdf - if signals are multimodal, information in one modality may rescue performance when the other modality is affected by environment (e.g. visual signals under low light)
Leonard AS, Dornhaus A, Papaj DR 2011 ‘Why are floral signals complex? An outline of functional hypotheses’, in: Evolution of Plant-Pollinator Relationships. Patiny, S. (ed.) Cambridge University Press - pdf - review of adaptive effects of multi-modal stimuli, specifically flowers on pollinators
Leonard AS, Dornhaus A, Papaj DR 2011 ‘Forget-me-not: complex floral signals, inter-signal interactions and pollinator cognition’ Current Zoology 57: 215-224 (invited contribution to special issue on complex signaling) - pdf - review of adaptive effects of multi-modal stimuli, specifically flowers on pollinators