Social Relationships and Experimental Pain Sensitivity in Healthy Adolescent Girls: An Exploratory Preliminary Study

ABSTRACT

Introduction

Greater understanding of relationships between social factors and pain responses can provide new modifiable targets for pain interventions in adolescents. This observational preliminary exploratory study aims to examine the associations between experimental pain sensitivity and social factors in healthy adolescents.

Methods

Sixty‐three healthy girls (age 11.9 ± 1.3) completed comprehensive quantitative sensory testing assessments of ten experimental pain measures as well as three open‐ended self‐identity qualitative questions and the Patient‐Reported Outcomes Measurement Information System peer‐relationships family‐relationships surveys. Linguistic Inquiry Word Count (LIWC) was used to quantify the proportion of socially oriented words. Higher LIWC scores indicate a greater emphasis on social aspects in responses. Associations between experimental pain measures and social variables were examined using 10 regression models and 20 correlations.

Results

Most associations were not statistically significant. Contrary to the hypothesis, a greater use of socially oriented language (higher LIWC social score) was associated with greater pain sensitivity reflected by lower cold pain thresholds (p = 0.013) and higher heat pain ratings (p = 0.020). As hypothesized, greater social scores were related to greater pain inhibitory responses, with LIWC scores related to the conditioned pain modulation response (p = 0.020), and better peer relationships were related to greater offset analgesia response (r = −0.276, p = 0.048). No other significant associations were observed.

Conclusions

Social experiences may have a minor effect on experimental pain sensitivity and modulation in healthy adolescent girls. These findings suggest that, in this population, social experiences may have only a minor influence on experimental pain responses relative to other biopsychosocial factors.

Significance Statement

This exploratory study identified only very few associations, suggesting that social experiences may have a minor effect on experimental pain sensitivity in healthy adolescent girls. However, the few associations observed may point towards a possible association between greater inhibitory response and greater social support from family and peers, which could guide future studies investigating the role of social factors in pain modulation during adolescence.

1. Introduction

It is well established that biopsychosocial factors impact pain (Ashton‐James et al. ²⁰²²; Fillingim ²⁰¹⁷; Gatchel et al. ²⁰⁰⁷; Nahman‐Averbuch et al. ²⁰²³), yet the role of social factors, particularly social support, has received less attention (Kapos et al. 2024). Evidence indicates bidirectional relationships between social factors and chronic pain, with greater social support from family and friends related to lower chronic pain severity and vice versa (Chan et al. 2017; Eccleston et al. ²⁰⁰⁸; Forgeron et al. ²⁰¹⁰; Kashikar‐Zuck et al. ²⁰⁰⁷; Palermo et al. ²⁰⁰⁷; Ross et al. ²⁰¹⁸; Wolock et al. ²⁰²⁵). In contrast, the effect of social support on experimental pain is less clear. One study in healthy adults found that larger social networks were related to greater experimental pain responses in women but lower responses in men (Vigil et al. 2013). However, little is known about the relationships between social factors, including social identity and social relationships, and experimental pain sensitivity in healthy adolescents, which could be critical to understanding changes in pain during adolescence.

Adolescence is a critical period involving extensive developmental biopsychosocial changes, which include alterations in experimental pain sensitivity and in the development and reinforcement of self‐identity (Blankenburg et al. 2010; Hagy et al. ²⁰²⁵; Hirschfeld et al. ²⁰¹²; Tsao et al. ²⁰¹³). Self‐identity includes the social identity of peer and family relationships and differentiation and autonomy from parents occurring as adolescents transition to having more social support needs met by peers (Berenbaum et al. 2015; Guassi Moreira et al. ²⁰¹⁸; Meeus et al. ²⁰⁰⁵; Pfeifer and Allen ²⁰²¹). A relatively novel approach to measuring psychological, affective and social aspects of people's experiences is the Linguistic Inquiry Word Count (LIWC) (Tausczik and Pennebaker 2010). The advantage of this method is that it allows for quantifying qualitative data and provides individuals with a more natural way to express themselves, which could reveal important social aspects (Pennebaker et al. 2003). Using the LIWC method, recent studies found that adolescents with and without chronic pain had similar LIWC social scores (Nimbley et al. 2021). In adult patients with chronic pain, pain catastrophizing scores were related to social personal scores (Junghaenel et al. 2017), and patients with greater pain intensity had lower social scores for ‘friends’ compared to patients with low pain intensity (Camerota et al. 2023). A better understanding of the impact of social factors on individual differences in experimental pain sensitivity and modulation could provide a mechanistic explanation for their interrelationships with chronic pain. Moreover, it could advance the development of social‐based interventions to prevent or reduce chronic pain severity, especially during puberty.

