Neuropsychological influences on visual photosensitivity and its treatment in traumatic brain injury

Abstract

Background

Visual photosensitivity is common following traumatic brain injury (TBI) and may be influenced by neuropsychological factors. We examined relationships between visual photosensitivity thresholds (VPTs) and these factors, evaluating tinted lens responses in persons with and without TBI.

Methods

Participants completed the Patient Health Questionnaire-9 (PHQ-9), 13-Item Pain Catastrophizing Scale-English Version (PCS-EN), Brief Test of Adult Cognition by Telephone (BTACT) and Short-Form 12-Item Survey Version 2. Lens response was calculated as the VPT difference from plano lens (PL) and categorised as mild, moderate or strong. Associations were analysed using multivariable linear and ordinal regression models adjusted for demographics.

Results

VPTs were lower in the TBI group (PL VPT: 1.5±0.9 vs 2.4±0.8 log lux; p<0.001). Among individuals with TBI, higher PCS-EN scores correlated with lower PL VPTs (β=−0.027; p=0.01). Higher PHQ-9 scores correlated with lower VPTs under FL-41 lens (FL) (β=−0.048; p=0.02) and grey-filtering lens (GL) (β=−0.051; p=0.02) conditions. Lower BTACT number series Z-scores correlated with lower VPTs under FL (β=0.349; p=0.01) and GL (β=0.376; p=0.009), suggesting greater visual photosensitivity with lower cognition. Ordinal regression indicated that higher PHQ-9 scores correlated with reduced odds of a stronger response to FL (OR=0.833; p=0.02) and GL (OR=0.829; p=0.047) lenses.

Conclusion

Pain catastrophising was linked to greater photosensitivity, while depression reduced lens benefit in TBI. These findings support the need for targeted interventions addressing neuropsychological factors in optimising therapeutic benefit.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Photophobia is a common and persistent symptom following traumatic brain injury (TBI), and tinted lenses such as FL-41 have shown promise in mitigating it. However, the role of neuropsychological factors in modulating visual photosensitivity and treatment response remains poorly understood.

WHAT THIS STUDY ADDS

• This study demonstrates that depression, pain catastrophising and decreased cognition are significantly associated with increased visual photosensitivity in individuals with TBI. Furthermore, higher depressive symptoms and cognitive deficits predict reduced benefit from light-filtering lenses, including FL-41 and grey-filtering filters.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings suggest that psychological and cognitive screening may be essential for optimising photophobia treatment in TBI populations. Personalising light-filtering interventions based on neuropsychological profiles may improve outcomes and guide multidisciplinary management strategies.

Introduction

Traumatic brain injury (TBI) is defined as any injury to the head, including a bump, blow or jolt, that affects the normal functioning of the brain.¹ Photophobia is a frequently reported post-TBI symptom, defined as increased sensitivity to light, experienced as light-induced pain or discomfort.² While most common in the subacute phase (1 week postinjury: 30.46%, 95% CI 20.05% to 40.88%), symptoms persist >1 year postinjury (14.85%, 95% CI 6.80% to 22.90%, n=27 942).³

Chronic headache is another frequent sequela of TBI, especially after mild TBI, occurring in approximately 40–60% of individuals 6–12 months postinjury.⁴ In a study of 30 individuals with post-traumatic headache and 35 controls without TBI or headache history, psychophysical assessments of light sensitivity were conducted by increasing light intensity up to patient-reported discomfort. Cases had lower light sensitivity thresholds (indicating higher light sensitivity) than controls (1.5 log lux vs 2.6 log lux, p=0.0006), with 61% of cases and 20% of controls classified as photophobic (defined as a light sensitivity threshold ≤1.8 log lux).⁵

Beyond pain, other TBI associated comorbidities include mental health disorders such as depression, anxiety and post-traumatic stress disorder (PTSD).^{6 7} A retrospective matched cohort study of Medicare Advantage enrollees reported increased incidence of depression among TBI enrollees (n=207 354) compared with those without a TBI history (n=414 708) (TBI: 79.5%, 95% CI 78.5% to 80.5% vs non-TBI: 33.5%, 95% CI 33.1% to 34.0% per 1000 person-years).⁶ Similarly, in the Transforming Research and Clinical Knowledge in TBI study, which enrolled patients within 24 hours of injury at 11 US trauma centres, the weight-adjusted prevalence of PTSD and/or major depression was higher among patients with mild TBI (n=1155, 20%) compared with orthopaedic trauma patients (n=230, 8.7%) at 3 months postinjury (p<0.001).⁷ Despite their relationship with TBI, the impact of depression and anxiety on light sensitivity remains largely unexamined.

