A cortical basis for perception of internal gut sensations
Omer Rafael-IzhakiiD, Stav ShtiglitziD, Juliet MilleriD, Yael PrilutskiiD, Itay TalpiriD, Ayal LaviiD, Yoav LivnehiD
Abstract
Interoception, the sensing of internal bodily signals, is essential for brain-body interactions and shapes emotion, cognition, and behavior1-5. Subconscious internal signals, including heartbeats or stomach fullness, can rise to conscious awareness, and this process can improve with practice. Disrupted interoception is emerging as a common deficit in diverse psychiatric disorders1,3,6,7. Nevertheless, we still lack a fundamental understanding of the neurobiological basis of perception and conscious reporting of internal sensations. Here, we combine genetic access to specific interoceptive primary afferents, with non-invasive optogenetics in mice, to establish a quantitative framework for studying internal perception. Specifically, we developed a behavioral task in which mice report detecting non-invasive optogenetic activation of vagal mechanosensory neurons, establishing "interoceptive psychophysics". We combine this approach with cellular-resolution imaging and manipulations to reveal the neuronal basis for perception of these internal sensations in the interoceptive insular cortex. While representations of sensory stimuli were consistently observed in insular cortex across different tasks, perceptual reports were only encoded during a more difficult psychophysics task, but not during simple detection. Accordingly, manipulations of insular cortex activity revealed that, although it is generally required for behavioral reports of internal signals, it plays an additional role in supporting perception under conditions of uncertainty. These findings reveal a neural basis for perception of internal sensations and provide a blueprint for future quantitative exploration of other interoceptive modalities.
02 /Specification & Classification
Click any tag to browse matching papers- Citations
- 1
- Organism
- Assay
- Research Focus
- Approach
In-depth summary
Claude Opus 5.5 · xhigh effort · 7 Nov 2026 · from the full text of version 3 (21 Oct 2026) · CC BY 4.0
Mice can report optogenetic activation of vagal gut mechanosensory neurons; insular cortex encodes perceptual report and causally affects detection only in a difficult psychophysics task, not simple detection.
The paper says quantitative study of interoception has been held back by the difficulty of varying internal stimuli while recording and manipulating neurons, and that human capsule studies cannot be paired with cellular-resolution tools. It builds a non-invasive optogenetic mouse assay for gut mechanosensation that allows psychophysics together with insular cortex imaging and inhibition.
Key results
ShownRetrogradely targeted vagal OxtR neurons behave as gut mechanosensors. Their non-invasive optogenetic activation suppresses licking and can be learned as a Go or NoGo cue, and the cue depends on these neurons being intact.
45/198 (23%) nodose OxtR neurons responded to intestinal balloon inflation (4 mice). Lick reduction appeared in ChRmine mice but not in wildtype or tdTomato controls. After near-complete ablation, hits and false alarms no longer differed (p = 0.129, n=3). · Non-invasive optogenetic activation of gut…; Behavioral report of optogenetic activation of…; Figure legends
ShownMice show sigmoidal psychometric detection curves for vagal stimulation frequency. Insular cortex neurons have sigmoidal or Gaussian frequency tuning, and population decoding detects stimuli the mice do not report.
Maximal correct rejection rate 98.5 ± 1.5%, sigmoid fit R^2 = 0.98 ± 0.01 (n = 10 mice). Sigmoidal neurons 14.5±2.6%, Gaussian 17.4±1.6%, together 32±3% of neurons. A decoder trained below the behavioral threshold gave a positive accuracy gap in all mice. · Cortical activity during gut mechanosensory…
ShownPerceptual report (detected vs undetected stimuli) is encoded in insular cortex in the psychophysics task but barely in the single-frequency detection task, at both single-neuron and population level.
Significant choice probability in 15.4±4.0% of neurons in psychophysics vs 6.2±2.4% in the single-frequency task. Population decoders separated detected from undetected trials only in psychophysics. Stimulus presence was decodable in both tasks. · A neural basis for detection of internal gut…; Figure legends
ShownOptogenetic inhibition of insular cortex raised detection of vagal stimulation only in the psychophysics task. An in silico inhibition of the most suppressed neurons reproduced the behavioral shift.
