The problem the researchers set out to solve was one of fit. Existing imagery questionnaires tend to be modality-specific, probing only how vividly people visualise objects or conjure smells in isolation, or they rely on generic prompts that were never designed with eating in mind. Food, however, is an inherently multisensory experience. A single bite of chocolate recruits orthonasal smell as the aroma drifts up from the plate, retronasal smell as volatile compounds travel from the back of the mouth, basic taste qualities such as sweetness and bitterness, the mechanical feel of texture, and the visual appearance of the food itself. A measure that captures only one of these channels risks missing the very individual differences that matter most for eating behaviour research, from craving susceptibility to appetite control.

The project began with a preliminary item-generation study designed to identify foods and drinks with distinct, recognisable sensory characteristics. The goal was to source items that would reliably evoke each sensory channel, so that a participant rating their mental image of brewed coffee, for example, would be engaging the olfactory system rather than guessing at a vague, undifferentiated food memory. The final scale asks respondents to imagine 25 items across five domains: smell, flavour, taste, texture and appearance. Each item is rated from zero, meaning no image at all, to ten, meaning an image as clear and vivid as real life. The published version includes prompts such as imagining the smell of freshly baked bread and cinnamon rolls, the flavour of parmesan cheese and dark chocolate, the bitterness of coffee, the sourness of a lemon, ice cream melting in the mouth, and the look of a red apple or a scoop of ice cream.

With the item pool in hand, the team subjected the scale to a four-experiment validation programme involving well over a thousand participants in total. Experiment 1, with 354 participants, used exploratory factor analysis and parallel analysis, two statistical techniques for uncovering the latent structure underlying a set of questionnaire items, to refine the pool and examine internal consistency. The results supported a correlated five-factor solution, meaning the data were best explained by five related but separable dimensions corresponding to the five sensory channels. The fit statistics were strong: a root mean square residual of 0.023, a root mean square error of approximation of 0.065, and a comparative fit index of 0.962, alongside high internal consistency, indicating that items within each factor measured the same underlying trait coherently.

Reliability over time is just as important as internal structure, and here the team turned to Experiment 2. Eighty-three participants completed the scale twice, three to four weeks apart, allowing the researchers to compute a test–retest reliability coefficient. The total score achieved an intraclass correlation of 0.75, with a 95 percent confidence interval ranging from 0.62 to 0.82. In practical terms, this means a person’s standing on the scale is reasonably stable over a month, which is precisely what one would expect if the instrument is measuring a genuine trait, a stable disposition of the mind, rather than a fleeting mood or situational appetite. For a tool intended to characterise individual differences in eating-related cognition, that temporal stability is a foundational credential.

Experiment 3 then put the five-factor structure to a stricter test. Using a new sample of 393 participants, the researchers applied confirmatory factor analysis, a method that tests a predefined structure rather than letting one emerge from the data, and examined correlations with established questionnaires of visual and olfactory imagery vividness. The five-factor structure replicated cleanly, with a chi-squared statistic of 503.15 on 265 degrees of freedom, comparative and Tucker–Lewis fit indices both at 0.998, RMSEA of 0.048 and a standardised root mean square residual of 0.040. The convergent validity results were particularly elegant. The appearance factor correlated most strongly with visual imagery vividness, at a Spearman coefficient of 0.47, while the smell factor correlated most strongly with olfactory imagery vividness, at 0.51, both highly statistically significant. In other words, the food-specific channels behave as they should: they align with their general-domain counterparts without collapsing into them, suggesting the FIS captures something genuinely food-tuned.

The final and arguably most consequential experiment addressed the question that motivated the whole project: does food imagery vividness relate to actual eating-related tendencies? Experiment 4 recruited 781 participants and examined criterion-related associations with trait food craving, the dispositional tendency to experience intense desires for specific foods. After adjusting for covariates, the researchers found small but reliable positive associations for two of the sensory channels. Smell imagery vividness correlated with trait food craving at r = 0.110, significant at p = 0.002, and flavour imagery showed an association of r = 0.080, significant at p = 0.026. The effect sizes are modest, as is typical for single-trait predictors of complex behaviour, but their pattern is telling: the chemosensory channels, smell and flavour, the ones most intimately tied to the reward and pleasure of eating, are the ones that link to craving.

That pattern has intriguing implications. Flavour, as the scale clarifies for respondents, refers to the combined in-mouth experience arising primarily from taste and retronasal smell, distinct from the basic taste qualities of sweetness, sourness, saltiness, bitterness and umami. If people who more vividly imagine smells and flavours also report stronger trait cravings, vivid multisensory imagery may act as a kind of mental rehearsal of eating, priming desire before any food is present. This connects to a broader literature in which imagery is understood to function as a weak form of perception, recruiting overlapping neural machinery. It also opens practical questions for interventions: if imagery vividness shapes craving, techniques that modulate imagery, from vividness training to competing-task strategies that load the sensory systems and blunt craving episodes, could be tailored according to an individual’s FIS profile.