Treatments for photophobia include optical tints which filter light at certain wavelengths and have been shown to reduce visual photosensitivity among a variety of conditions, including migraine, benign essential blepharospasm and TBI.8,12 FL-41 (a rose coloured lens) has been the tint most commonly used to treat photophobia in the setting of various conditions. In migraine, its effect has been demonstrated in a randomised crossover trial in which 37 participants wore a 480 nm optical notch filter, which closely approximates a FL-41 filter, and a 620 nm filter (orange tint) as a sham control, each for 2 weeks with a washout period in between.¹² Headache impact test (HIT-6) scores improved to a similar degree while wearing both filters (tint vs baseline: −4.30 for 480 nm and −5.01 for 620 nm).¹² In benign essential blepharospasm (n=20), FL-41 lenses (FLs) subjectively improved light sensitivity (scale 0–5 with 0 indicating no help and 5 indicating very significant improvement) in 27%, compared with only 3% reporting improvement with grey lenses (p=0.04).⁹ While FL-41 have not been specifically studied in TBI, tinted lenses have been shown to have therapeutic benefit. In a pilot study of 12 individuals with photophobia post mild TBI, 11 noted subjective improvements with a precision tint lens (lenses of different hues, saturations and luminance which individuals chose based on personal maximal comfort).¹³ Underscoring the complexity of treatment, a study of 39 individuals with post-TBI photophobia found that 85% (n=33) reported subjective improvement in light sensitivity with tinted lenses. However, tint preference varied among individuals, with blue being the most preferred, followed by green, red and purple.¹¹ Whether psychological and cognitive factors influence the effectiveness of photophobia treatment following TBI remains unclear.

To bridge these gaps, the current study examines relationships between visual photosensitivity and neuropsychological factors, specifically depression, pain catastrophising (a symptom frequently comorbid with PTSD) and cognition, in individuals with TBI. We also evaluate how these factors influence the effectiveness of light-filtering lenses in reducing visual photosensitivity. Understanding the comorbidities that influence visual photosensitivity and response to treatment in individuals with TBI may provide valuable insights that can be translated into optimisation of treatments for this often-debilitating condition.

Materials and methods

Eighty-seven adults (TBI, n=42; non-TBI controls, n=45) were enrolled between May 2021 and August 2023 at a tertiary academic centre. Exclusion criteria included ocular disease associated with photophobia or reduced vision (eg, iris defects, iritis or significant ptosis), prisoners, pregnant women, minors and those unable to consent. TBI severity was classified using a modified version of the 2019 DoD Standard Surveillance Case Definition for TBI Adapted for Armed Forces Health Surveillance Division Use.

All participants completed standardised instruments: 9-Item Patient Health Questionnaire (PHQ-9), 13-Item Pain Catastrophizing Scale-English Version (PCS-EN), Short-Form 12-Item Survey Version 2 (SF-12v2) with Physical Component Summary (PCS) and Mental Component Summary (MCS) scores, and Brief Test of Adult Cognition by Telephone (BTACT).14,17 Demographics and medical history were collected via structured questionnaire at the same visit; BTACT was performed at a separate telephone encounter.

