Effects of mTBI with loss of consciousness on neurobehavioral symptoms, depression, and insomnia in former collegiate and NFL football athletes

ABSTRACT

Objective

Considering that diagnostic decisions about mTBI are often predicated on clinical symptom criteria, it is imperative to determine which initial presentation features of mTBI have prognostic significance for identifying those at high risk for long-term functional impairment.

Setting

Zoom interview Participants: Male, former NCAA Division I, and professional-level National Football League (NFL) athletes (n = 177) between the ages of 27 and 85 (M = 54.1, SD = 14.7).

Design

Cross-sectional case-control. Main Measures: History of mild TBI, history of loss of consciousness (LOC), depression symptoms, insomnia, neurobehavioral symptoms.

Results

Number of mTBI exposures did not predict neurobehavioral symptoms (B = 0.21, SE = 0.18, p = 0.23), but number of mTBI + LOC events did (B = 2.27, SE = 0.64, p = <.001). Further analysis revealed that the number of mTBI + LOC events predicted neurobehavioral symptoms indirectly through both depression (B = 0.85, 95% CI = [0.27, 1.52) and insomnia (B = 0.81, 95% CI = [0.3, 1.4]). Further, the direct effect of mTBI + LOC events on neurobehavioral symptoms became non-significant when depression and insomnia were added to the model (B = 0.78, SE = 0.45, p = 0.08).

Conclusions

Findings support LOC at time of injury as an important predictor of long-term outcomes. Additionally, results suggest depression and insomnia as potential mediators in the association between mTBI + LOC and neurobehavioral symptoms. These findings provide justification for early depression and insomnia symptom monitoring following mTBI + LOC.

KEYWORDS:

• Mild traumatic brain injury
• loss of consciousness
• neurobehavioral symptoms
• insomnia
• depression
• American Football

Introduction

Worldwide, mild traumatic brain injury (mTBI) affects approximately 42 million individuals annually and has become a pressing health concern over the past several decades (1,2). Diagnosing mTBI or ‘concussion,’ has become increasingly controversial, as there is no universal consensus on a definition. Generally, mTBI is characterized by head injuries that result in disrupted brain function as evidenced by altered mental status (e.g. confusion and disorientation), loss of consciousness (LOC) (<30 min), post-traumatic amnesia (<24 hours), and/or transient neurological dysfunction (e.g. headache and dizziness) (3). These clinical features are typically accompanied by minimal impairment on objective neurological measures, e.g. Glasgow Coma score of 13–15 (4). Within a sports context, the Concussion in Sport Group has developed a conceptual definition of sports-related concussion as a “traumatic brain injury caused by a direct blow to the head, neck, or body resulting in an impulsive force being transmitted to the brain that occurs in sports and exercise-related activities (5). Clinical symptoms resulting from sports-related concussions may or may not involve loss of consciousness and will not present on standard clinical neuroimaging (computed tomography or magnetic resonance imaging T1- and T2-weighted images) as an abnormality (5). Research shows that while most post-concussive symptoms resolve after several months (6–9), a subset of individuals experience prolonged neurobehavioral symptoms that affect quality of life and daily functioning (e.g. headache, dizziness, fatigue, irritability, concentration difficulties, memory impairment, and disrupted emotional functioning) (10). Since mTBI diagnoses are often predicated on clinical symptom criteria, it is imperative to determine which initial presentation features of mTBI have prognostic significance for identifying those at high risk for long-term functional impairment.

Football athletes

Due to frequent head collisions that occur in American football and the pressure to return to play after injury, identifying prognostic markers for mTBI-related health outcomes is particularly relevant in football players. Over 70% of college-level football players have experienced acute symptoms of mTBI at some point in their athletic career, while 39% report having sustained an mTBI with LOC (mTBI + LOC) (11). Within a professional American football sample, 58% of players returned to play less than seven days after LOC (12), and LOC at time of injury was strongly associated with poor cognition-related quality of life, even decades after retiring from play (13).