The authors are careful about the boundaries of what they have established. They note that future work should establish the scale’s incremental predictive validity, that is, whether it predicts eating outcomes better than general imagery measures already exist, and should test its cross-cultural generalisability, since foods like miso soup, soy sauce and roast dinners carry very different familiarity across populations, and food familiarity is known to shape both imagery and craving. The project was funded by a grant from the Royal Society of New Zealand through the Marsden Fund, with the design, conduct, analyses and interpretation conducted independently of the sponsor, and the work received ethics approval from the University of Otago Human Ethics Committee.

Even with those caveats, the arrival of a reliable, food-specific, multisensory imagery measure is a notable addition to the psychologist’s toolkit. For decades, eating behaviour research has had to borrow instruments built for other purposes, averaging across senses that food never treats separately. The FIS treats the imagined meal as what it actually is, a symphony of smell, flavour, taste, texture and sight, and offers researchers a way to ask not just whether a person thinks vividly, but whether their culinary imagination in particular runs hot. Given the scale’s documented psychometric credentials across four experiments and more than 1,600 participants, and its open-access availability, it seems likely to find its way into studies of craving, dietary choice, appetite and sensory science for years to come.

Subject of Research: Development and validation of a self-report scale measuring trait vividness of multisensory food imagery for eating behaviour research

Article Title: Developing and validating the Food Imagery Scale: a trait measure of multisensory food imagery for eating behaviour research

Article References: Peng, M., McCormack, J., Golo, S., Hames, M., Roberts, R., & Spence, C. (2026). Developing and validating the Food Imagery Scale: a trait measure of multisensory food imagery for eating behaviour research. BMC Psychology. https://doi.org/10.1186/s40359-026-05526-x

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05526-x

Keywords: mental imagery, food imagery, eating behaviour, food craving, psychometrics, sensory perception, flavour, olfaction, questionnaire validation, BMC Psychology, Developing, validating

The research team focused on the IR8a-IR75k complex, a receptor assembly in which the co-receptor IR8a partners with the tuning receptor IR75k to form a functional channel. Using cryo-electron microscopy, the researchers captured this complex in four distinct states: the ligand-free, full-length closed conformation known as the apo state; a ligand-induced, full-length open state; and two structures of the extracellular ligand-binding domains bound to nonanoic acid and octanoic acid, two fatty acid odorants of different chain lengths. This state series is what gives the study its power, because it does not merely freeze the receptor in one pose but documents the conformational journey from closed to open, allowing the researchers to trace exactly how chemical recognition becomes channel activation.

The structural analysis revealed that ligand specificity is determined within a dynamic binding pocket located in the ligand-binding domain of IR75k. Two features of the pocket govern which molecules are accepted: the length of the odorant’s alkyl chain and the geometry of its carboxylate group. In practical terms, the pocket is tuned to accommodate fatty acids of a particular size and shape, and a carboxylic acid must present its charged head group in the correct orientation to make productive contacts. This explains how a receptor family with broadly similar architecture can be diversified across the mosquito genome into sensors with distinct chemical tuning, each responding to a different slice of the odor landscape.

When a ligand does bind, the study shows, it triggers a precise mechanical event: the bilobed ligand-binding domain of IR75k closes around its cargo, like a Venus flytrap snapping shut. This closure is not an isolated local change. The researchers found that the conformational shift propagates through conserved mechanical linkers that connect the extracellular domain to the transmembrane domain, the region of the receptor that spans the neuronal membrane and forms the ion-conducting pore. The linkers function as a molecular lever, converting the small-scale movement of domain closure into a larger-scale rearrangement of the membrane-spanning helices.

That lever action is the heart of the gating mechanism. As the conformational cascade reaches the transmembrane domain, the helices undergo an outward displacement that dilates the pore, opening a pathway for ions to flow across the membrane and generate the neuronal signal. The study’s molecular dynamics simulations provided independent support for this mechanism: when the ligand-bound structure was embedded in a lipid bilayer and subjected to an external electric field mimicking physiological conditions, sodium ions were observed permeating the channel, some in a partially dehydrated state as they squeezed through the narrowest region of the pore. When the same simulation was run starting from the apo-state structure, no sodium permeation occurred, confirming that ligand binding is required to trigger the conformational cascade necessary for gating.

Electrophysiological experiments rounded out the evidence, functionally validating the structural models by showing that the receptor complex behaves as the structures predict. Together, the cryo-EM structures, simulations, and functional assays form a mutually reinforcing argument that the authors have correctly identified both the chemistry of ligand recognition and the mechanics of channel opening. The approach, combining multiple structural states with computational and functional validation, represents a template for studying other members of the ionotropic receptor family, whose gating mechanisms have been even harder to capture experimentally.

The significance of the work extends well beyond basic structural biology. Ionotropic receptors enable mosquitoes to detect the carboxylic acids and other volatile cues that mark human hosts, and this detection underpins host-seeking, the behavior that puts mosquitoes in contact with the people they infect with malaria and other pathogens. By providing an atomic-resolution view of how IR-mediated host detection works, the study identifies concrete molecular features, such as the ligand-binding pocket of IR75k and the linker regions that transmit its motion, that could be targeted by chemists seeking to disrupt mosquito olfaction without harming other organisms.

