Photosensitivity Is Associated with Chronic Pain following Traumatic Brain Injury

Abstract

Individuals with a history of traumatic brain injury (TBI) report increased rates of chronic pain. Photosensitivity is also a common chronic symptom following TBI and is prevalent among other types of chronic pain. The aim of this study was to better understand the relationship between chronic pain, pain-related disability, and photosensitivity in a TBI population. We quantified participants' visual photosensitivity thresholds (VPT) using an Ocular Photosensitivity Analyzer and measured pressure-pain sensitivity using pressure algometry. Participants also completed a battery of self-report measures related to chronic pain, TBI history, and mental health. A total of 395 participants completed testing, with 233 reporting a history of TBI. The TBI group was divided into 120 symptomatic TBI participants (s-TBI), and 113 asymptomatic TBI participants (a-TBI) based on their Neurobehavioral Symptom Inventory (NSI) scores. Participants in the s-TBI group scored significantly higher on self-reported chronic pain measures compared with a-TBI and no-TBI participants, including the Symptom Impact Questionnaire Revised (SIQR; p < 0.001) and the Michigan Body Map (MBM; p < 0.001). Despite differences in chronic pain complaints, groups displayed similar pressure-pain thresholds (p = 0.270). Additionally, s-TBI participants were more sensitive to light (lower VPT, p < 0.001), and VPT was correlated with SIQR scores across all participants (R = -0.452, p < 0.001). These data demonstrate that photosensitivity is associated with self-reported chronic pain and disability in individuals with chronic TBI symptomatology. Photosensitivity could therefore serve as a simple, more highly quantitative marker of high-impact chronic pain after TBI.

Introduction

Individuals with a history of traumatic brain injury (TBI), including mild TBI, are disproportionately impacted by chronic pain. Although most TBI patients report symptoms resolving within a few weeks or months of the injury,^1,2 a subset will go on to have chronic symptoms and long-lasting sequelae, including persistent and widespread pain.^3,4 Chronic pain after TBI can have a significant impact on psychological and physical function, including disability and reduced quality of life.^5-7

The neurobiological basis for the high prevalence of chronic pain in individuals with chronic TBI symptomatology remains to be determined. One clue comes from the fact that many of these individuals report widespread pain in multiple body regions, including lower back, shoulder, neck, and limbs, although post-traumatic headache is also common.^3,8 The observation that chronic pain is not necessarily localized to the site of the initial injury in these individuals would argue that chronic pain after TBI is not due to ongoing tissue damage or activation of peripheral nociceptors, and would instead favor a central mechanism.^9–12 A central mechanism or mechanisms would also be more consistent with a general pattern of multi-sensory hypersensitivity seen in many of these individuals. For example, photosensitivity, or abnormal sensitivity to light, can persist for a number of years after a TBI.^7,13,14

One possible mechanism that could explain the high prevalence of both chronic pain and photosensitivity after TBI is alterations in descending pain-modulating systems. These systems, with projections from the brainstem to the dorsal horn of the spinal cord, can either suppress or amplify dorsal horn processing of sensory inputs. A maladaptive shift in the output from the descending modulatory system to favor sensory amplification is now thought to play a major role in many chronic pain states, including pain states once considered “functional,” or without an established neurobiological basis.^15,16

Studies in humans and in animal models document altered descending control following experimental TBI,^17-19 and a specific population of brainstem pain-facilitating neurons was recently shown to be activated by light exposure.²⁰ Because activation of these pro-nociceptive neurons leads to a lowering of the pain threshold,¹¹ their recruitment by light could cause innocuous somatic stimuli to be felt as painful, making light itself aversive. These lines of evidence suggest that photosensitivity could be used as an indicator of central mechanisms contributing to pain following TBI. However, although both chronic pain and photosensitivity are prevalent after TBI, whether photosensitivity is associated with chronic pain in a given patient has not been determined. The goal of the present study was to quantify photosensitivity in individuals with and without TBI, and test the hypothesis that photosensitivity is associated with chronic pain and pain-related disability.

