Do Positive Memory Characteristics Relate to Reckless Behaviors? An Exploratory Study in a Treatment-Seeking Traumatized Sample

Abstract

Reckless and self-destructive behaviors (RSDBs), common among traumatized individuals, are associated with negative health consequences. Gaining a stronger understanding of factors associated with an increased likelihood of RSDBs among traumatized individuals offers potential new avenues for research and treatment. Mounting evidence indicates relations between traumatic experiences and deficits/disturbances in characteristics of positive memories; however, relations between RSDBs and positive memory characteristics has been understudied. Using hierarchical multiple regression, we examined relations between positive memory characteristics (Memory Experiences Questionnaire-Short Form; MEQ-SF) and RSDBs, controlling for PTSD and depression severity, among a sample of treatment-seeking trauma-exposed individuals (N = 77; M_age = 33.96; 57.10% female). Results indicated that MEQ-SF subscales of Accessibility, Coherence, Emotional Intensity, and Sensory Details were significantly associated with engagement in RSDBs, above and beyond PTSD and depressive severity. Those who easily accessed emotionally evocative positive memories tended to engage in elevated RSDBs; those with less coherence and fewer sensory details in their positive memories were also more likely to engage in RSDBs. Theories related to emotion dysregulation and cognitive deficits may explain these obtained relations.

Introduction

Reckless and self-destructive behaviors (RSDBs; Lusk, Sadeh, Wolf, & Miller, 2017; Strom et al., 2012), including substance use, reckless spending/driving, problematic technology use, suicidal behavior, and aggression, are common among traumatized individuals and are associated with negative physical and mental health consequences (Drescher, Rosen, Burling, & Foy, 2003). Because of these established relations, RSDBs were added to the DSM-5 (APA, 2013) as a Criterion E symptom of PTSD. Existing theory helps to explain the presence of elevated reckless behavior among individuals with PTSD, including emotion dysregulation (Weiss et al, 2013) and cognitive deficit (Ben-Zur & Zeidner, 2009) models. Generally, these models focus on exploring why traumatic memory-related processes lead individuals to engage in RSDBs. One underexplored avenue links disruptions in the encoding, consolidation, and retrieval of both traumatic and positive memories, to increased PTSD symptom severity (Brewin, 2014; Contractor et al., 2018), and increased PTSD symptom severity to increased PTSD Criterion E RSDB severity (Armour et al., 2020). This is relevant because it suggests trauma-exposed individuals have broader and deeper disturbances in memory processes, beyond those associated with their trauma memories, and that this, in turn, increases the severity of Criterion E PTSD RSDBs.

Recent work has examined relations between positive memory processes and PTSD severity (Contractor et al., 2018); these relations appear to be bidirectional. Individuals with pre-existing deficits in positive memory processes may be more vulnerable to developing PTSD (e.g., Hauer, Wessel, Engelhard, Peeters, & Dalgleish, 2009) and individuals with PTSD symptoms may be less able to access positive memories (Williams et al., 2007). These broader memory-related deficits may extend to Criterion E RSDBs, whereby deficits in positive memory characteristics could put traumatized individuals at risk for RSDBs, in the same way that disturbances in traumatic memory characteristics may put individuals at risk for RSDBs. Thus, exploring the associations between positive memory characteristics and RSDBs offers a unique pathway to understand and potentially intervene on Criterion E RSDBs.

Existing theories linking PTSD and trauma memories to RSDBs can be extended to explore how positive memories link to RSDBs. Emotion regulation theories suggest individuals engage in RSDBs to regulate affect (Ben-Zur & Zeidner, 2009). Expanding upon this, individuals who easily access vivid and emotionally evocative positive memories may consequently experience positive affect, and, may engage in RSDBs to prolong/increase this positive affect. Alternatively, positive affect accompanying easily accessed emotionally evocative positive memories may be experienced as aversive. Some trauma-exposed individuals appraise positive emotions as undesirable and frightening (Weiss, Gratz, & Lavender, 2015), and dislike the physiological arousal accompanying positive affect (Weiss, Dixon-Gordon, Peasant, & Sullivan, 2018). Further, individuals may experience negative emotions (Frewen, Dean, & Lanius, 2012), maladaptive beliefs, and guilt cognitions (Norman, Wilkins, Myers, & Allard, 2014) in response to positive emotions. Such processes may then contribute to emotion dysregulation (Weiss, Contractor, Raudales, Greene, & Short, in press) and impulsivity (Weiss, Tull, Sullivan, Dixon-Gordon, & Gratz, 2015), thereby resulting in RSDBs. In essence, trauma-exposed individuals may engage in RSDBs to up- or down-regulate positive/negative affect, subsequent to recalling easily accessed, vivid, and emotionally evocative positive memories.