The present study aimed to examine the relationships between experimental pain sensitivity and social factors in healthy adolescent girls. We hypothesized that greater social scores, indicative of greater social support, would be related to lower experimental pain sensitivity. We focused on the social factors related to self‐identity and family and peer relationships, which represent different domains of social development (Jones et al. 2021).

2. Methods

This study was approved by the Institutional Review Board of Washington University in St. Louis. Flyers were sent to parents of school‐age adolescents via PeachJar and the research participant registry at Washington University in St. Louis. All eligible participants were recruited for the study. All participants and their parents or guardians provided written informed consent and assent before initiating the study procedures. This is a secondary analysis of data collected in our lab from September 2022 to September 2025 (Banerjee et al. 2024, 2025; Nahman‐Averbuch et al. ²⁰²⁵). Patients or the public were not involved in the research.

2.1. Participants

Participants were female (sex and gender), between the ages of 9 and 16 years, had no diagnosis of chronic pain, psychiatric or neurological disorders or conditions associated with pubertal maturation, had no regular use of medications that could affect their pain sensitivity (e.g., anxiolytics/sedatives), were not using hormonal contraception, were neither pregnant nor breastfeeding, and were capable of understanding basic instructions and communicating in English. Participants were instructed not to take pain medications 24 h before the study visit and to avoid certain activities that might cause persistent pain or soreness in the days leading up to the visit.

Sixty‐eight participants were enrolled and signed the consent and assent. Eligibility criteria were confirmed again after signing the consent/assent. Four participants were withdrawn after signing the consent/assent due to taking medications that deemed them ineligible to participate in the study. One participant decided not to complete the study visit after signing the consent/assent.

2.2. Study Design

Participants completed a study visit that included sensory testing and answering surveys. The study design is described in detail elsewhere (Banerjee et al. 2024, 2025; Nahman‐Averbuch et al. ²⁰²⁵). In summary, participants completed cold and heat pain thresholds at the nondominant volar forearm and pressure pain thresholds at the trapezius which were delivered in a pseudorandomized order based on the participant ID number. Mechanical temporal summation was assessed using a 6.45 von Frey filament (10 stimuli), with participants rating pain intensity/unpleasantness following the first and 10th stimuli. The temporal summation value was calculated as the difference in pain intensity ratings between the tenth pinprick in the series and the initial single pinprick. A positive value would indicate a facilitatory effect. Then, a noxious heat stimulus (30 s at 46°C) and an offset analgesia stimulus of a three‐temperature heat stimulus (46°C for 5 s, 47°C for 5 s, 46°C for 20 s) were delivered to the nondominant volar forearm. Participants rated their pain intensity using a continuous visual analog scale (Medoc, Israel) with a scale ranging from no pain intensity to the most intense pain sensation imaginable. The offset analgesia (OA) response was calculated as the difference between the average heat pain ratings from 13 to 23 s for OA and the noxious heat stimulus (Nahman‐Averbuch et al. ²⁰¹⁴). A negative value would indicate an inhibitory effect. Participants also completed two conditioned pain modulation (CPM) paradigms in which a test stimulus was delivered alone and together with a conditioning stimulus. In the heatCPM paradigm the test stimulus was a noxious heat stimulus (46°C for 30 s) and the conditioning stimulus was immersion of the foot into a cold‐water bath (8°C for 60 s). In the pressureCPM paradigm the test stimulus was pressure pain thresholds at the trapezius and the conditioning stimulus was immersion of the foot into a cold‐water bath (8°C for 60 s). The CPM response was calculated as the difference between the average heat pain ratings or pressure pain thresholds delivered alone and together with the conditioning stimulus. A negative value would indicate an inhibitory effect. Lastly, participants completed a cold pain tolerance task in which they immersed their foot in cold water (8°C) for as long as they could tolerate it or up to 120 s (participants were not informed about the cut‐off time). Participants were first familiarized with the experimental procedure and the visual analog scales. Surveys were completed by participants during breaks between tests. The survey scores were calculated after the quantitative sensory testing was completed thus, the investigator conducting the quantitative sensory testing was blinded to the survey scores. Only the social measures relevant to this secondary analysis are described below.

