{"paper_id":"1efc4dc9-cecb-49d6-b790-5d79c7869545","body_text":"Soft tactile stimulation engages parabrachial circuits traditionally \nassociated with aversion \n \nAnesten F1, Simfors S1, Ioneskou K1, Hezsö M1, Gündogdu B1, Tran A1, Stjernvall A1, Ratiglia \nV1, Almasri A1, Löken LS1*. \n \n1Department of Physiology, Institute of Neuroscience and Physiology, Sahlgrenska Academy, \nUniversity of Gothenburg. \n \n*Correspondence to: line.loken@gu.se \n \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\nAbstract: \n \nGentle tactile stimulation is associated with positive affect and social bonding, yet the central \ncircuits engaged by such stimuli remain incompletely understood. The lateral parabrachial \nnucleus (lePB) is a key hub in ascending affective sensory pathways and is robustly activated \nby aversive stimuli, including pain. Here, we examined neuronal activation in the lePB and the \nlikewise associated  subparafascicular nucleus, parv ocellular part (SPFp ), following different \ntactile stimulation paradigms in mice. Behavioral analyses confirmed that the soft touch stimuli \nused in this study were not aversive: mice displayed low aversive facial grimace scores during \nbrushing and von Frey stimulation compared with noxious heat, and showed a preference for a \nsoft tactile environment in a place preference assay. Neuronal activation was assessed using \nFos immunohistochemistry following exposur e to brushing -based soft touch, a fur -roll \nparadigm, innocuous punctate touch (von Frey), or noxious heat. Soft touch protocols robustly \nincreased Fos expression in the lePB compared with home cage controls, whereas innocuous \npunctate touch did not. Notably, the magnitude of activation produced by brushing -based \nstimuli was comparable to that induced by noxious heat. Using CalcaCre mice, we further found \nthat soft touch recruited a subset of CGRP -expressing neurons in the lePB. In contrast, tactile \nstimulation produced only modest activation in the SPFp and did not strongly increase overall \nFos expression in this region. Together, t hese findings demonstrate that affective tactile \nstimulation can engage neuronal populations within ascending parabrachial circuits, including \nCGRP neurons traditionally associated with nociceptive processing, suggesting that these \npathways may encode the salience or affective significance of somatosensory stimuli rather \nthan exclusively aversive input. \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\nIntroduction \n \nGentle tactile stimulation, such as slow stroking of the skin, is associated with positive affect, \nsocial bonding, and stress reduction (Packheiser, Hartmann et al. 2024) . In humans, these \nsensations are mediated by C -tactile (CT) afferents (Loken, Wessberg et al. 2009) , a class of \nunmyelinated low -threshold mechanoreceptors tuned to slow, gentle stroking  (Vallbo, \nOlausson et al. 1993) . CT afferents project to brain regions involved in affective and \ninteroceptive processing, including the posterior insular cortex (Olausson, Lamarre et al. 2002, \nBjornsdotter, Loken et al. 2009, Morrison, Loken and Olausson 2010). In rodents, the functional \nanalogue of CT afferents are C-low-threshold mechanoreceptors (CLTMRs), which innervate \nhairy skin and respond selectively to gentle dynamic touch. At the spinal level, CLTMR input \nis integrated through dorsal horn interneuron networks and can access ascending pathways \nwithin the ante rolateral system (Choi, Hachisuka et al. 2020, Liu, Qiao et al. 2022) . This \norganization suggests that signals associated with affective touch may converge with \nnociceptive pathways at early stages of somatosensory processing. \n \nThe parabrachial nucleus (PB) is a key node for the processing of sensory signals with affective \nand motivational value. In rodents, neurons in the lateral external parabrachial nucleus (lePB) \nand the parvocellular subparafascicular nucleus (SPFp) receive ascending input from spinal and \ntrigeminal pathways . Their projections reach  forebrain structures involved in emotional \nlearning, threat detection, and autonomic regulation, including the amygdala, hypothalamus, \nand bed nucleus of the stria terminalis (Bernard and Besson 1990, Cameron, Polgár et al. 2015) \n(Kang, Liu et al. 2022, Kang, Liu et al. 2025).  \nWithin these structures, neurons expressing calcitonin gene-related peptide (CGRP) have been \nstrongly implicated in aversive processing and defensive behaviors (Chiang, Bowen et al. 2019, \nPauli, Chen et al. 2022). Consistent with this role, a wide range of noxious and aversive stimuli \nrobustly activate neurons in the lePB and SPFp  (Palmiter 2018, Condon, Yu et al. 2024, \nPalmiter 2024).  \n \nRecent studies suggest that these nuclei may encode generalized sensory salience rather than \nmodality-specific nociceptive signals, as diverse aversive stimuli activate overlapping neuronal \npopulations within these regions (Kang, Liu et al. 2022, Kang, Liu et al. 2025) . This raises an \nimportant question: do these circuits exclusively signal aversion, or can they also be engaged \nby non-noxious tactile stimuli with positive affective significance? \nIn rodents, soft tactile stimuli such as brushing of the fur or contact with soft materials evoke \nbehaviors consistent with positive affect (Choi, Hachisuka et al. 2020, Liu, Qiao et al. 2022) . \nHere, we investigated how different forms of soft touch influence neuronal activation in the \nlePB and SPFp, and whether this activation involves CGRP-expressing neurons.  \n \nUsing Fos immunohistochemistry, we compared multiple soft touch protocols with innocuous \npunctate touch and noxious heat. To assess the affective valence of these stimuli, we \ncomplemented neural measurements with behavioral assays including place preference and \nfacial grimace scoring. Our results show that soft touch robustly activates parabrachial neurons, \nand to some extent also  activates the CGRP-expressing population within the lePB , despite \nlacking behavioral signatures of aversion. These findings suggest that parabrachial circuits \ntraditionally associated with negative valence may also participate in processing affectively \nsalient, non-aversive tactile signals. \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\nMethods \n \nAnimals \n \nMale and female heterozygous CalcaCre mice (8–12 weeks old; JAX #033168) mice express \nnuclear-localized cre recombinase:EGFP fusion protein targeted to mouse Calca gene (here \nreferred to as CGRP-GFP) and FosTRAP2 mice (JAX # 030323) on a C57BL/6J background \nwere bred in -house at the Facility for Experimental Biomedicine, Sahlgrenska Academy, \nUniversity of Gothenburg. FosTRAP mice were used without tamoxifen administration and \ntherefore functioned as wild-type mice. \nAnimals were group -housed under standardized conditions (12 h light/dark cycle; lights on at \n07:00; temperature 21 °C; humidity 50 –60%) with ad libitum access to standard chow and \nwater. \nAll experimental procedures were approved by the local Animal Ethics Committee at the \nUniversity of Gothenburg (ethical approval number 4534/22) and were conducted in \naccordance with national and European Union guidelines for the care and use of laboratory  \nanimals. \n \nBehavior Protocols \n \nAll behavior protocols were per formed between 9-11 am in the same room to which the mice \nhad previously been habituated over a course of 4 -5 days. Habituation consisted of remaining \nin the home cage in the test room for 15 mins with a permeable lid, for the mice to get used to \nthe smells and sounds of the room as well as the experimenter. After this habituation, mice were \ntransferred to the experimental cage (see each test for specifications) and allowed to acclimatize \nfor 15 mins. Mice who were exposed only to sitting in their home cage  were used as a control \ngroup. \n \nBrush \n \nThe hairy back skin was gently stroked with a hand -held soft brush, moving from the nape of \nthe neck to the lumbar enlargement region at constant speed (18-22 cm/s) and force (maximum \n23-25 mN) as described by Liu et al (ref). T hree bouts of 90 s of brushing, with a 300 s rest \nperiod in between each bout, were performed. The first of these bouts was filmed with a Sony \nHandycam CX405. Each mouse had its own brush to avoid odor contamination.  The protocol \nwas repeated for a period of ten days over the span of two weeks. On the tenth day  mice were \nperfused and sacrificed 90 min after the protocol was completed. \n \nBrush + Non-Noxious Heat and Blanket \n \nMice were placed on a hot plate set to 37oC. A soft blanket of similar coloration and texture to \nthe fur of a C57BL/6J mouse was placed on the hot plate, covering roughly half of it. Brushing \nwas performed as described above. Each mouse had its own brush and blanket to avoid odor \ncontamination. The protocol was repeated for a period of ten days over the span of two weeks. \nOn the tenth day mice were perfused and sacrificed 90 min after the protocol was completed. \n \nBlanket Cage \n \nMice were placed in a  plastic cage that was divided into two equal sides with a separator. \nIndividual cages were used for each mouse to avoid odor cross-contamination. One side of the \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\ncage was covered in a soft blanket as described above. The other side was left bare but with the \ntransparent cage bottom sitting atop the same kind of blanket. For habituation, the mouse was \nplaced alternatingly on the blanket side and the bare side, 15 min for each side. This protocol \nwas repeated daily over a period of five days. On the fifth day the separator was removed, \nallowing the mouse to freely choose which side to dwell in for 15 min. The test on the fifth day \nwas recorded as described above, and time spent by the mouse on either the blanket or the bare \narea was then measured. \n \nFur Roll \n \nMice were allowed free access to a cardboard cylinder clad in a blanket (as above) placed inside \na blanket cag e (as above) for a duration of 30 min. Each mouse had its own cage and roll to \navoid odor contamination. This protocol was repeated daily over a period of five days. On the \nfifth day, mice were first left inside the blanket cage with the blanket-clad roll for 15 min. The \nlatter 15 min of the test were recorded  as described above. Mice were perfused and sacrificed  \n90 min after the end of the experiment. \n \nNoxious Heat (Hot Plate) \n \nMice were placed on a hot plate covered by a sturdy piece of cardboard inside a plastic cylinder. \nMice were left to acclimatize for 15 min, and thereafter the temperature was slowly ramped up \nto 50oC over approximately 5 minutes. The cardboard plate was removed as the temperature \nreached close to target and the mouse left on the bare hot plate for up to maximum 15 s or until \npain behavior