Autonomic Nervous System and Possible Determinants of Social Behavior: A Systematic Review of its Relationship Among Psychopathy, Compassion, and Theory of Mind

Abstract

Objective

Sensitive alterations in the balance between physiological changes may be a relevant factor in the expression of social behavior and are systematically reported in a broad spectrum of psychopathological conditions. Among these, psychopathy, compassion and theory of mind hold a prominent role in social interaction. Biomarkers associated with such individual characteristics could represent an ambitious breakthrough in clinical practice, both diagnostically and therapeutically. The aim of this work is to evaluate the available evidence published in the past 20 years in order to investigate the relationship between the autonomic nervous system and the aforementioned individual characteristics.

Method

A literature search was conducted using dedicated search engines, including Pubmed, Scopus, and Google Scholar. The review focused on evidence published over the last two decades, from 2003 to 2024.

Results

Individuals with high trait psychopathy show reduced activation in skin conductance. This attenuation results in lower sensitivity to learning through fear conditioning mechanisms and an impaired ability to process affective stimuli. Studies focusing on compassion, while showing less variability in measures and proposed tasks, have revealed a possible effect of time in terms of efficacy, with therapeutic intervention programs (of at least 2 weeks' duration) also showing an impact on autonomic indices. Studies focused on theory of mind highlighted mixed findings and may require further investigation.

Conclusions

Interpretation of results may be limited by the numerosity and demographic characteristics of the participants, the plurality of autonomic indices used, and the heterogeneity of the proposed tasks. The ability to modify sensitivity to autonomic changes intentionally through possible specific therapeutic interventions could be a potential clinical goal.

Introduction

The autonomic nervous system (ANS) is a fundamental component of the human body, responsible for regulating involuntary physiological processes such as heart rate, blood pressure, digestion, respiratory rate, and thermoregulation. It operates through two main branches: the sympathetic nervous system, which facilitates “fight or flight” responses, and the parasympathetic nervous system, which supports “rest and digest” functions. In addition to these core roles, the ANS comprises several interconnected neural networks that regulate systemic responses across a wide range of conditions through interoceptive mechanisms. Consequently, the ANS also contributes to the control of other essential bodily functions, including bladder and gastrointestinal activity (Gibbons, 2019). Measurements of autonomic nervous system (ANS) activity have become essential tools in psychophysiological research, offering valuable insights into cognitive and emotional processes through physiological indicators. Numerous studies have explored the role of the ANS in various aspects of human health, employing a wide array of physiological measures to assess its function and dysfunction. For example, among the different available non-invasive methods to assess ANS activity, heart rate variability (HRV) and skin conductance response (SCR) have been widely employed in several research (Dawson et al., 2011; Shepherd et al., 2015; Sztajzel, 2004; Thomas et al., 2019).

Over the past decades, growing interest has emerged in understanding how individual psychological characteristics interact with autonomic nervous system (ANS) functioning. Since the late 1980s, researchers have sought to disentangle this relationship (e.g., Lester & Kimmel, 1989; Nardelli et al., 2015; Putnam et al., 1990), recognizing that traits such as personality, emotion regulation, and social cognition may exert a profound influence on psychophysiological processes. This line of research is rooted in the recognition that psychological differences significantly shape mental health outcomes and that investigating their physiological underpinnings may offer valuable clinical insights for the prevention and treatment of maladaptive and deviant behaviors. Among the many dimensions of psychological functioning, particular attention has been devoted to the role of the ANS in shaping social behavior. A substantial body of work has explored how physiological mechanisms support a range of social responses, including prosocial and aggressive behaviors (Fagan et al., 2017; Porges & Furman, 2011). Prosocial behavior is typically defined as voluntary behavior intended to benefit others (Eisenberg & Miller, 1987; Schroeder & Graziano, 2015), while proactive aggression is characterized by deliberate, goal-directed acts of harm toward others or objects without prior provocation (Poulin & Boivin, 2000; Puhalla & McCloskey, 2020). Recent work (e.g., Fanti et al., 2024) has examined the physiological mechanisms underlying aggression, particularly in the context of the hypoarousal hypothesis, which posits that proactive aggression may be associated with lower physiological arousal. However, findings in this domain remain inconsistent, and authors emphasize the need to account for moderating factors such as gender, parenting style, early adverse experiences, and broader contextual influences. These behavioral patterns are closely linked to core psychological constructs such as psychopathy traits, compassion, and Theory of Mind (ToM). Proactive aggression has been consistently associated with psychopathic traits (Porter et al., 2006; Woodworth & Porter, 2002), while individuals with such traits are often marked by a lack of prosocial behavior (Lehmann & Ittel, 2012; Lin & Xie, 2023). Conversely, prosocial behavior is positively associated with compassion—a motivational and emotional state characterized by feelings of care, concern, and a desire to alleviate others’ suffering (Leiberg et al., 2011). Moreover, ToM, defined as the ability to attribute mental states to others and predict their behavior accordingly (Baron-Cohen et al., 1985), plays a fundamental role in supporting prosocial actions. Indeed, individuals are more likely to act prosocially when they can mentally represent and understand the perspectives of others. Studies have shown that individuals with elevated trait psychopathy may exhibit some degree of perspective-taking (especially when employed for manipulative purposes), thus suggesting relatively intact cognitive mechanisms of understanding others (Carroll et al., 2021; Newman, 1998). Some authors suggest that perspective-taking could not be automatically adopted in social interaction by individuals with psychopathy traits, possibly due to an impairment in empathizing mechanisms (Drayton et al., 2018; Van Honk & Schutter, 2006).