The prospect of rationally designed mosquito repellents or behavior-modifying compounds is a major motivation for this line of research. Traditional repellents such as DEET act through mechanisms that remain incompletely understood, and resistance concerns continue to drive demand for new approaches. A structure-based strategy, in which compounds are designed to either block the ligand-binding pocket, jam the gating machinery in a closed state, or otherwise interfere with the conformational cascade documented in this study, offers a more precise route to controlling host-seeking behavior. The demonstration that specificity depends on alkyl chain length and carboxylate geometry gives medicinal chemists clear design constraints to work within or deliberately violate.

There are also broader evolutionary and comparative implications. Ionotropic receptors are conserved across insects, and the Anopheles stephensi structures now provide a reference point for understanding how receptor tuning varies among mosquito species and between disease vectors and non-vector insects. Sequence alignments included in the study’s supplementary data show conservation of IR75k and IR8a across mosquito species, suggesting that the structural principles uncovered here may generalize, though each species’ receptor repertoire reflects its own ecological niche. Comparative structural studies building on this foundation could reveal how receptor evolution tracks changes in host preference and habitat.

Technically, the study also showcases the maturing capability of cryo-electron microscopy to resolve full-length, membrane-embedded ion channels in multiple functional states, a feat that was largely out of reach for such complexes a decade ago. Capturing both the apo and ligand-bound full-length structures, alongside high-resolution views of isolated ligand-binding domains with two different fatty acids, required careful biochemical preparation and computational reconstruction. The inclusion of a molecular dynamics simulation reliability and reproducibility checklist in the supplementary materials reflects the growing rigor expected for computational claims in structural biology. As these methods continue to advance, the mosquito ionotropic receptors characterized here will likely be joined by structures of the remaining receptor families that govern insect olfaction, completing a molecular atlas of how disease vectors smell the world, and how that sense might be turned against them in the effort to control mosquito-borne disease.

Subject of Research: Structural mechanism of ligand recognition and gating in mosquito ionotropic olfactory receptors

Article Title: Structural basis of ligand recognition and gating in mosquito IRs

Article References: Xu, L., Wang, X., Quan, X., Ferguson, S. T., Liu, D., Chen, J., Wang, S., Liang, S., Huang, G., Huang, K., Liu, F., & Pan, X. (2026). Structural basis of ligand recognition and gating in mosquito IRs. Nature. https://doi.org/10.1038/s41586-026-11239-3

Image Credits: AI Generated

DOI: 10.1038/s41586-026-11239-3

Keywords: ionotropic receptors, cryo-electron microscopy, Anopheles stephensi, olfaction, ligand-gated ion channels, mosquito host-seeking, carboxylic acids, molecular dynamics simulations, electrophysiology, IR8a, IR75k, malaria vectors

Chemosensation is the invisible language of insect life. From locating a host plant across a forest clearing to recognizing a potential mate by pheromone, nearly every critical decision an insect makes depends on its ability to detect, bind, and transduce chemical signals. These functions are carried out by families of proteins expressed in sensory organs, chiefly the antennae. Odorant-binding proteins, or OBPs, and chemosensory proteins, or CSPs, capture volatile molecules in the fluid-filled pores of sensory hairs and shuttle them to receptors embedded in nerve cell membranes. Odorant receptors, or ORs, ionotropic receptors, or IRs, and gustatory receptors, or GRs, then convert chemical binding events into electrical signals the brain can interpret. Sensory neuron membrane proteins, or SNMPs, assist in the detection of fatty acid-derived compounds such as pheromones, while Niemann-Pick type C2 proteins, or NPC2s, contribute to the transport of hydrophobic molecules. Until this study, not a single one of these gene families had been reported for R. cyatheae, leaving researchers with no molecular handle on how the wasp tracks its unusual fern host.

To build that handle, the team assembled a transcriptome database from male and female adult wasps, sequencing the complete set of messenger RNA molecules being expressed in their bodies. The effort yielded 30,296 unigenes, the distinct assembled sequences representing the insect’s active genes, with an N50 length of 3,286 base pairs, a measure indicating that half of the assembled sequences were longer than that threshold and reflecting a high-quality assembly. The researchers then compared these sequences against six major public annotation databases: NR, Swiss-Prot, Pfam, eggNOG, GO, and KEGG. In total, 11,109 unigenes, or 36.67 percent of the full set, could be functionally annotated. The NR database proved the most productive, matching 10,774 unigenes, while the KEGG pathway database annotated the fewest, 6,300. This annotation step is essential because it allows researchers to distinguish genuine chemosensory genes from the thousands of other sequences involved in basic cellular housekeeping.

Screening the annotated transcriptome against known insect chemosensory gene families revealed 90 candidate chemosensory genes in R. cyatheae. The inventory includes 11 OBPs, 10 CSPs, 6 NPC2s, 24 ORs, 20 IRs, 15 GRs, and 4 SNMPs. Within the OR family, the researchers distinguished 23 typical odorant receptor genes and a single Orco gene, the highly conserved co-receptor that partners with typical ORs across virtually all insect species and is essential for odor detection. The numbers themselves tell an evolutionary story. Compared with insects that rely heavily on chemical communication, such as moths with dozens or hundreds of ORs, the wasp’s relatively compact receptor repertoire may reflect its narrow ecological niche, though the authors present the counts as a baseline for future comparison rather than drawing firm ecological conclusions.

