(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



A chemical signal in human female tears lowers aggression in males [1]

['Shani Agron', 'The Azrieli National Center For Human Brain Imaging', 'Research', 'Weizmann Institute Of Science', 'Rehovot', 'The Department For Brain Sciences', 'Claire A. De March', 'Department Of Molecular Genetics', 'Microbiology', 'Duke University Medical Center']

Date: 2024-01

( A ) Aggression ratings (APR) in Experiment 1, obtained after exposure to tears or saline. Each dot is a participant, n = 25. ( B ) The same data as in ( A ), presented in violin-plot. Each dot is a participant. The white dot represents the median, and the gray bar represents the quartiles. Saline in red and tears in blue. ( C ) Bootstrap analysis. Gray lines represent the 10,000 repetitions; the blue line represents the actual APR difference between saline and tears. ( D ) Scatter plots of the aggression ratings (APR) obtained in the MRI (Experiment 3) after exposure to tears or saline. Each dot is a participant, n = 26. ( E ) The same data as in ( A ) presented in violin-plot. Saline in red and tears in blue. The data in ( A ) and ( D ) are presented along a unit slope line (X = Y), such that if points accumulate above the line, this implies higher values after tears; if points accumulate below the line, this implies higher values after saline; and if points are distributed around the line, this implies no difference. Data used to generate graphs can be found in S1 Data .

In turn, we observed a remarkable reduction in aggression following exposure to tears. Whereas mean APR ± SD following trickled saline was 1.67±1.7, APR following tears was 0.94±0.92, or in other words, tears drove a 43.7% reduction in aggression (Shapiro–Wilk, W = 0.827, p < 0.001, implying a nonnormal distribution dictating a nonparametric test: Wilcoxon signed rank Z = 53, p = 0.031, effect size (r rb ) = 0.541, with no effect of order: Wilcoxon signed rank Z = 98, p = 0.555, effect size (r rb ) = 0.152. If we nevertheless use a parametric approach, the effect remains the same: t(24) = −2.68, p = 0.013, Cohen’s d = 0.527, with no effect of order: t(24) = −0.548, p = 0.59 (the effect remains even if we include the outlier in the analysis: Wilcoxon signed rank Z = 66, p = 0.05, effect size (r rb ) = 0.48) ( Fig 2A and 2B , S1 Data ). Finally, to further evaluate the robustness of this effect, we ran a bootstrap analysis. We randomly reassigned paired outcomes 10,000 times in order to generate a random distribution of results and then compared the actual result we obtained to this distribution. We observe that the chance probability to obtain this outcome is 2.9% ( Fig 2C ). These results suggest that, like in rodents, a primary chemosignaling function of human emotional tears may be a "stop aggression" signal. We next set out to ask whether the main olfactory system can respond to this perceptually odorless message.

Scatter plots of the normalized VAS ratings of tears and trickled saline for ( A ) pleasantness, ( B ) intensity, and ( C ) familiarity. Each dot is the average of 10 sniffs by a given participant; light-colored dots are from Experiment 1 (n = 22), and dark dots are from Experiment 3 (n = 24). The data in ( A - C ) are presented along a unit slope line (X = Y), such that if points accumulate above the line, this implies higher values after tears; if points accumulate below the line, this implies higher values after saline; and if points are distributed around the line, this implies no difference. Data used to generate graphs can be found in S1 Data .

In Experiment 1, we asked whether sniffing human emotional perceptually odorless tears reduces aggression in men as it does in male rodents. First, we harvested emotional tears from human female donors (6 regular donor women, age range 22 to 25 years) using methods previously described [ 13 ] (see Methods ). Because tears that trickled down the cheek and into the collection device may have collected skin-bound signaling molecules not originating from tear fluid, as a control substance, we trickled saline down the cheeks of the very same donors and collected it in a similar manner. Next, we used the point subtraction aggression paradigm (PSAP), a validated measure of aggression in response to provocation [ 31 , 32 ]. In brief, in the PSAP, participants play a monetary game with an opponent they are told is human but is, in fact, a computer algorithm. The game contains provocation events where money is “unfairly” taken from the participant, and revenge events, where the participant can deduct money from his opponent at no personal gain. Aggression is estimated by the aggression provocation ratio (APR), namely, the ratio between the number of revenge responses to the number of provocations the participant experienced. A higher APR reflects higher aggression. Before the PSAP, each participant went through a stimulus exposure procedure. Participants were told they are sniffing subthreshold concentrations of odors, but it was not stated at this stage what they were (they provided advanced consent for “assorted odors, including body odors”). A sniff jar containing 1 ml of stimulus was presented before the participant’s nose 13 times, with an approximately 35-second intersniff interval ( S1A Fig ). The first 3 sniffs were of saline solution (blank), and the following 10 sniffs were of the stimulus (either emotional tears or trickled saline). After each sniff, the participants used a visual analog scale (VAS) to rate the pleasantness, intensity, and familiarity of the stimulus. After this, a pad containing 100 μl of the stimulus (tears/trickled saline) was secured to the participant’s upper lip facing out, keeping the participant continuously exposed to the stimulus for the duration of the experiment. Participants (we recruited 31 but retained 25 men for analysis, age = 25.84 ± 3.46; see Methods for exclusion criteria) came to the lab on 2 consecutive days, at the same time of day, and engaged in a PSAP game, once after sniffing tears and once after sniffing trickled saline, counterbalanced for order, double-blind for condition. Consistent with previous results, we observed no perceptual differences between tears and trickled saline, which did not significantly differ in perceived intensity, pleasantness, and familiarity (Stimuli: F (1,20) = 2.53, p = 0.127; Descriptor: F (1,40) = 1.183, p = 0.317; Order: F (1,20) = 1.665, p = 0.212, with no interactions) (Figs 1A–1C and S2A–S2C and S1 Data ). Moreover, both stimuli did not perceptually differ from saline solution (blank), emphasizing the perceptually odorless nature of the stimulus (Blank versus Trickled Saline: F(1,21) = 1.815, p = 0.192, Descriptor: F(2,42) = 1.735, p = 0.189; Blank versus Tears: F(1,21) = 0.0004, p = 0.984, Descriptor: F(2,42) = 1.476, p = 0.24, without order effect or interactions for both stimuli) ( S2D–S2F Fig ).