Another explanation linking positive memory characteristics to RSDBs references a cognitive perspective. Deficits in executive functioning, narrowed attention span, and restricted information processing capacity following trauma may increase the likelihood of RSDBs (Ben-Zur & Zeidner, 2009) and may explain some individuals’ inability to recall positive memories (Dalgleish et al., 2007). For example, when recalling a self-defining memory, individuals with elevated anxiety have less vivid, coherent, and detailed recollections (Luchetti & Sutin, 2016). In this case, trauma-exposed individuals may have difficulties recalling coherent, vivid, positive memories with specific sensory details, due to underlying cognitive deficits, that are also associated with RSDBs. Thus, cognitive deficits may be an underlying common factor linking difficulties recalling positive memory characteristics to RSDBs. From an intervention perspective, remembering and sharing positive experiences may help contextualize traumatic experiences and integrate them within existing schemas (Resick & Schnicke, 1993) and with other autobiographical memories influencing one’s identity (Conway & Pleydell-Pearce, 2000; Ehlers & Clark, 2000), which may help to counteract the aforementioned cognitive deficits and aid in successful recovery post-trauma (Bryant et al., 2007; Ehlers & Clark, 2000), including reducing post-trauma RSDBs. Thus, individuals who readily share positive memories may be less likely to engage in RSDBs.

Despite theory linking positive memory characteristics and RSDBs, their associations have not been examined. Addressing this gap, we explored the associations between different facets of a recalled positive memory and post-trauma RSDBs. Prior network analyses examining the relations between positive memory characteristics and PTSD symptomatology (Contractor, Greene, Dolan, et al, 2020), identified several positive memory content-related characteristics relevant to include in our analyses (valence, vividness, sensory details, coherence). Based on emotion regulation models, we hypothesized individuals with easily accessible, positively valenced, and emotionally evocative positive memories would be more likely to engage in RSDBs, in an attempt to up or down-regulate experienced emotions. Based on cognitive deficits models, we hypothesized that individuals with positive memories that lacked coherence and sensory details would be more likely to engage in RSDBs, because of the common factor of cognitive deficits underlying both engagement in RSDBs and difficulties clearly recalling a coherent and detailed memory. Given inconsistent findings regarding the directionality of the impact of vividness on RSDBs, we did not have directional hypotheses for this memory characteristic. Finally, consistent with treatment research, we hypothesized that individuals who regularly shared positive memories with others would be less likely to engage in RSDBs. Potential covariates relevant to post-trauma RSDBs were examined, including PTSD and depression severity (Lusk et al., 2017; Strom et al., 2012).

Methods

Procedure and Participants

This study was approved by the [redacted] Institutional Review Board. From February 2017-October 2019, research staff approached 465 treatment-seeking adults at a community mental health center, within their first four therapy sessions (see Figure 1). Staff provided study information and assessed eligibility. Eligibility criteria included: 18+ years old, understands English, and states yes to: “Have you experienced at least one traumatic event such as a motor vehicle accident, abuse, loss of a loved one, a natural disaster etcetera?” Eligible participants provided informed consent and completed study measures via Qualtrics on a clinic tablet. Our final sample included 77 individuals (M_age = 33.96, SD = 10.30; 57.10% female); the Posttrauma Risky Behaviors Questionnaire (Contractor, Weiss, Kearns, Caldas, Dixon-Gordon, 2019) was added midway through the study, as it was not published until 2019; hence the sample size is restricted. Table 1 has detailed demographic information.

Figure 1. Study Sample.

Note. SLESQ is the Stressful Life Events Screening Questionnaire (Goodman, Corcoran, Turner, Yuan, & Green, 1998); MEQ-SF is the Memory Experiences Questionnaire–Short Form (Luchetti & Sutin, 2016).

* The Posttrauma Risky Behaviors Questionnaire (PRBQ) was developed recently and examined for psychometric properties (Contractor et al., 2019). Hence, we add the PRBQ mid-way through the study, explaining the extent of individuals missing > 30% data and the resulting truncation of the sample.

Table 1.

