Impulsivity and reduced quality of life in persons with paraplegia

Abstract

BACKGROUND:

Low Quality of Life (QoL) in persons with paraplegia may stem from impulsive behaviors. Impulsivity in persons with paraplegia and persistently low QoL has seldom been probed but could be targeted with cognitive behavioral therapies.

OBJECTIVE:

Determine how task-assessed and self-report impulsivity relate to quality of life (QoL) in adults with paraplegia.

METHODS:

In a preliminary study, 33 adults with paraplegia after traumatic SCI were administered verbal interviews on QoL from the PROMIS item bank at baseline and at six-month follow-up, along with several computerized metrics of impulsivity at baseline.

RESULTS:

A cluster of (n = 10) participants characterized by high levels of negative affect and low levels of resilience and life satisfaction across both baseline and follow-up showed significantly greater negative urgency impulsivity (p = 0.007) as well as significantly lower mindfulness and self-care in some domains (all p < 0.05), compared to the cluster of participants (n = 23) who showed higher life satisfaction and resilience. Behavioral metrics of delay discounting and rapid-response (motoric) impulsivity did not significantly differ (all p > 0.05) between the two clusters of participants.

CONCLUSIONS:

These data suggest that novel interventions that reduce trait impulsivity in other disorders could be applied to potentially reduce risk for reduced self-care and QoL in individuals with paraplegia.

Keywords

Paraplegia quality of life impulsivity spinal cord injury

1 Introduction

The costs associated with spinal cord injury (SCI) are very high (Lo et al., 2021) and include secondary health conditions (SHC) that are detrimental for social function and overall quality of life (QoL) that place individuals at greater risk of early mortality (Piatt et al., 2016). For example, a meta-analysis indicated that in North America, as many as 23% of individuals with SCI develop pressure-related injuries such as a pressure ulcer (PU) due to immobility (Chen et al., 2020), with rates as high as 50% in individuals with tetraplegia.

A critical determinant of QoL following SCI is the degree to which an individual undertakes proactive self-care behaviors. For example, due to being insensate, individuals with a complete SCI are recommended to proactively shift positions or raise themselves off their seat to relieve pressure and conduct frequent skin and surface inspections to prevent a PU. Also, in persons with SCI, physical activity reduces cardiovascular risk (Bailey et al., 2020) and has been found to make tissue in vulnerable areas more physiologically resistant to PU formation (Crespo-Ruiz et al., 2012).

Unfortunately, despite awareness of the importance of physical activity and preventive care, many individuals with SCI do not adhere to these measures. For example, veterans with reduced levels of SCI self-care behaviors had worse PU symptoms (LaVela et al., 2016). The COVID-19 pandemic has further reduced rates of physical activity in SCI (Bates et al., 2021). In attempts to remedy this, the near-universal daily internet use in SCI samples has been leveraged to create on-line curricula to promote physical activity and other self-care (Allin et al., 2020). Unfortunately, although these programs have increased knowledge of what individuals with SCI should do, and have even increased ratings of intent to increase physical activity, these programs have shown mixed findings of eliciting greater self-care behavior (Allin et al., 2020). Therefore, understanding motivation-related personality characteristics in individuals with SCI at risk for SHC and other low QoL matters for designing improved interventions.

One motivation-related characteristic that has (to our knowledge) received no research attention to date regarding risk for poor QoL with paraplegia is impulsivity. Impulsivity is a multi-faceted personality trait that centers on the lack of forethought, such as “acting without thinking” or “living for the moment.” The ability to inhibit behavioral responses in many laboratory impulsivity tasks is sometimes termed “rapid-response impulsivity” (Swann et al., 2002) and is thought to index a core latent trait of executive function (EF) (Friedman & Miyake, 2017). Impulsivity has been divided predominantly into decision-based impulsivity, rapid-response impulsivity as well as trait impulsivity typically indexed by paper-and-pencil self-report assessments (Cyders & Coskunpinar, 2011). Importantly, impulsivity facets have been linked to several risky health behaviors, such as substance abuse, seat-belt refusal, or poor medication compliance (Story et al., 2014).

In the context of SCI self-care, individuals with both paraplegia and high impulsivity would eschew the requisite sacrifice of time and leisure in the moment to engage in exercise or pressure-reducing maneuvers (or would be resistant to giving up smoking) in exchange for a future payoff of improved health, such as lower likelihood of a developing a PU. Fortunately, in persons with substance use disorder and other health conditions, impulsivity such as myopic tendency to devalue the future can be reduced with working memory training interventions that ostensibly bolster frontocortical neurocircuit function (Bickel et al., 2011) as well as episodic future thinking (EFT) prompts (Voss et al., 2021), wherein positive future outcomes (that require self-discipline in the moment) are made more salient.