Methods

The joint Oregon Health & Science University (OHSU) and VA Portland Health Care System (VAPORHCS) institutional review boards approved this study (VA IRB #3988), and all participants provided verbal and written informed consent. Participants were recruited from June 2018 through March 2020 using clinical referral from VAPORHCS outpatient clinics, flyers posted at OHSU, VAPORHCS, and throughout the local community, and with online advertisements. There was no requirement of a history of TBI to be enrolled in the study, however individuals with eye diseases (e.g., glaucoma, macular degeneration) or taking eye medications were excluded. A total of 433 participants were enrolled. Thirty-eight participants did not complete the TBI interview or could not be definitively diagnosed; therefore 395 subjects, with and without a history TBI, were included in the analyses. Data were verified independently by two researchers prior to analyses.

TBI evaluation

All participants underwent the Head Trauma Event Characteristics interview to screen for a history of TBI.²¹ Interviews were conducted in-person or over the phone by trained study personnel and included questions on participants' most severe and most recent head injuries, if they endorsed having had one. Details of the injury were elicited, including endorsement of altered mental status, loss of consciousness, and posttraumatic amnesia. All records were reviewed by a clinical neuropsychologist who made a diagnosis of no, mild, moderate, or severe TBI.

Survey instruments

Participants filled out the following battery of self-report surveys: Neurobehavioral Symptom Inventory (NSI) to assess post-concussive syndrome symptoms,²² Symptom Impact Questionnaire-Revised (SIQR) to assess overall chronic pain and its impact on function and disability over the previous 7 days,²³ Michigan Body Map (MBM) to assess body areas in which participants experienced chronic pain over the prior 3 months,²⁴ Post-Traumatic Stress Disorder (PTSD) Checklist for Diagnostic and Statistical Manual of Mental Disorders-5 (PCL-5) to assess PTSD symptom severity,²⁵ Patient Health Questionnaire 9 (PHQ-9) to assess depression severity,²⁶ Insomnia Severity Index (ISI) to assess insomnia severity, including difficulty initiating and staying asleep,²⁷ and World Health Organization Disability Assessment Schedule 2.0 (WHODAS) to assess activity limitations and disability severity.²⁸ A single question in the NSI pertaining to light sensitivity was not used in total score calculation due to its confound with our primary outcome.

Experimental pressure pain

Pressure-pain threshold and pressure-pain tolerance were determined for each participant using an Algomed Computerized Pressure Algometer (Medoc Ltd. Advanced Medical Systems, Israel) applied to the thumb of the non-dominant hand (except when the individual reported a recent injury near that area, in which case the dominant hand was used). Participants were first tested for pressure-pain threshold. Participants were told to press a button and say “stop” at the point at which they felt any pain from the pressure. This test was repeated three times, with a 2-min inter-trial interval. After each trial, the threshold pressure was recorded and participants were asked to rate the pain of the stimulus on a scale of 0-10, with 0 indicating “no pain” and 10 indicating “the worst pain imaginable.” Pressure-pain tolerance was assessed similarly except that participants were asked to report when “the pressure becomes intolerable.” Because thresholds can decrease with repeated testing,²⁹ the average of the second and third trials was used. The maximum level of pressure used was 1000 kPa; any trial that reached that level before the participant responded was terminated and scored as 1000 kPa. In all instances, pressure algometry testing was done after participants completed self-report surveys.

Photosensitivity

Visual photosensitivity thresholds (VPT) were determined using an Ocular Photosensitivity Analyzer (OPA; Bascom Palmer Eye Institute; Miami, FL). This instrument produces light stimuli ranging from 4 to 32,000 lux using 210 white light-emitting diodes focused 50 cm from the participant's eyes. We followed a previously developed protocol.³⁰ Following dim light adaption (10 min at 4 lux), participants were seated in front of the light panel and asked to stare forward for the entirety of the test. A computer-generated voice read test instructions. Participants were exposed to each light stimulus (2-sec duration, 4-sec inter-stimulus interval) and asked to respond with a button press if the light was uncomfortable. Stimulus intensity was adjusted using a Garcia-Perez staircase, with ascending and descending steps depending on the participant's previous response. VPTs were calculated based on the mean of 10 response reversals. In all instances, photosensitivity testing was done directly after participants completed pressure algometry testing.