The primary outcome was visual photosensitivity threshold (VPT), measured with the Ocular Photosensitivity Analyzer (OPA) under three lens conditions: plano lenses (PLs), FLs and grey-filtering lenses (GLs). The OPA is an automated test that delivered stepwise light stimuli after a 10 min adaptation at 4 lux, and participants pressed a button at discomfort threshold to quantify visual photosensitivity.¹⁸ Secondary outcomes were lens responses, defined as VPT differences from PL VPT log lux (control) and categorised as mild (<2 log lux), moderate (≥2–<3) or strong (≥3). These cutoffs were chosen to reflect clinically meaningful gradations in visual photosensitivity. Exposures and predictors included TBI status/severity, PHQ-9, PCS-EN, BTACT Z-scores and SF-12v2 scores. Potential confounders (age, sex, race, ethnicity) were identified a priori. We also examined TBI status as a potential effect modifier via stratified analyses.

OPA testing was conducted after participants removed their habitual refractive correction (glasses or contact lenses). During testing, participants wore custom-made light-filtering lenses (Optical Express, Bascom Palmer Eye Institute). Both FLs and GLs reduced transmittance by ∼40%, with spectral properties confirmed by spectrometer. The FL lenses filtered short wavelengths in the blue-green spectrum (∼480–520 nm), whereas GL lenses applied a quasi neutral density filter across the visible spectrum. VPTs were calculated as the mean of 10 response reversals.¹⁸

Potential selection bias was minimised by applying uniform eligibility criteria across groups. Measurement bias was reduced by using automated OPA protocols. Study size was determined by recruitment feasibility during the study period. Investigators had full access to all study data, which underwent routine range and consistency checks before analysis; no additional database linkage or data cleaning procedures were required.

Statistical analyses were performed using SPSS V.29 (IBM, Armonk, New York, USA). Descriptive statistics compared demographics and clinical measures. Between-group differences were tested with t-tests and χ² tests. Correlations were assessed with Spearman or Pearson coefficients, depending on scale properties. Multivariable linear regression modelled predictors of VPT, and ordinal regression modelled lens response categories, adjusting for demographic confounders. Subgroup analyses were performed by TBI status and lens type. There were no missing data and complete-case analysis was performed; no additional sensitivity analyses were conducted. Statistical significance was set at p<0.05 without multiple-comparison correction given limitations of Bonferroni adjustment.¹⁹

Patient and public involvement

Participants were recruited from a tertiary academic centre. Per BMJ definitions, patients and the public were not involved as active partners in the research process (eg, coproduction of research questions, study design, outcome selection or dissemination).

Results

Study population

Eighty-seven screened individuals were enrolled (42 TBI, 45 non-TBI); no individuals were excluded after initial screening (table 1). Participants with a history of TBI were more likely to be older (41.8±15.4 vs 31.8±12.6 years, p=0.001) and male (76.2% vs 44.4%, p=0.003) compared with the non-TBI control participants. Race was also significantly different across groups (p=0.007), but ethnicity was not (p=0.33). VPTs across all lens types were significantly lower, indicating greater light sensitivity, among TBI participants compared with non-TBI participants (For PL VPT: 1.5±0.9 vs 2.4±0.8 log lux, p<0.001, table 1, figures1 2).

Table 1. Demographics and questionnaire parameters in the study population.