To begin to fill this research gap, we sought to determine whether any facets of impulsivity are characteristic of persons with paraplegia and low QoL, who are at risk for SHC and other poor medical outcomes. Of particular interest was probing impulsivity in the subset of individuals who show consistently low QoL after a traumatic SCI at time points after which the bulk of psychological adjustment to paraplegia will have occurred in many individuals. Psychosocial and psychological reactions to SCI can vary, where in some individuals, psychological adjustment is robust soon after injury or improves over time while other individuals show enduring high rates of depression and anxiety (Bonanno et al., 2012). We recruited persons with traumatic SCI that occurred six months or more in the past, an interval past which many individuals with SCI have adjusted initially to their injury (Bonanno et al., 2012). We administered QoL metrics at two time points separated by six months, to identify a cluster of adults with traumatic SCI who evidenced consistently low adjustment and QoL. Doing so allowed us to identify a consistently less-resilient group of individuals with SCI to compare with a more resilient group.

Impulsivity (McTeague et al., 2016) and reduced motivation (Husain & Roiser, 2018) are trans-diagnostic features of several different mental illnesses, where for example some types of impulsivity are elevated in individuals with depressive symptoms (Saddichha & Schuetz, 2014). We therefore hypothesized that among adults with paraplegia after traumatic SCI, participants with consistently high scores of depression, grief and pain and low scores on resilience and life satisfaction, relative to a group of participants with the opposite pattern, would show greater scores on multiple impulsivity domains. Relatedly, we hypothesized that participants with higher QoL would have increased mindfulness (more reflective thought patterns), as a conceptual anti-correlate of impulsivity (Peters et al., 2011). We further hypothesized that in light of low motivation as a cardinal feature of depressive symptomatology, SCI self-care intensity would be lower in the lower QoL subset.

2 Methods

All recruitment and testing procedures were reviewed and approved by the Institutional Review Boards of Virginia Commonwealth University (VCU) and the Hunter Holmes McGuire Veterans Affairs Medical Center in Richmond, Virginia (RICVAMC).

2.1 Participants

We recruited 33 adults (n = 28 at VCU, n = 5 at RICVAMC) aged 18–50 who experienced traumatic SCI resulting in paraplegia at least six months prior. Individuals were required to self-report normal hand motor function but reliance on a wheelchair for mobility. Potential participants were ineligible if they were currently using opioid medication or reported current or previous problematic substance use other than nicotine, alcohol, or cannabis. Two additional participants underwent baseline testing but were lost to follow-up, so the analyses pertain to data from the 33 completers.

2.2 Procedure

Potential participants were screened by telephone. Due to COVID-19 distancing mandates, assessment procedures transitioned from face-to-face laptop PC based-assessments (n = 21) either in the lab or in a home-visit, to coached remote self-administered assessments on the participant’s own device (n = 12). Testing was performed using either Inquisit 5 Lab (face-to-face testing with a research assistant) or Inquisit 5 web (remote testing) neurobehavioral testing software (Millisecond Software LLC, Seattle WA) using the same core script code for in-person vs lab administration for each assessment. Remote participants were instructed by staff to download the Inquisit 5 Web app onto their device as well as the task testing scripts using a custom hyperlink transmitted by email or text message. Each participant underwent a ∼ 90 minute baseline assessment composed of: 1) verbal-interview based questionnaires about mood and pain symptomatology and QoL, 2) a ∼10 minute rest break, and 3) a series of neurobehavioral decision-making and performance tasks. Six months following the baseline assessment, each participant was phoned for a follow-up interview composed of the QoL and symptomatology questionnaires only.

2.3 Symptomatology and QoL assessments at baseline and follow-up

To capture mood and pain symptomatology as well as QoL, we administered scales from the PROMIS SCI-QOL item bank Version 1.0: Pain Interference, Pain Behavior, Depression, Grief and Loss, Positive Affect and Well-Being, and Resilience. These assessments were administered by verbal interview at baseline (either in-person or by phone post-COVID) to harmonize with (uniformly) phone-based follow-up repeat assessments of these six instruments.

2.4 Trait impulsivity and behavior self-report assessments

To probe different facets if impulsivity, we administered the Urgency, Perseverance, Premeditation, Sensation-Seeking, and Positive Urgency Scale (UPPS-P) questionnaire. This instrument indexes each of negative urgency, positive urgency (lack of) perseverance, (lack of) premeditation, and sensation-seeking, as distinct subscales. Each UPPS-P subscale scores is the average value of its individual item endorsements (range 1–4). Juxtaposed with trait impulsivity, we indexed mindfulness (propensity to act with awareness or “live in the moment” in a reflective way) with the Mindful Attention and Awareness Scale (MAAS)(Brown & Ryan, 2003). Finally, to assess individual differences in self-care behavior, we administered the Spinal Cord Injury lifestyle Scale (SCILS) (Pruitt et al., 1998), which includes cardiovascular, genitourinary, neuromusculoskeletal, skin care, and psychosocial subscales. Its external validity is suggested by findings that SCILS scores have correlated negatively with incidence of SHC (Mashola & Mothabeng, 2019). Substance use behavior was assessed with the Fagerstrom Test for Nicotine Dependence (smoking), the AUDIT (alcohol use severity), and the PhenX toolkit lifetime and past-30-day substance use modules (other substances).