Statistical analysis

All analyses were performed using R version 3.3.2.³¹ For all tests, p < 0.05 was considered statistically significant. Differences in numerical variables between groups were assessed using t-test or one-way analysis of variance with Tukey's honestly significant difference post hoc analysis, or a Kruskal-Wallis Test with Dunn's multiple comparison test, depending on the distribution of data. Analysis of covariance (ANCOVA) was used to test for group differences while controlling for sex for some outcomes. Categorical data were analyzed using a χ² test with Bonferroni correction for multiple comparisons. Pearson's correlation coefficients of different groups were compared based on their confidence intervals using the open-source R software package cocor.^32,33 A multi-variate linear regression was employed to analyze potential demographic variables and testing measures related to chronic pain.

Results

NSI scores and demographics

We first determined whether TBI severity was associated with any significant differences in chronic TBI symptomatology, measured by the NSI score. Participants with mild (n = 186, p < 0.001), moderate (n = 34, p < 0.001), and severe (n = 11, p < 0.001) all scored significantly higher on the NSI than participants without TBI (no-TBI group, n = 162). However, there was a substantial overlap across these groups, and NSI did not differ among the three TBI severity groups (p = 0.242; Fig. 1A). Given this, and the low frequency of moderate and severe TBIs in our sample, all participants with any history of TBI were combined into a single “TBI” group.

FIG. 1.

Distribution of Neurobehavioral Symptom Inventory (NSI) scores in traumatic brain injury (TBI) participants. (A) Raincloud plots illustrating the distribution on NSI scores across TBI severity (n = 231, two participants were discarded for missing data). Whisker plots mark the lowest/highest observations, upper/lower quartiles, and median scores within each group, while raincloud plot displays probability density. All three TBI groups scored significantly greater than the no-TBI group (p < 0.001, n = 162) but did not differ from one another. (B) Frequency histogram depicting the range of NSI scores in participants with confirmed TBI. Dashed lines mark lower quartile, median, and upper quartile scores. Due to the skewness of the distribution, TBI participants were categorized as symptomatic or asymptotic using a median split. NSI, Neurobehavioral Symptom Inventory, range 0-80.

The distribution of NSI scores in the TBI group revealed that many of these individuals reported few or no post-concussive symptoms (Fig. 1B). For some analyses, we therefore subdivided TBI participants into asymptomatic (a-TBI, NSI score <23) and symptomatic (s-TBI, ≥23) based on a median split. These scores are consistent with previous studies in this population.³⁴ The a-TBI and s-TBI groups did not differ in demographic variables, including gender, age, race, or veteran status (Table 1). There was also no difference in TBI severity between the a-TBI and s-TBI groups (χ² test with Bonferroni correction, p > 0.167), nor was there a difference in TBI recency between these groups (p = 0.382). There was a significantly greater proportion of female participants in the no-TBI group compared with the a-TBI and s-TBI groups. However, NSI scores in females were almost identical to those in males (males, mean [M] = 21.4, standard deviation [SD] = 16; females, M = 20.9, SD = 18.1; p = 0.710). We nonetheless controlled for sex when examining group differences on key analyses by performing an ANCOVA with sex as a covariate.

Table 1.