Demographics	TBI (n=42)	Non-TBI control (n=45)	P value
Age, years, mean±SD, range	41.8±15.4, 21 to 79	31.8±12.6, 21 to 70	0.001*
Gender, No. (%)			0.003*
Male	32 (76.2)	20 (44.4)
Female	10 (23.8)	25 (55.6)
Race, No. (%)			0.007*
American Indian	0 (0.0)	0 (0.0)
Asian	0 (0.0)	11 (24.4)
Black/African American	9 (21.4)	3 (6.7)
Other/Multiracial	4 (9.5)	3 (6.7)
Hawaiian	0 (0.0)	0 (0.0)
White/Caucasian	28 (66.7)	27 (60.0)
Unknown	0 (0.0)	0 (0.0)
Prefer not to answer	1 (2.4)	1 (2.2)
Ethnicity, No. (%)			0.33
Non-Hispanic	25 (59.5)	28 (62.2)
Hispanic	15 (35.7)	17 (37.8)
Unknown	2 (4.8)	0 (0.0)
Severity code, No. (%)			<0.001*
0—Non-TBI	0 (0.0)	45 (100.0)
1—Mild	12 (28.6)	0 (0.0)
2—Moderate	4 (9.5)	0 (0.0)
3—Severe	20 (47.6)	0 (0.0)
4—Unclassifiable	6 (14.3)	0 (0.0)
PL VPT (log lux), mean±SD, range	1.5±0.9, 0.0 to 4.2	2.4±0.8, 0.3 to 4.3	<0.001*
FL VPT (log lux), mean±SD, range	1.8±0.9, 0.0 to 4.5	2.7±0.8, 0.5 to 4.5	<0.001*
GL VPT (log lux), mean±SD, range	1.8±1.0, 0.0 to 4.5	2.7±0.8, 0.8 to 4.3	<0.001*
PHQ-9, mean±SD, range	8.6±7.2, 0 to 26	3.5±5.0, 0 to 21	<0.001*
PCS-EN, mean±SD, range	16.7±13.7, 0 to 52	7.0±10.1, 0 to 43	<0.001*
SF-12v2, mean±SD, range
PCS	47.2±11.8, 21.0 to 64.0	50.9±9.4, 24.0 to 62.7	0.12
MCS	44.8±10.3, 26.3 to 65.2	49.3±11.1, 19.1 to 65.0	0.06
BTACT Z-scores, mean±SD, range
Word list recall	−0.4±0.9, −1.8 to 3.2	0.4±1.0, −1.8 to 2.8	<0.001*
Digits span backward	−0.2±1.1, −2.3 to 1.3	0.2±0.9, −0.9 to 2.0	0.05*
Number series	−0.3±1.1, −2.1 to 1.1	0.3±0.8, −2.0 to 1.1	0.001*
Category fluency	−0.2±0.9, −2.2 to 1.6	0.1±1.1, −1.7 to 3.8	0.17
Backward counting	0.0±1.1, −3.3 to 3.0	0.0±0.9, −2.6 to 1.8	0.96
Composite	−0.3±1.1, −2.9 to 2.5	0.3±0.9, −1.8 to 2.4	0.005*

SF-12v2, a health survey measuring physical and mental health whose domains contribute to PCS and MCS scores.

Significance (p<0.05) denoted by *.

BTACT, Brief Test of Adult Cognition by Telephone; FL, FL-41 lens; GL, grey-filtering lens; MCS, Mental Component Summary; PCS, Physical Component Summary; PCS-EN, Pain Catastrophizing Scale; PHQ-9, Patient Health Questionnaire-9; PL, plano lens; SF-12v2, Short-Form 12-Item Survey-version 2; TBI, traumatic brain injury; VPT, visual photosensitivity threshold.

Figure 1. VPT (log lux) for TBI participants under each lens condition. FL, FL-41 lens; GL, grey-filtering lens; PL, plano lens; TBI, traumatic brain injury; VPT, visual photosensitivity threshold.

Figure 2. VPT (log lux) for non-TBI control participants under each lens condition. FL, FL-41 lens; GL, grey-filtering lens; PL, plano lens; TBI, traumatic brain injury; VPT, visual photosensitivity threshold.

TBI participants scored significantly higher on the PHQ-9 (8.6±7.2 vs 3.5±5.0, p<0.001) and PCS-EN (16.7±13.7 vs 7.0±10.1, p<0.001) compared with non-TBI controls, indicating greater severity of depression and catastrophising symptoms, respectively. On the SF-12v2, MCS (44.8±10.3 vs 49.3±11.1, p=0.06) and PCS (47.2±11.8 vs 50.9±9.4, p=0.12) scores were lower in TBI cases versus controls, indicating lower health related quality of life, but these differences did not reach statistical significance. On the BTACT, TBI participants scored lower on word list recall (−0.4±0.9 vs 0.4±1.0, p<0.001), digits span backward (−0.2±1.1 vs 0.2±0.9, p=0.05), number series (−0.3±1.1 vs 0.3±0.8, p=0.001) and composite cognitive scores (−0.3±1.1 vs 0.3±0.9, p=0.005), indicating lower cognitive functioning.