Repeated mTBI exposure

While most individuals recover from a single mTBI within four weeks, repeated mTBI exposure is associated with prolonged recovery and chronic symptoms in adolescent and young adult athlete populations (14). There is no clear evidence regarding changes in recovery time upon repeated mTBI in professional athlete populations. Although controversial, evidence suggests a link between repeated mTBI and chronic neurobehavioral symptoms. For example, in rodents, sustaining multiple mTBIs is linked with neuropathology that is distinct from isolated exposure (15–18), including increased long-term neuroinflammation and white matter degradation (16,19). In humans, athletes with a history of multiple mTBIs perform significantly worse on neurocognitive assessments (20–22) and are at much higher risk of long-term mild cognitive impairment when compared to a control group of athletes (23).

Loss of consciousness

While repeated mTBI exposures may have compounding effects, LOC at time of injury may independently impact long-term neurobehavioral outcomes (24–27). Although only a subset of mTBIs are associated with LOC, it is used as a diagnostic criterion to distinguish mTBI from moderate-to-severe TBI (28). LOC indicates greater severity of injury; neuroimaging reveals that mTBI + LOC has increased neuropathological biomarkers of dementia (e.g. amyloid-β load, gross infarctions; microinfarctions) (25) and decreased connectivity within prefrontal white matter tracts (26,27). Importantly, groups with more lifetime mTBI + LOC events show both decreased white matter connectivity and decreased cognitive flexibility on neurocognitive tests, indicative of long-term deficits in executive functioning (26). Although there is compelling evidence for the mechanisms by which mTBI + LOC may affect long-term clinical symptoms and recovery trajectory after mTBI, LOC’s utility as a prognostic marker remains controversial (21,24). It is imperative to consider how LOC at time of injury may uniquely impact lifetime neurobehavioral symptomatology.

Depression and insomnia

Critically, mTBI exposure is also associated with long-term psychiatric symptoms. Sustaining one or more mTBIs increases risk of depression at >11 months post-injury (29,30), and those with a history of multiple mTBIs or mTBI + LOC have an even higher incidence of depression (31–37). The relationship between mTBI + LOC and increased depressive symptoms may be partly explained by decreased white matter connectivity in mTBI-exposed samples (26,27,34); a systematic review by Medeiros et al. (2022) report associations between DTI measures of white matter damage and depression severity after TBI (38). Major Depression is also a significant predictor of neurobehavioral symptoms (39,40), suggesting that neurobehavioral symptoms may be attributable to post-mTBI onset of depression. Despite these findings, the potential mediating effect of depression on the relationship between mTBI + LOC and long-term neurobehavioral symptoms has not been explored.

Nearly all patients with depression report sleep disturbance, and insomnia is associated with an increased risk for depression, suggesting a bi-directional relationship between the two (41). Long-term sleep disturbance is a prevalent sequelae following mTBI exposure (42), with exposure to mTBI + LOC further increasing risk (43–45). In adolescents, disrupted sleep following mTBI exposure is associated with a three-to-four-fold increase in recovery time (46). Insomnia has also been independently linked to deficits in cognition (47,48), including impairments in working memory and episodic memory (49). The impact of sleep disturbance on neurobehavioral symptoms has only briefly been examined post-mTBI, where sleep quality accounted for 32% of the variance in neurobehavioral symptoms post-injury (50). Pronounced sleep disruption after mTBI + LOC could explain the relationship between mTBI + LOC and long-term neurobehavioral symptoms. Prior research has documented the mediating effect of sleep quality on risk for depression in mTBI-exposed individuals (51), however, the effect of mTBI + LOC on this relationship remains unexamined.

It is important to note that many of the existing studies examining the relationship between mTBI and long-term depression, insomnia, and neurobehavioral symptoms are conducted in Veteran and community samples and may not generalize to athlete samples. Evidence suggests that the mechanism of injury (e.g. blast versus non-blast) and context of injury (e.g. in conjunction with combat-related trauma) may influence pathology and outcomes (52–54).