2.3. Social Surveys

2.3.1. Social Scores of Self‐Identity Questions

Participants were asked three open‐ended questions about their self‐identity in the present and future. The questions were developed for this study by the study team (A.J) and related to the Present (‘If you met someone new today, what are the main things you would like them to know about you?’), Expected Future‐Reality (‘What do you think your life will look like in the future?’) and Expected Future‐Hope (‘What do you hope your life will look like in the future?’). For the future questions, participants were told to think about their hobbies, who they live with, what type of job they might have, or anything else that is important to them to provide a baseline framework for their response. Participants were informed to write as much as possible to help improve understanding of their thoughts and feelings for each question.

Prior to analysing the responses, a data quality check was performed, and identifiable information or protected health information such as participants' names, school names, etc., was removed. To improve interpretability and maximize dictionary matching, responses were minimally edited for spelling and punctuation without changing content. Participant responses were then analysed using Linguistic Inquiry and Word Count (LIWC) (Tausczik and Pennebaker 2010), a text analysis tool that can be used to understand the social state of an individual. The LIWC scoring identified social aspects found within participant responses. Sorting through language, and word choice, the scores provide a numerical value corresponding to terms found in the responses. The LIWC provides scores by counting words that are stored in its' dictionary (i.e., school, friend, parent, play) and then provides a calculated percentage of the words in the response that match LIWC's internal dictionary.

The overall social component was calculated separately for each question (Present, Expected Future‐Reality, Expected Future‐Hope) with a higher score indicating a greater emphasis of the participant on the social aspect in her answers to the self‐identity questions.

2.3.2. Patient‐Reported Outcomes Measurement Information System (PROMIS) Peer Relationships Survey

This survey assesses self‐reported quality of relationships with friends and includes aspects such as the quality of the relationships with friends, social involvement and friendship experience (Dewalt et al. 2013). The survey includes 7 questions such as ‘In the past 7 days I felt accepted by other kids my age’. Answers were based on a scale of 1–5, and the sum of all questions was calculated and used to determine the T score. A higher score indicates a stronger peer relationship (Dewalt et al. 2013). Overall, the survey has a good reliability (0.67–0.81) and weak to moderate validity (0.21–0.44) (Devine et al. 2018; Varni et al. ²⁰¹⁴).

2.3.3. PROMIS Family Relationships Survey

PROMIS Family Relationships is a social measure that assesses individuals' experiences with their family members such as family communication, trust, support and acceptance (Bevans et al. 2017; Flynn et al. ²⁰¹⁸). The survey includes 7 questions such as ‘In the past 4 weeks, I felt I had a strong relationship with my family’. Answers are based on a scale of 1–5, and the sum of all questions is calculated and used to determine the T score. A higher score indicates a stronger family relationship (Bevans et al. 2017). Overall, the survey has a strong reliability (0.84–0.99) and validity (0.52–0.77) (Bevans et al. 2017). The PROMIS Family Relationships survey was included only at a later stage of the study, and thus, only 50% of the sample completed this survey.