Demographic Information and Self-Reported TBI Sequelae Scores across Groups

	no-TBI	a-TBI	s-TBI	All TBI	Statistic	p value
Demographic variables
Total participants, n	162	113	120	233
Gender					14.69	< 0.001
Male, n	83 (51.2%)	80 (70.8%)†	86 (71.7%)†	166 (71.2%)†
Female, n	74 (45.7%)	32 (28.3%)†	33 (27.5%)†	65 (39.2%)†
Non-binary, n	3 (1.9%)	0 (0.0%)	1 (0.8%)	1 (0.0%)
Declined to answer, n	2 (1.2%)	1 (0.9%)	0 (0.0%)	1 (0.0%)
Race					12.460	0.255
White/Caucasian, n	131 (80.7%)	86 (76.2%)	88 (73.4%)	174 (74.67%)
Black/African American, n	6 (3.78%)	5 (4.4%)	6 (5.0%)	11 (4.7%)
Asian, n	10 (6.12%)	5 (4.4%)	3 (2.5%)	8 (3.4%)
Native American or Alaskan Native, n	2 (1.2%)	1 (0.9%)	6 (5.0%)	7 (3.0%)
Mixed, n	7 (4.3%)	11 (9.7%)	12 (10.0%)	23 (10.1%)
Other or missing, n	6 (3.7%)	5 (4.4%)	5 (4.2%)	10 (4.3%)
Veteran, n	84 (51.9%)	74 (65.5%)	99 (82.5%)	173 (74.2%)	4.383	0.112
Age, years ± SD	51.3 ± 16.5	51.4 ± 14.5	52.7 ± 14.4	52.1 ± 14.4	0.313	0.731
TBI severity					6.69	0.035
Mild, n	NA	96 (85.0%)	90 (75.0%)	186 (79.8%)
Moderate, n	NA	15 (13.3%)	19 (15.8%)	34 (14.6%)
Severe, n	NA	2 (1.8%)	11 (9.2%)	13 (5.6%)
TBI recency, years ± SD	NA	23.1 ± 19.0	19.8 ± 17.8	21.4 ± 18.0	1.369	0.172
Self-report measures
ISI	9.11 ± 6.60	7.98 ± 5.22	15.30 ± 5.41 *	11.97. ± 6.44*	60.06	< 0.001
PHQ-9	5.63 ± 2.87	3.65 ± 2.87 †	12.64 ± 5.50*	8.26 ± 6.32*	104.95	< 0.001
WHODAS	8.38 ± 9.15	6.23 ± 5.81	17.87 ± 7.81*	12.20 ± 8.96*	74.78	< 0.001
PCL-5	15.33 ± 17.81	12.32 ± 10.42	40.05 ± 16.92*	26.61 ± 19.80*	114.56	< 0.001

^†p < 0.001 vs. no-TBI; ^*p < 0.001 vs. no-TBI and a-TBI.

Bolded data indicates variables that were significantly different between groups.

Data presented as n (% total) or mean ± standard deviation (SD). Participants endorsing more than one race were classified as “Mixed.” Differences in categorical variables (gender, race, and veteran status) were tested between no-TBI and TBI groups, as well as across the no-TBI, a-TBI, and s-TBI groups, using a χ² test with a Bonferroni-corrected post hoc test. Differences in continuous variables (age, ISI, PHQ-9, WHODAS, PCL-5) were tested between no-TBI and TBI groups, as well as across the no-TBI, a-TBI, and s-TBI groups, using t-tests and one-way analyses of variance with a Tukey post hoc contrast. Differences in TBI severity and TBI recency were only tested between a-TBI and s-TBI groups using a χ² test and t-test.

TBI, traumatic brain injury; a-TBI, asymptomatic traumatic brain injury; s-TBI, symptomatic traumatic brain injury; SD, standard deviation; ISI, Insomnia Severity Index; PHQ-9, Patient Health Questionnaire 9; WHODAS, World Health Organization Disability Assessment Schedule; PCL-5, Post-Traumatic Stress Disordr (PTSD) Checklist for Diagnostic and Statistical Manual of Mental Disorders 5.