To place these genes in evolutionary context, the team constructed phylogenetic trees relating the chemosensory genes of R. cyatheae to their counterparts in other insect species. Phylogenetic analysis groups genes by sequence similarity, allowing researchers to infer which genes are orthologs, descended from a common ancestor, and which may have expanded or contracted in particular lineages. Such trees also help flag genes that cluster with well-characterized relatives in other insects, providing hypotheses about function that can later be tested experimentally. For a species whose chemosensory biology had never been examined at the molecular level, this comparative framework is the first map connecting the wasp’s sensory genes to the broader landscape of insect olfaction research.

One of the most intriguing findings emerged when the researchers compared gene expression between male and female adults. Thirteen of the candidate chemosensory genes were differentially expressed between the sexes, meaning their activity levels differed significantly. Sex-biased expression in chemosensory genes often points to roles in behaviors that differ between males and females, such as mate location through pheromone detection in males or host plant selection for oviposition in females. Because R. cyatheae infestations coincide with the sprouting period of its host fern, understanding which sex drives host finding, and which molecular sensors are involved, could reveal weak points in the infestation cycle that might be targeted by behavioral interventions or attractant-based monitoring tools.

To confirm that the transcriptome data accurately reflected biology rather than sequencing artifacts, the team verified the tissue expression profiles of the differentially expressed genes using reverse transcription quantitative polymerase chain reaction, or RT-qPCR. This laboratory technique measures the abundance of specific messenger RNA molecules with high sensitivity and is the standard method for validating transcriptome-wide expression patterns. The RT-qPCR results corroborated the expression differences observed in the sequencing data, strengthening confidence that the 13 sex-biased genes are genuinely regulated differently in males and females. This kind of validation is a critical quality checkpoint in genomics studies, where computational identification of candidate genes must be backed by independent experimental evidence before functional work can proceed.

The significance of the study extends beyond a single pest species. Tree ferns of the genus Alsophila are living fossils, and A. spinulosa is protected within the Chishui Alsophila National Nature Reserve, whose administration supported the fieldwork. Conservation of an endangered fern depends on managing the insects that attack it, and effective management in turn depends on understanding the chemical ecology of the pest. The gene catalogue now makes it possible to pursue questions that were previously unanswerable: which odorant receptors respond to fern volatiles, whether males detect female-produced pheromones, and whether synthetic lures could be designed to monitor or disrupt mating. Similar transcriptome-based approaches have enabled the development of semiochemical-based control tools for other forest pests, and the new data provide the raw material for analogous efforts against R. cyatheae.

The work also fills a taxonomic gap. Hymenoptera, the order that includes wasps, bees, and ants, is enormously diverse, yet chemosensory gene repertoires have been characterized for only a fraction of its members, and herbivorous hymenopterans that attack ferns are especially poorly studied. By documenting OBPs, CSPs, NPC2s, ORs, IRs, GRs, and SNMPs in a fern-feeding sawfly relative, the study offers comparative data that will help researchers trace how chemosensory gene families evolve as insects shift onto unusual host plants. The authors note that their findings lay a solid molecular foundation for future investigations of gene function and the mechanisms of olfactory perception in R. cyatheae.

From a technical standpoint, the study demonstrates a now-standard but powerful pipeline for working with non-model insects: assemble a transcriptome, annotate it against multiple databases, mine it for gene families of interest, place the candidates in phylogenetic context, and validate expression patterns experimentally. The resulting catalogue of 90 chemosensory genes, anchored by a high-quality assembly of 30,296 unigenes and confirmed sex-specific expression patterns, transforms R. cyatheae from a molecularly unknown pest into a tractable research system. For the endangered tree fern it threatens, that transformation may ultimately matter most: every gene identified is a potential target for understanding, predicting, and eventually managing the outbreaks that strike when the fern’s tender new fronds emerge. The research was funded by a project on pest resistance mechanisms of different Cyathea species based on multi-generation transcriptomes and protein metabolomes, and by a Science and Technology Innovation Talent Team project of Guizhou Province, reflecting sustained regional investment in protecting this unique fragment of prehistoric flora.