The normalized luminescence from the OR response to tears or trickled saline, ranging in concentration from 1% to 3.16% (in CD293 simulation medium). A dose response to tears but not trickled saline was evident in receptors ( A ) OR11H6, ( B ) OR2AG2, ( C ) OR5A1, and ( D ) OR2J2. ( E ) No dose response was seen in the control empty vector—pCI. Each dot is the mean of 3 repetitions for either tears (blue) or saline (red), and the error bar is the standard error across 3 replications. A two-way ANOVA followed by a Sidàk’s multiple comparison test was performed at each concentration between the OR response to tears and saline (*** = p < 0.0001, ** = p < 0.001, * = p < 0.05, no symbol = not significantly different). Data used to generate graphs can be found in S3 Data .

To ask if the human olfactory system can process signals from tears, in Experiment 2, we expressed 62 human olfactory receptors (ORs) ( S1 Table ) in Hana3A cells and monitored their real-time activation by tears or saline using a luciferase-based assay as previously described [ 33 , 34 ]. In this initial screening, we observed 21 ORs activated by tears and not by trickled saline (quadruplicates for each receptor type. all T > 2.24, all uncorrected p < = 0.05) ( S3 Fig and S1 Table and S2 Data ). To further probe for a typical sensory response profile in these 21 candidate receptors, we repeated the experiment with 6 serial dilutions of emotional tears (between 1% and 3.16% v/v). This confirmed the OR response in 4 of these 21 ORs: OR2J2, OR11H6, OR5A1, and OR2AG2 (all done in triplicates or quadruplicates, (F(1,28) > 5.827, p < 0.023) ( Fig 3 and S3 Table and S3 Data ). In other words, human emotional tears, although perceptually odorless, activate specific human ORs in vitro, and this may provide the molecular basis for human social chemosignaling through tears. Having verified that this stimulus has pronounced impact on behavior and the potential for generating a response through the main human olfactory system, we next set out to ask how this is reflected in the brain.

Sniffing tears coordinates the brain response in reactive aggression

To gauge the brain response to sniffing tears in the context of aggression, in Experiment 3, we performed functional magnetic resonance imaging (fMRI) in participants playing the PSAP in the MRI scanner (we recruited 33 but retained 26 men for analysis, age = 27 ± 3.2; see Methods for exclusions), day after day, once exposed to tears and once to trickled saline, double-blind for condition (i.e., 52 scans in total). Again, we observed no perceptual difference between trickled saline and tears (Stimuli: F(1,22) = 1.4, p = 0.25; Descriptor: Mauchly’s sphericity test p < 0.5, Huynh–Feldt correction F(1,44) = 2.283, p = 0.124; Order: F(1,22) = 1.117, p = 0.3, with no interactions) (Figs 1A–1C and S2A–S2C). In turn, in this challenging experiment, the behavioral effect of tears on aggressive behavior was only subtle. The mean APR ± SD following trickled saline was 1.306 ± 1.6, and APR following tears was 0.967 ± 1.357 (Shapiro–Wilk, W = 0.857, p < 0.002, implying a nonnormal distribution dictating a nonparametric test). Given the results of Experiment 1 where tears significantly reduced aggression, we apply a one-tailed hypothesis: Wilcoxon signed rank Z = 67, effect size (r rb ) = 0.42, p = 0.048, one-tailed (Fig 2D and 2E and S1 Data). We think that this weaker (effect size of 0.541 in Experiment 1 versus effect size of 0.42 in Experiment 3) result reflects the psychological dynamics of the day-after-day MRI experiment, as participants were more aggressive on the second day regardless of condition. We detail this in S4 Fig.