was observed (jumping, hindpaw  lifting, paw licking). The cardboard plate was \nthen reintroduced and the mouse allowed 90 s of recovery time. This was repeated seven to \neight times. Mice were perfused and sacrificed 90 min after the end of the experiment. \n \nvon Frey \n \nMice were acclimatized to the testing room and the raised metal grid where they were placed \nfor the experiment over a period of four days . On the fifth day, mice were left in their home \ncage in the testing room for 15 min after which they were transferred to plastic cylinders atop \nthe raised metal grid and left to acclimatize for 15 min. The right hind paw was stimulated using \na 0.4 g von Frey filament, enough to be felt by the mouse but not be painful (Deuis, Dvorakova \nand Vetter 2017). Stimulation went on until the mouse exhibited paw withdrawal, and repeated \nover a period of 15 min interspersed with resting periods. Mice were perfused and sacrificed 90 \nmin after the end of the experiment. \n \nTissue preparation and Immunohistochemistry \n \nMice were deeply anaesthetized with a mixture of domitor and ketamine, and perfused \ntranscardially with 10 mL PBS, followed by 30 mL 4% paraformaldehyde. The brain was \nremoved and post fixed in 4% paraformaldehyde for 3 h. Tissue was then transferred to a 30% \nsucrose solution in PBS until sectioning. Coronal 30 μm thick serial sections of the midbrain \nand thalamus were cut using a Leica CM3050S cryostat ( Leica CM3050 S, Leica Biosystems, \nWetzlar, Germany) and stored in PBS. \nSections were mounted on SuperFrost slides and blocked for 1.5 h with 10% normal goat serum \n(abcam, Cambridge, UK) in PBS with 0.3% Triton-X-100 (Perkin Elmer, Waltham, MA, USA). \nSections were incubated with primary antibodies (R abbit anti-cfos 1:2000, \nRRIID:AB_2247211; Ch icken anti-GFP 1:3000, RRID: AB_300798) overnight at room \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\ntemperature. After rinsing with PBS, sections were incubated for 2 h with secondary antibodies \n(Goat Anti-chicken Alexa Fluor 488 1:1000 and Goat Anti-rabbit Alexa Fluor 555 1:1000, A-\n10680, ThermoScientific, Waltham, MA, USA) with 1% normal goat serum in PBS with 0.3% \nTriton-X-100. Sections were rinsed with PBS, mounted with Fluoromount G and coverslipped. \n \nMicroscopy and Cell Counting \n \nSections were imaged using a Nikon Eclipse Ti2 inverted microscope with NIS -Elements \nsoftware using a 10x or 20x lens, using the same settings for exposure time and laser intensity \nfor all images . Images were adjusted for brightness and contrast in FIJI (version 2.1.0 ). \nRepresentative images from each of the two nuclei were obtained, one from each mouse for \neach distance from bregma. The appropriate extention of the nucleus was manually delineated \nand the cells were then counted cells manually. A researcher blind to the behavior protocol then \nrecounted a sample of slides  from each group  to ensure unbiased counting. Only cells with a \nlight intensity between 130-300 that had a uniform shape were counted. \n \nPainFace Analysis \n \nMice were filmed during behavior protocols as described above. Movies were cut using VLC \nMedia Player. The resulting clips had the mouse facing the camera during the behavioral \nstimulus with its face in clear view. Video clips were then stitched together using QuickTime \nPlayer and uploaded to the PainFace website (McCoy, Park et al. 2024). For analysis using the \nPainFace algorithm the following settings were used: fau model ID default(20250513-general), \npain-mgs model ID default(20221115-black) and Sample Rate High (1 Frame/Second). Scores \n(0-2) from the four face parts analyzed by the algorithm (Orbital, Nose, Ears, Whiskers) were \nextracted from the completed analyses and imported into Excel. After removing frames where \nthe face area was not visible (defined as a score of -1), average scores were obtained from each \nmouse and behavior test for each face area, as well as a total score for all four face areas. \n \nStatistics \n \nAll statistical analyses were conducted using IBM SPSS Statistics (Version 30.0.0.0, IBM \nCorp., Armonk, NY, USA). Data were first tested for normality using the Shapiro –Wilk test \nand for homogeneity of variances using Levene’s test. One -way ANOVA was perfo rmed to \nexamine the effect of treatment (Control, Brush, Brush+, Fur roll, Noxious heat, and von Frey) \non Fos expression, as well as proportional overlap between Fos and CGRP, and vice versa. \nTukey’s HSD post hoc tests were used for pairwise group comparis ons when the assumption \nof equal variances was met. When the assumption of homogeneity of variance was violated, \nWelch’s ANOVA was used instead, followed by Games –Howell post hoc comparisons. Post \nhoc analyses were conducted for all treatment groups. All results are presented as mean ± SEM, \nwith statistical significance set at p < 0.05. Graphs were generated using Prism (version 10.6.1). \n \nResults \n \nSoft touch stimuli are not aversive to mice. \n \nBecause the neural responses reported below compare activity evoked by soft touch and \nnoxious heat, we first sought to determine whether the soft touch stimulus might be perceived \nas aversive. As unpleasant stimuli in mice are known to increase facial grimacing, we quantified \ngrimace scores during the different behavioral protocols. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\nMice were filmed during the Brush, Noxious Heat, and von Frey protocols. As a baseline, mice \nwere filmed during unstimulated conditions in the test cage used for the Brush and von Frey \nassays (see figure 1 A). Because the stimulation periods were brief (~15 –90 s), frames were \nextracted during ongoing stimulation when the mouse was facing the camera. This resulted in \nrelatively few frames in which all four facial action units were clearly visible; therefore, partial \nframes in which one to three facial regions were visible were also included. Scores for each \nindividual facial action unit were summed to obtain an average grimace score. \nWhen comparing average grimace scores across conditions using a one -way ANOVA, mice \nexposed to noxious heat exhibited significantly higher grimace score s than mice in the Brush \n(p = 0.026), von Frey (p = 0.005), or Baseline (p = 0.002) conditions (figure 1B). \nWe next examined whether grimace scores during the Brush protocol changed over time, as we \nhypothesized that the unfamiliarity of the stimulus might initially increase grimacing in the \nearly days of habituation. Grimace scores were extracted from days 2, 4, 6, 8, and 10 of Brush \nprotocol habituation. We found  a significant difference between early and late days of \nhabituation, such that mice displayed higher average grimace scores on day 2 compared with \nday 8 (p = 0.026) and day 10 ( p = 0.022), and on day 4 compared with day 8 ( p = 0.029) and \nday 10 (p = 0.025) (figure 1C). No other comparisons were significant. \n \nBecause facial recordings could not be obtained during the fur-roll condition, we next assessed \nwhether mice showed avoidance  or preference  for a soft tactile environment using a place \npreference assay. Mice were habituated as described in Methods.  On the test day,  mice spent \nsignificantly more time on the blanket-covered side than on the bare floor (69% on the blanket \nside and 31% on the bare side (p < 0.001, figure 1D)). \nTaken together with previously published findings that investigate the same kind of protocols, \nwe conclude that the soft touch protocols are non-aversive. \n \n \nSoft touch engages neurons in the lateral external parabrachial nucleus. \n \nThe lateral external parabrachial nucleus (lePB) (figure 2 A) is an early hub in ascending \naffective sensory pathways and is robustly activated by aversive stimuli (Palmiter 2018, \nCondon, Yu et al. 2024, Palmiter 2024) . We therefore asked whether soft touch stimuli would \nalso engage neurons in this region. \nTo test this, neuronal activation was assessed using immunohistochemistry for the immediate \nearly gene Fos following exposure to our battery of  tactile stimulation protocols. Mice were \nexposed to one of three soft touch conditions: brushing of the fur (Brush), brushing combined \nwith mild non -noxious warmth and a blanket  (37 °C, Brush + ), and a Fur Roll condition in \nwhich animals were allowed to enter a fur -clad roll. As a comparison condition representing \ninnocuous non-affective touch, another cohort of mice were stimulated with a low -force von \nFrey filament (0.4 g). Another cohort of mice were exposed to a noxious heat condition to serve \nas a positive control.  Fos–positive neurons were quantified at three rostrocaudal levels of the \nlePB (−5.0, −5.1, and −5.2 mm from bregma), and mean counts across levels were used for \nstatistical comparisons. \n \nOne-way ANOVA revealed a significant effect of condition (F (5, 42) = 8.17, p < 0.001). Post \nhoc Tukey tests showed that all soft touch protocols and noxious heat significantly increased \nthe number of cfos–positive neurons in the lePB compared with home cage controls: Brush (p \n< 0.001), Brush + (p < 0.001), Fur Roll (p = 0.003) Noxious heat (p = 0.004). In contrast, low-\nforce von Frey stimulation did not increase Fos expression relative to home cage mice (p = 0.5) \n(figure 2 B-H). \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\n \nThe post-hoc comparison of the stimulation protocols further revealed that brushing-based soft \ntouch produced significantly more Fos expression compared to von Frey stimulation (Brush, p \n= 0.029; Brush +, p = 0.006), whereas Fur Roll did not differ significantly from von Frey ( p = \n0.166). Consistent with previous reports, noxious heat significantly increased Fos expression \nrelative to home cage controls (p = 0.039). Notably, Fos expression evoked by noxious heat did \nnot differ from those induced by the soft touch protocols (Brush p = 0.982; Brush + Heat, p = \n0.378; Fur Roll, p = 0.998).   \n \nTaken together, these findings confirm that noxious heat activates the lePB and demonstrate \nthat soft touc h, particularly brushing-based stimulation, also robustly activates lePB neurons, \nexceeding the activation produced by innocuous punctate touch. \n \n \nSoft touch activates a subset of CGRP neurons in the lateral external parabrachial nucleus. \n \nTo determine whether neurons activated by soft touch included CGRP-expressing neurons, two \nof the soft touch protocols (Brush and Fur Roll) were repeated in CalcaCre mice. This allowed \nus to assess the overlap between Fos –positive neurons and CGRP -expressing neurons in the \nlePB. \n \nWe first quantified the proportion of Fos –positive neurons that also expressed CGRP. A one -\nway ANOVA revealed a significant effect of stimulation condition on this proportion (F (4, 21) \n= 4.9, p = 0.006). Post hoc Tukey tests showed that