Subject of Research: Identification of chemosensory genes in the adult transcriptome of the tree fern pest wasp Rhoptroceros cyatheae

Article Title: Analysis and identification of chemosensory genes in the transcriptome of adult Rhoptroceros cyatheae

Article References: Zhou, M., Jiang, Y., Zhang, X., Wu, G., Peng, T., Liang, S., Liu, B., Yang, S., Chen, H., Liu, F., Sun, Q., Song, Y., & Yang, W. (2026). Analysis and identification of chemosensory genes in the transcriptome of adult Rhoptroceros cyatheae. BMC Genomics. https://doi.org/10.1186/s12864-026-13393-4

Image Credits: AI Generated

DOI: 10.1186/s12864-026-13393-4

Keywords: Rhoptroceros cyatheae, chemosensory genes, transcriptome, odorant-binding proteins, odorant receptors, Alsophila spinulosa, tree fern, olfaction, RT-qPCR, phylogenetic analysis, BMC Genomics, entomology

The central obstacle the team faced was the sheer complexity of the honey bee’s sense of smell. Bees rely on more than 200 odor receptors to navigate their world, detecting and responding to a vast array of volatile compounds in their environment. That sensitivity is what makes bees such effective foragers, but it also makes their olfactory behavior extraordinarily difficult to predict. Finding a scent that reliably repels a bee, rather than attracting it or doing nothing at all, has long been considered a major challenge. Anandasankar Ray, a professor of molecular, cell and systems biology at UC Riverside and an expert on insect olfactory behavior, led the interdisciplinary effort to overcome it.

“Bees rely heavily on their sense of smell to forage, but that sensitivity makes it tough to find odors that push them away instead of drawing them in,” Ray said. “Our goal was to flip that script and find a way to use scent as a deterrent — safely and effectively.” To pursue that goal, Ray’s team collaborated with honey bee researchers in the laboratory of Boris Baer, a professor of entomology at UC Riverside, combining expertise in computational modeling with hands-on knowledge of bee behavior.

The approach the researchers developed was a machine-learning model trained on two kinds of information: the chemical structures of odorant molecules and previously recorded behavioral responses of bees to those odorants. By learning the relationship between molecular features and behavioral outcomes, the model could begin to predict how a bee would respond to a compound it had never encountered. Crucially, the team did not stop with the initial training data. They refined the model with new behavioral data generated in their own laboratory, gathered from both honey bees and Drosophila, the fruit fly that serves as a workhorse of insect olfaction research. That iterative improvement step allowed the system to make increasingly accurate predictions of insect olfactory responses.

One of the most striking aspects of the work is that it succeeded without the enormous datasets usually assumed to be necessary for machine learning. “It is generally thought you need abundant data to do any kind of machine learning, but that’s not true for olfaction,” Ray said. “You simply need good-quality data and iterative improvement steps.” That insight has implications well beyond bee research, suggesting that carefully curated behavioral measurements can substitute for sheer data volume when the underlying structure of the problem is learnable. Once the model was optimized, the researchers put it to work on a truly industrial scale, screening a library of more than 50 million compounds. From that vast chemical space, the system identified roughly 130 candidates predicted to have strong potential as bee repellents.

Predictions, of course, mean little until they are tested against living insects. The team reports in eLife how they put the top-performing candidates through exactly that gauntlet. In laboratory assays, honey bees exposed to the candidate compounds exhibited clear avoidance behaviors, and those responses aligned closely with what the model had predicted. That agreement between computation and behavior is the heart of the study’s significance: it demonstrates that a model trained on limited but high-quality data can generalize to new molecules and correctly forecast how bees will react to them in the real world.

The validation did not stop in the laboratory. The researchers then moved to field experiments with freely foraging bees, a far more demanding test because wild conditions introduce wind, competing odors, and the full complexity of natural foraging behavior. All seven of the compounds tested in the field reliably repelled bees from honey combs, and importantly, they did so without harming the insects. A repellent that injured the very pollinators it was meant to protect would defeat its own purpose, so the demonstration that these odorants deter bees safely is a critical part of the result.

“This is a powerful demonstration of how machine learning can help solve real-world ecological problems,” Ray said. “By keeping bees away from harmful pesticides, we can potentially reduce their risk of exposure without compromising the protection of crops.” The logic of the approach is straightforward: instead of banning or restricting every pesticide that poses a risk to bees — a regulatory path that has already removed some of the most harmful products from use — farmers could blend repellent compounds into their pesticide applications. The crop would retain its protection against pests, while bees, sensing the deterrent odor, would simply choose to forage elsewhere. Unintended exposure to pesticides is widely considered a contributing factor to bee population decline and colony collapse, and minimizing bee contact with treated crops addresses that exposure directly.

The researchers also point to applications that extend well beyond agricultural spraying. “In certain public environments — hospitals, office buildings, and residential areas, for example — reducing the formation of beehives can help avoid human-bee conflicts,” Ray said. In those settings, a safe repellent could discourage bees from establishing colonies where they are unwelcome, without resorting to lethal control. There are agricultural uses as well: some farming practices, particularly with seedless fruit varieties, aim to avoid pollination altogether, and a reliable repellent could help keep bees away from crops where pollination is undesirable rather than essential. According to Ray, the research is a step toward developing bee-friendly pesticide formulations — products that safeguard pollinators while still meeting the needs of modern agriculture.