2.5 Neurobehavioral tasks

To measure myopic propensity for living for the moment with little consideration of the future, we administered a titrating Delay Discounting Task (DDT)(Bickel et al., 2012). In each trial, the participant had to choose between two alternative hypothetical monetary rewards: a lesser amount available now and a larger standard amount ($50 in one task block, $5000 in the other, counterbalanced) given at various points in the future (6 hours to 5 years). We quantified each participant’s proclivity toward devaluation of rewards with delay by using the area under the curve (AUC) defined by his/her indifference points between smaller-sooner and larger-later rewards across the different delay intervals, where lower AUC indicates more severe discounting (impulsivity).

We also administered the Probability Discounting Task (PDT) (Madden et al., 2009) to measure risk-taking, specifically how the individual devalues (discounts) prospective rewards as a function of reduced probability of their delivery. The PDT consists of 30 hypothetical choices between receiving a certain reward (valued from $20 to $40) versus receiving a larger ($60 to $100) but uncertain reward. For example, “Which would you prefer: Winning $20 for sure, or a 1-in-10 chance (10%) of winning $80?” For three of the questions, the reward alternatives had the same expected value (EV; reward amount X probability), for 12 questions, the guaranteed amount had the higher EV (optimal), and for the remaining half of questions (n = 15), the gamble had the higher EV. We calculated the proportion of choices of the risky reward option across trials.

To assess rapid-response impulsivity, we administered an Emotional Go-Nogo Task (EGNGT) (Tottenham et al., 2011). The EGNGT presented a series of face photographs singly for 500 ms with a 1.5- 4 s (jittered) intertrial interval. The participant was instructed to respond to happy faces but withhold responses to each calm (expressionless) face in the HC blocks, or to respond to fearful faces but withhold responses to calm faces in FC blocks, with the converse instruction in CH and CF blocks, respectively. The key impulsivity metric was percentage of “false alarm” commission error responses to non-targets across blocks.

2.6 Follow-up interview

Six months (±2 weeks) after the baseline assessment, participants were telephoned for a repeat administration of the PROMIS assessments.

2.7 Data analysis

First, we derived clusters of participants based on positive and negative symptomology measured by the PROMIS QoL interview over time. Then, questionnaire and neurobehavioral metrics were compared between the two symptom-defined cluster groups using one-sided t-tests, in accord with our expectation of uniformly greater impulsivity, lower mindfulness, and reduced self-care in those participants with consistently lower QoL.

3 Results

3.1 Derivation of participant clusters

To identify emergent clusters of participants based on stable patterns vs. changes in PROMIS symptomatology/QoL from baseline to six-month follow-up, we performed a k-means cluster analysis in JMP-SAS (SAS Institute, Cary, NC). First, scores from each of the six-scales at each time point were standardized with z-transformation. Next, to reduce data dimensionality, for each of the baseline and follow-up assessments, each participant’s z-scaled: 1) Depression and Grief and Loss scores were averaged to create a negative outlook composite, 2) Resilience and Positive Affect scores were averaged to create a positive outlook composite and 3) Pain Interference and Pain Behavior scores were averaged to create a pain composite. All three composite variables were then entered into K-means clustering with up to four potential cluster solutions envisioned: 1) persistent high positive, low negative symptom participants, 2) persistent low positive, high negative symptom participants, 3) participants with worsening symptoms from baseline to follow-up, and 4), participants with improving symptoms.

The two-cluster solution was shown to be optimum based on the minimal Cubic Clustering Criterion (CCC) value produced in the k-means analysis platform. This was composed of one predominant cluster of participants (Higher QoL group, (HQL): n = 23) scoring high in the positive outlook composite and low in negative outlook and pain composites across both baseline and follow-up, and a second cluster (Lower QoL group (LQL): n = 10) of participants scoring high in negative outlook and pain across time but low in positive outlook across both baseline and follow-up. Characteristics of these groups are shown in Table 1. Repeated-measures ANOVA indicated no significant change from baseline to follow-up in any of these three symptom composites in either group. The groups did not differ in proportion of participants who self-tested remotely (Chi Sq p = 0.789).

