Members Clinical Corner

Autonomic Arousal and Emotion in Victims of Interpersonal Violence: Shame Proneness But Not Anxiety Predicts Vagal Tone

The role of shame is increasingly recognised to be major in the experience of clients with complex trauma-related conditions. Indeed, Frewen and Lanius (2015: 106) describe shame as `the core affective counterpart to extreme negative self-referential cognitions’. In the following commentary on a paper by Freed and D’Andrea published in JTD in the same year, Ken Benau shares his insights about the salience of shame and the challenges it poses to clinicians and researchers as well as to the clients who suffer with it.

Happy reading!

Pam Stavropoulos

Focus article:

Freed, S. & D’Andrea, W.  (2015). Autonomic Arousal and Emotion in Victims of Interpersonal Violence: Shame Proneness But Not Anxiety Predicts Vagal Tone.  Journal of Trauma & Dissociation. 16, pp. 367-383.

Expert Contributors: Ken Benau, Ph.D.

Ken Benau, Ph.D. has been a licensed clinical psychologist and psychotherapist in private practice in the San Francisco Bay Area for twenty-seven years. Dr. Benau works with children, adolescents and adults in individual, couple and family therapy from an integrative perspective, paying special attention to attachment, emotion, somatic experience, and constructions of intrapersonal and interpersonal reality. For many years, he has studied and written about shame personally, from the inside out, and professionally, from the outside in. Dr. Benau views understanding shame and its counterpoint, authentic (non-hubristic) pride, as essential not only to our work as psychotherapists but to our shared humanity.

With the introduction of the DSM-5, shame is included in the diagnostic criteria of PTSD for the first time.  Specifically, under Criterion D, the diagnosis of PTSD states:

Criterion D: negative alterations in cognitions and mood

Negative alterations in cognitions and mood that began or worsened after the traumatic event(s)… as evidenced by two or more of the following:

2. Persistent and exaggerated negative beliefs or expectations about oneself, others, or the world (e.g., “I am bad,”…).

3. Persistent, distorted cognitions about the cause or consequence of the traumatic event(s) that lead the individual to blame
himself/herself or others.

4. Persistent negative emotional state (e.g., fear, horror, anger, guilt, or shame)”,  (Desk Reference to the Diagnostic Criteria from DSM-
, 2013, p. 145).

Note that PTSD criterion D2 refers to negative beliefs about oneself, D3 refers to self-blame, and criterion D4 specifically mentions shame as an example of a negative, trauma-related emotion.  All of these speak to the operations of shame on either the trauma survivor’s post-trauma beliefs about self and others, or directly to the survivor’s emotional experience.

With these additional criteria, Freed and D’Andrea (2015) recognize that shame is not only an important aspect of PTSD generally but particularly in instances of interpersonal violence (IPV) where there is a dominant perpetrator or “shamer” and submissive victim or “shamed” person.  These authors wisely appreciate that in IPV, the so-called victim experiences both a threat to her physical survival and a threat to her social survival.1   Based upon my clinical observations, but not stated by the authors, the threat of bodily harm tends to be more associated with terror/anxiety, whereas social threat is more often associated with shame in that shame is fundamentally about inclusion in or exclusion from one’s social group.  The authors do hypothesize, however, that “… the biological challenges arising from fear and from shame may differ” (Freed & D’Andrea, 2015, p. 369). This is also consistent with my clinical observations.

Freed & D’Andrea (2015) observe that with regard to neurophysiology, PTSD is associated with hyperarousal and reduced parasympathetic (PNS) tone, with the PNS serving, under optimal conditions, to modulate post-traumatic hyperarousal.  It follows that reduced PNS tone would be associated with respiratory sinus arrhythmia (RSA) in response to trauma-related stimuli.  As the authors observe:

“Studies of PTSD have found reduced parasympathetic tone, measured by respiratory sinus arrhythmia (RSA) at baseline and in response to trauma-related information…” (Freed & D’Andrea, 2015, p. 369).

The authors go on to suggest:

“Although chronic fear-related ANS changes are well established in

DSM–IV PTSD, less is known about the physiological concomitants of chronic shame.  Self-esteem, related to shame proneness (Tangney & Dearing, 2002), is inversely related to resting RSA (Martens et al., 2010)”, (Freed & D’Andrea, 2015, p. 370).

Thus, disruption in RSA is hypothesized to be associated with shame – especially shame evoked post-trauma that I refer to as “shame states” – to be differentiated from non-traumatic shame as emotion (Benau, 2017, pp. 6-9).