“Protecting pollinators doesn’t have to come at the expense of food security,” Ray said. “With the right tools, we can strike a balance — and this model helps us get there. We believe our work will help guide further exploration of machine learning-guided solutions in environmental protection and sustainable farming practices.” Ray and Baer were joined in the research by Joel Kowalewski, Barbara Baer-Imhoof, Tom Guda, Matthew Luy, and Payton DePalma. The work was funded by a grant from the California Research Alliance by BASF. The research paper, titled “Machine learning of honey bee olfactory behavior identifies repellent odorants in free flying bees in the field,” was published in eLife on 23 September 2026, with the DOI 10.7554/eLife.104831.3. The authors note that Ray is founder and president of Sensorygen and Remote Epigenetics and holds equity in both companies, Kowalewski holds equity in Sensorygen, and several of the authors are inventors in a patent application covering the compounds discussed in the article. If the approach scales from honey combs to whole orchards, the humble trick of making treated flowers smell unappealing could become one of the more elegant tools in the effort to keep the world’s most important pollinators alive.

Subject of Research: Machine learning-guided discovery of odorant compounds that repel honey bees from pesticide-treated crops

Article Title: Machine learning helps identify chemicals that repel honey bees from pesticides

Article References: Machine learning helps identify chemicals that repel honey bees from pesticides. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: honey bees, machine learning, pollinator protection, pesticides, olfaction, repellents, eLife, UC Riverside, colony collapse, sustainable agriculture, chemical ecology, Drosophila

The team focused on the olfactory system because it offers an unusually clean experimental window into memory. Odors are high-dimensional chemical stimuli, yet the brain rapidly compresses them into compact internal representations. When an animal learns that a particular odor predicts a reward or a punishment, the neural code for that odor changes. The researchers set out to determine exactly how those changes unfold at the level of entire populations rather than single cells.

Using large-scale recordings from mice as the animals learned to associate specific odors with outcomes, the investigators tracked activity in key nodes of the olfactory memory circuit, including the piriform cortex and connected structures. Dimensionality reduction techniques revealed that odor representations occupy smooth, low-dimensional manifolds embedded in the high-dimensional space of population activity. Learning did not simply add noise or scatter the responses; instead, it systematically reorganized the geometry of these manifolds.

Specifically, as animals became better at discriminating rewarded from unrewarded odors, the manifolds corresponding to different odor categories moved apart, increasing the margin between them. This geometric separation mirrors the objective functions used in machine learning classifiers, which seek to maximize the distance between classes. In other words, the biological network appeared to solve an optimization problem: reshape its internal representation space so that behaviorally relevant distinctions become as easy as possible to read out.

The study also examined how this optimization is implemented mechanistically. Plasticity at synapses within the olfactory cortical network provides the natural substrate for manifold reshaping. By adjusting connection strengths in a way that depends on task demands, the circuit can rotate, stretch, and translate the manifolds on which odor memories reside. Computational models in the paper demonstrated that a simple learning rule acting on recurrent connections is sufficient to reproduce the observed geometric changes and the accompanying improvements in behavioral performance.

One of the most striking findings is that manifold reorganization predicted behavior on a trial-by-trial basis. In sessions where the geometric separation between odor categories was larger, animals performed the discrimination task with greater accuracy. This tight coupling between representational geometry and action strengthens the argument that the manifold is not an epiphenomenon of recording analysis but a functional unit of memory itself, something the brain actively constructs and maintains.

The work also addresses a long-standing puzzle in memory research: why memories remain stable even as individual neurons change their firing properties. If a memory were stored in the activity of particular cells, drift in those cells should degrade the memory. But if the memory is stored in the shape and position of a manifold, the system can tolerate turnover and drift at the single-cell level as long as the global geometry is preserved. The olfactory network appears to exploit exactly this robustness, stabilizing the manifold while individual neurons come and go from the active ensemble.

These results resonate with a broader theoretical movement in neuroscience that treats population activity through the lens of geometry and dynamics. Rather than decoding the activity of single neurons, this framework asks how entire trajectories and subspaces of activity support computation. The new findings provide some of the clearest evidence yet that learning sculpts these subspaces deliberately, and that the rules governing this sculpting can be described with the same mathematical language used in machine learning.

The implications reach beyond olfaction. Manifold optimization may be a general principle of representational learning in the brain, applying to motor skills, decision-making, and perhaps even higher cognitive functions. If so, therapeutic strategies for disorders of memory and perception might one day aim not at individual synapses but at restoring healthy manifold geometry in malfunctioning circuits. The study also suggests that artificial neural networks, which already borrow heavily from brain-inspired principles, could benefit from architectures that explicitly optimize manifold structure the way biological networks appear to do.

Future experiments will aim to identify the precise plasticity mechanisms and neuromodulatory signals that guide manifold optimization, and to test whether similar geometric learning rules operate in other sensory and memory systems. For now, the study offers a compelling synthesis: memory is not a static snapshot written into cells, but a dynamically optimized shape in the brain’s representational space, continuously refined by experience until the world’s important distinctions stand out in sharp relief.

The piriform cortex, the largest olfactory cortical area, occupies a distinctive position among sensory cortices. Unlike primary visual or auditory regions, it receives direct input from the olfactory bulb without an intervening thalamic relay, and its recurrent collateral connections are extraordinarily dense. Pyramidal cells in this structure broadcast axonal branches widely across the network, meaning that any learning rule acting at these recurrent synapses has access to an almost associative memory-like architecture. This anatomical arrangement has long suggested that the piriform cortex functions as a pattern completion and pattern separation device, and the manifold optimization account fits naturally within that tradition, extending it from single-cell response changes to the collective geometry of ensembles.