Table 1
Characteristics of the participants

Higher QoL Lower QoL

Mean (SD) Mean (SD) |t|/Chi-Sq p-value

Sex 18 M, 5 F 6 M, 4 F 0.0292 0.279

Age 32.5 (7.7) 32.5 (7.7) 0.007 0.500

PROMIS scores at baseline

Pain Interference 36.7 (12.3) 53.7 (25.8) 1.991 0.036

Pain Behavior 13.5 (5.7) 18.5 (7.0) 1.996 0.033

Depression 35.2 (9.8) 58.9 (16.6) 4.209 0.0006

Grief and Loss 25.5 (12.0) 49.2 (13.5) 4.791 0.0001

Positive Affect and Well-Being 129.2 (12.7) 103.0 (17.5) 4.258 0.0004

Resilience 97.5 (8.5) 78.2 (11.3) 4.846 0.0003

PROMIS scores at 6-mo follow-up

Pain Interference 36.7 (14.4) 62.8 (16.1) 4.411 0.0002

Pain Behavior 13.5 (5.4) 22.4 (6.7) 3.70 0.001

Depression 33.2 (6.1) 63.7 (14.6) 6.365 <0.0001

Grief and Loss 23.0 (7.8) 59.3 (12.2) 8.693 <0.0001

Positive Affect and Well-Being 129.9 (11.6) 99.9 (14.4) 5.815 <0.0001

Resilience 97.8 (7.1) 75.4 (9.8) 6.494 <0.0001

Fagerstrom-current smoker 9 Y, 24 N 5 Y, 5 N 3.376 0.053

Fagerstrom-score 0.6 (1.7) 2.8 (3.2) 1.889 0.044

AUDIT 3.9 (3.4) 3.1 (4.0)a 0.504 0.689

Phenx-Past 30 Days Cannabis 9 Y, 14 N 6 Y, 4 N 1.225 0.268

	Higher QoL	Lower QoL
	Mean (SD)	Mean (SD)	\|t\|/Chi-Sq	p-value
Sex	18 M, 5 F	6 M, 4 F	0.0292	0.279
Age	32.5 (7.7)	32.5 (7.7)	0.007	0.500
PROMIS scores at baseline
Pain Interference	36.7 (12.3)	53.7 (25.8)	1.991	0.036
Pain Behavior	13.5 (5.7)	18.5 (7.0)	1.996	0.033
Depression	35.2 (9.8)	58.9 (16.6)	4.209	0.0006
Grief and Loss	25.5 (12.0)	49.2 (13.5)	4.791	0.0001
Positive Affect and Well-Being	129.2 (12.7)	103.0 (17.5)	4.258	0.0004
Resilience	97.5 (8.5)	78.2 (11.3)	4.846	0.0003
PROMIS scores at 6-mo follow-up
Pain Interference	36.7 (14.4)	62.8 (16.1)	4.411	0.0002
Pain Behavior	13.5 (5.4)	22.4 (6.7)	3.70	0.001
Depression	33.2 (6.1)	63.7 (14.6)	6.365	<0.0001
Grief and Loss	23.0 (7.8)	59.3 (12.2)	8.693	<0.0001
Positive Affect and Well-Being	129.9 (11.6)	99.9 (14.4)	5.815	<0.0001
Resilience	97.8 (7.1)	75.4 (9.8)	6.494	<0.0001
Fagerstrom-current smoker	9 Y, 24 N	5 Y, 5 N	3.376	0.053
Fagerstrom-score	0.6 (1.7)	2.8 (3.2)	1.889	0.044
AUDIT	3.9 (3.4)	3.1 (4.0)a	0.504	0.689
Phenx-Past 30 Days Cannabis	9 Y, 14 N	6 Y, 4 N	1.225	0.268

We sought to characterize the clinical significance of depression levels in the two groups. Because the PROMIS SCI-QOL Depression 28-item configuration has not been population normed, we calculated the PROMIS Depression 10a scale score of each participant from those 10 individual item responses from the full scale. The median reconstructed PROMIS Depression 10a score of the LQL group (19) corresponds to a scaled t-score of 55.4 (Tulsky et al., 2015), where a t-score of 55 or higher in several PROMIS depression variants was found to be an optimal cut point for correspondence with major depression diagnosis per structured clinician interview in patients with chronic pain or stroke (Kroenke et al., 2021). Therefore, many if not most LQL participants were likely clinically depressed. Conversely, the median Depression 10a score of the HQL group (10) corresponds to a t-score of only 38.3, over 1 SD below population norms.

In subsequent analyses, we compare the following metrics of self-care, mindfulness, and impulsivity between the HQL and LQL groups.

3.2 Self-report questionnaire assessments

LQL participants showed greater Fagerstrom (nicotine dependence) scores and higher levels of negative urgency impulsivity, but lower levels of perseverance and mindfulness (See Fig. 1 and Table 2). The LQL participants also showed reduced self-reported genitourinary and psychosocial self-care.

Fig. 1

Shown are mean (±SEM) scale scores of the UPPS-P (which are defined as the mean item-level score for that subscale) as well as the mean item score of the MAAS in each of the LQL and HQL groups. *denotes significant difference per one-tailed t-test P < 0.05.