Within this framework, the present review focuses on psychopathy traits, compassion, and ToM as three theoretically interrelated constructs that reflect distinct yet complementary aspects of socio-emotional functioning. While they may initially appear heterogeneous, they converge in their relevance to moral behavior, empathy, and interpersonal relationships. Their integration in this review is guided by the hypothesis that individual differences in ANS activity may underpin variations in social engagement, empathy, and the regulation of aggressive or prosocial behaviors. The concurrent investigation of psychopathy, compassion, Theory of Mind (ToM), and autonomic nervous system (ANS) functioning is crucial, as these domains are deeply interconnected in shaping emotional, social, and physiological processes. Psychopathy has been consistently associated with impairments in both compassion and ToM, capacities fundamental for the understanding and appropriate response to others’ emotions and intentions, ultimately influencing moral reasoning and social functioning (Campos et al., 2021; Johanson et al., 2020). These psychological deficits are paralleled by alterations in ANS functioning, notably a reduction in physiological reactivity to stress and fear stimuli, which constitutes a core psychophysiological feature of psychopathy. Furthermore, emotional intelligence has been posited as a critical mediator linking ANS activity to social behavior, thus elucidating the pathway through which physiological dysregulation may contribute to affective and interpersonal difficulties (Ling et al., 2018). Given the shared neural and cognitive bases of compassion and ToM, deficits in these domains may synergistically contribute to the maladaptive social and moral behaviors observed in individuals with psychopathic traits (Ishikawa, 2023; Song et al., 2023). Integrating these constructs within a unified framework allows for a more comprehensive understanding of the complex interplay between psychological characteristics and physiological processes, thereby advancing theoretical models and informing intervention strategies (Ali & Chamorro-Premuzic, 2010; Preckel et al., 2018; Thomson et al., 2018). Despite increasing interest in this field, several issues remain to be addressed. Studies on the association between ANS, social behavior and its determinants have pointed out issues that needed to be address to the debate, including the need to extend the body of evidence with further investigations on this research field. For instance, De Looff (2022) conducted a meta-analysis aimed to investigate the association between ANS and antisocial behavior, revealing that results from studies which investigated the aforementioned association were heterogeneus and bidirectional, pointing out the needed to further explore the aforementioned relationship. Insights encouraging further studies have been provided also by Zammuto et al., (2021) that review the available evidence on the association between ToM and HRV in individuals during development. Their findings revealed small to medium size association between HRV and performance on ToM tasks, and the absence of intervening variables in this association. Another meta-analysis (Di Bello et al., 2020) found a positive association between compassion and vagally-mediated HRV, with a medium-sized effect that appears robust to publication bias. However, the strength of the association varied depending on the specific methods used to assess HRV. Despite these findings, the results are limited by high heterogeneity among studies and overall poor methodological rigor. Finally, a more recent review summarizing available evidence between proactive aggression behavior, biology and physiology (Belfry & Kolla, 2021), provides useful insight for future research. This review shows that proactive aggression is linked to reduced resting heart rate and dampened sympathetic reactivity. It also highlights heritability of proactive aggression, amygdala dysfunction, and altered default mode network activity as key biological correlates. The findings underscore the complexity and multiple endophenotypes of proactive aggression. Aiming to offer a unified overview of the association between the ANS and traits such as psychopathy, proactive behavior, compassion, and Theory of Mind could also allow for improvements in treatment strategies. It could help clarify how certain ANS parameters may play a role in predisposed risk factors or protective factors, ultimately advancing our understanding and intervention practices. In order to offer a systematic summary and to provide insights and reflections on the possible interplay between ANS and certain psychological characteristics (i.e., psychopathy trait, compassion, and Theory of Mind) the present study was carried out. In line with our purpose, a comprehensive literature search, and a systematical evaluation of the recorded studies was performed. A separate section will be dedicated to each of the aforementioned psychological characteristics, with a table presenting the main findings from the selected studies. Each section will be followed by a brief discussion of the main findings, and the paper will conclude with a general discussion.

2. Materials & methods

2.1. Literature search

A comprehensive literature search, updated until November 2024 was conducted using the following databases: PubMed, ScienceDirect and Google Scholar. Common search strategies (TEXT and/or MeSH) were employed separately for each research construct: for psychopathy queries, terms like "Psychopathy AND heart rate”, “Psychopathy AND hrv”, “Psychopathy AND Respiratory sinus arrhythmia”, “Psychopathy AND Galvanic Skin Response”, “Psychopathy AND Electrodermal activity”, “Psychopathy AND Skin Conductance”, “Psychopathy AND ans” have been employed; for compassion queries, terms applied were "Compassion AND heart rate”, “Compassion AND hrv”, “Compassion AND Respiratory sinus arrhythmia”, “Compassion AND Galvanic Skin Response”, “Compassion AND Electrodermal activity”, “Compassion AND Skin Conductance”, “Compassion AND ans” ; for theory of mind queries, terms like "Theory of Mind AND heart rate”, “Theory of Mind AND hrv”, “Theory of Mind AND Respiratory sinus arrhythmia”, “Theory of Mind AND Galvanic Skin Response”, “Theory of Mind AND Electrodermal activity”, “Theory of Mind AND Skin Conductance”, “Theory of Mind AND ans” were used.

Searches were conducted on the overall content of English-language published papers, without specifying preferences for title or abstract. Additional articles were added by consulting the reference bibliography of the articles initially included and any published reviews on the selected topics.

2.2. Selection criteria

Papers included in the final systematic review followed these inclusion criteria: a) article written in English; b) publication of the article between 2003 and 2024 in a peer-reviewed journal; c) adult age of the study human participants (≥18 years); d) use of a research-validated index for measuring autonomic activity; e) use of measure(s) for quantifying investigated trait (psychopathy, compassion, theory of mind). Exclusion criteria were: animal studies, paediatric age of the participants and a diagnosis of a psychiatric disorder.