The statistical structure of odor space itself provides important context for why the olfactory system might be particularly suited to geometric reorganization. Natural odorants are mixtures of many volatile molecules, and the relationships between odorants are smooth: chemically similar molecules tend to smell alike, and perceptual similarity decays gradually with molecular distance. The early stages of the olfactory pathway, from receptors in the nose through the glomerular map of the olfactory bulb, already impose a dimensionality reduction on this chemical space. What the new findings add is evidence that later, experience-dependent stages continue this compression but do so adaptively, stretching the dimensions that matter for current behavioral goals while letting irrelevant dimensions collapse.

Classical work on olfactory learning in rodents emphasized changes in single-neuron selectivity, with individual piriform neurons broadening or sharpening their tuning after conditioning. Those observations were sometimes difficult to reconcile, because different cells appeared to change in inconsistent directions. The manifold perspective resolves this apparent inconsistency: heterogeneous single-cell changes can produce a coherent geometric shift if they collectively move or reshape the ensemble’s low-dimensional subspace. A neuron increasing its responses to one odor while a neighbor decreases its responses may look contradictory at the single-cell level, yet both changes can contribute to expanding the distance between odor-category manifolds, exactly what improved discrimination requires.

The distinction between representational drift and manifold stability also connects to an active debate about how the brain balances flexibility and permanence. Longitudinal recordings in several systems have documented that the identities of neurons participating in a code can change over days and weeks, even when behavioral performance is fully preserved. Explanations of this drift have ranged from passive turnover to active consolidation processes. The finding that geometric structure is preserved while its cellular substrate rotates suggests that the relevant invariant for memory is relational rather than compositional: what matters is how representations sit relative to one another, not which particular neurons carry them. This reframing gives theorists a concrete target for models of memory maintenance over long timescales.

Connections to machine learning deepen the significance of the results. Classification algorithms such as support vector machines explicitly maximize margins between categories, and deep networks trained with error-driven rules are known to linearize their internal representations as performance improves, spreading class clusters apart in hidden-layer activity spaces. The apparent convergence between these engineered objectives and the geometry observed in olfactory cortex suggests that margin maximization may be a general computational principle that biological networks arrived at independently. It also raises the question of whether the brain’s learning rules approximate gradient-based optimization over a representational objective, or whether local synaptic plasticity merely produces margin expansion as an emergent byproduct. The computational models in the study, which reproduce the geometric changes with simple recurrent plasticity, favor the latter possibility, indicating that no explicit global error signal is required.

Methodologically, the study illustrates the growing power of combining large-scale electrophysiology or imaging with tools from applied mathematics. Dimensionality reduction methods, including approaches that preserve local neighborhood structure and those that track low-dimensional trajectories over time, allow researchers to ask quantitative questions about representational geometry that were previously unanswerable. Trial-by-trial alignment between geometric measures and behavior exemplifies a broader trend: rather than treating population analyses as descriptive, investigators now use them to generate predictions that can be tested against the animal’s choices on individual trials, tightening the link between neural data and computation.

Several questions remain open. Whether manifold optimization operates during passive exposure to odors or requires explicit reinforcement is unresolved, as is the role of top-down signals from prefrontal or hippocampal structures that inform the olfactory cortex about task context. The timescale of geometric consolidation, and whether reshaped manifolds persist during sleep-related replay, would clarify how learning is stabilized. Answering these questions will require the same combination of longitudinal population recording, behavioral quantification, and geometric analysis demonstrated here, applied across circuits and species. The study thus serves both as an empirical advance for olfactory neuroscience and as a methodological template for testing whether representational optimization is a universal grammar of learned neural computation.

Subject of Research: Representational learning through geometric optimization of neural manifolds in the olfactory memory network

Article Title: Representational learning by optimization of neural manifolds in an olfactory memory network

Article References: Hu, B., Temiz, N. Z., Chou, C.-N., Rupprecht, P., Meissner-Bernard, C., Titze, B., Chung, S., & Friedrich, R. W. (2026). Representational learning by optimization of neural manifolds in an olfactory memory network. Nature Neuroscience. https://doi.org/10.1038/s41593-026-02429-3

Image Credits: AI Generated

DOI: 10.1038/s41593-026-02429-3

Keywords: neuroscience, olfaction, neural manifolds, memory, representational learning, piriform cortex, neural plasticity, population coding, dimensionality reduction, behavioral learning, machine learning, Nature Neuroscience

Hygiene is a cornerstone of public health. Poor hygiene is closely tied to infection, disease transmission, and preventable illness, which is why public health researchers have long sought to understand the psychological drivers of behaviors such as handwashing, housecleaning, and personal grooming. Decades of work on disgust—the emotion thought to have evolved as a behavioral defense against pathogens—have established that people who are more easily disgusted tend to engage in more protective hygiene behavior. At the same time, olfaction has repeatedly been implicated in this picture: the sense of smell is our most direct chemical sensor of the environment, capable of detecting spoilage, contamination, and body odors that signal potential health threats.