Other chronic TBI sequelae include sleep disturbances, mood disorders, and increased disability.^7,35,36 We therefore examined group differences in these domains using self-report surveys. The TBI group scored significantly higher on the ISI (p < 0.001), PHQ-9 (p < 0.001), WHODAS (p < 0.001), and PCL-5 (p < 0.001) compared with the no-TBI group. However, following subgroup analyses, we found the s-TBI group was driving these differences as they scored significantly higher than both the a-TBI and no-TBI groups on all measures (p < 0.001), indicating greater levels of insomnia, depression, disability, and PTSD symptom severity. With the exception of PHQ-9 scores (p = 0.005), the a-TBI and no-TBI groups did not significantly differ on any of these measures (ISI, p = 0.327; WHODAS, p = 0.068; PCL-5, p = 0.266). Thus, the s-TBI group reported significantly more behavioral, functional, and mood-related symptoms than participants classified as a-TBI or no-TBI.

Group differences in self-reported chronic pain

SIQR and MBM scores were used to assess chronic pain status. TBI participants had significantly higher SIQR and MBM scores compared with the no-TBI group, even while controlling for sex (Fig. 2A, 2B), demonstrating more severe and more widespread pain in participants with TBI which is consistent with previous studies.³ Subgroup analyses using a one-way ANCOVA showed that this difference was driven primarily by the s-TBI group, who scored significantly higher on both measures compared with a-TBI or no-TBI participants (Fig. 2C, 2D).

FIG. 2.

Symptomatic traumatic brain injury (TBI) group reports more intense and more widespread chronic pain. (A) Participants in the TBI group reported significantly higher Symptom Impact Questionnaire Revised (SIQR) scores than the no-TBI group (no-TBI n = 162, TBI n = 233; *** p < 0.001). (B) The TBI group also endorsed significantly more body areas affected by chronic pain (***p < 0.001). When broken down into symptomatic (s-TBI) and asymptomatic (a-TBI) groups, we found the s-TBI group scored significantly higher on the SIQR (C) and the MBM (D) than both the a-TBI and no-TBI groups. The no-TBI and a-TBI groups did not significantly differ on either measure (SIQR, p = 0.061; MBM, p = 0.985). Due to demographic differences across groups, sex was used as a covariate in these analyses. Data are presented as mean ± standard error.

No differences in pressure-pain threshold and tolerance

Despite differences in self-reported chronic pain and pain impact, there were no differences in acute pressure-pain thresholds or tolerance in individuals with and without TBI (Fig. 3A, 3B). We again employed a series of one-way ANCOVAs to test for differences between the no-TBI, a-TBI and s-TBI subgroups, which again showed no significant differences (Fig. 3C, 3D). In addition, there were no group differences in average pain ratings during the pressure threshold trials (p = 0.272) or in pressure tolerance trials (p = 0.888).

FIG. 3.

No differences in pressure pain levels between traumatic brain injury (TBI) groups Despite the TBI group reporting significantly higher levels of chronic pain, we found no differences between groups on either experimental pressure-pain thresholds (A; p = 0.224) or pressure-pain tolerance levels (B; p = 0.300). This was also true when we analyzed a-TBI and s-TBI groups separately on both measures (C; p = 0.503; D, p = 0.490). Due to demographic differences across groups, sex was used as a covariate in these analyses. Data are presented as mean ± standard error.

Group differences in photosensitivity

Participants with TBI exhibited significantly lower VPTs compared with the no-TBI group (Fig. 4A). As with chronic pain measures, this difference was driven by the s-TBI participants, as a-TBI participants were not different from the no-TBI group (Fig. 4B).

FIG. 4.

Symptomatic traumatic brain injury (s-TBI) participants have lower light-evoked discomfort levels. (A) TBI participants had significantly lower visual photosensitivity thresholds compared with the no-TBI group (***p < 0.001). (B) This difference was almost entirely due to the s-TBI participants, which were significantly lower compared with both the no-TBI group and the aymptomatic TBI (a-TBI) group (***p < 0.001). The a-TBI participants were not significantly different from the no-TBI participants (p = 0.670). Due to demographic differences across groups, sex was used as a covariate in these analyses. Data are presented as mean ± standard error.