2.4. Statistical Analysis

The data were stored in the Research Electronic Data Capture system (REDCap, version 14.0.7), and statistical analyses were conducted using JMP Pro (Version 16.0.0, SAS Institute Inc., Cary, NC). Since this was an exploratory secondary analysis of the data, there was no power calculation and correction for multiple comparisons.

Ten separate regression models were conducted to assess the relationships between social scores of self‐identity (Present, Expected Future‐Reality, Expected Future‐Hope; independent variables) and the experimental pain sensitivity measures (dependent variables). The social scores of the self‐identity variables were either not related or weakly correlated, which did not present a computational problem due to collinearity; thus, all three social scores of self‐identity (Present, Expected Future‐Reality, Expected Future‐Hope) were included in the same regression model (see 3.1). In addition, since only 50% of the sample completed the PROMIS Family Relationships survey, it was not included in the regression models, and separate 20 Pearson correlations were used to identify the relationships between the PROMIS scores and experimental pain sensitivity. Normal distribution of the residuals was verified for all models. All extreme observations were determined to be valid and were therefore retained in the analyses (Karch 2023). Cook's distance, which assesses the influence of individual observations on the estimated regression coefficients, was calculated for all models (Belsley et al. 1980; Cook ¹⁹⁷⁷). All observations had Cook's distance values < 1, except for one observation. To evaluate its influence, the corresponding analysis was repeated after excluding this observation as a sensitivity analysis. A p‐value of < 0.05 was considered significant.

3. Results

Sixty‐three healthy girls participated in the study (11.9 $\pm$ 1.3). Most participants identified as White (76.1%), followed by Black (15.2%), mixed (6.5%) or Asian (2.2%). Descriptive information of the participants and the results is presented in Supplementary Table S1.

3.1. Social Aspects of Self‐Identity

Answers to the open‐ended self‐identity questions ranged from a few words to several sentences (mean = 17.6). For the social component of self‐identity Present question, participants often mentioned their hobbies and friends or family (‘That I like to run and do art (painting, drawing, etc.)’; ‘How close and grateful I am of the people I'm family and friends with’). For the questions related to their future, common answers focused on future career and status (‘I hope it's me going to college’; ‘I hope I am a vet’; ‘I would like to live in a mansion and have as much money as I could possibly get’; ‘Successful with a happy family’).

Using the LIWC, scores for each social component were calculated. The present social scores were not correlated with future social scores (r ² = 0.024, p = 0.219 and r ² = 0.057, p = 0.062 for Expected Future‐Reality and Expected Future‐Hope, respectively). The social scores of Expected Future‐Reality were significantly correlated with Expected Future‐Hope (r ² = 0.167, p = 0.001).

Of the 10 regression models tested, three showed significant associations between the LIWC social‐content scores and experimental pain sensitivity responses (Table 1). A greater social component in the Expected Future‐Reality responses was significantly related to lower cold pain thresholds (indicating a higher temperature and higher pain sensitivity, p = 0.013), higher pain ratings of noxious heat (p = 0.020) and a more inhibitory pressure CPM response (p = 0.020). Although no correction for multiple comparisons was applied because of the exploratory nature of this study, these associations are unlikely to remain significant after such correction. No significant relationships were found between any of the social factors and the other experimental pain measures including heat and pressure pain thresholds, pain ratings to cold stimuli, cold pain tolerance, temporal summation, offset analgesia and the heat CPM responses.

TABLE 1.

Experimental pain sensitivity and social scores for self‐identity questions.