The new research set out to connect these threads more rigorously. Rather than treating olfaction as a single, undifferentiated capacity, the team distinguished between olfactory metacognition—how people think and feel about their own sense of smell—and olfactory function measured objectively in the laboratory. The central hypothesis was that olfactory functioning relates to hygiene behavior at least partly through disgust sensitivity, the tendency to experience revulsion in response to potential contaminants. To test this, the researchers conducted two studies with distinct designs and measurement approaches.

Study 1 recruited 270 participants and focused on three indicators of olfactory metacognition: odor awareness, the subjective importance people attach to their sense of smell, and self-rated olfactory ability. Each of these indicators was positively associated with self-reported hygiene behavior. Crucially, when the researchers used mediation analysis to examine indirect pathways, each indicator showed a significant indirect effect through disgust sensitivity. In other words, people who were more attuned to odors, who valued their sense of smell more highly, and who rated their olfactory ability as stronger also tended to report higher disgust sensitivity—and that heightened disgust, in turn, was associated with more frequent and thorough hygiene practices.

These findings align with a conceptual framework in which disgust functions as part of the so-called behavioral immune system, a suite of psychological mechanisms that guard against disease threat before pathogens ever reach the body. According to this view, sensory inputs—especially smells—provide cues about potential contamination, and disgust converts those cues into avoidance and cleaning behavior. The Study 1 results suggest that individual differences in how prominently smell figures in daily awareness may shape the intensity of this disgust-based defense, and thereby the vigor of one’s hygiene routine.

But subjective self-reports are not the same as measured sensory ability. Study 2 therefore extended the investigation with 66 participants who completed psychophysical odor-threshold testing and suprathreshold ratings for one pleasant and one unpleasant odor. This design allowed the researchers to ask whether the ability to detect faint odors—rather than merely one’s beliefs about one’s own nose—also relates to hygiene behavior, and whether the valence of the odor matters. The results were more nuanced here. Odor sensitivity itself was not directly associated with hygiene behavior, a finding that echoes earlier work suggesting that disgust sensitivity relates to affective responses to odors rather than to the raw capacity to detect them.

Yet the indirect pathway through disgust sensitivity did appear in Study 2, though selectively. A significant indirect effect through disgust sensitivity was observed for the pleasant odor but not for the unpleasant odor. That an unpleasant smell failed to show the same pathway may seem counterintuitive, but the authors caution that a bootstrap comparison did not support a statistically reliable difference between the two indirect effects. In other words, the apparent pleasant-versus-unpleasant asymmetry at the threshold level should be interpreted with care, as the data could not confirm that the two odor conditions truly differed in the strength of this mediated pathway.

Where odor-condition differences did emerge clearly was at the suprathreshold level—when participants judged clearly perceivable odors. The associations of hygiene behavior with odor pleasantness and familiarity differed significantly between the pleasant and unpleasant odor conditions, with positive associations observed specifically for the pleasant odor. This suggests that how people appraise and contextualize everyday smells—their emotional coloring and their sense of familiarity—may relate to hygiene behavior differently depending on the hedonic character of the odor in question, whereas simple detection ability appears less consequential.

Taken together, the two studies paint a differentiated picture: subjective and psychophysical aspects of olfactory functioning are related to hygiene behavior in partly different ways. Disgust sensitivity emerges as one affective correlate of these associations, providing a plausible psychological route by which smelling and thinking about smells translates into washing, cleaning, and grooming. Meanwhile, odor valence effects were more clearly supported for conscious appraisal of suprathreshold odors than for basic odor sensitivity, implying that the emotional meaning of smell, rather than the acuity of the nose per se, drives much of the connection to hygiene.

The findings carry practical weight for public health. If disgust sensitivity mediates the link between olfaction and hygiene behavior, interventions aimed at improving hygiene might benefit from strategies that engage the sense of smell and the emotion of disgust together—approaches that earlier experimental and field-based work on hand hygiene has begun to explore. At the same time, the authors emphasize limits: the indirect pathway was not uniformly observed across odor conditions, the pleasant-versus-unpleasant asymmetry in threshold-based mediation was not statistically reliable, and the data underlying the study will be shared only on reasonable request to the corresponding author. Future research with larger and more diverse samples will be needed to determine when, and for whom, the nose-disgust-hygiene pathway operates most strongly—and whether it can be harnessed to promote healthier behavior at the population scale.

Subject of Research: The relationship between olfactory functioning, disgust sensitivity, and hygiene behavior

Article Title: Disgust sensitivity as an indirect pathway linking olfactory functioning and hygiene behavior

Article References: Jiang, C., Modi, U., & Han, P. (2026). Disgust sensitivity as an indirect pathway linking olfactory functioning and hygiene behavior. Current Psychology, 45(17), Article 1487. https://doi.org/10.1007/s12144-026-10057-0

Image Credits: AI Generated

DOI: 10.1007/s12144-026-10057-0

Keywords: olfaction, disgust sensitivity, hygiene behavior, odor awareness, olfactory metacognition, odor threshold, odor valence, behavioral immune system, public health, psychology, sensory perception, handwashing