Descriptive information on demographics and psychopathology constructs.

	Mean	SD
Age (n = 77)	33.96	10.30
Years of schooling (n = 77)	12.14	2.56
PCL-5 total (n = 75)	43.49	27.67
PHQ-9 total (n = 77)	12.88	8.02
PRBQ total (n = 76)	12.72	12.57
		n, (%)*
Gender (n = 77)	Male	33 (42.9%)
	Female	44 (57.1%)
Employment Status (n = 77)	Part-time	4 (5.2%)
	Full-time	41 (53.2%)
	Retired	1 (1.3%)
	Unemployed	28 (36.4%)
	Unemployed Student	3 (3.9%)
Current annual household income (n = 76)	Less than $24,999	37 (38.7%)
	$25,000 or higher	39 (61.3%)
Current Relationship Status (n = 77)	Single	39 (50.6%)
	Living with significant other	23 (29.9%)
	Married	13 (16.9%)
	Divorced, separated, or widowed	2 (2.6%)
Ethnicity (n = 77)	Hispanic or Latino	2 (2.6%)
	Not Hispanic or Latino	54 (70.1%)
	Unknown	21 (27.3%)
Race (could endorse multiple choices; n = 77)	Caucasian	63 (81.8%)
	African American	11 (14.3%)
	Asian	2 (2.6%)
	American Indian/Alaskan Native	2 (2.6%)
	Unknown	5 (6.5%)

Note.

^*All reported percentages are valid percentages to account for missing data; PCL-5 = PTSD Checklist for DSM-5; PHQ-9 = Patient Health Questionnaire – 9.

Measures

The Stressful Life Events Screening Questionnaire (SLESQ; Goodman, Corcoran, Turner, Yuan, & Green, 1998) is 13-item self-report measure assessing the presence/absence of lifetime traumas. It has good psychometric properties (Goodman et al., 1998).

The PTSD Checklist for DSM-5 (PCL-5; Weathers et al., 2013) is a 20-item self-report measure assessing PTSD symptom severity in the past month, referencing the most traumatic event endorsed on the SLESQ. Response options range from 0 (not at all) to 4 (extremely). The PCL-5 has excellent psychometric properties (Blevins, Weathers, Davis, Witte, & Domino, 2015; Bovin et al., 2016). In this study, Cronbach’s α was .98.

The Patient Health Questionnaire-9 (PHQ-9; Kroenke, Spitzer, & Williams, 2001) measures depressive symptom severity. Response options range from 0 (not at all) to 3 (nearly every day). Internal consistency and test-retest reliability are high (Kroenke et al., 2001). In this study, Cronbach’s α was .92.

The Posttrauma Risky Behaviors Questionnaire (PRBQ; Contractor et al., 2019) is a 16-item self-report measure evaluating past month engagement in RSDBs (PTSD’s E2 criterion). 14 PRBQ items measure frequency of engagement in specific RSDBs, ranging from 0 (never) to 4 (very frequently). Supplemental items measure functional impairment and association between RSDB frequency and onset of the worst trauma. Higher PRBQ summed scores (1^st 14 items) indicate greater engagement in RSDBs. The PRBQ has excellent internal consistency, good construct validity, and good convergent and incremental validity (Contractor et al., 2019). In this study, Cronbach’s α was .89.

The Memory Experiences Questionnaire–Short Form (MEQ-SF; Luchetti & Sutin, 2016) is a 31-item self-report measure evaluating 10 memory phenomenological domains. Prior to completing the MEQ-SF, participants wrote two paragraphs about a positive memory (Boyacioglu & Akfirat, 2015; Sutin & Robins, 2007): “Please describe a specific memory of the most positive event experienced by you. Specific memories are defined as memories referring to events that happened at a particular time and place and did not last longer than a day. Examples of positive memories are getting married, achieving a promotion at work, and having a child. You will then rate how much you agree with each of the following items for the memory you described.” Memories were coded using the Coding and Assessment System for Narratives of Trauma (CASNOT) dimensions of emotional tone and emotional valence (Fernández-Lansac & Crespo, 2017); scores of 3-4 on a scale from 0 (completely positive) to 4 (completely negative) for either emotional tone or valence rendered a memory to be coded as non-positive. Percentage agreement for coding memories as positive versus non-positive was 87.80%.