Associations between VPT while wearing plano lenses and questionnaire scores

In TBI participants, PL VPT was inversely correlated with PHQ-9 (ρ=−0.385, p=0.01) and PCS-EN scores (ρ=−0.401, p=0.009) (table 2), implying higher light sensitivity in individuals with more severe symptoms of depression and catastrophising. Positive correlations between PL VPT and the Z-scores of the digits span backward (r=0.312, p=0.05) and number series (r=0.417, p=0.007) portions of the BTACT were also found, implying higher light sensitivity in individuals with lower cognitive functioning. Among non-TBI participants, there were no significant correlations between questionnaire and PL VPT scores (table 2).

Table 2. Correlation coefficients (Pearson and Spearman) between PL VPT (log lux) and questionnaire scores among TBI and non-TBI control participants.

	TBI (n=42)		Non-TBI controls (n=45)
	Spearman ρ	P values	Spearman ρ	P values
PHQ-9	−0.385	0.01*	0.151	0.32
PCS-EN	−0.401	0.009*	−0.128	0.40
SF-12v2
PCS	−0.023	0.89	−0.036	0.82
MCS	−0.040	0.81	−0.151	0.34

	Pearson r	P values	Pearson r	P values
BTACT Z-scores
Word list recall	0.077	0.64	−0.177	0.24
Digits span backward	0.312	0.05*	−0.098	0.52
Number series	0.417	0.007*	−0.166	0.28
Category fluency	0.261	0.10	−0.016	0.92
Backward counting	0.111	0.50	0.015	0.92
Composite	0.310	0.05	−0.128	0.40

SF-12v2, a health survey measuring physical and mental health whose domains contribute to PCS and MCS scores.

Significance (p<0.05) denoted by *.

BTACT, Brief Test of Adult Cognition by Telephone; MCS, Mental Component Summary; PCS, Physical Component Summary; PCS-EN, Pain Catastrophizing Scale; PHQ-9, Patient Health Questionnaire-9; PL, plano lens; SF-12v2, Short-Form 12-Item Survey-version 2; TBI, traumatic brain injury; VPT, visual photosensitivity threshold.

Multivariable models to examine predictors of VPT while wearing plano lenses

Controlling for demographics, a multivariable linear regression model was examined with PL VPT as the dependent variable and all questionnaires (PHQ-9, PCS-EN, SF-12v2 PCS and MCS), BTACT Z-scores and TBI severity as independent variables (table 3). Questionnaire scores were found to predict 16.3% of PL VPT variance (R²=0.163). The most influential factor was pain catastrophising, with each point increase in PCS-EN corresponding to a 0.027 log lux decrease in PL VPT (95% CI −0.048 to −0.006, p=0.01). Similar relationships were not noted in the non-TBI population.

Table 3. Multivariable regression examining factors predictive of VPTs among TBI participants controlling for demographics (Models 1–3) and multivariable ordinal regression examining factors predictive of magnitude of response expressed as ‘mild’, ‘moderate’ or ‘strong’ response to FL or GL as compared with PL tints among TBI participants while controlling for demographics (Models 4–5).

Variable	Model 1: PL VPT	Model 2: FL VPT	Model 3: GL VPT	Model 4: FL VPT magnitude of response	Model 5: GL VPT magnitude of response
PHQ-9	–	β=−0.048 (p=0.02*)	β=−0.051 (p=0.02*)	OR=0.833, 95% CI 0.717 to 0.969 (p=0.02*)	OR=0.829, 95% CI 0.689 to 0.998 (p=0.047*)
PCS-EN	β=−0.027 (p=0.01*)	–	–	–	–
SF-12v2
PCS	–	–	–	–	OR=1.201, 95% CI 1.029 to 1.403 (p=0.021*)
MCS	–	–	–	–	OR=1.166, 95% CI 1.031 to 1.318 (p=0.02*)
BTACT Z-scores
Word list recall	–	–	–	–	–
Digits span backward	–	–	–	–	–
Number series	–	β=0.349 (p=0.01*)	β=0.376 (p=0.009*)	–	–
Category fluency	–	–	–	–	OR=0.138, 95% CI 0.023 to 0.836 (p=0.03*)
Backward counting	–	–	–	–	–
TBI severity code	–	–	–	–	–

SF-12v2, a health survey measuring physical and mental health whose domains contribute to PCS and MCS scores.