2.3. Data extraction

Once the selection process was completed, all included studies were synthesized by extracting the available data. Figure 1 follows PRISMA statement and illustrates selection process (Page et al., 2021).

Figure 1.

PRISMA flowchart illustrating selection process

Relevant information such as year of publication, sample size, grouping of participants, characteristics of participants (gender and age), type of task employed in the study and autonomic activity index have been reported in each paragraph in separate tables.

The included literature was finally discussed among the authors and organized by type of construct investigated. Tabulation allowed better visualization of the data, in order to assess the effect of the autonomic activity indices on the traits under review.

3. Results

After filtering through inclusion and exclusion criteria, a grand total of 69 studies were eligible for selection in the present review: 39 studies on psychopathy, 18 on compassion, 12 on theory of mind. The results from studies will be discussed separately per each section.

3.1. Psychopathy

The study of psychophysiology related to proactive aggression and psychopathy is relevant to understanding how individual differences in ANS functioning may be related to behavioural problems or the emergence of criminal behaviour. Since the early 2000 (Lorber, 2004), one of the most replicated findings in the field of psychopathy has been the reduction in electrodermal activity, which is also present in more recent reviews (De Looff, 2022). However, the uncertainty regarding the absence of other significant findings that can identify specific patterns of autonomic activity in individuals with elevated psychopathy using other indices may be attributed to the heterogeneity of measures used or to factors that may influence these measures (such as type of innervation: exclusively sympathetic, parasympathetic or the co-presence of both branches of the peripheral nervous system). A total of 39 articles have been included in the present review, with a major prevalence of studies employing gender-specific samples (22 with a male population, 2 with a female only population). Around half of articles included employed at least 2 different autonomic indexes (N=21). Most studies utilized electrodermal activity as main index for autonomic function (N=31), followed by cardiac indexes (N=24). Only 5 studies registered an absence of effect on autonomic indices (Deming et al., 2022; Kavish et al., 2019b; Ragsdale et al., 2013; Render et al., 2024; Sörman et al., 2016). The characteristics of the studies selected for review have been summarized and shown in table 1.

Table 1.

Studies included for trait psychopathy

Authors	Year	Groups	Participants Nr.	Mean (± std) age	Gender %M	Task / Condition	Autonomic index
Alshukri et al.	2025	2 (low vs high psychopathy)	49	25 (7.3)	47	Empathy for pain + pain induction	SCR
Armenti & Babcock	2018	Single group	135	32.02 (10.21)	100	Conflictual discussion with partner by offenders with history of domestic abuse	SCR HR
Babcock et al.	2005	3 (non-violent; low grade violent; clinically violent)	21; 37; 35	32 (9.6)	100	3 conditions: baseline, rage induction, conflictual partner discussion	HR
Babcock & Michonski Bare et al.	2019 2004	Single group Single group	79 76	29.9 (8) 18.9 (3)	100 22	Emotion labeling Guided imagery	SCR SCR HR
Benning et al.	2005	Single group	70	20 (N/A)	100	Emotional labeling (images)	SCR
Birbaumer et al.	2005	2 (controls vs criminal offenders with elevated psychopathy)	7; 5	33.08 (7.14)	100	Fear conditioning	SCR
Casey et al.	2013	Single group	61	40.5 (11.11)	100	Emotional regulation after exposure to emotionally impacting images (different strategies as condition)	HR
de Barros et al.	2013	Single group	30	18 (N/A)	100	Exposure to emotionally impacting images	SCR
Deming et al.*	2022	Single group	88	38.4 (7.6)	100	Exposure to emotional facial expressions	SCR HR
Dindo & Fowles	2011	Single group	115	19 (N/A)	100	Aversive conditioning	SCR
Florez et al.	2017	2 (low vs high psychopathic criminals)	91; 26	40.9 (11.2)	76	Implicit association task (IAT), IGT (Iowa Gambling Task)	HRV HRV-LF HRV-HF
Gao et al.	2012	2 (low vs high psychopathy)	45; 44	35.52 (10.2)	100	Social stress task (public discourse on personal faults)	SCR HR
Goulter et al.	2019	2 (factor1 psychopathy vs factor2 psychopathy PPI-SV)	101 (64;37)	19.02 (1.5)	0	Social rage induction / reciprocation task	HR HRV
Hansen et al.	2007	Single group	53	32.07 (7)	100	Cognitive test for executive functions	HR HRV
Hong et al.	2018	2 (low vs high psychopathy)	20; 20	(N/A)	65	Concealed information test	SCR HR RLL
Ibáñez et al.	2016	Single group	220	(N/A)	37.3	Trust game	SCR HR
Kavish et al.	2019a	Single group	292	18 (N/A)	100	HR recorded at rest	HR
Kavish et al.*	2019b	Single group	453	20.21 (2.85)	32	Social stress task (public discourse on personal faults)	SCR HR
Konicar et al.	2021	Single group	14	43.14 (11.52)	100	Neurofeedback training (EEG-mediated)	SCR
Kyranides et al.	2017	Single group	88	19.92 (0.99)	50	Arousal induction (emotional videos exposure)	SCR HR
Ling et al.	2018	Single group	156	35.72 (8.61)	100	Social stress task (public discourse on personal faults)	SCR HR
Osumi et al.	2007	4 (detached+antisocial, detached/non antisocial, non-detached/antisocial, control)	32 (8; 7; 7; 9)	20 (1)	57.5	Arousal induction (emotional video exposure)	HR
Osumi & Ohira	2010	2 (low vs high psychopathy)	28	19 (1)	53.6	Ultimatum Game	SCR
Osumi	2019	Single group	35	19.23 (1.22)	31.4	Ultimatum Game/Dictator Game	SCR
Pastor et al.	2003	3 (low vs medium vs high psychopathy)	10; 20; 18	30.67 (N/A)	100	Exposure to neutral and emotional images	SCR HR
Pfabigan et al.	2015	3 (community vs incarcerated low psychopathy vs incarcerated high psychopathy)	15; 16; 14	35.6 (11.58); 33 (11.5); 38.5 (13.4)	100	Exposure to unpleasant video (pain induction), empathy rating	SCR
Ragsdale et al.*	2013	Single group	54	20.41 (4.83)	50	Arousal induction (emotional images)	SCL
Raine et al.	2003	2 (controls vs participants with high psychopathy)	25; 15	28.8 (6.5); 31.6 (6.6)	100	Social stress task (public discourse on personal faults)	SCR HR
Render et al.*	2024	3 (sexual arousal vs neutral vs calm condition)	21, 20, 18	23.75 (4.21)	44	Temporal binding before and after condition induction	SCR HR
Rosenberger et al.	2019	2 (prisoners vs controls)	25; 23	35.41 (10.19); 34.65 (10.12)	100	Trust game	SCR
Rothemund et al.	2012	2 (controls vs criminals with high psychopathy)	11; 11	28 (6.7); 31 (6.4)	100	Fear conditioning	SCR HR
Seeger et al.	2024	2 (factor1 Psychopathy vs factor2 Psychopathy)	24; 17	26.54 (8.34); 24.18 (5.74)	100	Cyberball social exclusion paradigm	SCL HR
Sorman et al.*	2016	Single group	61 (26 for HR)	31.7 (17.6)	81.1	Empathy rating after exposure to painful stimuli	SCR HR
Stanger et al.	2012	Single group	66	19.95 (1.61)	60.6	Exposure to neutral and emotional images	SCR HR
Thomson et al.	2018	Single group	103	19.65 (1.22)	31	Horror VR simulation / Exposure to emotional images	SCR RSA
Thomson et al.	2019	Single group	83	19.57 (1.1)	0	RSA recorded at rest	RSA
Verschuere et al.	2005	Single group	37	39 (11)	100	Concealed information test	SCR HR RSA
Zimak et al.	2014	2 (low vs high psychopathy)	79	19.32 (1.54)	100	Iowa Gambling task, Exposure to emotional images	SCR