	Unstandardized B	St. E	Standardized beta	B 95% confidence interval	T ratio	p
Pain thresholds Heat Pain Thresholds (n = 61)
Present	0.033	0.044	0.100	−0.055 to 0.121	0.76	0.453
Future‐Reality	−0.051	0.066	−0.109	−0.184 to 0.081	−0.77	0.443
Future‐Hope	−0.063	0.071	−0.126	−0.206 to 0.081	−0.88	0.385
Cold Pain Thresholds (n = 61)
Present	−0.200	0.116	−0.214	−0.431 to 0.032	−1.73	0.090
Future‐Reality	0.447	0.175	0.338	0.097 to 0.798	2.56	0.013
Future‐Hope	0.026	0.189	0.019	−0.352 to 0.405	0.14	0.891
Pressure Pain Thresholds (n = 61)
Present	−1.769	0.929	−0.247	−3.629 to 0.090	−1.91	0.062
Future‐Reality	−1.293	1.404	−0.127	−4.105 to 1.518	−0.92	0.361
Future‐Hope	−0.207	1.516	−0.019	−3.244 to 2.830	−0.14	0.892
Suprathreshold pain ratings to heat stimulus (n = 52)
Present	−0.028	0.430	−0.009	−0.892 to 0.836	−0.07	0.948
Future‐Reality	1.510	0.630	0.365	0.243 to 2.776	2.40	0.020
Future‐Hope	−0.567	0.642	−0.135	−1.858 to 0.724	−0.88	0.381
Pain ratings to cold stimulus (n = 61)
Present	−0.008	0.036	−0.027	−0.081 to 0.065	−0.21	0.836
Future‐Reality	0.090	0.055	0.230	−0.020 to 0.201	1.63	0.108
Future‐Hope	−0.034	0.060	−0.083	−0.154 to 0.085	−0.58	0.566
Cold pain tolerance (n = 61)
Present	0.298	0.592	0.068	−0.887 to 1.484	0.50	0.616
Future‐Reality	0.434	0.895	0.069	−1.358 to 2.226	0.49	0.629
Future‐Hope	0.362	0.967	0.055	−1.574 to 2.297	0.37	0.710
Pain modulation Mechanical Temporal Summation (n = 61)
Present	−0.008	0.008	−0.126	−0.024 to 0.008	−0.95	0.349
Future‐Reality	−0.008	0.012	−0.091	−0.032 to 0.016	−0.64	0.527
Future‐Hope	0.009	0.013	0.101	−0.017 to 0.035	0.69	0.490
PressureCPM (n = 61)
Present	−0.910	0.718	−0.160	−2.347 to 0.527	−1.27	0.210
Future‐Reality	−2.605	1.085	−0.322	−4.777 to −0.432	−2.40	0.020
Future‐Hope	−0.096	1.172	−0.011	−2.443 to 2.250	−0.08	0.935
HeatCPM (n = 51)
Present	0.025	0.280	0.013	−0.537 to 0.588	0.09	0.929
Future‐Reality	−0.282	0.410	−0.111	−1.107 to 0.542	−0.69	0.494
Future‐Hope	−0.095	0.418	−0.037	−0.936 to 0.746	−0.23	0.821
Offset Analgesia (n = 50)
Present	−0.173	0.304	−0.080	−0.784 to 0.439	−0.57	0.572
Future‐Reality	−0.887	0.437	−0.310	−1.757 to 0.003	−2.01	0.051
Future‐Hope	0.120	0.441	0.042	−0.768 to 1.007	0.27	0.788

Note: Ten separate regression models were conducted. All three social scores of self‐identity (Present, Expected Future‐Reality, Expected Future‐Hope) were included in one regression model.

Abbreviation: CPM, conditioned pain modulation.