Current study

In the current study, we investigated the long-term impact of repeated mTBI with and without LOC on neurobehavioral symptoms in National Collegiate Athletic Association (NCAA) Division I former college-level and professional football athletes. Consistent with prior studies, we hypothesized that the number of mTBI + LOC events is associated with long-term neurobehavioral symptoms. Less is known, however, about whether the effects of mTBI + LOC events are still associated with neurobehavioral symptoms when controlling for the effects of depression and insomnia. Considering depression and insomnia are directly associated with mTBI and are independently associated with neurobehavioral symptoms (43,55–57), we predicted that mTBI + LOC would predict neurobehavioral symptoms indirectly through both depression and insomnia. Indirect effects may suggest insomnia and depression as potential mediators in the relationship, although further longitudinal investigation is needed. The practical consequences of this investigation will guide the implementation of short-term screening of depression and insomnia in high-risk patients following sports-related mTBI, for example, in those who lost consciousness at time of injury.

Methods

Participants

A sample of 211 of Male former contact and non-contact athletes were recruited via social media, recruitment flyers, and recruitment e-mails to participate in a larger neuroimaging study. Participants were primarily contacted through NFL alumni chapters, university athletic departments, and NCAA football team offices. For the present study, 30 former NCAA Division I non-contact sport athletes were excluded from the sample. One hundred and seventy-seven former elite football athletes between the ages of 27 and 85 (M = 54.1, SD = 14.7) were included. The sample consisted of former collegiate (NCAA Division I; n = 65) and professional-level (National Football League [NFL]; n = 112) football players. Participants completed a health survey administered virtually over Zoom. The questionnaire assessed mTBI history and included several measures of general health outcomes.

All participants were fluent in English, had no history of neurological injury or disease and no substance use disorder. Participants provided informed consent, and all study procedures were approved by the Duke University Institutional Review Board and the George Washington University Institutional Review Board. For full inclusion and exclusion criteria, see Appendix A.

History of mTBI and mTBI+LOC

Five retrospective questions were used to determine (1) number of mTBI exposures and (2) number of mTBI + LOC events during each phase of the participant’s athletic career (Table 1). Participants responded on a scale of 0–30 (number of mTBI exposures) for each question. Total number of sports-related mTBI exposures and mTBI + LOC events were determined by summing the response from these 5 questions, with total scores ranging from 0 to 120 mTBI exposures (M = 7.9, SD = 14.1) and 0 to 8 mTBI + LOC events (M = 1.1, SD = 1.5), out of a possible 150. To address outliers in a limited sample size, we performed 90% Winsorization on both mTBI exposures and mTBI + LOC events, whereby all data below the 5th percentile were set equal to the 5th percentile (mTBI = 0, mTBI + LOC = 0), and data above the 95th percentile set to the 95th percentile (mTBI = 12.9, mTBI + LOC = 3.9) (58). A total of 26 mTBI observations and 14 mTBI + LOC observations were adjusted for Winsorization. After Winsorization, mTBI exposures ranged from 0 to 12.9 (M = 5.3, SD = 4.42), and mTBI + LOC events ranged from 0 to 3.9 (M = 1.03, SD = 1.3).

Table 1. Health Survey questions on athlete sport-related concussion history

	Question
History of mTBI	(1)‘How many sports-related concussions did you have before high school?’
	(2)‘How many sports-related concussions did you have in high school?’
	(3)‘How many sports-related concussions did you have in college?’
	(4)‘How many sports-related concussions did you have in the armed forces?’
	(5)‘How many sports-related concussions did you have in your professional career?’
History of mTBI+LOC	(1)‘Before high school, how many times did you lose consciousness (“knocked out”) due to sports-related concussion?’
	(2)‘In high school, how many times did you lose consciousness (“knocked out”) due to sports-related concussion?’
	(3)‘In college, how many times did you lose consciousness (“knocked out”) due to sports-related concussion?’
	(4)‘In the armed forces, how many times did you lose consciousness (“knocked out”) due to sports-related concussion?’
	(5)‘In your professional career, how many times did you lose consciousness (“knocked out”) due to sports-related concussion?’