the Fur Roll and noxious heat protocols \nsignificantly increased the proportion of Fos -positive neurons that were CGRP -positive \ncompared with home cage controls ( p = 0.008 and p = 0.004, respectively), whereas Brush \nshowed a similar trend (p = 0.055). Although von Frey stimulation produced fewer Fos-positive \nneurons overall, the proportion of Fos –CGRP co-localization did not differ significantly from \nthat observed in the Brush, Fur Roll, noxious heat, or home cage conditions. \nNext, we examined the proportion of CGRP neurons that were Fos -positive following \nstimulation. Stimulation condition again significantly affected this measure (one-way ANOVA, \nF (4, 21) = 4.8, p = 0.007). Post hoc comparisons revealed that both Brush and Fur Roll \nsignificantly increased the proportion of CGRP neurons expressing Fos compared with home \ncage controls (Brush, p = 0.014; Fur Roll, p = 0.011) (figure 3 A-G). \n \nTo determine whether this recruitment was specific to affective touch, we compared these \nresults with the von Frey protocol, a measure of innocuous punctate touch. In contrast to Brush \nand Fur Roll, neither von Frey nor noxious heat significantly increased the proportion of CGRP \nneurons that were Fos -positive compared with home cage controls, and these proportions did \nnot differ significantly from those observed in the other conditions.  Together, these results \nindicate that soft touch stimuli recruit a subset of CGRP neurons in the lePB. \n \n \nParvocellular subparafascicular nucleus  neurons are activated by noxious heat but modestly \nby tactile stimulation. \n \nThe SPFp (figure 4A) also contains a population of CGRP -expressing neurons and receives \nsensory inputs similar to those of the lePB (Kang, Liu et al. 2022, Kang, Liu et al. 2025) . We \ntherefore performed the same analyses in the SPFp using tissue from CalcaCre mice. \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\nWe first quantified the number of Fos -positive neurons following tactile stimulation. Because \nhomogeneity of variance was violated, Welch’s ANOVA was used. Welch’s ANOVA revealed \na significant effect of stimulation condition on the mean number of Fos-positive neurons in the \nSPFp (Welch’s F (4, 8.4) = 13.7, p < 0.001) (figure 4G). Games–Howell post hoc comparisons \nshowed that the noxious heat group had significantly higher Fos counts than home cage controls \n(p = 0.008), whereas none of the tactile stimulation conditions differed significantly from \ncontrols. \nWe next examined whether neurons activated by tactile stimulation in the SPFp also expressed \nCGRP. A one -way ANOVA revealed no significant effect of stimulation condition on the \nproportion of Fos-positive neurons that co-expressed CGRP (F (4, 19) = 1.8, p = 0.176). \nFinally, we assessed the proportion of CGRP neurons that expressed Fos following stimulation. \nA one-way ANOVA revealed a significant effect of stimulation condition on this measure ( F \n(4, 19) = 6.5, p = 0.002) (figure 4H-I). Post hoc comparisons showed that the Brush protocol \nsignificantly increased the proportion of CGRP neurons that were Fos-positive compared with \nhome cage controls ( p = 0.003), whereas Fur Roll did not ( p = 0.211). Similar increases were \nobserved following von Frey ( p = 0.022) and noxious heat ( p = 0.003) stimulation. Together, \nthese results suggest that tactile stimulation recruits a small subset of CGRP neurons in the \nSPFp without producing a strong increase in overall neuronal activation in this region. \n \n \nDiscussion \n \nWe here show that several forms of tactile stimulation, including soft affective touch,  recruit \nneuronal populations in the lePB  and modestly in the SPFp . Importantly, innocuous punctate \ntouch did not significantly increase neuronal activation in the lePB compared with home cage \ncontrols, suggesting that soft touch protocols carries an affective component or are more salient \nstimuli. Consistent with previous studies, noxious heat also increased activation of lePB \nneurons (Chiang, Nguyen et al. 2020). \nBecause CGRP neurons in these regions are typically linked to aversive processing, we first \nconsidered whether the soft touch stimuli might be perceived as unpleasant by the animals. \nHowever, several observations argue against this interpretation. Previous  studies have shown \nthat mice prefer both soft brushing and blanket -lined environments over unstimulated \nconditions (Liu, Qiao et al. 2022, Liu, Rahman et al. 2025). Consistent with these reports, mice \nin our place preference assay spent significantly more time on the blanket -covered side of the \ncage than on the bare floor. There has been considerable work done in the field of mouse \ngrimace analysis in the con text of stimuli with both positive and negative valence (Langford, \nBailey et al. 2010, Dolensek, Gehrlach et al. 2020, Le Moene and Larsson 2023). In our hands, \nfacial grimace analysis revealed no increase in aversive grimacing during brushing compared \nwith baseline or punctate non-affective touch conditions. Grimacing in mice also decreased \nprogressively during habituation  to brushing.  This potentially indicates a change towards \npositive valence .  