Table 2

Performance metrics from the EGNGT

	Higher QoL	Lower QoL
	Mean (SD)	Mean (SD)	\|t\|/Chi-Sq	p-value
Fagerstrom-current smoker	9 Y, 24 N	5 Y, 5 N	3.376	0.053
Fagerstrom-score	0.6 (1.7)	2.8 (3.2)	1.889	0.044
AUDIT	3.9 (3.4)	3.1 (4.0)a	0.504	0.689
Phenx-Past 30 Days Cannabis	9 Y, 14 N	6 Y, 4 N	1.225	0.268
UPPS-P
Negative Urgency	1.9 (0.5)	2.4 (0.4)	2.675	0.007
Positive Urgency	1.8 (0.6)	1.8 (0.4)	0.073	0.528
(Lack of) Premeditation	1.9 (0.6)	2.1 (0.2)	1.532	0.069
(Lack of) Perseverance	1.7 (0.3)	2.2 (0.5)	3.057	0.005
Sensation-Seeking	3.1 (0.7)	2.9 (0.4)	1.058	0.851
MAAS	4.7 (0.7)	3.7 (1.0)	2.885	0.006
SCILS
Cardiovascular	9.2 (3.2)	8.9 (3.5)	0.247	0.404
Genitourinary	11.1 (3.4)	9.0 (2.9)	1.801	0.043
Neuromuscular	19.9 (5.9)	16.6 (5.2)	1.594	0.064
Skin care	22.3 (3.7)	20.4 (6.8)	1.045	0.159
Psychosocial	7.7 (0.7)	6.5 (1.4)	2.743	0.001
TOTAL	70.7 (10.1)	61.4 (15.7)	1.721	0.055
Delay Discounting Task
Area-under-curve $50 reward	226.6 (78.9)	213.4 (45.6)	0.594	0.279
Area-under-curve $5000 reward	28032 (5991)	24460 (6728)	1.439	0.084
Probability Discounting Task
Proportion risky choices	0.57 (0.2)	0.46 (0.2)	1.343	0.901
Emotional Go-Nogo Task
Hit rate (%)	92.5 (0.1)	89.9 (0.2)	0.337	0.360
Commission error rate (%)	19.9 (0.1)	18.8 (0.1)	0.247	0.404
Median reaction time to targets	464.5 (63.5)	467.5 (60.4)	0.129	0.551

3.3 Behavioral assessments

Regarding decision-based impulsivity, LQL participants did not show significantly increased delay discounting in the DDT or increased risk-taking on the PDT. Performance metrics from the EGNGT were similar between groups, including commission error rates (c.f. Table 2).

In light of the limited sample size, we performed post hoc tests of minimum sample size required to detect a significant group difference in task performance, based on observed mean differences and variability. These indicated that 2029 participants would be needed to detect a group difference in motoric impulsivity in the EGNGT in either direction, and 514 participants would be needed to detect a group difference in delay discounting of smaller rewards. Although post hoc tests of power for null results warrant caution, these analyses nevertheless suggest that if differences in these domains exist between persons with high QoL and low QoL, they are of trivial magnitude. Conversely, a more feasible sample of 71 participants would have detected significantly higher risk-taking (i.e. lower probability discounting) in the HQL group. In addition, post hoc G*power analysis indicated that with 80% power (1 - β), the current sample size of 33 participants was enough to detect all large-sized effects and medium-sized effects > ρ= 0.40 in these t-tests, but not small effects.

4 Discussion

We sought to determine whether persons with paraplegia who experience consistently low psychological, psychosocial and pain-related QoL (LQL) across a six-month time span that began well after injury were characterized by significantly greater facets of impulsivity compared to persons with higher QoL (HQL). Using composites of PROMIS SCI questionnaires, cluster analysis revealed two groups with distinctly different mood, attitude and pain scores that remained consistent over assessments. To these groups, we administered a variety of impulsivity probes. These ranged from personality trait characteristics, to increased preference for immediate rewards versus deferred gratification, to rapid-response impulsivity as a potential marker of lower EF (Friedman & Miyake, 2017), and also impulsivity in terms of SCI self-care, where low levels of self-care could reflect preference for physical ease in the moment over deferred health benefits of preventive action.

Analyses suggested significantly increased negative urgency and lack of perseverance, as well as poorer self-care, in the cluster of participants defined by persistently low QoL and high pain and depression symptoms. Conversely, there appeared to be minimal differences in decision-based or motoric impulsivity. Finally, the LQL participant cluster also showed lower levels of mindfulness on the MAAS. Parenthetically, across all participants, both negative urgency (r = –0.571, p = 0.003) and lack of perseverance (r = –0.679, p < 0.0001) scores correlated negatively with MAAS scores, in accord with negative correlations found between mindfulness and other impulsivity scales (Peters et al., 2011).

Our data suggest then that persons with traumatic SCI who are consistently depressed and who have more grief and loss over their injury are also more prone to acting impulsively while in a negative mood state, as measured in UPPS-P negative urgency. This scale contains items such as: “When I feel bad, I will often do things I later regret in order to make myself feel better now.” This significantly greater emotion-related impulsivity in the LQL group also replicates previous findings of greater impulsivity in the clinically depressed (Saddichha & Schuetz, 2014), and could possibly relate to their greater smoking as indexed by Fagerstrom scores. In addition, LQL had higher UPPS-P (lack of) perseverance scores, endorsing items such as: “Sometimes there are so many little things to be done that I just ignore them all.”