The symbol ‘–’ denotes predictors that are not significant.

Significance (p<0.05) denoted by *.

BTACT, Brief Test of Adult Cognition by Telephone; FL, FL-41 lens; GL, grey-filtering lens; MCS, Mental Component Summary; PCS, Physical Component Summary; PCS-EN, Pain Catastrophizing Scale; PHQ-9, Patient Health Questionnaire-9; PL, plano lens; SF-12v2, Short-Form 12-Item Survey-version 2; TBI, traumatic brain injury; VPT, visual photosensitivity threshold.

Multivariable models to examine predictors of VPT while wearing FL lenses

Controlling for demographics, a similar multivariable linear regression model was examined with FL VPT as the dependent variable (table 3). The model explained 30.4% of FL lenses VPT variance (R²=0.304). In this model, the most influential factors were PHQ-9 and BTACT Number Series Z-scores; with each point increase in PHQ-9 corresponding to a 0.05 log lux decrease in FL VPT (95% CI −0.088 to −0.008, p=0.02) and with each point increase in Number Series Z-score corresponding to a 0.35 log lux decrease in FL VPT (95% CI 0.081 to 0.0617, p=0.01). Similar relationships were not observed in non-TBI participants.

To examine extent of treatment response to FL lenses in TBI participants, a multivariable ordinal regression model was created, with the dependent variable expressed as ‘mild’ difference <2 log lux; ‘moderate’ difference ≥2 and <3 log lux; or ‘strong’ difference ≥3 log lux based on the difference between FL and PL VPTs. Unadjusted associations are presented in tables1 2, with adjusted regression models summarised in table 3. In this model, higher PHQ-9 scores predicted a less robust response to FL lenses (less decrease in light sensitivity) (OR=0.833, 95% CI 0.717 to 0.969, p=0.02). Taken together, these models demonstrate that higher depressive symptoms and less cognitive dysfunction reduced the effectiveness of FL lenses in alleviating visual photosensitivity.

Multivariable models to examine predictors of VPT while wearing GL lenses

A similar analysis was performed using GL VPT as the dependent variable (table 3), with the model explaining 32.2% of the variance and with PHQ-9 and Number Series Z-score being most influential. For each point increase in PHQ-9, GL VPT decreased by 0.051 log lux (95% CI −0.092 to −0.010, p=0.02) and for each point increase in Number Series Z-score, GL VPT increased by 0.376, (95% CI 0.101 to 0.652, p=0.009). These relationships were not seen in non-TBI participants.

A multivariable ordinal regression model was next examined, with the dependent variable expressed as ‘mild’, ‘moderate’ or ‘strong’ response to GL as compared with PL tints (table 3). A higher PHQ-9 score predicted a less robust response to GL lenses (OR=0.829, 95% CI 0.689 to 0.998, p=0.047), as did lower PCS (OR=1.201, 95% CI 1.029 to 1.403, p=0.02) and MCS (OR=1.166, 95% CI 1.031 to 1.318, p=0.02) scores. Furthermore, a higher BTACT categorical fluency Z-score predicted a less robust response to GL lenses (OR=0.138, 95% CI 0.023 to 0.836, p=0.03). In other words, these models demonstrate GL lenses had a reduced effect in alleviating visual photosensitivity in individuals with higher depressive symptoms, lower physical and mental quality of life, and less cognitive dysfunction.

Discussion

Our results suggest that individuals with TBI exhibited greater visual photosensitivity, depression, pain catastrophising and cognitive dysfunction compared with those without TBI. Additionally, neuropsychological symptoms emerged as key predictors of both baseline visual photosensitivity and treatment response to light-filtering lenses, with higher depression scores and higher cognitive performance associated with a poorer response to filtering lenses. These findings highlight the importance of considering mental health and cognitive status in the evaluation and treatment of visual photosensitivity in individuals with a history of TBI.

