Assessing Post-Traumatic Tonic Immobility Responses: The Scale for Tonic Immobility Occurring Post-Trauma

Tonic Immobility

Biological indicators consistent with TI states in mammals have now been reported in humans and are proposed to be homologous.⁹ For example, following exposure to a non-escapable biological stressor (20-s inhalation of 20% CO₂/balance O₂), participant-reported feelings of immobility more than doubled.²³ Moreover, whereas trauma-exposed community members demonstrated heart rate deceleration and decreased body sway when shown threat-based pictures depicting no means of escape, amplitude of heart rate and sway increased when a clear escape route was depicted.²⁴ In addition to significant reductions in body sway,^19,24–30 TI responses have been associated with increased muscle stiffness.²⁵ Changes in heart rate during TI responses have also been reported, although results remain equivocal. Specifically, whereas one body of evidence suggests TI is characterized by tachycardia or heart rate acceleration,³⁰ another suggests that it is associated with bradycardia or heart rate deceleration.³¹ It is possible these conflicting data reflect the simultaneous activation of two opposing parts of the autonomic nervous system, sympathetic and parasympathetic, as observed in animal models of TI.^1,22,31

Dissociation is defined as a disturbance in the normal integration of consciousness, memory, identity, emotion, perception, body representation, motor control, and behavior.³² Critically, dissociation is thought to increase alongside parasympathetic activity as an organism progresses through the cascade, from active to passive responses.^5,33,34 Here, dissociation serves a highly adaptive purpose in TI: an individual may endure unremitting threat, and injury, by becoming unresponsive to the environment.^4,35

Measures of Post-Traumatic TI

Current measures assessing peri-traumatic TI^6,36 do not assess responses that continue to occur post-traumatically outside the context of an acute traumatic incident. A few measures assay aspects of post-traumatic TI; however, a comprehensive self-report measure is presently lacking. Whereas the Somatoform Dissociation Questionnaire (SDQ-20) evaluates the severity of somatoform manifestations of structural dissociation,³⁷ the STOP focuses on the post-traumatic TI response itself, as an expression of the human defense cascade. Although the SDQ-20 assesses some symptoms broadly consistent with TI (e.g., item 19: I am paralyzed for a while; item 20: I grow stiff for a while), other items pertain to symptoms that extend beyond what is expected to occur during post-traumatic TI (e.g., item 2: I dislike tastes that I usually like; item 10: I feel pain in my genitals; item 17: I cannot sleep for nights on end, but remain very active during daytime). Furthermore, the SDQ-20 fails to adequately assess relevant aspects of depersonalization thought to occur during TI responses (i.e., out of body experience).

The Shutdown Dissociation Scale (Shut-D)³³ represents the first theoretically derived measure to focus on post-traumatic “shut down” symptomatology within a dissociative framework. The Shut-D assesses dissociative symptoms thought to occur alongside parasympathetic dominance (“shut down”) during late-stage defense responding.³³ Unfortunately, the Shut-D is administered as a structured clinical interview, limiting its feasibility across settings. Conversely, the STOP represents the first self-report measure to capture TI specifically as part of this parasympathetic cascade.

Critically, neither the SDQ-20 nor the Shut-D considers the potential residual symptoms or affect associated with post-traumatic TI. Given patient report and emergent literature, it is essential that tools assessing post-traumatic TI include items that measure non-fear primary responses to trauma, including guilt and shame.^38,39

The Present Study

In the present study, we evaluate the factor structure and psychometric properties of STOP, designed as the first self-report measure of post-traumatic TI. STOP assesses for the presence and severity of altered sensorimotor and perceptual experiences associated with TI occurring more than one month after acute trauma in response to traumatic reminders. Unlike the SDQ-20 and Shut-D, STOP items additionally assess the physical and emotional consequences of chronic TI, further facilitating clinical recognition and future research. Specifically, STOP was developed to inform treatment planning, guide the development of targeted interventions, and result in a better understanding of this altered state.

We predicted that STOP scores would be strongly and positively associated with other measures of post-traumatic symptom severity and dissociation. As persistent TI responses have been linked with state²⁰ and trait²³ anxiety as well as severe depression,⁹ we also expected STOP scores to be positively associated with measures of anxiety and depression.

Participants

North American participants (N_pooled = 713) were recruited using two validated web-based, crowd-sourcing platforms (CrowdFlower; Amazon’s Mechanical Turk) across two independent rounds of data collection (N_Crowdflower = 340; N_MTurk = 373).^40,41 Participants voluntarily responded to an online advertisement (see Supplemental Information; SI), self-identifying based on the following eligibility criteria: (i) no history of head injury (no loss of consciousness due to physical trauma, lack of oxygen, or electric shock), (ii) at least one episode of TI in response to a past traumatic event occurring more than one month prior, and (iii) at least one episode of TI in the past month, in response to reminder(s) related to the original traumatic event (i.e., post-traumatic TI).