Together, these findings suggest that the soft touch protocols used in the \npresent study are unlikely to be aversive. \n \nPrevious work has demonstrated that CGRP neurons in both the parabrachial nucleus and the \nSPFp contribute to the affective and motivational dimensions of pain  (Palmiter 2018, Kang, \nLiu et al. 2025) . In the present study, noxious heat, and soft touch produced similar levels of \nactivation within CGRP neuronal populations in the lePB. In contrast, activation within the \nSPFp was more limited, with overall neuronal activation primarily driven by noxious heat. \nThese findings suggest that although both regions contain CGRP neurons ass ociated with \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\naffective processing, the lePB may be more responsive to non -aversive tactile input than the \nSPFp.  \n \nAn additional consideration when interpreting Fos expression is that immediate early gene \ninduction often reflects relatively strong or sustained neuronal activation. Classic studies have \nshown that noxious stimulation reliably produces robust Fos inductio n throughout pain \npathways, whereas weaker or transient sensory inputs may not consistently engage these \ntranscriptional responses (Hunt and Mantyh 2001). In this context, the observation that gentle \ntactile stimulation was sufficient to induce Fos expression in a subset of CGRP neurons in the \nlePB suggests that these neurons can be recruited even by relatively mild sensory input.  One \npossible explanation for the recruitment of CGRP neurons by brushing is that this stimulus \nengages C-low-threshold mechanoreceptors (CLTMRs), unmyelinated afferents tuned to gentle \ndynamic stimulation of hairy skin. Activation of CLTMR pathways in mice has been shown to \npromote behavioral preference for gentle tactile stimulation, suggesting that these afferents \ncontribute to the positive valence of social touch (Liu, Qiao et al. 2022, Liu, Rahman et al. \n2025). The ability of brushing to recruit CGRP neurons therefore raises the possibility that \nCLTMR-driven signals can access ascending circuits that are typically associated with \nnociceptive and aversive processing. This convergence may reflect a broader role f or \nparabrachial neurons in signaling the salience or motivational significance of somatosensory \nstimuli, rather than exclusively encoding aversive input. Our data suggest that such affective \ntouch signals are more strongly represented in the parabrachial n ucleus than in the SPFp, \nconsistent with the robust activation of lePB neurons observed here. \n \nBoth the SPFp and lePB send projections to the amygdala, although they target distinct \nsubregions. CGRP neurons in the SPFp project predominantly to the lateral amygdala, whereas \nthose in the lePB project strongly to the central amygdala. These amygdala nu clei are known \nto play key roles in threat learning and defensive behaviors. The recruitment of these circuits \nby non-aversive tactile stimulation may therefore reflect the involvement of these pathways in \nbroader affective or salience processing rather than aversion alone (Kang, Liu et al. 2022, Pauli, \nChen et al. 2022) . For example, activation of inhibitory neurons within the central amygdala \nhas been shown to suppress pain responses  (Hua, Chen et al. 2020) , suggesting that a \nsubpopulation of parabrachial inputs may contribute to the modulation of nociceptive \nprocessing. \nIn addition to CGRP neurons in the lePB, our stimuli also activated neurons in the lateral dorsal \nparabrachial nucleus (ldPB), which contains a population of prodynorphin-expressing neurons \nthat project to targets distinct from those of CGRP neurons (Huang, Grady et al. 2021). Whether \nthe neurons activated in the present study correspond to this population remains to be \ndetermined. It is possible that soft tactile stimulation engages parallel parabrachial circuits \ninvolving both CGRP and prodynorphin neurons. \n \nFinally, parabrachial circuits may influence sensory processing through descending pathways. \nProdynorphin neurons in the ldPB send descending projections to the spinal dorsal horn and \nhave been implicated in the development of mechanical allodynia  (Huo, Du et al. 2023) . \nAlthough the descending projections of CGRP neurons remain less well characterized, \nparabrachial outputs to hindbrain and brainstem structures may similarly participate in the \nmodulation of sensory processing. \nIn summary, our findings demonstrate that CGRP neuronal populations in the lePB and SPFp, \npreviously associated primarily with aversive signaling, can also be engaged by non -aversive \ntactile stimulation. In particular, soft affective touch robustly activated neurons in the lePB, \nwhereas the SPFp showed only limited activation. These results suggest that ascending affective \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\npathways involving CGRP neurons may integrate both nociceptive  signals and tactile signals \nwith positive valence , potentially allowing shared circuits to encode the motivational or \nemotional significance of somatosensory stimuli. Future studies will be required to determine \nhow these circuits contribute to the perception of pleasant and unpleasant touch and how they  \ninteract with established pain-processing pathways. \n \nAcknowledgements:  \nThis work was supported by the Swedish Research Council (Vetenskapsrådet; grant number \n2021-01109 to LSL), The Wilhelm and Martina Lundgren Science Fund Foundation  (grant \nnumber 2025-SA-5082 to LSL ), and The Swedish Brain Foundation (Hjärnfonden, grant \nnumber FO2023-0399 to LSL) and The Mary von Sydow, née Wijk, Donation Fund (stipends \nto MH and AS; grant nr 2025-343 to FA). \n \nReferences \n \nBernard, J. F. and J. M. Besson (1990). \"The spino(trigemino)pontoamygdaloid pathway: \nelectrophysiological evidence for an involvement in pain processes.