The MEQ-SF has acceptable psychometric properties (median alpha = .79, Luchetti & Sutin, 2016) and each short-form subscale is highly correlated with the corresponding MEQ (i.e., long-form) subscale (median r = .95; Luchetti & Sutin, 2016). MEQ-SF items were rated on a 1 (strongly disagree) to 5 (strongly agree) Likert scale. Consistent with prior research (Luchetti & Sutin, 2016), median Cronbach’s α in the current study was .76; we excluded the MEQ-SF Vividness and Distance subscales in our analyses due to their lower internal consistency (alpha = .55 and .62, respectively). Consistent with prior network analyses establishing relevant positive memory characteristics (Contractor et al., 2020), Visual Perspective and Time Perspective subscales were excluded due to a lack of a prior hypotheses regarding these variables. Internal consistency for included variables are: Accessibility (.71), Emotional Intensity (.84), Coherance (.72), Sensory Details (.65), Sharing (.66), and Valence (.65); internal consistencies are congruent (within about .05 points) with the MEQ-SF internal consistency results reported in the development of the MEQ-SF (Luchetti & Sutin, 2016).

Statistical Analyses

First, we examined normality (skewness < 2; kurtosis < 7; Curran, West, & Finch, 1996); the assumption of normality was not violated for the primary variables of the current study. Next, we examined the assumption of multicollinearity using the Variance Inflation Factor (VIF) ≥ 10 and tolerance value < .01 rules (Hair, Black, Babin, & Anderson, 2009); multicollinearity was not violated (for the variables of PTSD and depression severity; the VIF was 3.07 and tolerance was .33). Finally, we computed a hierarchical multiple regression examining RSDBs as the dependent variable. Step 1 included PTSD and depression severity; Step 2 added MEQ-SF dimensions of valence, coherence, accessibility, sensory details, emotional intensity, and sharing. Effect sizes were interpreted using standardized beta (β) estimates.

Results

In Step 1, PTSD and depression severity were significantly associated with engagement in RSDBs (p = .05 and p = .005, respectively) and the total variance explained by the model was significant (r² = .49, p < .001). In Step 2, adding positive memory characteristics resulted in significant improvements in model fit (ΔR² = .11, p = .02). Specifically, MEQ-SF subscales of Accessibility, Coherence, Emotional Intensity, and Sensory Details, were significantly associated with engagement in RSDBs, controlling for PTSD and depression severity (Table 2). Other memory characteristics were not significantly associated with RSDBs.

Table 2.

Results of the multiple regression analyses for the dependent variable of engagement in reckless and self-destructive behaviors.

	B	SE	β	t	R2	ΔR2	F	ΔF
Step 1					.49		33.68 p < .001
PTSD severity	.13	.07	.30	1.97 p = .05
Depression severity	.66	.23	.44	2.93 p < .01
Step 2					.60	.11		2.89 p = .02
PTSD severity	.10	.07	.22	1.50 p = .14
Depression Severity	.85	.22	.56	3.88 p < .001
MEQ-SF Accessibility	1.61	.58	.33	2.78 p < .01
MEQ-SF Emotional Intensity	1.16	.41	.31	2.81 p < .01
MEQ-SF Coherence	−1.09	.39	−.32	−2.82 p < .01
MEQ-SF Sensory Details	−.78	.40	−.23	−1.96 p = .05
MEQ-SF Sharing	−.38	.38	−.10	−1.00 p = .32
MEQ-SF Valence	−.24	.63	−.04	−.38 p = .70

Note. MEQ-SF is Memory Experiences Questionnaire-Short Form. Significant p values are bolded.

Discussion

This is the first study, to our knowledge, to examine how RSDBs relate to positive memory characteristics in a trauma-exposed sample. As expected, engagement in RSDBs was significantly associated with greater PTSD and depression severity (Contractor, Weiss, Dranger, Ruggero, & Armour, 2017). As hypothesized, multiple positive memory characteristics were significantly associated with RSDBs. MEQ-SF Accessibility (ease of accessing the positive memory), Emotional Intensity (evocation of strong emotions when recalling the positive memory), Coherence (recall of consistent/logical details of a positive memory), and Sensory Details (“I can hear it, I can see it” regarding the positive memory) were significantly associated with RSDBs. Specifically, those who easily accessed emotionally intense positive memories tended to engage in elevated RSDBs; those with less coherent and less sensorially-detailed positive memories were also more likely to engage in elevated RSDBs. Theoretical/empirical foundations as well as implications for findings are discussed in detail.