Photosensitivity is associated with chronic pain

No-TBI, a-TBI, and s-TBI groups each showed a significant negative correlation between VPT and SIQR scores (no-TBI: R = −0.379, p < 0.001; a-TBI: R = −0.258; p = 0.007; s-TBI: R = −0.306, p < 0.001; Fig. 5). Group correlations did not significantly differ from one another (p = 0.843). Regardless of TBI status, there was a significant negative correlation between VPT and SIQR scores (R = -0.452, p < 0.001; Fig. 5) as well as between VPT and MBM scores (R = −0.27, p < 0.001; not shown) across all participants. Photosensitivity was thus strongly associated with widespread chronic pain, independent of TBI status.

FIG. 5.

Strong correlation between photosensitivity and chronic pain complaints. (A) Visual photosensitivity threshold (VPT) levels were strongly correlated with Symptom Impact Questionnaire Revised (SIQR) scores across all participants, regardless of traumatic brain injury (TBI) status (black; R = -0.447, p < 0.001). (B) When broken down by group, we found that no-TBI participants (green; R = −0.379, p < 0.001), s-TBI participants (purple; R = −0.306, p < 0.001) and a-TBI participants (orange; a-TBI: R = −0.258; p = 0.007) all displayed strong negative correlations between VPT and SIQR scores, and these did not differ statistically from one another.

Multiple linear regression reveals VPT, PTSD, and age are predictors of chronic pain scores

A multi-variate linear regression model was next used to explore the relationship between chronic pain and relevant variables, including TBI characteristics (severity, recency, and number of TBIs) and other variables that could relate to TBI exposure, chronic pain, and/or light sensitivity (age, sex, eye color, and PTSD status). Pressure-pain tolerance and VPT were included to determine whether either of these sensory tests was independently associated with chronic pain. This model accounted for more than 44% of the variance of SIQR scores with age, VPT, and PTSD diagnosis as significant predictors (Table 2).

Table 2.

Multi-Variate Regression Analysis Predicting SIQR Scores

Outcome Variable	Predictor variables	B	β	T value	p value	Adjusted R2
SIQR score					< 0.001	0.437
	Age	0.251	0.175	2.810	0.006
	Sex (Female)	1.122	0.024	0.427	0.670
	Eye color	-0.180	-0.105	-0.199	0.842
	Pressure tolerance level (kPa)	0.001	0.002	0.31	0.976
	VPT (lux log10)	-5.976	-0.279	-4.622	< 0.001
	PTSD diagnosis	22.204	0.489	8.451	< 0.001
	TBI recency	-0.137	-0.121	-1.920	0.056
	Number of TBIs	-0.311	-0.217	-0.685	0.494
	TBI severity
	Mild	-3.309	-0.625	-0.685	0.494
	Moderate	-1.027	-0.169	-0.188	0.852
	Severe	3.309	0.366	0.685	0.494

Bolded variables were significant predictors of SIQR scores.

Multi-variate linear regression predicting SIQR scores. TBI diagnosis and gender were included as categorical variables while age, eye color, pressure tolerance, VPT, TBI recency, and number of TBIs were included as continuous measures. A significant regression was found [F(10, 214) = 17.31, p < 0.001] with age, VPT, and PTSD being the only significant predictors.

SIQR, Symptom Impact Questionnaire-Revised; VPT, visual photosensitivity thresholds; PTSD, post-traumatic stress disorder; TBI, traumatic brain injury.

Discussion

TBI is associated with high prevalence of both chronic pain and complaints of sensitivity to light.^3,4,37 The present study quantified visual photosensitivity in individuals with and without a history of TBI to directly examine the relationship between chronic pain complaints and photosensitivity, and to analyze the relationship between photosensitivity and other chronic TBI sequelae. The primary finding was that individuals with a history of TBI, the majority of whom experienced mild TBI, exhibit a significantly lower threshold for light-evoked discomfort, even many years after the injury. Individuals who had sustained at least one TBI also reported higher levels of chronic pain compared with those with no history of TBI. Finally, there was a significant correlation between photosensitivity and chronic pain.