\" J Neurophysiol  63(3): \n473-490. \nBjornsdotter, M., L. Loken, H. Olausson, A. Vallbo and J. Wessberg (2009). \"Somatotopic \norganization of gentle touch processing in the posterior insular cortex.\" J Neurosci  29(29): \n9314-9320. \nCameron, D., E. Polgár, M. Gutierrez-Mecinas, M. Gomez-Lima, M. Watanabe and A. J. Todd \n(2015). \"The organisation of spinoparabrachial neurons in the mouse.\" Pain 156(10): 2061 -\n2071. \nChiang, M. C., A. Bowen, L. A. Schier, D. Tupone, O. Uddin and M. M. Heinricher (2019). \n\"Parabrachial Complex: A Hub for Pain and Aversion.\" J Neurosci 39(42): 8225-8230. \nChiang, M. C., E. K. Nguyen, M. Canto -Bustos, A. E. Papale, A. M. Oswald and S. E. Ross \n(2020). \"Divergent Neural Pathways Emanating from the Lateral Parabrachial Nucleus Mediate \nDistinct Components of the Pain Response.\" Neuron 106(6): 927-939.e925. \nChoi, S., J. Hachisuka, M. A. Brett, A. R. Magee, Y. Omori, N. -u.-A. Iqbal, D. Zhang, M. M. \nDeLisle, R. L. Wolfson, L. Bai, C. Santiago, S. Gong, M. Goulding, N. Heintz, H. R. Koerber, \nS. E. Ross and D. D. Ginty (2020). \"Parallel ascending spinal pathways for affective touch and \npain.\" Nature 587(7833): 258-263. \nCondon, L. F., Y. Yu, S. Park, F. Cao, J. L. Pauli, T. S. Nelson and R. D. Palmiter (2024). \n\"Parabrachial Calca neurons drive nociplasticity.\" Cell Rep 43(4): 114057. \nDeuis, J. R., L. S. Dvorakova and I. Vetter (2017). \"Methods Used to Evaluate Pain Behaviors \nin Rodents.\" Front Mol Neurosci 10: 284. \nDolensek, N., D. A. Gehrlach, A. S. Klein and N. Gogolla (2020). \"Facial expressions of \nemotion states and their neuronal correlates in mice.\" Science 368(6486): 89-94. \nHua, T., B. Chen, D. Lu, K. Sakurai, S. Zhao, B. X. Han, J. Kim, L. Yin, Y. Chen, J. Lu and F. \nWang (2020). \"General anesthetics activate a potent central pain -suppression circuit in the \namygdala.\" Nat Neurosci 23(7): 854-868. \nHuang, D., F. S. Grady, L. Peltekian and J. C. Geerling (2021). \"Efferent projections of Vglut2, \nFoxp2, and Pdyn parabrachial neurons in mice.\" J Comp Neurol 529(4): 657-693. \nHunt, S. P. and P. W. Mantyh (2001). \"The molecular dynamics of pain control.\" Nat Rev \nNeurosci 2(2): 83-91. \nHuo, J., F. Du, K. Duan, G. Yin, X. Liu, Q. Ma, D. Dong, M. Sun, M. Hao, D. Su, T. Huang, J. \nKe, S. Lai, Z. Zhang, C. Guo, Y. Sun and L. Cheng (2023). \"Identification of brain -to-spinal \ncircuits controlling the laterality and duration of mechanical allodynia in mice.\" Cell Rep 42(4): \n112300. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\nKang, S. J., S. Liu, J. H. Kim, D. I. Kim, T. G. Oh, J. Peng, M. Ye, K. F. Lee, R. M. Evans, M. \nGoulding and S. Han (2025). \"Thalamic CGRP neurons define a spinothalamic pathway for \naffective pain.\" Proc Natl Acad Sci U S A 122(28): e2505889122. \nKang, S. J., S. Liu, M. Ye, D.-I. Kim, G. M. Pao, B. A. Copits, B. Z. Roberts, K.-F. Lee, M. R. \nBruchas and S. Han (2022). \"A central alarm system that gates multi-sensory innate threat cues \nto the amygdala.\" Cell Reports 40(7). \nLangford, D. J., A. L. Bailey, M. L. Chanda, S. E. Clarke, T. E. Drummond, S. Echols, S. Glick, \nJ. Ingrao, T. Klassen-Ross, M. L. Lacroix-Fralish, L. Matsumiya, R. E. Sorge, S. G. Sotocinal, \nJ. M. Tabaka, D. Wong, A. M. van den Maagdenberg, M. D. Ferrari, K. D. Craig and J. S. \nMogil (2010). \"Coding of facial expressions of pain in the laboratory mouse.\" Nat Methods \n7(6): 447-449. \nLe Moene, O. and M. Larsson (2023). \"A New Tool for Quantifying Mouse Facial \nExpressions.\" eNeuro 10(2). \nLiu, B., L. Qiao, K. Liu, J. Liu, T. J. Piccinni-Ash and Z. F. Chen (2022). \"Molecular and neural \nbasis of pleasant touch sensation.\" Science 376(6592): 483-491. \nLiu, D., M. Rahman, A. Johnson, R. Amo, I. Tsutsui-Kimura, Z. A. Sullivan, N. Pena, M. Talay, \nB. L. Logeman, S. Finkbeiner, L. Qian, S. Choi, A. Capo -Battaglia, I. Abdus -Saboor, D. D. \nGinty, N. Uchida, M. Watabe-Uchida and C. Dulac (2025). \"A hypothalamic circuit underlying \nthe dynamic control of social homeostasis.\" Nature 640(8060): 1000-1010. \nLoken, L. S., J. Wessberg, I. Morrison, F. McGlone and H. Olausson (2009). \"Coding of \npleasant touch by unmyelinated afferents in humans.\" Nat Neurosci 12(5): 547-548. \nMcCoy, E. S., S. K. Park, R. P. Patel, D. F. Ryan, Z. J. Mullen, J. J. Nesbitt, J. E. Lopez, B. \nTaylor-Blake, K. A. Vanden, J. L. Krantz, W. Hu, R. L. Garris, M. G. Snyder, L. V. Lima, S. \nG. Sotocinal, J. S. Austin, A. D. Kashlan, S. Shah, A. K. Trocinski,  S. S. Pudipeddi, R. M. \nMajor, H. O. Bazick, M. R. Klein, J. S. Mogil, G. Wu and M. J. Zylka (2024). \"Development \nof PainFace software to simplify, standardize, and scale up mouse grimace analyses.\" Pain \n165(8): 1793-1805. \nMorrison, I., L. S. Loken and H. Olausson (2010). \"The skin as a social organ.\" Exp Brain Res \n204(3): 305-314. \nOlausson, H., Y. Lamarre, H. Backlund, C. Morin, B. G. Wallin, G. Starck, S. Ekholm, I. Strigo, \nK. Worsley, A. B. Vallbo and M. C. Bushnell (2002). \"Unmyelinated tactile afferents signal \ntouch and project to insular cortex.\" Nat Neurosci 5(9): 900-904. \nPackheiser, J., H. Hartmann, K. Fredriksen, V. Gazzola, C. Keysers and F. Michon (2024). \"A \nsystematic review and multivariate meta-analysis of the physical and mental health benefits of \ntouch interventions.