Inhibition lowered false alarm rates at all three tested frequencies in psychophysics, but had no effect in single-frequency tasks. It also had no effect in an odor-only Go/NoGo task (correct rejections p = 0.44, n=6). Imaging data in silico: inhibiting the top 20% most suppressed neurons improved decoding. · A neural basis for detection of internal gut…; Partial psychophysics with InsCtx inhibition; In Silico Neural Perturbation Analysis; Figure legends
ShownInsular cortex representations of the vagal stimulus and of reward are distinct and stable across psychophysics, NoGo and Go tasks, with little consistent spatial organization.
Fractions of neurons responding to stimulus, reward or both were similar across tasks, and stimulus and reward were similarly separable at the population level. Spatial clustering was seen in 4/5 mice in the NoGo task but 1/5 in psychophysics. · Distinct stable representations of internal gut…; Figure legends
Methods, data and reproducibility
Vagal OxtR sensory neurons were made light-sensitive with a retrograde ChRmine virus and activated through the neck, and mice reported this in head-fixed Go/NoGo tasks with varied pulse frequency. Layer 2/3 insular cortex was imaged by two-photon microscopy and optogenetically inhibited with stGtACR2, with auROC, choice probability and SVM or Naive Bayes decoding.
- Behavioral psychophysics cohort (vagal opto NoGo then 7-frequency testing) · mouse · n = n = 10 mice
- Two-photon calcium imaging of layer 2/3 insular cortex during psychophysics, NoGo and Go tasks · mouse · n = n = 5 mice
- Two-photon imaging of nodose ganglion during intestinal balloon inflation · mouse · n = 198 neurons from 4 mice
- Lick reduction (sucrose consumption) test, ChRmine vs control mice · mouse · n = males 12 ChRmine+, 5 controls; females 5 ChRmine+, 2 controls
- Caspase ablation of nodose OxtR neurons (near-complete and partial) · mouse · n = n=3 near-complete, n=3 partial
- Methods & tools
- ChRmine, stGtACR2, GCaMP7f, Suite2p, DeepInterpolation, MonkeyLogic2, Linear SVM, Naive Bayes classifier, auROC / choice probability, pyPhotometry, MATLAB
- Reproducibility gaps
- The text gives no code or data availability statement; analysis is described as custom MATLAB and Python code without a repository link.
Limitations (5)
- The 23% of nodose OxtR neurons responding to balloon inflation is likely an underestimate, because the balloon covered only about 3-6% of the small intestine length. · stated by the authors
- Optogenetic excitation of insular cortex caused seizures in pilot experiments, so causal tests used inhibition only; inhibition effects were described as mild and not a global silencing. · stated by the authors
- Heart rate fell in response to vagal stimulation in some anesthetized mice but not others; the authors report that task performance was similar regardless. · stated by the authors
- Imaging and population analyses rest on 5 mice, the ablation groups have n=3, and psychophysics was run for only 1-2 days after training at 15 Hz alone. · noted in this summary
- Stimulation intensity varied across mice (15-102 mW/mm^2), which complicates comparison of thresholds across animals. · noted in this summary
Who should read:Read this if you study interoception, cortical correlates of perceptual report, or want a mouse psychophysics assay with imaging and optogenetic inhibition.
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Mechanosensation of the heart and gut elicits hypometabolism and vigilance in mice.
Karen A. Scott, Yalun Tan, … Eric G. Krause·Nature MetabolismPeer-reviewed·Jan 17, 2025
- Cardiac and gastric interoception have distinct neural substrates
Yusuke Haruki, Kenji Ogawa·eNeuroPeer-reviewed·Jan 1, 2023
- Neural divergence and convergence for attention to and detection of interoceptive and somatosensory stimuli
Aleksandra Maria Herman, Clare E. Palmer, … Manos Tsakiris A·CortexPeer-reviewed·Dec 8, 2020
- Differences in layer-specific activation of the insula during interoceptive vs. exteroceptive attention
Matthias Müller-Schrader, Frederike H. Petzschner, … Klaas Enno Stephan·bioRxiv·Oct 8, 2025
- Reliable sensory processing of superficial cortical interneurons is modulated by behavioral state
Carolyn G. Sweeney, Maryse E. Thomas, … Anne E. Takesian·Cell ReportsPeer-reviewed·May 1, 2025
- Interoception vs. Exteroception: Cardiac interoception competes with tactile perception, yet also facilitates self-relevance encoding
Marie Loescher, Patrick Haggard, Catherine Tallon‐Baudry·Proceedings of the National Academy of SciencesPeer-reviewed·Dec 2, 2025