Freed & D’Andrea (2015) go on to outline clearly the relationship between social status, shame proneness, and the neuroendocrine correlates of stress:

“Social hierarchy research has established that lower social status is associated with the presence of stress hormones indicative of SNS activation in both human community samples (Gruenewald, Kemeny, & Aziz, 2006; Rohleder, Chen, Wolf, & Miller, 2008) and among members of stigmatized groups where shame proneness corresponds to poor health outcome (Persons, Kershaw, Sikkema, & Hansen, 2010 ).  These endocrine findings suggest that shame may be autonomically arousing; however, autonomic correlates of shame in PTSD have not been examined.” (Freed & D’Andrea, p.  370).

It is here that the authors and I both agree and part company.  Freed & D’Andrea (2015) contend, as indicated above, that shame is associated with SNS activation and thus may be hyper-arousing.  In my clinical observations, shame is both hyper-arousing in the early phases of psychosocial threat, and then rapidly down-regulating and hypo-arousing, in the latter phases of what I refer to as “shame proper” (Benau, 2017, pp. 3-4).  Part of what makes shame states so disorganizing is that when the traumatic interpersonal event is recalled, the trauma survivor will likely experience rapid cycling between states of hyper-arousal (“I’m about to be socially excluded!”) and hypo-arousal (“I’m out, and I have no way back into my tribe!”).  This is congruent with the disorganized attachment (child) and unresolved attachment (adult) paradigms, where the traumatized person experiences fright without escape (Liotti, 2004).

For their research, Freed & D’Andrea (2015) hypothesized that while shame along with fear and anxiety would each heighten arousal, shame would cause even greater post-traumatic stress arousal.  Specifically, they posit:

“Hypothesis 1 (H1): Trait and state anxiety, fear, and shame will be associated with increased physiological activation, indicated by a negative correlation with RSA and a positive correlation with SCL [skin conductance].

Hypothesis 2 (H2): Trait and state shame will contribute significant variance to physiological arousal after the effects of anxiety and fear are accounted for.” (Freed & D’Andrea, 2015, p. 370).

The subjects of this study were 27 adult women, each with a history of physical or sexual abuse in childhood, and 52% with both (Freed & D’Andrea, 2015, p. 371).  This resulted in one important but unspecified limitation of the study, in that the subjects had distinct abuse histories that could have differentially affected the results.

These researchers measured fear (“afraid”, “scared”), anxiety (“nervous”, “jittery”), and shame (“ashamed”) using the Positive and Negative Affect Scale (PANAS) (p. 371).  Unfortunately, ratings merged the self-conscious emotions, i.e. guilt and shame, making the findings regarding shame somewhat misleading.

The researchers also assessed the subjects’ respiratory sinus arrhythmia (RSA) during a two-minute baseline, stress (5 second exposure to trauma related, violent images) and recovery conditions. RSA refers to the alteration of heart rate due to breathing, and is used as an index of activity of the parasympathetic nervous system (PNS), that is the organism’s capacity, following stress, to slow heart rate and return to homeostasis or relative calm following threat (Freed & D’Andrea, 2015, pp. 371-372).  Likewise, the researchers measured skin conductance (SCL) during baseline, stress and recovery conditions, as SCL is associated with the tonic level of sympathetic nervous system (SNS) (Freed & D’Andrea, 2015, p. 372).

This study’s findings demonstrated the following:

  1. Trait shame was significantly correlated with both fear and anxiety (Freed & D’Andrea, 2015, p. 372). This suggests, to me, that shame may induce anxiety by reminding the subject of how social exclusion or negative judgment has already occurred, and could occur again.
  2. Higher trait shame was related to lower RSA at baseline, during trauma reminders and during recovery, indicating that shame proneness was predictive of PNS withdrawal (Freed & D’Andrea, 2015, p. 373). Said another way, subjects who were generally prone to experiencing shame had higher arousal levels and slower recovery to baseline than those who were not shame prone.
  3. Higher state shame was associated with significantly lower RSA at baseline and during recovery (Freed & D’Andrea, 2015, p. 373). Likewise, those in the grips of shame following violent imagery or traumatic triggers also tended toward higher arousal, showing again that shame was negatively correlated with higher arousal and slower recovery to baseline.
  4. “Though trait and state shame were the only significant affective correlates of physiology, a series of regression analyses were performed predicting RSA from affect in epochs in which fear or anxiety approached significance”, (Freed & D’Andrea, 2015, p. 375). Again, this showed that shame more than fear or anxiety contributed to disruptions in physiological signs of stress, specifically lower RSA, suggestive of less parasympathetic nervous system modulation of arousal.
  5. “PNS deactivation was associated with fear and shame at baseline, anxiety and shame during trauma reminders, and shame alone during recovery. Shame marginally predicted larger SNS responses when participants were viewing trauma-related information, but fear and anxiety did not”, (Freed & D’Andrea, 2015, p. 376).  Thus, while for IPV trauma survivors there was heightened arousal for fear and shame at baseline, and for anxiety and shame during trauma reminders, only shame significantly weakened arousal recovery (RSA).