Neurobehavioral symptoms

The Neurobehavioral Symptom Inventory (NSI) is a 22-item questionnaire assessing the current severity of post-concussive symptoms associated with mTBI, including dizziness, poor concentration, and slowed thinking. All items were measured on a five-point scale, with self-reported symptom ratings between 0 (absent) and 4 (severe) (59). The NSI has excellent internal reliability (α = 0.93) and includes a total score (sum of all items) and subscale scores for three symptom categories: cognitive, affective, and somatic (60). In addition, the Validity-10 scale can be derived to screen for potential overreporting of symptoms by summing ten NSI items rarely endorsed by individuals exposed to TBI (61). As recommended for clinical utility (62), we excluded participants with a Validity-10 score >18 (n = 4), which may suggest symptom exaggeration (63).

Depression

Current depressive symptom severity (symptoms experienced within 2 weeks of study participation) was measured using the Beck Depression Inventory II (BDI-II), a 21-item self-report questionnaire (64). Participants responded using a four-point scale (0 to 3), with 0 and 3, respectively, indicating lowest and highest severity of symptoms. Total severity scores were determined by summing the 21 items.

Insomnia

Current insomnia symptom severity (symptoms experienced within 2 weeks of study participation) was measured using a modified Insomnia Severity Index (ISI) (65). The ISI is a 7-item questionnaire measuring the severity of sleep disturbance and its related waking symptoms. Participants respond using a 5-point scale (0–4) (66,67). By error, a modified version of ISI was administered for all participants in this study, which included 6 of the 7 items [excluded item: ‘To what extent do you consider your sleep problem to INTERFERE with your daily functioning (e.g. daytime fatigue, mood, ability to function at work/daily chores, concentration, memory, mood, etc.) CURRENTLY?’]. Total severity score was determined by summing the 6 items.

Statistical analyses

Statistical modeling was carried out using R (v. 4.2.1). First, bivariate relationships between mTBI and mTBI + LOC with neurobehavioral, depression, and insomnia symptom severity were evaluated using linear regression. In all cases, diagnostic tests were conducted to ensure that all data met the necessary assumptions for linear regression. Descriptions and results of diagnostic tests can be found in Appendix B.To follow-up on significant effects of mTBI + LOC and neurobehavioral symptom severity, we then assessed mTBI + LOC as a predictor of neurobehavioral symptom severity, indirectly through depression and insomnia. Analysis was performed using a parallel structural equation model with the package lavaan (68). For all indirect and total effects, 95% confidence intervals (CI) were estimated. CIs excluding 0 were considered statistically significant. The structural equation model was tested using 10,000 bootstraps resampled with replacement using bias-corrected confidence estimate.

Results

Sample characteristics

In the total sample (n = 177), race/ethnicity composition was 32% Black, 61% white, 1% Hispanic/Latino, 1% Asian, 4% multi-racial, and <1% other or not reported. Ages ranged from 27 to 85 years (M = 54.1, SD = 14.7). All participants completed at least some college. For sample characteristics, see Table 2.

Table 2. Sample demographics and symptom characteristics

	NCAA Division 1 (n = 69)	Professional (n = 108)	Total Sample (n = 177)
Age (SD)	47(15.4)	59(12.2)	54(14.7)
Race
White	63.8%	57.9%	60.9%
Black	27.5%	34.6%	32.1%
Mexican	1.4%	-	<1%
Other Latino	-	<1%	<1%
Asian	1.4%	<1%	1.1%
More than one	2.9%	4.7%	4%
Other	-	<1%	<1%
Not Reported	2.9	<1%	1.7%
Education
Some college	1.4%	10.1%	6.8%
Associate’s	-	<1%	<1%
Bachelor’s	52.2%	55.6%	54.2%
Postgraduate	46.4%	33.3%	38.4%
Mom Education
8^th grade or less	5.8%	10.2%	8.5%
Some HS	1.4%	4.6%	3.4%
HS graduate	28.9%	45.3%	38.9%
Some college	5.8%	8.3%	7.3%
Associate’s	2.9%	7.4%	5.6%
Bachelor’s	34.8%	12.9%	21.5%
Postgraduate	20.3%	11.1%	14.7%
LOCs (SD)	0.79(1.2)	1.3(1.6)	1.1(1.5)
mTBIs (SD)	6.9(14.1)	8.7(14.2)	7.9(14.1)
Winsorized LOCs(SD)	0.78(1.2)	1.2(1.3)	1.0(1.3)
Winsorized mTBIs(SD)	3.0(3.9)	5.9(4.6)	5.3(4.4)
NSI (SD)
Affective	5.6(5.0)	4.3(4.4)	4.8(4.7)
Somatic	2.3(2.7)	3.5(4.6)	3.3(4.0)
Cognitive	2.8(2.8)	2.5(3.2)	2.6(3.1)
Total NSI	10.8(8.7)	10.3(11.1)	10.5(10.2)
BDI (SD)	8.2(7.3)	6.5(6.9)	7.2(7.1)
ISI (SD)	6.4(5.5)	6.5(5.9)	6.5(5.8)