In the context of experiencing easily accessible and emotionally intense positive memories, traumatized individuals may have difficulties regulating positive emotions that arise (Weiss, Dixon-Gordon, Peasant, & Sullivan, 2018) and become disinhibited, thereby increasing the likelihood of engaging in RSDBs. Indeed, greater impulsivity in the context of positive emotions, higher non-acceptance of positive emotions, and difficulties engaging in goal-directed behavior when experiencing positive emotions, is present among those with more severe PTSD symptomatology (Weiss et al, 2018). Thus, attempts to up- or down-regulate positive emotions, following engagement with positive memories, helps to explain why MEQ-SF subscales of Accessibility and Emotional Intensity are related to RSDBs; those who quickly and easily engage with positive memories at a high level of emotionality may be most likely to engage in RSDBs to re-regulate their consequent emotions.

There are several potential explanations underlying the relations between lower positive memory coherence and sensory details with increased RSDBs. First, there may be common factors, like cognitive deficits, linking RSDBs to lower coherence and sensory details of positive memories. Trauma experiences relate to reduced attention span and information processing capacity, thereby resulting in less coherent positive memories with fewer sensory details; similarly, deficits in executive functioning, narrowed attention span, and restricted information processing increase the likelihood of RSDBs (Ben-Zur & Zeidner, 2009). Second, we can draw from trauma intervention research to understand the current study findings. Individuals receiving trauma-focused therapy exhibit improvements in trauma memory coherence (Moulds & Bryant, 2005) and report increased sensory details in their trauma memories (Bedard-Gilligan, Zoellner, & Feeny, 2017) post-treatment. Parallel to these findings, perhaps, improvements in coherence and sensory details for positive memories would relate to improvements in post-trauma outcomes such as Criterion E RSDBs. Third, individuals who are unable to recall sensory details for their positive memories may experience blunted positive affect; thus, they may engage in RSDBs to attempt to increase positive affect (Baker , Piper, McCarthy, Majeskie, & Fiore, 2004).

Limitations must be considered when interpreting study findings. First, we utilized self-report assessments (e.g., of RSDBs); future research would benefit from using semi-structured diagnostic assessments (e.g., RSDBs assessed on the Clinician-Administered PTSD Scale for DSM-5; Weathers et al., 2013) or behavioral tasks (e.g., Balloon Analogue Risk Task; Lejuez et al., 2002). Second, women were overrepresented, and minorities underrepresented in the current study; future work would benefit from examining these patterns within more diverse samples. Third, consistent with findings from the development of the MEQ-SF (Luchetti & Sutin, 2016), internal consistency was low for the Vividness and Distance MEQ-SF subscales, likely due to the limited number of items included in the subscale. Fourth, it would be worthwhile to examine the interaction effects between other key features of PTSD beyond Criterion E RSDBs (e.g., Criterion B Intrusions, Criterion C Avoidance), RSDBs, and positive memory characteristics in a future, well-powered study, given the strong theoretical/empirical relations of PTSD symptom clusters, including RSDBs (Lusk, Sadeh, Wolf, & Miller, 2017; Strom et al., 2012) and positive memory characteristics (Bryant, Sutherland, & Guthrie, 2007; Hauer, Wessel, Engelhard, Peeters, & Dalgleish, 2009). Indeed, replication and extension of findings in larger and more diverse samples is necessary. Fifth, longitudinal research is needed to further understand whether these relations are causal or bidirectional; indeed, it is possible that RSDBs have an impact on individuals’ engagement with positive memories. Finally, we did not assess characteristics of trauma memories. Given research indicating relations between recall/vividness of trauma memories and urges to engage in RSDBs (Coffey, Saladin, Drobes et al., 2002), future research would benefit from examining the incremental predictive effects of positive memory characteristics, beyond trauma memory characteristics, on RSDBs.

Despite these limitations, the current study offers new directions, both for research and treatment, to address some of the underlying positive memory deficits and difficulties experienced by trauma-exposed individuals. Specifically, results indicate the importance of addressing positive memories and their characteristics in trauma assessment (Contractor et al, 2019) and treatment (Contractor, Banducci, Jin, Keegan, & Weiss, 2020), in relation to post-trauma symptoms such as RSDBs. Broadly, the current study and its findings reflects a shift from the traditional focus on negative memories in trauma work to understanding how memory-related processes more broadly impact PTSD symptomatology, including Criterion E RSDBs.