Note: * indicates reported absence of effect between autonomic indexes and psychopathy.

Legend: HR = heart rate, HRV = heart rate variability, HRV-LF= low frequency heart rate variability, HRV-HF= high frequency heart rate variability, RSA = respiratory sinus arrhythmia, RLL = respiration line length, SCR = skin conductance response, SCL = skin conductance levels, M=males.

So far, several studies have hypothesized that specific activation patterns of ANS are a basis for psychopathy and related antisocial behaviors, depending in part on the instruments used to determine these constructs (Armenti & Babcock, 2018; Casey et al., 2013; Goulter et al., 2019; Hansen et al., 2007; Kavish et al., 2019a; Ling et al., 2018; Verschuere et al., 2005).

The studies selected for the review employed different methods to assess differences in autonomic activation of participants with high trait psychopathy: ranging from decision making tasks, emotion induction, fear conditioning and tasks with social stress induction.

Most recent data about fear conditioning (Rothemund et al., 2012) show that participants with high psychopathy learn only partially about the association between conditioned stimulus and unconditioned response, without processing the emotional meaning of the stimuli: evidence for this can be found in poor anticipatory SCR (skin conductance response) and from a lack of responsiveness to the startle reflex in the association phase and during the extinction phase. These deficits do not appear to be attributable to insufficient assessment or responsiveness to the aversive stimulus, as there were no significant differences between groups in the ability to correctly perceive electrical stimuli or to rate them as painful. However, the authors underline that in the habituation phase individuals with high psychopathy would evaluate the stimuli as less painful, suggesting possible differences in the phasic evaluation of the nociceptive stimulus.

In the studies included in the review, regarding economic games, individuals with high psychopathy showed a pattern of attenuated physiological activation compared to controls or similar to their resting condition, associated with a tendency to make rational economic choices (even accepting unfair but convenient offers (Osumi & Ohira, 2010; Osumi, 2019) or by not reciprocating the trust in investments made by others in their favor (Ibáñez et al., 2016; Rosenberger et al., 2019). Similarly to what happens in aversive conditioning, no change of behaviour by individuals with high psychopathy is observed in economic tasks that punish for engaging in unfair behavior, even in case of repeated interactions (Osumi, 2019).

In studies that have dealt with social stress (Gao et al., 2012; Ling et al., 2018), the lack of somatic awareness in response to such contexts could be a predisposing factor toward the development of the psychopathic trait. Studies that focused on exposure or induction of emotional states in participants with high psychopathy show an effective reduction in the intensity of emotions at self-report level (Zimak et al., 2014). This reduction would be confined to certain emotions, such as sadness and fear (Flórez et al., 2017). Recent findings highlight a blunted nociceptive response associated with reduced empathy for pain regarding others in participants with high psychopathy (Alshukri et al., 2025).

Prominent evidence, entailing most recent (Deming et al., 2022) and largest sample (Kavish et al., 2019b) among studies included, have not found a relationship between high psychopathy and changes in physiological measures. In case of Kavish and colleagues (Kavish et al., 2019b), the authors illustrate how, although initial analyses show alterations between autonomic reactivity and trait psychopathy, correction for age and race then nullifies the interaction. This interpretation shows that there may be several factors to take into consideration in the evaluation of autonomic indices and therefore it would be relevant to consider potential confounders operating at an individual level. Regarding evidence from Deming and colleagues (Deming et al., 2022), limitations regarding electrophysiological data recordings and sample characteristics showing low median scores on PCL-R (total and Factor 1) may hamper the interpretation of null findings obtained.