\" Nat Hum Behav 8(6): 1088-1107. \nPalmiter, R. D. (2018). \"The Parabrachial Nucleus: CGRP Neurons Function as a General \nAlarm.\" Trends Neurosci 41(5): 280-293. \nPalmiter, R. D. (2024). \"Parabrachial neurons promote nociplastic pain.\" Trends Neurosci  \n47(9): 722-735. \nPauli, J. L., J. Y. Chen, M. L. Basiri, S. Park, M. E. Carter, E. Sanz, G. S. McKnight, G. D. \nStuber and R. D. Palmiter (2022). \"Molecular and anatomical characterization of parabrachial \nneurons and their axonal projections.\" Elife 11. \nVallbo, A., H. Olausson, J. Wessberg and U. Norrsell (1993). \"A system of unmyelinated \nafferents for innocuous mechanoreception in the human skin.\" Brain Res 628(1-2): 301-304. \n \n \n  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\nFigures \n \n \n \n \n \nFigure 1. Soft touch stimuli are not aversive to mice.  (A) Timelines for the behavior \nprotocols from top to bottom : Noxious Heat, Brush, Brush +, Fur Roll. (B) Average grimace \nscores of mice exposed to noxious heat are significantly higher in mice exposed to noxious heat \n(n=5, M=0.94, SE=0.07) as compared with Brush (n=7, M=0.71, SE=0.03, p = 0.026), von Frey \n(n=4, M=0.61, SE=0.09, p = 0.005), or Baseline (n=5, M=0.59, SE=0.05, p = 0.002) conditions. \n(C) Mice (n=7) exposed to the Brush protocol display lower average grimace scores over time. \nMice displayed higher average grimace scores on day 2 (M=0.97, SE=0.06) compared with day \n8 ( M=0.72, SE =0.07, p = 0.026) and day 10 ( M=0.71, SE=0.03, p = 0.022), and on day 4  \n(M=0.96, SE=0.04) compared with day 8 ( p = 0.029) and day 10 ( values, p = 0.025). (D) In a \nconditioned place preference test, mice spent significantly more time on the blanket -covered \nside than on the bare floor (n= 12. Bare side M=31, SE=3.91, Blanket side M=69, SE=3.91. p \n< 0.001). * = p < 0,05, ** = p < .005, *** = p < .001.  \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\n \n \nFigure 2. Soft touch engages neurons in \nthe lateral  external parabrachial \nnucleus. (A) Anatomical overview of the  \nmouse lateral parabrachial nucleus , \nshowing its dorsal (ldPB) and external \n(lePB) subdivisions , as well as  the \nsuperior cerebellar peduncle (scp).  (B-G) \nRepresentative micrographs showing Fos \nexpression (red)  after (B) noxious heat, \n(C) Brush, (D) Brush +, (E) blanket roll, \n(F) von Frey or (G) homecage conditions (scalebars 100m). (H) Noxious heat (n=8, M=43.75, \nSE=6.06, p = 0.004), Brush (n=14, M=49.86, SE=7.10, p < 0.001), Brush + ( n=4, M=67.75, \nSE=15.77,  p < 0.001), Fur Roll ( n=6, M=48.17, SE=5.10, p = 0.003) all show significantly \nincreased Fos expression as compared with homecage  (n=7, M=4.86, SE=0.63). In contrast, \nlow-force von Frey stimulation did not increase Fos expression  (n=9, M=22.56, SE=4.80, \np=0.5). * = p < 0,05, ** = p < .005.  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\n \n \nFigure 3. Soft touch activates a subset of CGRP neurons in the l ateral external \nparabrachial nucleus. (A-E) Representative micrographs showing Fos expression (red) and \nCGRP-GFP (green) after (A) noxious heat, (B) Brush, (C) blanket roll, (D) von Frey or (E) \nhomecage conditions (scalebars 100m). (F) Noxious heat (n=6, M=38.94, SE=5.77, p=0.004), \nand blanket roll (n=6, M=35.95, SE=4.31, p=0.008) show a significantly increased proportion \nof Fos-expressing neurons that also express CGRP , as compared with homecage (n= 3, M=0, \nSE=0). No such difference was found for Brush (n=6, M=27.71, SE=4.72) or von Frey (n=5, \nM=23.72, SE=8.71).  (G) Brush (M= 13.63, SE= 2.69, p=0.014) and blanket roll ( M=14.01, \nSE=2.67, p=0.011) show a significantly increased proportion of CGRP-expressing neurons that \nalso express Fos, as compared with homecage (M=0, SE=0). No such difference was found for \nnoxious heat (ME=9.91, SE=1.96) or von Frey (M=5.79, SE=1.85). * = p < 0,05, ** = p < .005. \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint \n\n \n \n \nFigure 4. Parvocellular subparafascicular nucleus neurons are activated by noxious heat \nbut modestly by tactile stimulation. (A) Anatomical overview of the mouse parvocellular \nsubparafascicular nucleus (SPFPC) and the medial lemniscus (ml). (B-F) Representative \nmicrographs showing Fos expression (red) and CGRP-GFP (green) after (A) noxious heat, \n(B) Brush, (C) blanket roll, (D) von Frey or (E) homecage conditions (scalebars 100m). (G) \nNoxious heat (n=6, M=6.72, SE=0.83, p=0.008) shows significantly increased Fos expression \nas compared with homecage (n=3, M=1.61, SE=0.15). Brush (n=6, M=7.44, SE=1.58), \nblanket roll (n=5, M=5.00, SE=0.87) and von Frey (n=4, M=5.13, SE=1.36) showed no such \ndifferences. (H) No difference in the proportion of Fos-expressing neurons that also expressed \nCGRP was found in any of the conditions, as compared to homecage. (I) Noxious heat \n(M=22.44, SE=2.90, p=0.003) and Brush (M=22.62, SE=2.53, p=0.003) show a significantly \nincreased proportion of CGRP-expressing neurons that also express Fos, as compared with \nhomecage (M=1.65, SE=0.87. No such difference was found for blanket roll (M=12.64, \nSE=1.99) or von Frey (M=19.19, SE=5.29). * = p < 0,05, ** = p < .005. \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 18, 2026. ; https://doi.org/10.64898/2026.03.15.711870doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}