Photosensitivity associated with “high-impact” chronic pain

The increased photosensitivity and pain reported in our TBI cohort were driven in large part by the subset of the TBI group that endorsed chronic symptomatology. Thus, photosensitivity was greatest in the subset of the TBI participants endorsing an array of chronic symptoms, as measured by the NSI (“symptomatic” TBI group). The symptomatic TBI group also scored high on the SIQR, a measure of chronic pain impact. In addition, these individuals also reported greater levels of depression, PTSD, sleep disturbances, and a low quality of life due to disability (schematization in Fig. 6). These symptoms are inter-related, mutually reinforcing, and respond poorly to conventional treatments. As a whole, they represent what is now referred to as “high-impact chronic pain.”^38,39

FIG. 6.

Radar plot depicting self-report and behavioral scores between groups. Points farther out from the center represent higher severity and high (worse) scores. The symptomatic TBI (s-TBI) group has the highest severity in all measures, except for pressure-pain threshold and tolerance levels. VPT, visual photosensitivity threshold; SIQR Symptom Impact Questionnaire Revised scores (chronic pain); SI, Insomnia Severity Index scores (sleep disurbances); WHODAS, World Health Organization Disability Assessment Schedule 2.0 scores (disability); PCL-5 Post-Traumatic Stress Disorder (PTSD) Checklist for Diagnostic and Statistical Manual of Mental Disorders-5 scores (PSTD symptoms); PHQ-9, Patient Health Questionnaire 9 scores (depression).

By contrast, individuals who had never sustained a TBI and the “asymptomatic” TBI group reported less chronic pain, depression, and overall disability, and fewer sleep disturbances. Our data suggest that photosensitivity, a simple and more highly quantitative measure collected without reference to pain and without the potentially emotionally charged questions found in self-report surveys of mental/emotional health and disability, could serve as a marker for the population experiencing high-impact chronic pain, a group that should be a scientific and clinical focus. Indeed, although photosensitivity is commonly reported following TBI, visual photosensitivity thresholds were strongly correlated with self-reported chronic pain regardless of TBI status. By contrast, experimental pressure pain was not associated with chronic pain measures. Even though some chronic pain disorders such as fibromyalgia feature increased pressure hypersensitivity compared with healthy controls, experimental pain is not necessarily correlated with clinical complaints.⁴⁰

One striking finding was the increased report of PTSD symptoms in the symptomatic TBI group, which along with photosensitivity, was a significant predictor of SIQR scores in our multiple linear regression model. We were unable to directly analyze the association between PTSD scores and photosensitivity/pain, due to our small sample of participants with PTSD but without TBI. Nevertheless, photosensitivity pointed to a symptom complex (Fig. 6) that closely aligns with the idea of a “clinical polytrauma triad” comprising pain, PTSD, and TBI, as a better predictor of impaired function than TBI alone.^41–44 In any case, since both TBI and PTSD were both independently linked to chronic pain and sensory sensitivity, and are highly prevalent in the Veteran population, future studies should examine PTSD-only and comorbid TBI+PTSD patients on these measures.

It is important to note that the present study did not directly analyze other sensory modalities, but sensitivity to other non-somatic stimuli is also likely to be associated with chronic pain. Multiple studies have found increased sensitivity to sound after TBI or more generally with nociplastic pain states such as fibromyalgia.^45-47 Although continued investigation of the relationships between pain and other sensory modalities is important for understanding chronic pain, and a future direction for research, more highly quantitative photosensitivity testing is likely to prove simple and relatively reliable.

Potential mechanisms linking photosensitivity and chronic pain after TBI

Although the neurobiological etiology of TBI-related chronic pain is still poorly understood, it is likely that the increased prevalence seen here is related to changes in the central nervous system, and not directly linked to ongoing peripheral tissue damage or the initial injury.^9–12 This follows from the observation that many individuals experienced pain at multiple sites and at sites remote from the original injury.⁴ Additional direct evidence for central dysfunction as an important factor in chronic pain after TBI comes from demonstrations of altered activity and reduced connectivity among pain-related brain regions such as the thalamus, pons, prefrontal cortex, and anterior cingulate,⁴⁸ and from work in humans and animal models documenting changes in pain-modulating circuits following TBI.^17-19 Given the potential for abnormal engagement of brainstem pain-modulating neurons by light,²⁰ the present correlation between chronic pain and photosensitivity also points to a possible role for alterations in pain-modulating circuits.