Participants who completed less than one third of the survey battery were excluded from analyses. Data were manually inspected, and a total of 251 participants were excluded based on one or more of the following reasons: incomplete data (n = 170), incoherent response (n = 72), suspected random responding (e.g., entered 1’s for all responses even when not applicable; n = 4), duplicate entry (n = 3), or self-reported head injury (n = 2). Excluded participants did not differ significantly from those retained on any demographic measure.

Remaining participants (N_pooled = 462) completed demographic information (Table 1). Participants (54.8% female) were generally of middle age (M = 34.33, SD = 10.29), ranging from 18 to 65 years old. Ninety participants reported no post-traumatic TI responses over the past month and were therefore excluded from these analyses pertaining to post-traumatic TI responses. Thus, 372 participant responses were retained for subsequent analyses (N_Crowdflower = 181; N_MTurk = 191).

Measures

Procedure

This study was approved by the Western University Ethics Board for Health Sciences Research (#108288). To preserve participant anonymity and confidentiality, data collection occurred on a secure, encrypted website (Qualtics) independent of crowd-sourcing access sites.

Participants who self-identified as eligible for the study were presented with a letter of information and provided written consent prior to completing survey material. Consenting participants completed demographic information, followed by the STOP and additional measures of PTSD, dissociation, anxiety, and depression. The location of items and response scales were varied to ensure participants were attending to item content and response quality. Participants included in the final sample responded to validity items accurately (e.g., “What year is it?”). Participants were provided with a code to tender $1 compensation comparative to rates offered by similar studies.

Statistical Analysis

We used a stepwise approach to determine the ideal item-count and factor structure of the 41-item STOP. First, exploratory factor analysis (EFA) was conducted with a principal axis extraction and Promax (oblique) rotation. Second, exploratory structural equation models were conducted with a maximum likelihood estimator and a target rotation with cross-loadings specified to approximately zero. Using these two methods, items with strong cross-loadings (i.e., > .40) were removed. Finally, we tested the suitability of a simple structure to STOP with confirmatory factor analysis (CFA) with maximum likelihood estimation and robust standard errors. EFA procedures were conducted using SPSS software, and exploratory structural equation model and CFA procedures were performed in MPlus version 7.4.⁴⁹

Factor structure was explored initially using EFA and ESCM in Sample 2 (n_MTurk = 191) and testing the resultant model for fit (CFA) in Sample 1 (_Crowdflower = 181). The opposite procedure was also conducted (i.e., EFA on Sample 1, CFA Sample 2). Results of these cross-validation analyses yielded consistent factor structures in Samples 1 and 2. Since a near-identical factor structure emerged when considering the total (pooled) sample, the results reported here will focus on the total sample.¹

Exploratory Analysis

Factor loadings and intercorrelations of EFA conducted on the 41-item STOP were examined. Six latent variables emerged with an eigenvalue greater than one, rendering a solution accounting for 65.3% of the variance. Each eigenvalue was compared to the 95th percentile of a sampling distribution of 100 randomly generated eigenvalue correlation matrices. Three factors emerged before the randomly generated eigenvalues exceeded those from our data. A fourth factor (eigenvalue > 1) was retained for theoretical reasons, together accounting for 59.7% of the variance. Pattern and structural matrices were used to interpret the factor structure, with items supporting the following factor labels (% variance explained): (1) Sensorimotor and Perceptual Alterations (45.5%), (2) Somatic Detachment and Amnesia (6.2%), (3) Residual Symptoms (4.8%), and (4) Negative Affect (3.1%; Figure 1).

Due to inconsistent factor loadings, three items were removed from the final scale versions (b2, b22, and b40) and excluded from CFA analyses. Factors were strongly and positively correlated in the factor correlation matrix (r’s between .4 and .7). Therefore, in CFA analyses, a hierarchical structure was tested.

Confirmatory Factor Analysis

Using a hierarchical four-factor structure, CFA metrics suggested an acceptable model fit. Model fit was assessed by evaluating the global fit indices, including the Root Mean Square Error of Approximation, Standardized Root Mean Residual, Tucker-Lewis Index, and Comparative Fit Index.^50,51 Minimal modifications were made to improve overall fit by allowing one pair of items to correlate residuals (Table 2). Each of the four factors loaded strongly on its higher order parent factor, supporting the rationale for a global score calculation, in addition to subscale scores (see SI for scoring procedure).

Reliability and Validity

The Onset and Maintenance of TI

The Sensorimotor and Perceptual Alterations and the Somatic Detachment and Amnesia subscales contain items that describe alterations in experience that accompany the onset and maintenance of TI. This transition may include disturbances in somatic experience (e.g., b1: “I cannot move”), emotional experience (e.g., b12: “I feel numb”), and perception of time (e.g., b20: “it feels like the experience lasts for a long time”).⁴³ Items describe loss of speech as this function is secondary to respiration (i.e., b5: “I take short quick breaths”; b8: “I cannot speak even if I wanted to”), along with seemingly adaptive alterations in vision (i.e., b10: “my vision becomes narrowed”; b9: “I am still able to move my eyes”).⁵² The final stages in the defense cascade, beyond TI, are thought to result in eventual loss of consciousness.^5,33 Two items pertaining to a flag and faint response (b26: “my muscles go limp”; b27: “I collapse to the floor”) were retained as part of the Somatic Detachment and Amnesia subscale, providing an indicator of response severity.