Results of one-way ANOVAs indicated neurobehavioral symptoms did not significantly differ by race (f = .015, p = 0.9). There were no associations between neurobehavioral symptoms and participant age. Neurobehavioral symptoms differed by at least two maternal education groups (f = 4.04, p = 0.05). Neurobehavioral symptoms differed between at least two participant education groups (f = 10.34, p = 0.001). Depressive symptoms did not significantly differ by race (f = 0.23, p = 0.99) or maternal education (f = 0.24, p = 0.62). There was a significant correlation between depressive symptoms and age (r = −0.18, p = 0.01). Depressive symptoms differed significantly between at least two participant education groups (f = 4.7, p = 0.03). Insomnia did not significantly differ by race (f = 1.09, p = 0.36), education (f = 0.36, p = 0.54), or maternal education (f = 2.47, p = 0.12). There was no association between insomnia and age (r = −8.9e-4, p = 0.36). For mean NSI scores by group, see Table 3.

Table 3. Demographic-level descriptive statistics for neurobehavioral, depression and insomnia questionnaires

	NSI(SD)	BDI(SD)	ISI(SD)
Race
White	10.7 (10.5)	7.4(8.1)	6.1(5.6)
Black	10.5 (9.8)	7.1(5.6)	7.5(5.9)
Mexican	13.0 (NA)	9.0(NA)	5.0(NA)
Other Latino	2.0 (NA)	0.0(NA)	2.0(NA)
Asian	4.0 (5.7)	6.5(9.2)	1.5(2.1)
More than one	13.0 (12.7)	6.3(4.3)	8.3 (7.8)
Other	4.0 (NA)	3.0(NA)	0(NA)
Education
Some college	14.2 (17.0)	9.7(8.9)	7.0(6.1)
Associate’s	2.0 (NA)	4.0(NA)	0.0(NA)
Bachelor’s	11.6 (10.2)	7.9(7.8)	6.8(5.8)
Postgraduate	7.9 (7.8)	5.7(5.9)	6.0(5.7)
Mom Education
8^th grade or less	13.5 (12.4)	7.3(6.4)	8.0(7.8)
Some HS	10.3(8.0)	6.3(5.6)	6.5(7.6)
HS graduate	12.0(12.6)	7.1(7.9)	6.9(6.2)
Some college	10.3(6.5)	7.4(6.9)	6.1(5.4)
Associate’s	5.8(4.3)	2.7(4.4)	7.4(6.3)
Bachelor’s	8.1(7.0)	7.0(7.8)	5.5(4.4)
Postgraduate	10.0(8.6)	8.6(5.6)	5.7(6.7)

Bivariate associations

To evaluate the direct effect of the number of mTBI exposures on neurobehavioral symptoms, we conducted a linear regression while covarying for age, race, education, and maternal education. When controlling for these demographic factors, the association between the number of mTBI exposures and neurobehavioral symptoms was not significant (B = 0.21, SE = 0.18, p = 0.23).

When the same linear regression was run to determine if number of mTBI + LOC events predicted neurobehavioral symptoms while covarying for age, race, education, and maternal education, and number of mTBI exposures, there was a significant effect of mTBI + LOC (B = 2.27, SE = 0.64, p = <.001).