The LQL group also showed reduced levels of some facets of self-care, such as in the psychosocial domain, with item content such as “I am with or talk to other people at least once a day.” This absence of social contact may be accounted for by the higher levels of depression and pain symptomatology in the LQL group. The trend toward lower self-care overall in the LQL group, coupled with their lack of perseverance, suggests a general motivational deficit or inertia. As noted previously, motivational decrements are also characteristic of populations selected for psychiatric disorder (Husain & Roiser, 2018).

Conversely, there were no consistent indications of increased decision-making or rapid-response impulsivity in the LQL participants. Of note, discounting of smaller rewards (such as the $50 delayed reward trials herein) is typically more sensitive to individual differences than discounting of larger rewards because most individuals discount larger rewards less severely (Johnson & Bickel, 2002), potentially restricting range. Thus, it is unexpected that LQL would show a trend toward more severe discounting than HQL only for the larger amount. The LQL group also showed a trend toward fewer risky choices in the PDQ than the HQL group. This could have perhaps stemmed from comorbid anxiety in the LQL group, where participants with anxiety show increased aversion to uncertainty (Gu et al., 2020). Moreover, the results may have been driven by how risk-taking was actually (economically) advantageous in half of the choice trials, and may have been dependent on how probabilities were explicitly presented.

Finally, there were no group differences in motor impulsivity, indexed by “false alarm” commission error responses to non-target stimuli in the EGNGT. Ability to withhold inappropriate motor responses on sub-second time scales has been found to load onto a common g-like cognitive ability factor in bifactor modeling of human cognitive performance (Friedman & Miyake, 2017). Thus, this similarity between groups loosely suggests that differences in QoL are not associated with differences in raw executive function. However, this would be more properly tested with a fuller cognitive battery. We note that in general, cognitive tasks have lower test-retest reliability than trait questionnaire-based assessments, due to transient factors such as fatigue, and this is thought to account in part for poor within-subject correlations between task-based and questionnaire-based assessments of impulsivity (Cyders & Coskunpinar, 2011).

These findings should be considered in light of study limitations, chief of which was the low sample size. These findings should be replicated in full-scale studies sufficiently powered to explore sex, gender or other demographic moderators of impulsivity relationships. However, we nevertheless believe these preliminary results are of practical significance for neurobehavioral research in SCI in that they demonstrate that trait-based questionnaire metrics are much more potent at detecting greater impulsivity as a risk factor in individuals who persist in high levels of negative mood symptomatology well after traumatic SCI. In addition, to reduce participant burden, we did not collect a formal structured clinical interview for mental disorders. Finally, data collection was an admixture of face-to-face interview and laptop administration and remote testing on participant-owned computers or other devices.

5 Conclusion

These preliminary findings indicate that individuals with paraplegia who show enduring depressed mood with a negative outlook may be at increased risk for poor medical and psychosocial outcomes, including SHC. Replicated findings would suggest that propensity to act rashly in a negative mood state and propensity to not persevere and consider the future could be directly targeted in populations with paraplegia in novel trials of post-injury counseling. These behaviors are being be successfully targeted, for example, in certain cognitive behavioral therapies (CBT) to treat substance use disorder (Polak et al., 2019). Notably, relapse to substance use is often driven by acute negative affect without regard to the future. Finally, mindfulness interventions in SCI have shown greater efficacy for reducing depression than reducing pain (Hearn & Cross, 2020), suggesting the utility of this intervention in SCI populations defined more by negative affect.

Footnotes

Acknowledgments

The authors thank Ms. Lisa Straub for project management.

Conflict of interest

None of the authors has any financial conflict of interest regarding the study findings to disclose.

Funding

This research was funded by award #W81XWH1910032 from the Congressionally Directed Medical Research Programs of the Department of Defense.

Data availability

De-identified data are available on request to qualified investigators by contacting the first author.

References

Allin,

, Shepherd,

, Thorson,

, Tomasone,

, Munce,

, Linassi,

, McBride,

C. B.

, Jiancaro,

, & Jaglal,

(2020) Web-Based Health Coaching for Spinal Cord Injury: Results From a Mixed Methods Feasibility Evaluation. JMIR Rehabilitation and Assistive Technologies, 7(2), e16351. https://doi.org/10.2196/16351

Bailey,

D. P.

, Withers,

T. M.

, Goosey-Tolfrey,

V. L.

, Dunstan,

D.W.