STOP items reflect an updated conceptualization of the human TI response as an evolutionarily adaptive defense response. TI and dissociation are conceptualized traditionally as separate constructs that co-occur frequently in response to acute stress.⁶ Some authors caution against a unified conceptualization of these constructs (i.e., attributing dissociative symptoms to the cognitive experience of TI).⁵³ Schauer and Elbert (2010) hypothesize that sympathetic and parasympathetic nervous system activity underpin the hyperarousal and “shut-down” symptomatology observed, respectively, in PTSD and its dissociative subtype. In our own work, we posit that TI is characterized by parasympathetic dominance and by associated dissociative symptoms. As a late-stage defensive response, these dissociative symptoms function to disengage and protect the individual from injury.^5,54 Accordingly, a number of items within this subscale represent depersonalization symptoms (e.g., b16: “I view my body from outside myself”). Dissociating the body from conscious awareness is inherent to TI and can function as an adaptive strategy under imminent threat (e.g., reduced pain). This state, however, also serves to disturb integration of the traumatic experience with one’s sense of self.⁵⁵ For example, TI was endorsed as evoking feelings of safety (b17: “I feel safe”) and amnesia (b34: “I have difficulty remembering the experience”), possibly reflecting the protective function of TI.

Residual Symptoms of TI

The Residual Symptoms subscale refers largely to the aftermath of TI (e.g., b33: “After unfreezing my muscles hurt”). Here, items reflect the heightened impact post-traumatic TI has on the brain and body, following transitions between TI responses and normal-waking states of consciousness (e.g., b30: “After unfreezing, I find it difficult to recover”).

The aftermath of post-traumatic TI was further captured on the Negative Affect subscale, which indexes social-moral feelings of guilt, shame, and/or embarrassment that may emerge following an episode (e.g., b37: “After unfreezing, I feel shame”). TI involves a loss of agency and a failure to act, both of which may represent socially violating behavior capable of eliciting negative social-moral emotion.⁵⁶ Guilt is elicited by negative evaluation of specific action(s), motivating approach behavior intended to repair the wrongdoing; shame serves to condemn the self thereby motivating self-evaluation and withdrawal.⁵⁷ Both shame and dissociation have been identified as forms of interpersonal avoidance,⁵⁸ serving ultimately to increase social distance (e.g., alienation). We propose that shame, like TI, serves as a functional and protective response to threat. Any social avoidance that may result is better viewed as a secondary effect of a response intended to prevent further social harm, ridicule, and/or condemnation. Moreover, embarrassment is related directly to the response of an actual or perceived audience who witnesses the humiliating behavior.⁵⁹ An individual who immobilizes in front of others, for example, may worry about their social image thereby motivating efforts to preserve reputation, rather than the reparative and avoidant efforts seen in guilt and shame, respectively.⁵⁷

Limitations

Although provocative, the present study has a number of limitations including a reliance on retrospective, self-report data. As participants were invited to self-select based on at least one previous TI experience, selection bias may contribute to the present findings, limiting generalizability to other samples. Further limiting these findings, qualitative information about the measure was not collected in this sample. Crowd-sourcing platforms are increasingly used in studies of PTSD; however, it is not well understood how data quality compares to in-vivo methods and results should be interpreted with caution. Although the STOP is designed to measure post-traumatic TI responses, we are unable to definitively link this experience to trauma without future examination. In the present study, the STOP was validated in two independent, trauma-exposed (post-trauma) samples. Research utilizing multi-method and multi-informant approaches is required to determine convergent and discriminant validity and to strengthen evidence for the structural model developed in this study.

Future Directions

The present findings suggest a need to consider TI items in measures of post-traumatic symptomatology and dissociation used in clinical and research settings. It will be important to understand the role of shame and guilt in relation to TI moving forward. Relations between sex, gender, and TI, in addition to clinical and cultural differences in the prevalence and expression of TI responses, should be examined.⁶⁰ Evidence suggests that aversive stimuli can produce TI rather than active defensive action, as the former is more adaptive than the latter.²⁵ Future research should therefore examine the relation between intrusive symptoms and post-traumatic TI responses in individuals with and without a prior history of peri-traumatic TI. Future work is also needed to explore the intersection between TI and the dissociative subtype of PTSD, and other dissociative phenomena. Finally, dissociative symptoms consistent with post-traumatic TI responses (e.g., depersonalization, analgesia, and amnesia) present transdiagnostically across psychiatric disorders including personality disorders, schizophrenia, and eating and mood disorders.^61–63 Symptoms relevant to TI may present clinically as conversion disorder^60,64,65 or somatoform disorder.^66,67 Future research is needed to investigate the intersection between post-traumatic TI and these disorders.