We then conducted multiple linear regressions to explore the proposed structural equation model. As hypothesized, there were significant direct effects of number of mTBI + LOC events on depression (B = 1.26, SE = 0.44, p = 0.005), and insomnia (B = 1.34, SE = 0.35, p= <.001) when controlling for age, race, education, and maternal education. Additionally, both depression (B = 0.91, SE = 0.09, p =<.001) and insomnia (B = 1.12, SE = 0.12, p = <.001) significantly predicted neurobehavioral symptoms.

Indirect effects

Results of exploratory structural equation analysis suggest mTBI + LOC predicted neurobehavioral symptoms indirectly through both depression (B  = 0.85, 95% CI = [0.27, 1.52) and insomnia (B = 0.81, 95% CI = [0.3, 1.4]). Additionally, the direct effect of mTBI + LOC group on neurobehavioral symptoms became non-significant (B = 0.78, SE = 0.45, p = 0.08) when depression and insomnia were added to the model (Figure 1). Significant indirect effects suggest that insomnia and depression should be further explored as potential mediators in the relationship between mTBI + LOC and neurobehavioral symptoms, using longitudinal data.

Figure 1. mTBI+LOC significantly predicts affective symptoms indirectly through depression (B  = 0.39, 95% CI = [0.15, 0.65]) and insomnia (B = 0.44, 95% CI = [0.17, 0.71]).

Overlapping content between the BDI-II, ISI, and NSI questionnaires could contribute to associations between the measures. Specifically, the BDI-II and NSI both include items about indecisiveness, fatigue, sadness, and irritability. The ISI and NSI both include an item about difficulty falling asleep. See Table 4 for correlations between measures. As a result, we ran further exploratory analyses using the three subscales of the NSI (affective, somatic/sensory, and cognitive) to examine more granular indirect effects.

Table 4. Correlations between symptom measures

Correlations
		1	2	3	4	5	6
1	NSI Total Score	–
2	NSI Affective Subscale	.91**	–
3	NSI Somatic/Sensory Subsclae	.83**	.60**	–
4	NSI Cognitive Subsclae	.85**	.71**	.55**	–
5	BDI-II Total Score	.65**	.68**	.33**	.68**	–
6	ISI Total Score	.63**	.69**	.42**	.48**	.47**	–

**Correlation is significant at the 0.01 level (2-tailed).

Affective subscale

Results of exploratory mediation analysis revealed mTBI + LOC predicted affective neurobehavioral symptoms indirectly through both depression (B  = 0.39, 95% CI = [0.15, 0.65]) and insomnia (B = 0.44, 95% CI = [0.17, 0.71]). The direct effect of mTBI + LOC on affective symptoms became non-significant when depression and insomnia were added to the model (B = 0.24, SE = 0.17, p = 0.15) (Figure 2).

Figure 2. mTBI+LOC significantly predicts somatic/sensory symptoms indirectly through insomnia (B  = 0.25, 95% CI = [0.06, 0.49]), but not depression (B = 0.12, 95% CI = [−0.04, 0.34]).

Somatic/Sensory subscale

Results of exploratory mediation analysis revealed mTBI + LOC predicted somatic/sensory neurobehavioral effects indirectly through insomnia (B  = 0.25, 95% CI = [0.06, 0.49]), but not depression (B = 0.12, 95% CI = [−0.04, 0.34]). In addition, the direct effect of number of mTBI + LOC events on somatic/sensory symptoms is non-significant (B = 0.43, SE = 0.25, p = 0.8) when insomnia was added to the model (Figure 3).

Figure 3. mTBI+LOC significantly predicts cognitive symptoms, indirectly through depression (B  = 0.33, 95% CI = [0.09, 0.65]) and insomnia (B = 0.13, 95% CI = [0.04, 0.25]).

Cognitive subscale

Results of exploratory mediation analysis revealed mTBI + LOC predicted cognitive neurobehavioral symptoms indirectly through both depression (B  = 0.33, 95% CI = [0.09, 0.65]) and insomnia (B = 0.13, 95% CI = [0.04, 0.25]). In addition, the direct effect of mTBI + LOC on cognitive symptoms is non-significant (B = 0.11, SE = 0.13, p = 0.38) when depression and insomnia are added into the model (Figure 4).

