Bumblebees produce distinct facial behaviours in response to pleasant and aversive tastes — behaviours that, according to a new study in Proceedings of the National Academy of Sciences, parallel the hedonic responses documented in mammals.

The research, a collaboration between Macquarie University in Australia and Southern Medical University in China, used slow-motion video to record bees as they were presented with five solutions: 60% sugar, 20% sugar, plain water, 5% salt, and quinine at 1 millimolar concentration. After tasting sweet solutions, bees extended their glossa and continued licking — a behaviour the authors term "post-consumption glossa." After encountering salt or quinine, the bees shook their heads and wiped their mouths. The researchers characterise these as analogues to the "liking" and "disliking" facial action patterns described in the primate and rodent affective neuroscience literature, per The Guardian.

The critical methodological move was ruling out a hardwired chemosensory reflex. The team ran trials across 18 colonies under manipulated conditions — heat stress, satiation states, and graded drug doses. Under heat exposure, bees shifted their response to water or mildly saline solutions from neutral or aversive to positive. That context-dependence is what the authors say distinguishes the responses from simple reflexive tropisms: the animal's internal state modulates the valence of a stimulus, a defining feature of affect-like processing.

Professor Andrew Barron of Macquarie University, one of the paper's co-authors, put it plainly: "It tells us there is an inner life to the insect."

Associate Professor Thomas White, an entomologist at the University of Sydney who was not involved in the work, called the study notable for directing attention toward positive affective states in invertebrates — pleasure and preference — rather than the pain and fear responses that have dominated the field's welfare debate.

Why the Methodology Matters

The distinction between reflex and affect is not semantic. In regulatory and ethical frameworks, reflex-based nociception and genuine affective experience carry very different weight. The invertebrate welfare debate has largely turned on whether an animal's aversive reaction involves subjective suffering or simply a protective withdrawal circuit — the latter requiring no inner experience at all.

What makes the Macquarie/Southern Medical study potentially useful to that debate is its positive-affect arm. Most prior invertebrate research has probed nociception or stress. Documenting that bees show context-sensitive, state-modulated approach behaviours toward rewarding stimuli — and that the geometry of those behaviours maps onto mammalian hedonic-response taxonomies — adds a dimension that is harder to dismiss as mere reflexology.

The cross-species comparison to primates and rats is doing real theoretical work here, not decoration. Those mammalian facial-action patterns are associated, in the literature, with activation of the nucleus accumbens and opioid signalling — substrates that insects do not share in homologous form. The researchers are therefore making a functional analogy, not a claim of mechanistic equivalence. That is a defensible position, but it is also the point where critics will push back hardest: analogous outputs from non-analogous hardware do not settle questions of phenomenal experience.

What Comes Next

The study does not resolve whether bumblebees are sentient in any philosophically robust sense, and the authors do not claim it does. What it does is move the evidentiary baseline. If context-dependent hedonic-response patterns can be reliably elicited and quantified in a species with roughly one million neurons, the methodological toolkit for probing insect inner states becomes considerably more tractable.

For applied entomology and pollinator-welfare policy, the practical questions follow quickly. Commercial bumblebee colonies are already a regulated commodity in several jurisdictions for their role in greenhouse pollination. Evidence of state-dependent affective processing — even functionally defined — will likely be cited in future debates over colony transport conditions, pesticide exposure thresholds, and captive-rearing standards.

The broader scientific interest, though, is phylogenetic. Bees and vertebrates diverged roughly 600 million years ago. If core features of hedonic valuation evolved independently in lineages that separated so early, that suggests either strong convergent selection pressure for affect-like processing, or that precursor mechanisms are more ancient and more widespread than the vertebrate-centric tradition of consciousness research has assumed. Neither possibility is small.