3.2. Compassion

Compassion represents a multidimensional construct implicated in mammalian motivational systems of care. This construct is based on the sensitivity to perceive suffering in oneself and in others and a general commitment to actively try to alleviate and prevent it. A total of 18 studies on this topic have been included in the present review. Most studies investigated the role of heart rate variability (HRV) in modulating the expression of compassion, while fewer address the role of RSA (Comes-Fayos et al., 2024; Maeda, 2022; Oveis et al., 2009; Stellar et al., 2015). A significant majority of confirmed a relationship between HRV and compassion. Among included articles, two studies employed a female-only sample (Beck et al., 2017; Kemper & Shaltout, 2011) and three studies selected only male participants (Comes-Fayos et al., 2024; Luo et al., 2018; Maeda, 2022). Remaining studies included a mixed gender sample, with greater female prevalence. For the purpose of the review, studies included have not discriminated between “compassion” as a dispositional trait or as an outcome target for dedicated interventions, such as compassion focused programs or trainings. Studies included have been summarized in table 2.

Table 2.

Studies included for compassion

Authors	Year	Groups	Participants Nr.	Mean std) age (±	Gender %M	Task / Condition	Autonomic index
Beck et al.	2017	2 (mindfulness vs control)	33 (16; 17)	21.2 (N/A)	0	Mindfulness intervention - pre and post evaluation	HRV
Comes-Fayos et al.	2024	3 (Intimate partner violence vs General Offenders vs Controls)	124 (39; 42; 45)	39.18 (11.53); 39.14 (10.98); 33.93 (14.10)	100	VR Compassion-eliciting task	HRV RMSSD RSA
Di Bello et al.	2023	2 (active vs sham)	93 (51; 46)	24.57 (8.32), 23.52 (7.94)	21.5	tDCS (anodal/sham right Insula + empathic induction, Redistribution game)	HRV
Dor-Ziderman et al.*	2021	2 (“affective empathy”, “compassion)	74 (38; 36)	24.16 (3.17)	50	Emotion induction (video)	HR SCR
Gao et al.	2022	2 (Awareness training program vs waitlist)	85 (42;43)	45 (8); 46.7 (7.8)	N/A	7-week training ATP vs control on negative emotional stimuli induction and compassion/wisdom meditation	HRV SCR
Greig et al*	2024	2 (biofeedback vs no biofeedback)	60 (30;30)	22.5 (4.32); 23.77 (5.21)	46.5	VR single session self-compassion training	HR
Kemper & Shaltout	2011	Single group	5	43.6 (10.8)	0	LKM (Loving Kindness Meditation)	HR/ HRV
Kim et al.	2020	Single group	40	22 (0.49)	32.5	Compassion Focused Therapy (2 Week program)	HRV
Luo et al.	2018	2 (high vs low self-compassion)	34 (17; 17)	19.7 (0.7)	100	TSST (Trier social stress test)	RMSSD (HRV)
Maeda*	2022	2 (self-compassion condition vs control)	67	20.86 (1.72)	100	Maastricht Acute Stress Test	HR HRV RSA
Matos et al.	2017	2 (active vs control)	93 (56; 37)	23.34 (4.16)	10	Compassion meditation training (2 Week program)	HRV
Matos et al.	2022	2 (CMT-T, waitlist)	95 (55;40)	51.54 (8.21); 51.16 (6.09)	7.5; 6.6	CMT for Teachers (8 weeks)	HRV
O'Brien et al.*	2023	Single group	81	20 (3)	46	TSST	HR HRV
Oveis et al.	2009	Single group	80	20 (N/A)	25	Positive emotion induction (awe, pride, compassion)	RSA
Petrocchi et al.	2017	Single group	86	26.3 (7.8)	46.5	Mirror compassion induction	RMSSD (HRV)
Rockliff et al.	2008	Single group	22	N/A	N/A	Compassion induction through emotional images	HRV
Steffen et al.	2021	Single group	31	22	37.5	Compassion focused therapy (12 Week program)	HRV
Stellar et al.	2015	Single group	300	N/A	33	Mirror compassion induction	RSA

Note: * indicates reported absence of effect between autonomic indexes and compassion. Legend: HR = heart rate, HRV = heart rate variability, RMSSD = root mean square successive difference, HRV-LF = low frequency RSA = respiratory sinus arrhythmia, M=males.

The results of the included studies show how cardiac variability indices can be considered a biomarker of the individual ability to connect to social environments through a care-oriented approach, a peculiar feature of compassion (Di Bello et al., 2020) and a predictive factor in implementation of prosocial behaviors (van Kleef & Lelieveld, 2022). The relationship between RSA and compassion still appears to be uncertain, despite having the review article consisting of 4 different studies and with the largest number of participants (Stellar et al., 2015). Regarding cardiac autonomic indexes, most recent studies have focused on evaluating systematic interventions, with exercise programs to cultivate compassion of varying duration (from 2 to 12 weeks). In the case of individual exercise sessions as in Petrocchi and colleagues (Petrocchi et al., 2017), the degree of negative affect was not modified by the different experimental conditions. However, modulation of HRV was present only in the experimental condition, which required the conduction of the exercise of compassion in front of a mirror: the compassion practice could therefore enhance a socially-mediated type of response influencing the autonomic activity. Conflicting evidence has emerged in more recent studies (Dor-Ziderman et al., 2021; Greig et al., 2024; Maeda, 2022; O'Brien et al., 2023), highlighting possible differences in the operalization of “compassion” and a different outcome when comparing single session trainings or compassion as a dispositional trait and multiple session training.