, Leicht,

C. A.

, Champion,

R. B.

, Charlett,

O. P.

, & Ferrandino,

(2020) Acute effects of breaking up prolonged sedentary time oncardiovascular disease risk markers in adults with paraplegia. Scandinavian Journal of Medicine & Science in Sports, 30(8), 1398–1408. https://doi.org/10.1111/sms.13671

Bates,

L. C.

, Conners,

, Zieff,

, Adams,

N. T.

, Edgar,

K. M.

, Stevens,

, Faghy,

M. A.

, Arena,

, Vermeesch,

, Joseph,

R. P.

, Keith,

, & Stoner,

(2021) Physical activity and sedentary behavior in people with spinal cord injury: Mitigation strategies during COVID-19 on behalf of ACSM-EIM and HL-PIVOT. Disabilityand Health Journal 101177. https://doi.org/10.1016/j.dhjo.2021.101177.

Bickel,

W. K.

, Jarmolowicz,

D. P.

, Mueller,

E. T.

, Koffarnus,

M. N.

, & Gatchalian,

K. M.

(2012) Excessive discounting of delayed reinforcers as a trans-disease process contributing to addiction and other disease-related vulnerabilities: Emerging evidence. Pharmacol Ther, 134(3), 287–297. https://doi.org/10.1016/j.pharmthera.2012.02.004.

Bickel,

W. K.

, Yi,

, Landes,

R. D.

, Hill,

P. F.

, & Baxter,

(2011) Remember the future: Working memory training decreases delaydiscounting among stimulant addicts. Biol Psychiatry, 69(3), 260–265. https://doi.org/10.1016/j.biopsych.2010.08.017.

Bonanno,

G. A.

, Kennedy,

, Galatzer-Levy,

I. R.

, Lude,

, & Elfström,

M. L.

(2012) Trajectories of resilience, depression, and anxiety following spinal cord injury. Rehabilitation Psychology, 57(3), 236–247. https://doi.org/10.1037/a0029256.

Brown,

K. W.

, & Ryan,

R. M.

(2003) The benefits of being present: Mindfulness and its role in psychological well-being. Journal of Personality and Social Psychology, 84(4), 822–848.

Chen,

H.-L.

, Cai,

J.-Y.

, Du,

, Shen,

H.-W.

, Yu,

H.-R.

, Song,

Y.-P.

, & Zha,

M.-L.

(2020) Incidence of Pressure Injury in Individuals With Spinal Cord Injury: A Systematic Review and Meta-analysis. Journal of Wound, Ostomy, and Continence Nursing: Official Publication of The Wound, Ostomy and Continence Nurses Society, 47(3), 215–223. https://doi.org/10.1097/WON.0000000000000633.

Crespo-Ruiz,

, del-Ama,

A. J.

, Jiménez-Díaz,

F. J.

, Morgan,

, de a Peña-González,

lA.

, & Gil-Agudo,

Á. M.

(2012) Physical activity and transcutaneous oxygen pressure in menwith spinal cord injury. Journal of Rehabilitation Research andDevelopment, 49(6), 913–924. https://doi.org/10.1682/jrrd.2011.05.0087.

10.

Cyders,

M. A.

, & Coskunpinar,

(2011) Measurement of constructs using self-report and behavioral lab tasks: Is there overlap in nomothetic span and construct representation for impulsivity? Clinical Psychology Review, 31(6), 965–982. https://doi.org/10.1016/j.cpr.2011.06.001.

11.

Friedman,

N. P.

, & Miyake,

(2017) Unity and diversity of executive functions: Individual differences as a window on cognitive structure. Cortex; a Journal Devoted to the Study of the Nervous System and Behavior, 86, 186–204. https://doi.org/10.1016/j.cortex.2016.04.023.

12.

Gu,

, Gu,

, Lei,

, & Li,

(2020) From Uncertainty to Anxiety: How Uncertainty Fuels Anxiety in a Process Mediated by Intolerance of Uncertainty. Neural Plasticity, 2020, 8866386. https://doi.org/10.1155/2020/8866386

13.

Hearn,

J. H.

, & Cross,

(2020) Mindfulness for pain, depression, anxiety, and quality of life in people with spinal cord injury: A systematic review. BMC Neurology, 20(1), 32. https://doi.org/10.1186/s12883-020-1619-5

14.

Husain,

, & Roiser,

J. P.

(2018) Neuroscience of apathy and anhedonia: A transdiagnostic approach. Nature Reviews. Neuroscience, 19(8), 470–484. https://doi.org/10.1038/s41583-018-0029-9.

15.

Johnson,

M. W.

, & Bickel,

W. K.

(2002) Within-subject comparison of real and hypothetical money rewards in delay discounting. JExp Anal Behav, 77(2), 129–146. https://doi.org/10.1901/jeab.2002.77-129

16.

Kroenke,

, Stump,

T. E.

, Kean,

, Krebs,

E. E.

, Damush,

T. M.

, Bair,

M. J.

, & Monahan,

P. O.

(2021) Diagnostic operating characteristics of PROMIS scales in screening for depression. Journal of Psychosomatic Research, 147, 110532. https://doi.org/10.1016/j.jpsychores.2021.110532

17.