Biopsychosocial elements associated with TBI are also likely to play a significant role. For example, an increase in the prevalence of insomnia and sleep disturbances following TBI, confirmed in the present study, has been associated with increased pain, as have depression and anxiety.^49,50 Biological and psychosocial changes associated with TBI are likely mutually reinforcing, and together contribute to the observed high prevalence of chronic pain after TBI. Mechanistic follow-up studies will be needed to better examine the changes in pain circuity related to increased photosensitivity.

Last, autonomic dysfunction following TBI can affect the pupillary light reflex,⁵¹ which could affect light sensitivity. If pupillary constriction is slowed in this population, it might explain the lowered VPT in our s-TBI group. However, the fact that VPT and SIQR scores were strongly correlated in all subjects, including our no-TBI group, suggests the recruitment of pain processing circuitry and altered descending control. There is some evidence of altered autonomic dysregulation in certain chronic pain disorders⁵²; therefore, further examination of the pupillary light reflex is worth consideration to elucidate the role of autonomic dysfunction and photosensitivity in these populations.

Limitations

One limitation of the current study is the reliance on participants' subjective responses to photosensitivity testing. VPT is not an entirely objective neurophysiological marker and is still dependent on the participant's subjective response to stimuli; however, the use of standard stimuli and stimulation protocols as applied here allowed us to better quantify photosensitivity and improve on self-report measures. The current study also lacks longitudinal data and due to the cross-sectional design, and we cannot directly attribute causality to the relationship between increased photosensitivity or chronic pain impact and TBI. Moreover, while the sample tested was large, it may not be representative of the larger population, with a large subset of participating Veterans receiving care at a single VA system. This resulted in a predominantly male sample and may have contributed to the high prevalence of PTSD in our s-TBI group. PTSD was also a significant predictor of SIQR scores in our multiple linear regression model, but due to the lack of participants who had PTSD without TBI, we could not directly analyze the association of PTSD with pain and photosensitivity on its own. Since TBI and PTSD are both independently linked to chronic pain and sensory sensitivity, future studies should examine PTSD-only and comorbid TBI+PTSD populations on these measures.

Significance

These data demonstrate that photosensitivity is associated with self-reported chronic pain in individuals with chronic TBI symptomatology. This marker is more highly quantitative than most subjective pain measures, non-invasive, easily measured, and does not refer specifically to pain as part of the test. It also avoids some of the emotionally charged elements inherent in self-report survey instruments. It has high clinical tractability and has the potential to guide future treatment and help utilize targeted therapies for chronic pain patients. For example, treatments that would be less effective in situations where central plasticity contributes to the pain state could be avoided, such as surgery directed at a site of perceived pain. Perhaps most importantly, the simple test utilized in this work with TBI could be extended to other populations with chronic pain.

Authors' Contributions

NMB, SDM, KDJ, MLC, MPB, MML, and MMH participated in study concept and design. NMB, MPB, MML, and MMH contributed to the drafting of manuscript. NMB, AAM., MLC participated in data acquisition. NMB performed the statistical analyses. All authors participated in data interpretation and critical revision of the manuscript.

Funding Information

This work was supported by the Office of the Assistant Secretary of Defense for Health Affairs under Award #W81XWH-17-1-0423 to MMH. MML was supported by VA Merit Award #I01 CX002022 and the Portland VA Research Foundation. NMB was supported by NIH TL1TR002371, NIH T32AT002688, and a VA ORD Research Supplement to Promote Diversity (to #I01 CX002022).

Authors' Disclosure Statement

No competing financial interests exist.