In our study, individuals with TBI reported more severe depression and pain catastrophising symptoms and displayed lower cognitive functioning than non-TBI controls, findings consistent with prior literature. In a meta-analysis that compared individuals with (n=724 842) and without (n=323 666) TBI, cases were more than twice as likely to have self-reported feelings of depression or a diagnosis of depression compared with controls (relative risk=2.10, p<0.01).²⁰ Prior studies have also shown high pain catastrophising scores in some individuals with mild TBI.²¹ In one study, individuals were grouped by the number of physical (eg, headaches, dizziness), cognitive (eg, poor concentration) and emotional (eg, irritability, depression) symptoms reported on the 16-item Rivermead Postconcussion Symptoms Questionnaire.^{21 22} Individuals with ≥6 symptoms 1 and 2 months post mild TBI had higher PCS-EN scores compared with those with <6 symptoms at these time points (17.47±10.62 vs 10.29±9.69, p<0.001).²¹ Finally, with respect to cognition, a multicentre, cross-sectional study demonstrated impaired cognition in many adults with mild to severe TBI (n=498, BTACT composite scores −0.70±1.16, 40% considered impaired based on score < −1).²³ Overall, our findings, along with prior research, highlight the substantial mental health and cognitive burden experienced by individuals with a history of TBI.

Our results build on the literature by demonstrating that psychological and cognitive comorbidities are not only more prevalent in individuals with TBI but are also meaningfully linked to visual photosensitivity. First, we demonstrated more severe visual photosensitivity (lower VPTs) in individuals with TBI compared with controls, findings indirectly supported by prior studies. In a meta-analysis of 2084 individuals with TBI and 2233 controls, TBI cases were more likely to self-report photophobia compared with controls (OR=5.75, 95% CI 4.66 to 7.10, p<0.001).³ Furthermore, on psychophysical testing using the OPA, TBI cases (n=233) were found to have greater light sensitivity (lower VPT) compared with 162 controls (526.96±88.92 lux vs 919.37±180.19 lux, p<0.001).²⁴ The same study noted that depression severity was also higher in TBI cases as compared with non-TBI controls (PHQ-9 scores: 8.26 ± 6.32 vs 5.63 ± 2.87, p<0.001).²⁴ These findings support our noted relationships demonstrating that psychological conditions can impact TBI associated visual photosensitivity.

Our data further highlight that psychological conditions and cognitive dysfunction may influence response to tinted lenses as a treatment of visual photosensitivity, a novel observation not previously examined in prior studies. However, a prior study had examined individual preference to tints in active-duty military and veterans following TBI (n=392). The study first determined each individual’s preferred filter and subsequently assessed photophobia concern (scale 0–5 with 0 indicating no concern and 5 indicating profound concern) with and without the selected filter.²⁵ Participants reported a reduction in photophobia-related concerns while wearing their preferred tinted lenses (pre: 4.44±0.90 → post: 0.06±0.43). Interestingly, self-reported mental health (pre: 4.12±0.83 → post: 2.04±0.04) and short-term memory (pre: 4.02±1.26 → post: 2.89±1.67) scores also improved, suggesting associations between photophobia, its treatment (light-filtering lenses), mental health and cognition.²⁵

Besides tinted lenses, therapies such as botulinum toxin A (BoNT-A) and calcitonin gene-related peptide receptor (CGRP) monoclonal antibodies (mAbs) are used to treat photophobia, although these entities have been robustly studied in migraine and not TBI. Interestingly, there is evidence suggesting that psychological factors may influence response to these treatments. In one study, response to BoNT-A was evaluated in 40 patients with chronic migraine, with responders defined by a >50% reduction of headache days per month. Responders (20%, n=5) had a lower frequency of depressive symptoms (assessed with the PHQ-4 prior to BoNT-A treatment) than non-responders (60%, n=9, p=0.002).²⁶ Similar findings have been noted for CGRP mAbs. In individuals with migraine, CGRP mAbs responders (>50% reduction of monthly headache days, n=53) had lower anhedonia scores on the Personality Inventory for DSM5 compared with non-responders (<30% reduction in monthly headache days, n=31) (0.91±0.68 vs 1.25±0.76, p=0.03).²⁷ These findings support the premise that psychological conditions may impact treatment of photophobia across various treatments (tints, injections) and diagnoses (TBI, migraine).

