Generalization and extinction of learned fear alter primary sensory input to the brain

Background

Fear learning is a fundamental process that allows the brain to refine its ability to discriminate between threat-predictive and neutral stimuli. When this process becomes dysregulated, it can lead to maladaptive fear generalization, a core feature of clinical conditions such as post-traumatic stress disorder (PTSD) and generalized anxiety disorder (GAD)¹. Classical models attribute fear learning and generalization to higher-order brain regions such as the amygdala, hippocampus, and prefrontal cortex. Hence, a central challenge in neuroscience is to understand how associative learning shapes sensory representations, and how early along the sensory hierarchy these changes can occur².

Emerging evidence suggests that aversive learning may influence neural processing much earlier in the sensory pathway than previously assumed. The olfactory system provides a powerful model to examine this observation³. Odor information is initially detected by olfactory sensory neurons (OSNs), whose axons converge onto glomeruli, allowing researchers to observe odor-specific activity at the first synapse of the sensory pathway. This raises a critical question: How do olfactory sensory neurons (OSNs), as primary sensory neurons, exhibit plasticity that tracks the generalization and extinction of learned fear, even for stimuli never directly paired with threat?

In this preprint⁵, Rosenthal and colleagues combine odor-based fear conditioning with in vivo imaging of olfactory sensory neuron (OSN) terminals (see preprint Fig. 1) to test whether early sensory activity reflects the generalization and extinction of learned fear in mice. Specifically, they examine whether OSNs show enhanced responses to novel odors perceived as threatening through generalization, and whether these responses decrease when those odors are no longer associated with danger. The authors hypothesize that OSN output parallels a mouse’s perception of threat, such that early sensory activity dynamically reflects learned threat and safety signals.

Key Findings

Fear learning amplifies sensory input at the first synapse in the olfactory system

Using an odor-cued fear conditioning paradigm in which the conditioned stimulus (CS; the odor methyl valerate, MV) was paired with foot shocks (the unconditioned stimulus), the authors demonstrated that synaptic output from OSNs to the olfactory bulb increased following fear learning in the “Fear Conditioned” group. This increase in OSN activity, as measured by changes in fluorescence (reflecting the magnitude of neurotransmitter release), was not observed in the “Never Shocked” group (where the CS was repeatedly presented without shock). These findings suggest that fear learning amplifies sensory input at the first synapse of the olfactory system.

Fear learning generalizes across odors

Through freezing behaviour and OSN activity, the authors investigated whether fear learning occurs not just for the CS, but for novel odors. They measured freezing behaviour and OSN activity when different odors were presented. OSN fluorescence was measured after pairing MV with foot shocks, followed by the presentation of novel odors that were never paired with a shock, revealing an increase in OSN activity for both the CS and the novel odors. These findings indicate that both OSN responses and freezing behaviour increased in response to all tested odors, suggesting that fear learning generalizes across all odors.

Sensory amplification spreads across glomeruli

The authors used in vivo imaging of OSNs to measure changes in epifluorescence in glomeruli before and after conditioning, to determine whether sensory amplification spreads across glomeruli. Comparing before and after conditioning showed that, in addition to the glomeruli that normally respond to the CS, those that originally responded only to the novel odors also exhibited increased responses to the CS following conditioning. These findings suggest that sensory amplification in OMP-spH mice spreads across glomeruli.

Fear learning changes how odors are represented in the olfactory bulb

To determine whether fear learning changes how odors are represented at the level of the olfactory bulb, the experimenters used in vivo epifluorescence imaging of OSNs and an analysis of glomerular activity patterns in the “Fear Conditioned” OMP-spH mice. Following conditioning, the researchers observed that the patterns of glomerular activity had become similar to those produced by the CS. These findings suggest that fear learning alters the representation of odors in the olfactory bulb.

Extinction reverses both the sensory and behavioral effects of fear learning

Following the fear conditioning paradigm (see preprint Fig. 1), the experimenters employed various extinction protocols to determine whether the behavioural and neural consequences of fear learning could be reversed. Through these extinction sessions, where no foot shocks were paired, the experimenters presented either the conditioned stimulus (CS extinction), a panel of odors including the CS and all novel odors (Odor Panel Extinction), a single novel odor (ET or 2-Hex; Novel Odor Extinction), or no odor (Procedural Extinction). Afterwards, freezing behaviour and glomerular activity were assessed using these extinction paradigms. These experiments showed that extinction training – over all extinction paradigms – reduced OSN responses and freezing behaviour in mice, with Odor Panel Extinction producing the most significant reductions across all odors.