Figure 4. mTBI+LOC significantly predicts cognitive symptoms, indirectly through depression (B  = 0.33, 95% CI = [0.09, 0.65]) and insomnia (B = 0.13, 95% CI = [0.04, 0.25]).

Discussion

In a sample of former collegiate and professional football athletes, there was a direct effect of mTBI + LOC – but not mTBI – exposures on current neurobehavioral, depression, and insomnia symptoms. Further analyses revealed mTBI + LOC predicted neurobehavioral symptoms indirectly through both insomnia and depression. Additionally, the direct effect of mTBI + LOC on neurobehavioral symptoms became non-significant when depression and insomnia were added into the model. Indirect effects suggest that insomnia and depression may mediate the relationship between mTBI + LOC and neurobehavioral symptoms, however, we are limited by the cross-sectional nature of our data. In clinical science, mediation analysis on cross-sectional data is limited because causal inferences are not possible (69). Due to the range of time between mTBI exposure (M = 29.07, SD = 13.91) and mTBI + LOC exposure (M = 32.85, SD = 15.19) and the timeframe for reporting of depression, insomnia, and neurobehavioral symptoms, we can conclude that mTBI exposure precedes the symptoms reported in this study, providing strong justification for further mediation analyses. Future studies should examine longitudinal effects to provide more insight on the causality of these associations.

In exploratory analyses of subscale measures from the NSI, a number of mTBI + LOC events predicted cognitive and affective neurobehavioral symptoms indirectly through both depression and insomnia. In both models, the direct effect became non-significant when depression and insomnia were included in the model, suggesting a potential full mediation. In addition, a number of mTBI + LOC events predicted somatic/sensory neurobehavioral symptoms indirectly through insomnia alone. It is important to note that these findings are specific to mTBIs that include loss of consciousness, as the number of mTBIs alone did not predict neurobehavioral symptoms.

Prior research reveals an association between mTBI + LOC and increased sleep disturbance, as well as mediating effects of mTBI-related sleep disturbance on depression symptoms (46). Our findings expand on the unique effect of mTBI + LOC on depression and insomnia symptoms, and their potential mediating influence on neurobehavioral symptoms. The neurocognitive effects of both insomnia and depression are often acute; neurocognitive effects of disrupted sleep may be reversed with as little as one eight-hour sleep period (70), while cognitive functioning typically improves following treatment for depression (71). This underscores the importance of frequent screening for depression and insomnia in patients who experience LOC at the time of mTBI, as they may be a useful target of intervention to alleviate neurobehavioral symptoms.

Understanding the neurophysiological mechanisms associated with these findings will provide further insight toward prevention of persistent post-mTBI symptoms. Sleep is regulated by complex neurophysiological mechanisms, with various neurohormones, neurotransmitters, and brain regions facilitating the sleep-wake cycle (57). This cycle is crucial for maintaining metabolic homeostasis, involving the glymphatic system which removes neurotoxic waste from intercellular space (72). It is possible that mTBI + LOC damages neural pathways crucial to this system, disrupting this maintenance cycle. Elevated biomarkers of neurodegeneration (neurofilament light) have been identified among sleep-disrupted mTBI patients in comparison to non-sleep disrupted mTBI patients though the direction of this relationship remains unclear (73).

Although the neurobiological underpinnings of depressive symptoms are not fully understood, glymphatic dysfunction from disrupted sleep may also influence depressive symptoms (74). Some studies suggest that depression is associated with increased inflammation, which is particularly relevant to the distinct inflammatory response associated with mTBI exposure (75,76). Our present study suggests a need for further investigation toward the neurophysiological effects that mTBI + LOC has on mechanisms for sleep regulation and depression.

We also found that insomnia and depression may fully mediate the relationship between mTBI + LOC and subjective cognitive post-concussive symptoms. This expands on prior findings that mTBI + LOC, especially in patients exposed to repetitive head impact, corresponds to worse prolonged cognitive function and increased depression severity (77). It is also imperative for future work to include more objective measures of cognition, although we excluded participants with possibly exaggerated symptoms based on established cutoffs (61), additional study is needed to determine whether this relationship is specific to subjective cognitive deficits or also translates to objective neuropsychological test performance.