Lavela,

S. L.

, Etingen,

, & Miskevics,

(2016) FactorsInfluencing Self-Care Behaviors in Persons with Spinal Cord Injuriesand Disorders. Topics in Spinal Cord Injury Rehabilitation, 22(1), 27–38. https://doi.org/10.1310/sci2201-27.

18.

Lo,

, Chan,

, & Flynn,

(2021) A Systematic Review of the Incidence, Prevalence, Costs, and Activity and Work Limitations of Amputation, Osteoarthritis, Rheumatoid Arthritis, Back Pain, Multiple Sclerosis, Spinal Cord Injury, Stroke, and Traumatic Brain Injury in the United States: A 2019 Update. Archives of Physical Medicine and Rehabilitation, 102(1), 115–131. https://doi.org/10.1016/j.apmr.2020.04.001.

19.

Madden,

G. J.

, Petry,

N. M.

, & Johnson,

P. S.

(2009) Pathological gamblers discount probabilistic rewards less steeply than matched controls. Experimental and Clinical Psychopharmacology, 17(5), 283–290. https://doi.org/10.1037/a0016806.

20.

Mashola,

M. K.

, & Mothabeng,

D. J.

(2019) Associations betweenhealth behaviour, secondary health conditionsand quality of life in people with spinal cord injury. AfricanJournal of Disability, 8, 463. https://doi.org/10.4102/ajod.v8i0.463

21.

McTeague,

L. M.

, Goodkind,

M. S.

, & Etkin,

(2016) Transdiagnostic impairment of cognitive control in mental illness. Journal of Psychiatric Research, 83, 37–46. https://doi.org/10.1016/j.jpsychires.2016.08.001.

22.

Peters,

J. R.

, Erisman,

S. M.

, Upton,

B. T.

, Baer,

R. A.

, & Roemer,

(2011) A Preliminary Investigation of the Relationships Between Dispositional Mindfulness and Impulsivity. Mindfulness, 2(4), 228–235. https://doi.org/10.1007/s12671-011-0065-2.

23.

Piatt,

J. A.

, Nagata,

, Zahl,

, Li,

, & Rosenbluth,

J. P.

(2016) Problematic secondary health conditions among adults with spinal cord injury and its impact on social participation and daily life. The Journal of Spinal Cord Medicine, 39(6), 693–698. https://doi.org/10.1080/10790268.2015.1123845.

24.

Polak,

, Burroughs,

T. K.

, Reisweber,

, & Bjork,

J. M.

(2019) Four-Session Transcending Self Therapy for Substance use, Depression, and Treatment Retention among Veterans with Substance use Disorders: A Pilot Study. J Addict Res Ther, 10(2), 380.

25.

Pruitt,

S. D.

, Wahlgren,

D. R.

, Epping-Jordan,

J. E.

, & Rossi,

A. L.

(1998) Health behavior in persons with spinal cord injury: Development and initial validation of an outcome measure. Spinal Cord, 36(10), 724–731.

26.

Saddichha,

, & Schuetz,

(2014) Impulsivity in remitted depression: A meta-analytical review. Asian Journal of Psychiatry, 9, 13–16. https://doi.org/10.1016/j.ajp.2014.02.003.

27.

Story,

G. W.

, Vlaev,

, Seymour,

, Darzi,

, & Dolan,

R. J.

(2014) Does temporal discounting explain unhealthy behavior? A systematic review and reinforcement learning perspective. Article. Frontiers in Behavioral Neuroscience, 8, 76. https://doi.org/10.3389/fnbeh.2014.00076

28.

Swann,

A. C.

, Bjork,

J. M.

, Moeller,

F. G.

, & Dougherty,

D. M.

(2002) Two models of impulsivity: Relationship to personality traits and psychopathology. Biol Psychiatry, 51(12), 988–994.

29.

Tottenham,

, Hare,

T. A.

, & Casey,

B. J.

(2011) Behavioral assessment of emotion discrimination, emotion regulation, and cognitive control in childhood, adolescence, and adulthood. Frontiers in Psychology, 2, 39. https://doi.org/10.3389/fpsyg.2011.00039

30.

Tulsky,

D. S.

, Kisala,

P. A.

, Kalpakjian,

C. Z.

, Bombardier,

C. H.

, Pohlig,

R. T.

, Heinemann,

A. W.

, Carle,

, & Choi,

S. W.

(2015) Measuring depression after spinal cord injury: Development and psychometric characteristics of the SCI-QOL Depression item bank and linkage with PHQ-. The Journal of Spinal Cord Medicine, 38(3), 335–346. https://doi.org/10.1179/2045772315Y.0000000020.

31.

Voss,

A. T.

, Jorgensen,

M. K.

, Murphy,

J. G.

(2021). Episodic future thinking as a brief alcohol intervention for heavy drinking college students: A pilot feasibility study. Experimental and Clinical Psychopharmacology. https://doi.org/10.1037/pha0000451.