Why we highlight this preprint

This preprint stood out to us because it uses a convincing approach that combines odor-based fear conditioning/extinction with longitudinal optical imaging of early odor representations, offering a powerful alternative to traditional sound–shock paradigms. Together, these methods support a real perspective shift: generalized fear may be represented much earlier in the sensory pathway, rather than being confined to higher-order brain regions like the amygdala and hypothalamus. We think this matters because it reframes generalization as more than a simple stimulus-similarity problem; early sensory processing itself may be tuned by learned threat expectation. We also appreciated the clinical relevance of the extinction experiments, since they map onto key challenges in exposure therapy and show that training design can narrow or broaden fear updating. Overall, by addressing how fear generalizes beyond the original trigger, this work highlights early sensory circuits as promising intervention targets for maladaptive fear generalization in PTSD and GAD. On a personal note, it resonated with us because this is a problem many of us have encountered, directly or indirectly. What made this preprint especially compelling is that it shows fear as more than a response to what is perceived. Through an innovative yet convincing method, it suggests that learned threat can create an anticipatory state that reshapes sensory processing itself, even before a stimulus is consciously recognized as dangerous. In other words, helping people with anxiety may not only mean reducing fear of the original trigger, but also working on the expectation that something safe might become threatening.

Questions for the authors

1. The increase in olfactive sensory neurons (OSN) output is seen as evidence that associative learning strengthens signaling for a threatening odor. How could we dissociate odor-specific associative plasticity from global fear-induced sensory gain? Would combining unpaired shock-odor control and measuring OSN terminals isolate associative effects?
2. The study also reports increased responses in glomeruli that were not initially responsive to the conditioned odor. While it is generally taught that glomeruli encode stimulus features, does this suggest that OSNs may also encode threat value, or instead reflect higher-order modulation of sensory processing? Could this be further examined by varying threats across odors in a fear learning protocol and measuring corresponding OSN activity?
3. During extinction experiments, exposure to the full odor panel produced a stronger suppression of neural responses than exposure to the conditioned stimulus alone. Does this broader suppression reflect the formation of a generalized safety rule? If so, could this effect be driven by higher brain regions such as the prefrontal cortex or amygdala influencing early sensory processing? Could this be tested by blocking feedback signals to the olfactory bulb and measuring changes in OSN activity?
4. If fear learning can amplify activity at the first sensory stage in olfaction, do you think that similar mechanisms could occur in vision or audition, potentially contributing to cross-sensory hypervigilance after trauma? Could this be tested by adapting similar paradigms to other modalities and examining whether fear learning alters responses in early sensory relays such as the thalamus, and possibly in primary sensory cortices?

Bibliography

1. Traina G, Tuszynski JA. The Neurotransmission Basis of Post-Traumatic Stress Disorders by the Fear Conditioning Paradigm. Int J Mol Sci. 2023 Nov 15;24(22):16327. doi: 10.3390/ijms242216327. PMID: 38003517; PMCID: PMC10671801.
2. Weinberger NM. Auditory associative memory and representational plasticity in the primary auditory cortex. Hear Res. 2007 Jul;229(1-2):54-68. doi: 10.1016/j.heares.2007.01.004. Epub 2007 Jan 17. PMID: 17344002; PMCID: PMC2693954.
3. Li W, Howard JD, Parrish TB, Gottfried JA. Aversive learning enhances perceptual and cortical discrimination of indiscriminable odor cues. Science. 2008 Mar 28;319(5871):1842-5. doi: 10.1126/science.1152837. PMID: 18369149; PMCID: PMC2756335.
4. Mombaerts P. Axonal wiring in the mouse olfactory system. Annu Rev Cell Dev Biol. 2006;22:713-37. doi: 10.1146/annurev.cellbio.21.012804.093915. PMID: 17029582.
5. Michelle C. Rosenthal, Alper K. Bakir, John P. McGann bioRxiv 2026.02.25.708099; doi: https://doi.org/10.64898/2026.02.25.708099

Tags: anxiety disorders, behavioral neuroscience, extinction learning, fear conditioning, fear generalization, learning and memory, mouse model, neuroplasticity, neuroscience, olfaction, olfactory bulb, sensory neuroscience, sensory processing, synaptic transmission

(No Ratings Yet)