3.3. Theory of Mind

Theory of mind (ToM) represents one of the main skills in the field of social cognition that systematically influences social interactions, as it is responsible for the attribution, inference and interpretation of the mental states of others. A total of 12 articles have been included in this review, that mostly investigate HRV as an autonomic index. Among those studies, several have reported null findings about the autonomic relationship (Blasberg et al., 2022; Cugnata et al., 2018; Deuter et al., 2018; Iorfino et al., 2015; Okruszek et al., 2017). Studies included have been summarized in table 3.

Table 3.

Studies included for Theory of mind

Authors	Year	Groups	Participants Nr.	Mean (± std) age	Gender %M	Task / Condition	Autonomic index
Blasberg et al.*	2022	Single group	118	40.1 (9.03)	45	EmpaToM, TSST (Confession about a simulated crime, audio-registered)	HR HF-HRV
Bögl et al.	2024	Single group	50	30.04 (5.1)	48	Affective reasoning in a gambling task with HR as cue	HR
Cugnata et al.*	2018	Single group	91	26.78 (N/A)	52	RMET (Reading the mind in the eyes)	RMSSD (HRV)
Deuter et al.*	2018	Single group	90	23.5 (3.5)	50	Multifaceted Empathy Test	RMSSD (HRV) / SCR
Engelen et al.	2023	Single group	32	24.72 (4.89)	47	Self/Other emotion rating (emotion induction through images)	SCR
Iorfino et al.*	2015	Single group	25	23.96 (2.19)	100	RMET (Reading the mind in the eyes, Baron-cohen)	HRV
Kalbe et al.	2007	Single group	17	22.4 (2.4)	23.5	ToM Stories (affective, cognitive, neutral)	SCR
Lischke et al.	2017	Single group	42	26.2 (0.43)	100	Resting state recording	HRV
Lyyra et al.	2018	2 (being seen in the interaction task vs not being seen)	60 (30; 30)	25.17 (5.2)	33.3	Mentalizing Task	SCR HR
Okruszek et al.*	2017	Single group	25	35.9 (12.5)	52	Hinting task / facial emotion identification	HRV
Quintana et al.	2012	Single group	65	20.91 (6.16)	46	RMET (Reading the mind in the eyes)	HRV
Wang et al.	2022	Single group	42	21.74 (2.19)	26	Negative video emotion induction	HRV

Note: * indicates reported absence of effect between autonomic indexes and Theory of mind.

Legend: HR = heart rate, HRV = heart rate variability, HRV-HF = high frequency heart rate variability, RMSSD = root mean square successive difference, RSA = respiratory sinus arrhythmia, SCR = skin conductance response, M=males.

Considering the included studies individually, only two show significant results (Lischke et al., 2017; Quintana et al., 2012). Both studies used the same task (RMET, Reading the Mind in the Eyes Test), traditionally used to assess Theory of Mind skills by attributing the correct emotion to affective images in which only the eyes of the expression are visible. Individuals with high HRV might be more adept at picking up cues related to approach and affiliation behaviors, so they would be more likely to enact these behaviors than people with low HRV. The ability to approach other people promotes the possibility of positive interactions by both individuals, due to a better regulatory ability by individuals with higher HRV (Quintana et al., 2012). Individuals with better ability to identify positive mental states show a greater high frequency pattern (HF-HRV) (Lischke et al., 2017), while also reflecting effortful emotional regulation associated with phasic oscillations on HRV (Zammuto et al., 2021).

Most recent evidence on this field (Bögl et al., 2024; Engelen et al., 2023) suggest a possible role of autonomic activity when accounting for perspective taking (self versus other), highlighting a possible role of neurovisceral information as cues for orienting social behaviour.

The absence of significant interactions, reported in several studies, poses a challenge for the interpretability of the results: features related to task difficulty in terms of cognitive resources between the two experimental conditions (Cugnata et al., 2018; Lyyra et al., 2018), the congruence between the investigated construct and the chosen task (Deuter et al., 2018), the effectiveness of the manipulation of the independent variable and integrity of the measurements (Blasberg et al., 2022; Iorfino et al., 2015), the ability to associate changes in online HRV with the task (Okruszek et al., 2017).

4. Discussion

The purpose of the present study was to carry out a systematic review of the available evidence on the association between autonomous nervous system indices (ANS) and several psychological outcomes. Direct or indirect stimulation of ANS may contribute significantly in the development process of new treatment interventions (Kamboj et al., 2023; Liebmann et al., 2023; Sousa et al., 2023; Weng et al., 2021) or as a relevant biomarker for diagnosis (Branchadell et al., 2023; de Araujo et al., 2023; Liu et al., 2023). Overall, this review encompassed the current literature on the ANS and its association with psychopathy, compassion, and theory of mind. Relevant findings have been separated for clarity in the paragraphs below.

4.1. Discussion: psychopathy

Studies investigating psychopathy employed a wide range of experimental tasks to assess differences in autonomic activation. It was found that most of the studies in this area had gender-specific samples. This gender specificity may limit the generalizability of the findings. Additionally, studies varied in participant characteristics, which further adds to the potential limitation in generalizing the results.