I understand the previous (#5) finding in the following way:  Shame, borrowing from the work of social psychologist Thomas Scheff (2007), tends to be recursively recalled by the IPV trauma survivor.  That is, not unlike other traumatic states, traumatic shame states repeatedly remind the survivor of the shaming event.  As a result, the IPV survivor who is filled with shame has a much harder time returning to post-trigger arousal levels, because even after the external trigger is removed the internal triggers (recurrent memories of the shaming event) are not.  One could also argue that the mind/brain repeatedly reminds the trauma survivor of the shaming events not so much to punish the survivor, as to warn him or her to ensure “this never happens again”.  Of course, the irony or tragedy is that these recurring internal reminders result in the shamed trauma survivor re-experiencing repeatedly the violent and social exclusion.

  1. “Trait shame was the only predictor of RSA during recovery, indicating that shame-prone individuals were less able to recover from the effects of trauma reminders.  State shame was the only significant predictor of RSA during the task and added significant variance over fear/anxiety at baseline.” (Freed & D’Andrea, 2015, p. 376).

Thus, shame as a generalized trait makes physiological arousal recovery more difficult, but shame as a state makes a trauma survivor, when triggered, more hyperaroused.  This makes perfect sense, as shame as trait establishes a certain emotional baseline, while shame as traumatic state makes the survivor more likely to be triggered by shame-inducing stimuli.

  1. The authors concluded that, “Although traditionally clinicians have looked for symptoms such as hyperarousal, in this population, shame may indicate a failure of emotion regulation that indexes physiological dysregulation”, (Freed & D’Andrea, 2015, p. 376). I take this to mean that shame is dysregulating not only because it increases arousal when a person is faced with a threat or re-imagined threat of social exclusion, but shame also rapidly decreases arousal in the form of submit or collapse response to threat.  Thus, the dysregulation these authors identified is due to both hyper- and hypo-arousing responses, in response to perceived or, following Porges (2011, p. 11) implicitly neuroceived threat.

The authors go on to interpret their research findings in the following ways:

“Shame may affect symptomatology in survivors of IPV by compromising the effectiveness of help-seeking behaviors (Black, Curran, & Dyer, 2013) or disrupting other interpersonal processes (Covert, Tangney, Maddux, & Heleno,

2003). Similarly, the tendency to withdraw inherent in shame may lead to social isolation (Kim, Thibodeau, & Jorgensen, 2011), which corresponds to lower RSA (Grippo, Lamb, Carter, & Porges, 2007).  Shame and RSA likely have a reciprocal influence on each other.” (p. Freed & D’Andrea, 2015, pp. 376-77).  I agree wholeheartedly with the authors here.  Re-stated more simply, “Shame begets hiding and isolation which begets recurrent shame”.

The authors further discuss their findings:

“Shame was the only significant correlate of SNS reactivity to trauma reminders. The distinctive relationships between shame, PTSD, and the PNS highlight the complexity of the relationship between affect, symptom, and physiology and suggest the importance of measuring trait affect in a variety of populations with PTSD, particularly as the PTSD diagnosis has been revised to incorporate negative affect (DSM–5; American Psychiatric Association, 2013)”, (Freed & D’Andrea, 2015, p. 377)

This quote speaks to the importance of exploring and working with shame states in our relational trauma survivors, particularly those patients who have experienced boundary violations such as IPV as well as emotional neglect.  Furthermore, this research suggests that when a person is reminded of having been violated, their shame state will likely be activated.  That is, when working with these traumatic memories (both bad things that happened and good, needed things that didn’t happen), these patients are inevitably reminded of having been socially excluded, of how this still causes them to feel socially-emotionally isolated, and finally that the social violence, even more than the physical violence, could happen again.  Said another way, the therapist working with shame prone trauma survivors must be prepared to help them work with and transform shame which will often be evoked by this work.