We next explored the brain response to provocation under tears versus saline. We generated a whole-brain voxel-wise statistical parametric map (p < 0.005, cluster-corrected for multiple comparisons). Provocation versus inactive time regardless of condition revealed a typical salience network activation, which included typical provocation-associated regions [35] such as the right inferior and middle frontal gyri (S5 Fig) (see S3 Table for full list of regions). This pattern suggests that we effectively recruited the neural substrates of aggression typically activated in this task. In turn, the ANOVA contrast of provocation with the added interaction level of saline versus tears (p < 0.005, cluster-corrected for multiple comparisons) revealed no region where provocation under tears was associated with a significant increase in activity, but several brain structures where provocation under tears was associated with a significant reduction in activity (S4 Table for full list of regions). Notably, dampening rather than increasing activity by tears is consistent with the one previous functional neuroimaging study we conducted with tears [13], and the extent of this effect here is convincing considering the very strict criteria applied, namely, a significant interaction after correcting for multiple voxel-wise comparisons. Out of these regions where tears had this dampening impact, 2 regions have been repeatedly implicated in aggression [35]: the left anterior insula cortex (left AIC) [36] and bilateral prefrontal cortex (PFC) [37] (Fig 4A). Tellingly, we observe that the difference in the beta values between conditions (tears and saline) in these regions was significantly correlated with the difference in the level of aggression expressed in the scanner as measured by APR (left AIC Spearman rank correlation: r = 0.54, p = 0.006, PFC Spearman rank correlation: r = 0.41, p = 0.046) (Figs 4B, 4C,S6D, and S6E) (S1 Data). These correlations suggest that we captured a parametric link between brain and behavior, whereby tears are associated with dampening provocation-induced activity in the brain aggression network. We observe that although we balanced our study for order, the later subject exclusions violated this balance. To assure that these brain activity patterns were not a result of this imbalance, we conducted 2 analyses. We have 9 fMRI participants who had saline-first. Thus, a balanced group from this perspective is reduced to 18 participants, which is borderline in power. To overcome this, using a bootstrap approach, we randomly selected balanced groups of 18 participants 10,000 times, and each time conducted the analysis to create a distribution of beta values in the left AIC and PFC. The reduced activation in these regions remained significant (left AIC: mean t (17) = 2.417, mean p-value = 0.036, and for the PFC: mean t (17) = 3.814, mean p-value = 0.0015 (S6A and S6B Fig). Second, we conducted a whole brain analysis on a group of 18 participants balanced for order. We removed 6 tears-first participants by removing those with a lower Aggression Questionnaire (AGQ) score, to create 2 groups balanced in this respect (mean AGQ score for: saline first group = 35.89 ± 11.73, tears first group = 46.67 ± 14.4, p = 0.1). Once again, we observed the same pattern of reduced brain response as in the larger group (z threshold > 2.31, refracting a p < 0.01 corrected for multiple comparisons), and the correlation between behavior and brain response in the left AIC and PFC was also maintained (S6C–S6E Fig).

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Tears reduced activation in the brain substrates of reactive aggression. (A) Statistical map of the GLM ANOVA Provocation > inactive time contrast with an added level of saline vs. tears (tears < saline in blue; tears > saline in red), n = 24. GLM z threshold > 2.58, cluster corrected to p = 0.05. Color bars represent z-values. (B, C) Correlation between differences in behavioral APR scores (saline -tears) and differences in beta values (saline- tears) of (B) left AIC and (C) PFC. Each dot represents a participant, n = 24. The continuous line represents the fit. The dashed line marks the confidence bounds. Spearman rank correlation coefficient and p-values are depicted. Data used to generate graphs can be found in S1 Data; fMRI data are available at https://openneuro.org/datasets/ds004274. https://doi.org/10.1371/journal.pbio.3002442.g004

Next, to investigate how these regions may be modulating aggression under tears, we probed their functional connectivity with the entire brain under tears versus saline. We applied whole-brain psychophysiological interaction (PPI) analysis [38] using the left AIC and PFC functional regions of interest (ROIs) as seeds (p < 0.005, cluster-corrected for multiple comparisons). We observed that tears significantly impacted functional connectivity only for the left AIC, which under tears significantly increased connectivity specifically with the right temporal pole (right TP) extending into the right amygdala and piriform cortex (Fig 5). These brain regions share structural connectivity and constitute a functional brain network repeatedly implicated in olfaction [39] and aggression [40]. We further observe that the greater the difference in aggression between tears and saline, the greater the increase in connectivity associated with tears between the left AIC and right amygdala (Spearman rank correlation: r = 0.407, p = 0.048) (Fig 5D) (S1 Data). We did not, however, observe such a link with the right TP (Spearman rank correlation: r = 0.26, p = 0.217). In conclusion, tears significantly increased functional connectivity within a network of brain structures associated with aggression and olfaction, and this increase was correlated with the individual behavioral impact of sniffing tears. Combining the 2 imaging results, (1) that tears reduce provocation-related activity in the neural substrates of reactive aggression and (2) that tears increase functional connectivity between the neural substrates of reactive aggression and the neural substrates of olfaction, we conclude that tears coordinate the brain aggression response.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002442

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.