When analyzing studies that have investigated fear or aversive conditioning, it should be noted that the results of the studies selected fall within earlier research works. In contrast to what happens to control individuals, in which threat signals trigger defensive actions and can generate behavioral responses such as the startle reflex, in psychopathic individuals this network could be altered, due to an alleged inability to learn from aversive experiences (Benning et al., 2005; Kyranides et al., 2017; Pastor et al., 2003; Stanger et al., 2012). Economic games represent a thriving context for the assessment of decision-making skills, as individuals with higher trait psychopathy can meet success in the professional field more easily, in terms of promotions and economic benefits (Eisenbarth et al., 2018). This result could be due to mere boldness or a reluctance in adopting low-risk behaviors to prevent losses, a mechanism well documented in previous research through the Iowa Gambling Task (Bechara et al., 1997) and present in most healthy controls. Results from the selected studies also confirm this trend, despite their relatively small number. The reduced physiological response would be a consequence of insensitivity to risk, which results in hazardous choices in decision-making and gambling tasks. Studies focused on task employing social stress showed that people with high psychopathy would be able to detect social pressure correctly and use that information to report emotional experiences similar to controls, based on societal expectation (Gao et al., 2012). This experience constitutes a possible schism between bodily and mental experience, which impacts the ability to detect signs of distress in others (Pfabigan et al., 2015). Such empathic capacity and its expression in terms of psychopathic behaviors and blunted autonomic activity would also be mediated in these individuals by emotional intelligence, a trait that could be an interesting target for the construction of therapeutic interventions (Ling et al., 2018). When tasks regarding induction or exposure to emotional stimuli are performed, individual with elevated trait psychopathy show blunted affective experience, that could also derive from decreased ability to regulate those emotions. The complex combination of less perceived fear in dangerous situations, less fear of being punished, the use of positive reappraisal strategies in fearful experiences and the attenuation of the physiological response would show how individuals with high psychopathy seem bold in risky situations and lack remorse in adopting predatory behaviors (Blair et al., 2004; Thomson et al., 2018). Among the main characteristics of psychopathy, some authors hypothesize that there is a deficit in the ability to feel emotions due to decreased sensitivity in perceiving somatic changes. One possible hypothesis concerns the appropriate ability to detect emotions through exteroceptive, social clues: although physiological changes are attenuated, it would be possible to observe congruent self-reported feelings (Gao et al., 2012). Individuals with elevated psychopathic trait may correctly evaluate guilt and the wrongness of immoral acts, but may be less sensitive to its deterrent power, experienced by control participants; this mechanism may explain characteristical boldness and lack of care typically observed in psychopathic individuals (Seara-Cardoso et al., 2016). Although the effects of autonomic activation are often antagonistic, some studies have hypothesized that patterns of synchronous inhibition may be distinctive for individuals with high psychopathy (Thomson et al., 2018). Paradoxical opposing action could also be such an effect, considering the results of the study. However, it would represent a major shift in the paradigm for studying psychopathy and the influence of the autonomic nervous system, to be investigated through more extensive research studies.

4.2. Discussion: compassion

Studies on compassion have investigated both trait-compassion and efficacy for intervention using compassion exercises. The results of the included studies show how cardiac variability indices can be considered a biomarker of the individual ability to connect to social environments through a care-oriented approach, a peculiar feature of compassion (Di Bello et al., 2020) and a predictive factor in implementation of prosocial behaviors (van Kleef & Lelieveld, 2022).

Among the included studies, a separate mention should be made for the only study conducted in China (Luo et al., 2018), which reported a greater effect between HRV and compassion than the others. In consideration of the methodological expedients adopted by the study through the use of a paradigm that measures both at rest and under stress conditions such as the TSST (Trier Social Stress Test) (Kirschbaum et al., 1993), it is appropriate to hypothesize a possible influence originated by sample’s origin culture: the tendency toward a society-centered approach, present in Eastern cultures as opposed to an individual-centered approach, more typical of Western cultures, could partially explain these results. On the other hand, studies employing RSA as index showed uncertain results. Although some degree of individual variability is present, RSA is actively used to self-regulate emotions, especially distress related to negative emotions (Beauchaine, 2015). A possible interpretation of the results is that differences in RSA may be triggered by compassion as a response to distress related to induced suffering; another hypothesis could link an active type of regulation, mediated through RSA (with increase during the experience of compassion), to modulate the stress of experienced suffering (Stellar et al., 2015).

Results from pilot studies on compassion training have shown several effects of practicing compassion through structured exercise programs on neurophysiological activation (i.e., HRV): significant improvement in experiences of compassion for self and compassion from others (Matos et al., 2017); individuals with lower baseline HRV benefited more from the intervention (Kim et al., 2020); longer-lasting intervention modifies HRV responsiveness, indicating less susceptibility to self-criticism (Steffen et al., 2021). In case of comparisons with affective-type empathy conditions, the experience of compassion would induce an approach oriented to emotional regulation and therefore less affected by phasic oscillations in autonomic response: these interesting results shed new light on contrasting findings regarding non- significant synchrony between autonomic indexes and compassion induction (Dor-Ziderman et al., 2021). A careful consideration about the significance of HRV should be made about baseline self-compassion measures that may not be associated with variations in high frequency HRV (O’Brien et al., 2023).

A new approach is offered by the possibility of combining the effect of neuromodulation techniques with self-compassion exercise sessions. A very recent study aimed to verify whether the influence of the anodal transcranial direct current stimulation (tDCS) on the frontotemporal lobe modulated the activity of the right insular cortex and influenced the ability to help others in adjunction to a self-compassion practice (Di Bello et al., 2023). Although there has not been a behavioral modification, the finding obtained through the modification of the HRV reactivity could represent a promising result that needs to be investigated in depth in order to evaluate its use for new therapeutic intervention protocols.