There are some limitations of Freed & D’Andrea’s (2015) research worth mentioning.  As noted above, the measure the authors used didn’t adequately differentiate shame versus guilt.  The authors themselves allude to this problem when they wrote:

“…there is reason to believe that participants may not be able to adequately differentiate a dispositional tendency toward feelings of shame from guilt using this method (see Chapter 3 in Tangney & Dearing, 2002, for a full discussion of the value of different measures of shame and guilt)”, (Freed & D’Andrea, 2015, p. 377).

I believe helping patients differentiate shame from guilt (i.e. “I am bad” from “I did bad”) is not only doable, it is a necessary part of the therapy with relational trauma survivors for many reasons. They include that these patients often confuse the two and are left feeling immobilized (shame-induced) rather than engaged in relational repairs (following guilt), when appropriate.

The authors have done admirable work teasing out the central role of shame in relational trauma survivors. This is especially given that prior to the DSM-5, PTSD privileged anxiety or fear over other emotions.  As the authors remark,

“Although shame may not be more important among trauma survivors than other affects, these data suggest that it be given significant weight.  These findings highlight the importance of measuring shame in this population in understanding the course and maintenance of symptoms in this group, consistent with the addition of negative trauma-related emotions such as shame to PTSD criteria in the DSM–5.  A purely fear-based model of PTSD overlooks other potential sources of arousal that may affect symptoms that range from startle response to social functioning”, (Freed & D’Andrea, 2015, p. 377).

The authors were correct to suggest that treating “…shame may require different approaches than treating arousal caused by fear”. They go on to say that shame “…may not be amenable to the gold-standard treatment of exposure therapy”, (Freed & D’Andrea, 2015, p. 378).

Not only is shame not amenable to exposure therapy, it is made worse by this approach.  To oversimplify, the treatment of shame requires that from a secure intra-relational and inter-relational base, the patient develops a new, curious and compassionate relationship with the part of themselves that holds the shame.  Re-exposure makes things worse. This is because it repeats what the patient is already doing to themselves both consciously and often outside their awareness, in effect serving to re-traumatize them.

My final comment reflects a limitation of many sophisticated clinical researchers, (as these authors appear to be) who are not as experienced as psychotherapists.  The authors observe that, “If shame represents a failure of emotion regulation, therapies that teach these skills may be more effective in targeting shame-related symptomatology”, (Freed & D’Andrea, 2015, p. 378).  While emotion regulation will occur in effective psychotherapy with shame-ridden trauma survivors, this work requires a much more complex understanding of intra- and inter-personal relations.

It should include the transformation of shame-driven ways of relating to self and others, as well as knowing how to help patients move from traumatic shame states toward states of authentic, non-hubristic pride.  Thus, treatment for shame must help the patient become re-associated with the parts of self that have been rejected and dissociated from, as well as connecting with non-shaming, loved ones in the world.  While this will certainly benefit emotion regulation, it also goes well beyond that.


  1. The pronoun “her” is used as Freed & D’Andrea’s (2015) research was done with female survivors of physical and/or sexual abuse.


American Psychiatric Association.  (2013). Desk Reference to the Diagnostic Criteria from DSM-5.  Washington, D.C.: American Psychiatric.

Benau, K. (2017).  Shame, Attachment, and Psychotherapy: Phenomenology,  Neurophysiology,
 Relational Trauma, and Harbingers of Healing.  Attachment Journal, 11(2): 1-27.

Freed, S. & D’Andrea, W.  (2015). Autonomic Arousal and Emotion in Victims of Interpersonal Violence: Shame Proneness But Not Anxiety Predicts Vagal Tone.  Journal of Trauma & Dissociation. 16: 367-383.

Liotti, G. (2004). Trauma, dissociation, and disorganized attachment: three strands of a single braid. Psychotherapy: Theory, Research, Practice, and Training, 41(4): 472–486.

Porges, S. W. (2011). The Polyvagal Theory: Neurophysiological Foundations of Emotions, Attachment, Communication, and Self-regulation. New York: Norton.

Scheff, T. S. (2007). A Socialemotional Theory of Depression.  Accessed January 31,2017.

© 2018 International Society for the Study of Trauma and Dissociation.  |  Home  |  Contact Us  |  About ISSTD  |