3.2. Peer and Family Relationships

The relationships between the 10 experimental pain responses and the PROMIS peer and family relationship scores were tested separately, yielding 20 correlation analyses. Only two significant associations were identified. Higher PROMIS peer relationship scores were associated with a more efficient inhibitory OA response (r = −0.276, p = 0.048, Table 2, Figure 1). Similarly, higher PROMIS family relationship scores were associated with a more efficient inhibitory OA response (r = −0.608, p = 0.003, Table 2, Figure 1). For this association, one influential observation was identified (Cook's distance = 1.2). A sensitivity analysis excluding this participant attenuated the association to a non‐significant trend (r = −0.380, p = 0.067). Furthermore, because no correction for multiple comparisons was applied due to the exploratory nature of the study, these associations are unlikely to remain statistically significant after such correction. No other relationships were found between peer or family relationship scores and experimental pain measures, including heat, cold and pressure pain thresholds, pain ratings to heat and cold stimuli, cold pain tolerance, temporal summation and the CPM responses.

TABLE 2.

Relationships between experimental pain sensitivity and peer and family relationships.

	PROMIS peer relationships	PROMIS family relationships
Pain thresholds
Heat pain thresholds	r = −0.134, p = 0.294 (n = 63)	r = −0.223, p = 0.227 (n = 31)
Cold pain thresholds	r = 0.186, p = 0.144 (n = 63)	r = 0.278, p = 0.130 (n = 31)
Pressure pain thresholds	r = −0.133, p = 0.298 (n = 63)	r = −0.290, p = 0.114 (n = 31)
Suprathreshold pain
Pain ratings to heat stimulus	r = 0.070, p = 0.614 (n = 54)	r = 0.137, p = 0.544 (n = 22)
Pain ratings to cold stimulus	r = 0.087, p = 0.497 (n = 63)	r = 0.055, p = 0.770 (n = 31)
Cold pain tolerance	r = 0.073, p = 0.570 (n = 63)	r = −0.073, p = 0.695 (n = 31)
Pain modulation
Mechanical Temporal Summation	r = 0.065, p = 0.615 (n = 63)	r = 0.028, p = 0.882 (n = 31)
PressureCPM	r = 0.104, p = 0.417 (n = 63)	r = 0.301, p = 0.100 (n = 31)
HeatCPM	r = 0.107, p = 0.445 (n = 53)	r = −0.029, p = 0.901 (n = 21)
Offset Analgesia	r = −0.276, p = 0.048 (n = 52)	r = −0.608, p = 0.003 (n = 21) a

^a Because one influential observation was identified, a sensitivity analysis excluding this participant was performed. This attenuated the association to a non‐significant trend (r = −0.380, p = 0.067).

FIGURE 1.

The relationships between offset analgesia response and peer relationships and family relationships. A more efficient OA response was significantly related to (A) a greater PROMIS peer relationships score (r = −0.276, p = 0.048) and (B) PROMIS family relationships score (r = −0.608, p = 0.003); however, because one influential observation was identified, a sensitivity analysis excluding this participant was performed. This attenuated the association to a non‐significant trend (r = −0.380, p = 0.067). Peer and family relationships are calculated T scores of the PROMIS surveys. Higher scores indicate greater peer and family relationships. Lower negative scores indicate greater offset analgesia and inhibitory pain modulation response.

4. Discussion

The present study assessed the relationships between experimental pain sensitivity and social factors, including peer and family relationships and social aspects of self‐identity, in healthy adolescent girls. Overall, no relationships were found. However, preliminary evidence shows that adolescents who place greater emphasis on social aspects when describing their expected future identity may have higher experimental pain sensitivity while also demonstrating a more efficient inhibitory pain modulation response. Likewise, our exploratory analyses suggest that adolescents with greater and better social support and relationships may have a more efficient pain modulatory response. Given the exploratory nature of the study and the limited number of significant findings, these results should be interpreted with caution and warrant confirmation in future studies.