The cause(s) of post-TBI associated visual photosensitivity are unclear; however, impaired neurogenesis, inflammation and microglial activation, and glutamate system dysfunction are processes that have all been implicated at the level of the central nervous system.²⁸ Similar mechanisms have been postulated to underlie post-TBI neuropsychological symptoms, potentially linking the two entities.²⁹ In humans, overlapping brain regions for photophobia, depression and anxiety have been demonstrated using functional MRI technology.^{30 31} For example, activation within the anterior cingulate cortex was noted in response to sad facial expressions in healthy participants³⁰ and on exposure to a white screen in individuals with ocular pain,³¹ linking this region to both processing of emotions and light. Intrinsically photosensitive retinal ganglion cells (ipRGCs), which respond to light via melanopsin may also link the two entities, as ipRGCs project to limbic regions that process emotion, such as the lateral habenula and medial amygdala.^{32 33}

In mice, light has been shown to influence cognition and mood through ipRGC activity.³⁴ One study demonstrated that a 10-minute light pulse in mice negatively affected cognition (spatial learning deficits noted), mood (decreased preference for novel objects) and stress (increased corticosterone levels).³⁴ Molecularly, significant increases in c-fos expression, an indicator of neuronal activity, were noted in brain regions involved in emotional processing that included the amygdala, the lateral habenula and the subparaventricular nucleus.³⁴ These results provide insight into potential mechanisms on why depressive symptoms and cognitive function may impact light sensitivity and modulate responses to tinted lenses.

Our study has several limitations that merit consideration. First, our sample size was modest, which reduced power for detecting smaller effect sizes or subgroup differences. Second, our study included both civilian and Veteran participants, groups that may differ in health experiences, comorbidity burden, injury mechanisms and psychosocial stressors. Third, not all psychological disorders (eg, anxiety, PTSD) were comprehensively assessed in our study. Fourth, although the OPA provides a standardised measure of light sensitivity, it may not capture all facets of visual photosensitivity relevant to daily life, such as visual discomfort under various lighting conditions (eg, sunlight). Additionally, psychological symptoms and cognitive performance were assessed via self-report, which are subject to bias and may not fully capture clinical diagnoses. This introduces risk of misclassification in both exposures and outcomes. Residual confounding from unmeasured factors such as anxiety or PTSD may also have influenced associations. The heterogeneity of TBI-related comorbidities (eg, headache) also poses challenges as they may influence mental health and sensory processing. Finally, only two different lens tints were tested, and we did not compare responses to other tints or non-optical interventions. Given our modest sample and multiple analyses, results should be interpreted with caution. Future studies in larger, younger and more diverse populations will be necessary to strengthen generalisability and to determine whether these relationships hold across differing injury contexts and demographic backgrounds.

Despite these limitations, this study provides novel insights into how mental health and cognitive comorbidities modulate both visual photosensitivity and treatment response in individuals with TBI. While many studies have assessed these domains in isolation, our work emphasises their interdependence and highlights the need for integrated approaches to treatment. Current implementation strategies to connect vision-related psychological distress are limited, leading to gaps in care for patients who would benefit from a collaborative and multidisciplinary approach to management.³⁵ Currently, it is not known if addressing mental health will improve visual photosensitivity and/or treatment responsiveness. Therefore, future research is needed to further explore these interconnectivities and examine whether addressing psychological symptoms through pharmacological or psychotherapeutic means could enhance responsiveness to visual interventions and ultimately may yield more personalised and effective care for patients with debilitating light sensitivity.

Data availability statement

Data are available upon reasonable request.