Limitations

This study has several limitations. An item from the ISI was excluded, potentially compromising replicability. Additionally, while the ISI has been validated in clinical settings (78), polysomnography may be more reliable for evaluating sleep-related disorders (79).

We are further limited by the cross-sectional nature of our data and thus, claims regarding causality are not possible. These results are part of an ongoing five year, longitudinal effort to evaluate the effects of high-level football participation on brain health, while the present results are cross-sectional, they provide strong justification for investigating the mediational effects of depression and insomnia on the relationship between mTBI + LOC and long-term neurobehavioral symptoms. Future analyses within this cohort, as well as other large-scale prospective longitudinal studies (e.g. TRACK-TBI) will be valuable for more reliably mapping pathways between mTBI + LOC, depression and insomnia symptoms, and neurobehavioral symptoms (80,81).

While current neurobehavioral, depressive, and sleep-related symptoms were measured at time of participation, previous (pre-mTBI) symptomatology was not assessed, and thereby not controlled for, in the current study. History of neurobehavioral, depressive, and sleep-related symptoms, in conjunction with mTBI and mTBI + LOC events could differentially impact long-term outcomes. Additionally, other previous and current morbidities were not assessed. Evidence suggests that factors such as post-traumatic stress, psychiatric symptoms, and headaches/migraines can impact long-term outcomes following mTBI (82,83). While outside of scope for the current study, future studies may consider administering a battery of medical and psychiatric symptom assessments or accessing medical records in order to control for important pre-mTBI and post-mTBI morbidities.

Medical records were not accessed to confirm mTBI history in participants and the reliability of self-reported mTBI history is controversial. While it is well established that current athletes are motivated to underreport or not disclose mTBI symptoms in order to return to play, it is unclear if retired athletes retrospectively underreport mTBI history One study in retired NFL athletes found moderate reliability of self-reported mTBI history (with and without LOC) across multiple timepoints of instrument administration (84). To increase validity of self-report data, our study provided a definition of ‘concussion’ to all participants, as suggested by prior literature (85,86).

Additionally, non-sports related mTBI was not assessed in the study, and subjects may have had mTBI + LOC events from non-sporting events (e.g. motor vehicle collision) that were not included in the current analyses.

Despite limitations, the current study makes a novel contribution to our understanding of the relationship between mTBI + LOC and long-term neurobehavioral outcomes. Findings reveal that LOC at time of injury has prognostic significance for identifying those at high risk for long-term impairment. Further, this study is among the first to provide empirical support that depression and insomnia may mediate relationship between mTBI + LOC and long term neurobehavioral symptoms. However, further investigating the relationships amongst LOC at mTBI exposure, depression, insomnia, and neurobehavioral symptoms is necessary to inform clinical care for mTBI patients. Longitudinal analysis is needed to determine true statistical mediation in the relationship between mTBI + LOC and long-term outcomes. Future research should utilize psychiatric and medical assessments to control for comorbid symptoms. Examining potential sex differences is also needed, as women are underrepresented in sport-concussion studies but may have greater prevalence of sports concussion and subsequent impact on neurobehavioral symptoms and cognitive function (87).

Conclusions

Overall, these findings provide insight into the factors influencing post-LOC neurobehavioral symptoms. Consistent with prior research, we found that a history of mTBI + LOC is directly associated with increased neurobehavioral symptom endorsement at >11 months post-injury. However, there has been little research investigating the factors that may influence this relationship. In the current study, mTBI + LOC predicted total neurobehavioral symptoms indirectly through depression and insomnia. Notably, the direct effect of mTBI + LOC on total neurobehavioral symptoms became non-significant when depression and insomnia were added into the model. Overall, findings support early intervention and symptom monitoring for depression and insomnia following mTBI + LOC exposure, as these factors may be the key indicators of poor functional outcomes.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/02699052.2024.2347552

Additional information

Funding

This study was funded by National Institute of Aging (Award No. K01AG054762) and the National Institute of Neurological Disorders (Award No. R01NS134043).