Journal: bioRxiv

Article Title: Cardiac signals shape insular cortex activity and emotion coding

doi: 10.64898/2026.03.18.712676

Figure Lengend Snippet: A. Schematic of the experimental timeline. Left: Mice were first trained in an appetitive conditioning paradigm to associate a CS+ app with the delivery of a reward (sucrose), and to discriminate between the CS+ app and CS- app . Appetitive training occurred once a day for 6-16 days. Once successfully trained, mice entered the first recording day and exposed to CS+ app and CS- app . They were further exposed to a third, novel CS+ (CS+ hab ). On that same day and after the recording session, mice were placed in a fear conditioning chamber where the CS+ hab was paired with a foot shock (US) 5 times. After fear conditioning, the CS+ hab was referred to as CS+ fear . On the next day, mice entered the second recording day and exposed to CS+ app , CS- app , as well as CS+ fear . B. Heatmap of the z-scored firing rate of all units that are significantly heartbeat tuned (*p<0.05, rayleigh test) during CS- app (left) and appetitive CS+ app (right). Units are sorted according to their peak firing rate in the normalized cardiac cycle. C. Quantification of the percentage of heartbeat tuned units during baseline, CS- app. , CS+ app., CS+ hab. , CS+ fear , and freezing. (Wilcoxon and Mann-Whitney U-tests, *p<0.05). Data represents mean ± s.e.m. D. Average heart rate during CS+ fear and freezing (student t-test, *p<0.05). Data represents mean variable ± s.e.m. N=4 mice. E. Average HRV during CS+ fear and freezing (student t-test, n.s.). Data represents mean variable ± s.e.m. N=4 mice. F. Percentage of heartbeat tuned pInsCtx units during baseline (black), periods of low heart rate (blue), periods of high heart rate (brown). Baseline periods were defined as three 15s epochs before any CS presentation. Low and high heart rate epochs were defined as three 15s continuous periods were heart rate was below or above the average heart rate respectively, and outside of any CS presentation, freezing event, or baseline epoch. Data represents mean variable ± s.e.m.

Article Snippet: Briefly, on day 1 after the first electrophysiological recording session, mice were placed in a commercial fear conditioning setup (Ugo Basile, Italy), consisting of a behavior box, an electric grid floor, and a light source located in a soundproof chamber.

Techniques: MANN-WHITNEY

Journal: bioRxiv

Article Title: Cardiac signals shape insular cortex activity and emotion coding

doi: 10.64898/2026.03.18.712676

Figure Lengend Snippet: A. Example heart rate (top) and pInsCtx neuronal population average firing rate (bottom) illustrating the effect of metoprolol administered at 180s. B. Heart rate before and after metoprolol administration (student t-test, *p<0.05). Data represents mean ± s.e.m. N=4 mice. C. pInsCtx single unit average firing rate before and after metoprolol administration (student t-test, n.s.). Data represents mean ± s.e.m. N=4 mice. D. Correlation between heart rate increases and the corresponding pInsCtx firing rate increases in control (n=237, day 1, dark green) versus metoprolol-treated (n=201, day 2, light green) mice. Each dot corresponds to an increase in heart rate. N=4 mice per group. E. Heart rate change relative to the CS+ app. in control (day1) versus metoprolol (day2) groups. Traces are normalized to 2s before the start of the CS. (***p<0.001, repeated measures ANOVA with Bonferroni correction). N=4 mice per group. Data represents mean heart rate ± s.e.m. F. Same as in (E) but for CS+ fear . Data represents mean heart rate ± s.e.m. N=4 mice per group. G. Firing rate of pInsCtx neurons relative to CS+ app. in control versus metoprolol groups (***p<0.001, repeated measures ANOVA with Bonferroni correction). N=4 mice per group. Data represents mean firing rate ± s.e.m. H. Same as in (G) but for CS+ fear . Data represents mean firing rate ± s.e.m. N=4 mice per group. I. Percentage of the total number of pInsCtx units that are heartbeat tuned during the different CS in control and metoprolol treated mice (student t-test, *p<0.05). J. Representation of the 2 first components after principal component analysis (PCA) of pInsCtx population activity in control mice after appetitive and fear conditioning during CS presentations and behavioral events. Ellipses represent covariance confidence intervals. K. Same as in (J) but for metoprolol treated mice. L. SVM classifier decoding accuracies for CS identity (CS+ app . vs CS- app .) from pInsCtx activity (FDR-corrected permutation tests, p < 0.05). M. SVM classifier decoding accuracies for CS+ app . vs baseline from pInsCtx activity (FDR-corrected permutation tests, p < 0.05). N. SVM classifier decoding accuracies for CS+ fear . vs baseline from pInsCtx activity (FDR-corrected permutation tests, p < 0.05). O. SVM classifier decoding accuracies for CS- app . vs baseline from pInsCtx activity (FDR-corrected permutation tests, n.s.).

Article Snippet: Briefly, on day 1 after the first electrophysiological recording session, mice were placed in a commercial fear conditioning setup (Ugo Basile, Italy), consisting of a behavior box, an electric grid floor, and a light source located in a soundproof chamber.

Techniques: Control, Activity Assay