4.3. Discussion: Theory of Mind

Lastly, this review examined evidence on ANS and theory of mind. Interpreting their results proved to be challenging due to the different types of tasks included in the studies and contrasting evidence. Individuals who exhibit high frequency heart rate variability (HF-HRV) may be more sensitive to subtle cue detection for positive emotions when interacting with others, thus inducing approaching behaviour (Lischke et al., 2017). However, in another study (Deuter et al., 2018) employing the Multifaceted Empathy Test, no correlation has been found between cognitive empathy and autonomic indexes (skin conductance, heart rate and heart rate variability). More recent evidence (Blasberg et al., 2022) indicates a more complex relationship when measuring autonomic indexes within the context of Theory of Mind tasks: findings show a blunted HF-HRV, that may be linked to the observation of stressful interaction chosen for the task (EmpaToM and Trier Social Stress Task). The lack of significant results in this area could be limited to the need to identify autonomic influences in more selective ways to discriminate sympathetic and parasympathetic activity. It will be necessary to clarify, through more robust methodological expedients (larger sample, recording of autonomic parameters at rest and during the task, and better operationalization of the Theory of Mind construct, with corresponding choice of a task suitable for the purpose), the possible relationship between autonomic nervous system and the abilities underlying Theory of Mind.

4.4. General discussion

Several autonomic indexes have been compared for efficacy, examining evidence from the selected studies. Electrodermal activity has gained prominent traction in the field of psychopathy and proactive aggression: findings regarding a blunted sympathetic response in conditions of stress and while processing relevant emotional information remain well replicated to this day. Skin conductance was the most utilized index used in studies entailing both psychopathy and proactive aggression. Tracking this type of data had limited application in the field of compassion and Theory of Mind: investigation of possible links between skin conductance response and these aspects may be a key component for evaluating efficacy in their therapeutic application. Cardiovascular activity, being modulated by the influences of the sympathetic nervous system (SNS) and parasympathetic nervous system (PNS), may be differentially involved in the processes implied in individuals with elevated psychopathy. In order to track the occurring changes in SNS and PNS activity in response to external stimuli, patterns of coactivation, coinhibition, or decoupled processes (with opposite activity) can be identified (Puhalla & McCloskey, 2020; Thomson et al., 2018; Thomson et al., 2019). The implication arising is the need to employ multiple indices of cardiac activity to isolate the different components in order to investigate the influence of each branch of the autonomic nervous system separately.

Based on the available evidence, neurophysiological changes induced by compassion practices appear to occur only in participants who are able to fully engage in self-compassion exercises (Kim et al., 2020; Steffen et al., 2021). In order to ensure effectiveness of the therapeutic intervention, it is crucial to identify the obstacles present to the person's engagement to adapt the practice. Daily activity through regular exposure to compassionate thoughts and images may have a key role. Regarding Theory of Mind, the role of neurophysiological markers needs to be clarified through standardization of practices and the implication of multiple indices per study. Efficacy in clinical practice of protocols based on Theory of Mind for the rehabilitation of individuals exhibiting high trait psychopathy or proactive aggression may be limited due to the possibility of completion through purely cognitive mechanisms, without tackling core aspects of these conditions, such as lack of remorse and fearlessness.

5. Conclusions

The current understanding of the interplay between psychopathy, compassion, and theory of mind and their respective links to autonomic nervous system activity reveals a complex nature that requires further investigation. Biological mechanisms could be accounted for the individual differences emerged between the presence of psychopathic traits, trait compassion and different capabilities of mentalizing ability: such variations may imply different expressions of social behavior exhibited. Validating autonomic activity as a possible variable of interest during assessment, routine monitoring and outcome could hold potential value in care planning, due to possible specific links to various mental disorders (De Looff, 2021). Careful consideration is needed when examining autonomic nervous system activity due to a plethora of relevant biological individual characteristics interfering with the interpretation of findings (Jeffrey et al., 2024). Future work should focus also on the possibility of accounting activation patterns instead of single autonomic indices (Kim & Hong, 2024; Qaiser et al., 2023; Thomson et al., 2018), in order to provide more insight on the complexities associated with social behaviour. Evaluation of possible biomarkers linked to autonomic activity may offer new diagnostic and therapeutic pathways that could impact both clinical and forensic field. Based on these findings, some studies have elaborated possible intervention programs tailored to target the treatment of psychopathy. For instance, Sousa et al. (2023) have conducted a study to test the efficacy of a 20-session individual intervention program based on compassion focused therapy (CFT). This program was focused in reducing psychopathic and aggressive behavior among offenders through the improvement of compassion. Its efficacy was tested through physiological indexes of emotion regulations, including hearth rate variability (HRV). Their findings suggested that people who have been involved in the intervention program had less difficulties in emotion regulation. Moreover, as showed by Kamboj et al. (2023), the activation of vagus nerve (VN) may have a key role in self-compassion. Overall broadening positive emotions, such as joy, compassion and pride can mitigate the physiological effects of negative emotions, promoting a return to a calmer, parasympathetic state (Fredrickson & Levenson, 1998; Kreibig et al., 2013). Some studies (Onwezen et al., 2013; Schneider et al., 2017) suggested that anticipating the feelings of pride may contribute to increased pro-social behavior and overall healthier lifestyle choices, and a recent meta-analytic review by Behnke et al. (2022) explored ANS responses to various positive emotions—including pride—highlighting a potential link between emotional experiences and physiological regulation. Therapeutic approaches such as Compassion-Focused Therapy (CFT), cognitive-behavioral interventions, and positive psychology strategies may help in promoting these adaptive emotional states. Further research may disentangle the existing links between patterns of autonomic nervous activity and singular facets of individual characteristics such as psychopathy, compassion and theory of mind through diverse population settings.