The few associations found between the self‐identity social scores and experimental pain measures and demonstrated unexpected relationships between greater social scores and greater pain sensitivity to experimental stimuli, as reflected by cold pain thresholds and pain ratings of noxious heat. Although contrary to our hypothesis, these findings are in line with a meta‐analysis showing that social support can have a wide impact on experimental pain including increasing experimental pain responses (Che et al. 2018). Moreover, a relationship between a greater social network and greater experimental pain responses in women was also demonstrated (Vigil et al. 2013). In addition, the self‐identity questions were open‐ended, and participants chose what to focus on in their answers. Thus, some participants did not mention social aspects at all. Interestingly, the associations were found exclusively with the social score for the Expected Future‐Reality question (What do you think your life will look like in the future?). The scores of the two questions about the future were significantly but surprisingly only weakly correlated with each other, while the scores of the question about the present were not correlated with the scores of the two questions about the future. This indicates that participants may have emphasized different social aspects when answering these questions. Since the self‐identity questions were not structured, the answers to future questions may have included the participants' views on optimism/pessimism about their future. More specifically, it is possible that the Expected Future‐Reality question evoked a pessimism‐dominant response while the Expected Future‐Hope question evoked an optimism‐dominant response. Both optimism and pessimism are related to experimental pain sensitivity (Costello et al. 2002; Geers et al. ²⁰⁰⁸; Goodin et al. ²⁰¹³). Interestingly, in a more detailed analysis, which included not only the total score of the Dispositional Optimism survey but also analyses of the sub‐scale scores, it was found that only the pessimism sub‐scale and not the optimism sub‐scale was related to experimental pain ratings (Hanssen et al. 2014). This could explain the relationships that were found in the present study, primarily between higher experimental pain sensitivity and the higher social score of the Expected Future‐Reality question, but not with scores of the Expected Future‐Hope question.

Our preliminary exploratory analyses found that social scores of the Expected Future‐Reality question were related to greater pain modulation of the CPM response. Similarly, peer and family relationships were related to pain modulation of the OA response. The social brain network is involved in processing complex social information and includes areas such as the medial prefrontal cortex, amygdala, anterior cingulate cortex, posterior superior temporal sulcus and temporo‐parietal junction (Becht et al. 2021; Burnett et al. ²⁰¹¹). These areas are also involved in pain processing and specifically in pain modulation of the CPM and OA (Alter et al. 2022; Apkarian et al. ²⁰⁰⁵; Nahman‐Averbuch et al. ²⁰¹⁴; Piche et al. ²⁰⁰⁹), which could explain the relationships found in the present study between greater social scores and greater pain modulation. Intriguingly, the relationships were found only with the pressure CPM but not the heat CPM. Previous studies found that CPM responses evoked by different test stimuli do not correlate with each other, suggesting the involvement of different mechanisms (Nahman‐Averbuch et al. ^{2019, 2013}; Sean et al. ²⁰²¹). Moreover, a meta‐analysis found that different psychological factors are related to different CPM responses, such that depression is related to heat CPM response, anxiety is related to pressure CPM response, and pain catastrophizing is related to electrical CPM response (Nahman‐Averbuch et al. ²⁰¹⁶). Thus, it is possible that different social factors would be differentially related to different CPM responses.

Importantly, the present study assessed correlations and not causations. Patients with chronic pain report more loneliness, being misunderstood, less social connections and participation in social activities, and lower family functioning (Forgeron et al. 2022, 2018, 2011; Kashikar‐Zuck et al. ²⁰⁰⁸; Lewandowski et al. ²⁰¹⁰; Linkiewich et al. ²⁰²³; Palermo ²⁰⁰⁰). The relationships between pain and social factors are bidirectional, with chronic pain impacting social experiences and social experiences impacting chronic pain (Block et al. 2018; Eccleston et al. ²⁰⁰⁸; Forgeron et al. ^{2013, 2010}; Jones et al. ²⁰²¹; Kashikar‐Zuck et al. ²⁰⁰⁷; Logan et al. ²⁰⁰⁶; Palermo et al. ²⁰⁰⁷; Ross et al. ²⁰¹⁸; Simons et al. ²⁰¹⁰; Wolock et al. ²⁰²⁵; Wurm et al. ²⁰¹⁸). Social factors can also impact acute experimental pain responses (Brown et al. 2003; Jackson et al. ²⁰⁰⁵; Karos et al. ²⁰²⁵; López‐Solà et al. ²⁰¹⁹; Lucas et al. ²⁰²⁴; McClelland and McCubbin ²⁰⁰⁸; Vervoort et al. ²⁰¹¹; Vigil et al. ^{2015, 2013, 2014}). Previous studies found that the presence of a significant other can reduce pain responses in participants with and without chronic pain (López‐Solà et al. ²⁰¹⁹; Montoya et al. ²⁰⁰⁴). In addition, a recent study found that the presence of a supportive figure (in person or online) reduced cold pain intensity (Karos et al. 2025). The literature in paediatric populations is more limited, and almost exclusively examines social impacts on experimental pain in the context of mother–child relationships. Healthy girls reported greater pain sensitivity to the cold pressor test when their mothers engaged with them in a pain‐promoting manner compared to the control group or to girls whose mothers engaged with them in a pain‐reducing manner (Boerner et al. 2017; Chambers et al. ²⁰⁰²). In addition, adolescent patients with cerebral palsy had lower pressure pain thresholds when they were with their mother compared to being alone or with a stranger; however, no differences between social conditions were found in healthy adolescents (Riquelme et al. 2021). In adolescents with and without migraine, greater heat pain thresholds were found when mothers were present at the study visit (Zohsel et al. 2006). Thus, social experiences can impact experimental pain response; though, unlike chronic pain, due to the brief and transient nature of the pain experience (in contrast to chronic pain), the reverse relationship is not expected to be present.

A strength of the study is the comprehensive assessment of experimental pain. In addition, the study evaluated complementary domains of social development by combining open‐ended self‐identity questions (Jones et al. 2021) analysed using linguistic methods, with validated questionnaires assessing peer and family relationships. There are several critical limitations of this study. First is the large number of regression models and associations that were conducted with no correction for multiple comparisons. Another limitation is the relatively small sample size for the family relationships survey, which was included only at a later stage of the study. In addition, the self‐identity questions were developed for this study by the study team, and the validity and reliability of these questions are unknown. The LIWC scoring system is optimal for long answers, but in the present study, answers ranged from a few words to a few sentences, which could impact the social scores. Moreover, this study focused only on girls. While social effects on experimental pain may be more salient for girls compared to boys (Boerner et al. 2017; Chambers et al. ²⁰⁰²), there is a need to study these processes in boys as well, and the results of the present study could not be generalized to boys. Lastly, the results of this study should be treated with caution as this is a secondary analysis and no power calculations for these analyses.

The relationships between social factors and chronic pain are well established, and social factors are recommended to be included in clinical and non‐clinical studies involving youth with chronic pain (Li et al. 2024). The present exploratory study identified only limited associations between social factors and experimental pain sensitivity and modulation in healthy adolescent girls, suggesting that other biopsychosocial factors may play a larger role in experimental pain sensitivity during adolescence. Future studies are warranted to replicate and confirm these preliminary findings, with particular attention to the potential influence of social factors on endogenous pain inhibitory mechanisms, which may represent a protective factor against the development of chronic pain during adolescence.

Author Contributions

H.N.‐A. takes responsibility for the integrity of the work. A.J. and H.N.‐A. conceived the project. A.M., G.B. and J.B. collected the data. E.M., P.M. and H.N.‐A. performed the analyses. E.M., R.C.‐O., K.E.B., A.J. and H.N.‐A. drafted the manuscript. G.B., J.B., A.M. and P.M. revised the manuscript. All authors approved the manuscript.

Funding

This study was funded by NIH (R01NS129742), the International Association for the Study of Pain (Early Career Research Grant) and the International Headache Society (Research into Headache in Children and Adolescents Seed Funding Grant).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Descriptive data.

Data Availability Statement

Data from the parent study are still being analysed for additional aims and are therefore not yet available; de‐identified data will be shared once analyses for all aims are complete.