Check your eyes before your theme

Question: A developer who ends most workdays with a headache after long terminal sessions wants to know what the published evidence says about the usual suspects: dark mode against light mode, pure white on black against a softened palette, line spacing, font size and typeface, blue light and warm screen tints, backlight flicker and refresh rate, blinking and break schedules, and what an eye examination can find that a theme change cannot. Which of these has evidence behind it, what should change first, and what defaults should a terminal-focused IDE ship as a result?

Short answer: the highest-yield first move is a comprehensive eye examination that explicitly includes refraction for small astigmatism, binocular (convergence) and accommodative testing, and a dry-eye assessment, not a change of editor theme. Every major review of digital eye strain puts refractive correction and binocular and ocular-surface management first, and no display-side variable (polarity, contrast, font, refresh rate, blue light, line spacing) has ever been tested with headache as a measured outcome. On the display side the one change with real mechanistic support is larger text with looser line spacing: reading speed collapses below an x-height of about 0.2 degrees, the light-mode legibility advantage grows as fonts shrink, and rows of text at the pitch of a typical 12-14 px terminal font on a laptop land close to the 3 cycles-per-degree stripe frequency that provokes the most visual discomfort in susceptible people. Dark against light mode is a draw on comfort: light mode wins every legibility experiment because its higher luminance shrinks the pupil, but fatigue trials find no significant difference and no study has measured headache. Blue-light filtering fails a 120-person double-masked trial and a Cochrane review; software warm tints have no eye-strain trial at all. Invisible 100 Hz lighting flicker more than halved headaches in a double-blind office crossover, but an M4 MacBook Pro backlight was bench-measured at 14,880 Hz, far above every published sensitivity threshold. A 5-minute break every hour has field-trial support that the 20-20-20 rule lacks, and complete blinks matter more than blink count. This is a synthesis of published evidence, not medical advice.

Headache is one of the core symptoms of what optometry calls computer vision syndrome or digital eye strain, and it is common enough that its presence says almost nothing about its cause. In a survey of 520 New York office workers, 22.3% reported headache at least half the time during computer use in the prior week [1], [2]. Pooled estimates of any visual symptom run from 50% to 90% of computer users, and the 2023 Tear Film and Ocular Surface Society (TFOS) consensus report notes figures "up to 97%" because there are no agreed diagnostic criteria [3], [4]. All of these are cross-sectional self-reports; none shows that screens cause the headache rather than co-occur with it.

What the reviews agree on is the order of attack. TFOS states that "interventions are not well established; patients experiencing digital eye strain should be provided with a full refractive correction for the appropriate working distances. Improving blinking, optimizing the work environment and encouraging regular breaks may help. Based on current, best evidence, blue-light blocking interventions do not appear to be an effective management strategy" [4]. Rosenfield's review identifies oculomotor anomalies and dry eye as the principal ocular causes and concludes that "the efficacy of proposed treatments to reduce symptoms of CVS is unproven" [3]. Both imply the same order: refraction and binocular status, then the ocular surface, then workstation ergonomics, with unproven add-ons such as tints last.

A useful triage tool comes from Sheedy's factor analysis of induced eye strain in 20 subjects. Symptoms split into an "external" cluster (burning, irritation, tearing and dryness, provoked by glare, up-gaze, small fonts and flicker, "highly related to dry eye") and an "internal" cluster (ache, strain and headache behind the eyes, provoked by close viewing distance, accommodative flippers and mixed astigmatism, "likely related to accommodative and vergence stress") [5]. A developer whose headache sits behind the eyes and arrives in the afternoon should suspect the internal cluster, which points to refraction and binocular vision, not to the colour of the editor background.

One framing point matters throughout: not a single study in this evidence base used a terminal, a monospaced code font or a multi-hour session. The polarity trials run 15 minutes to an hour on tablets and phones with prose, the typography work is on print, and the break studies are on data-entry clerks. Everything applied here to a terminal is extrapolation.

2.1 The light-mode advantage is a luminance effect that shrinks the pupil

The best-replicated finding in this literature is the "positive polarity advantage": dark text on a light background is read and proofread more accurately than light on dark. The Düsseldorf group found it in younger and older adults alike and recommended positive polarity "for all ages" [6]. The advantage grew linearly as character size shrank across 8-14 pt, leading the authors to conclude that "especially with small font sizes, negative polarity displays should be avoided" [7]. The mechanism was tested directly: pupils were smaller and proofreading better on positive-polarity displays, consistent with "the display luminance hypothesis" that a brighter screen produces a smaller pupil and hence a sharper retinal image [8]. The decisive experiment predates these: when overall display luminance was equated, "no positive polarity advantage was observed... Display polarity per se does not affect readability" [9]. The effect held in darkness and office light alike and could not be rescued by colour contrast [10].

An independent lab agrees: in a 23-lux room, lexical-decision thresholds were 85.3 ms for positive against 119.4 ms for negative polarity, about 40% longer for light on dark, attributed to "the dilation of the pupils under low illumination produc[ing] optical blurring" [11].

For a developer the operative variable is therefore mean screen luminance reaching the eye, not the colour arrangement: a dark theme on a bright panel with large glyphs narrows the gap, and a dim dark theme with small monospace glyphs is the documented worst case. The same mechanism is the only physically grounded version of the folk claim that dark mode is worse for people with astigmatism, since a larger pupil enlarges the blur circle from any uncorrected refractive error. The specific "halation" claim in accessibility writing comes from a vendor blog with no study citations, and targeted searches found no peer-reviewed test of polarity against uncorrected astigmatism [12]; it is plausible and untested.

2.2 Fatigue trials are small, inconsistent and mostly null

The best-controlled recent trial randomized 30 young women to one hour each of light and dark mode on an iPad at full brightness: asthenopia scores rose in both (p < 0.001 pre-post) with a between-mode p of 0.305, not significant. Secondary outcomes differed (critical flicker fusion p = 0.001, dry-eye score p = 0.048) and the authors concluded dark mode "may help reduce the risk of eye fatigue", but the paper reads a higher flicker-fusion frequency as more fatigue, contrary to convention, and reports no effect sizes [13]. An 18-person conference study found no main effect of polarity, only an interaction with ambient light: dark mode produced lower fatigue at 460 lux (p = 0.004) and no difference at 33 lux, with auto-brightness left on as a confound [14]. The same paper summarizes an earlier eye-tracking study by Xie's group in which blink rate and pupil measures indicated less fatigue in dark mode at night even though subjects rated light mode less tiring and preferred it; it was not read directly here [14]. A 36-person follow-up from the Xie group reported lower fatigue in dark mode at both 450 and 3 lux; only the abstract was accessible, so its statistics are unverified [15]. Against these, Sethi and Ziat found higher cognitive load under negative polarity for older adults in bright rooms and younger adults in dim ones, with the younger group preferring dark mode "for aesthetic reasons" [16]. The older Buchner and Baumgartner series measured breathing, heart rate, skin conductance and self-reported fatigue, eyestrain and headache and found none "var[ied] as a function of any of the independent variables" [10].

Put together: performance measures favour light mode, one group's eye-tracking favours dark mode, and subjective fatigue is null or an interaction with room light and screen size. No study shows a large, consistent fatigue advantage for either mode on a desktop monitor in office lighting, the developer's actual situation. The one consistent thread is that a large luminance mismatch between screen and surround is the bad cell in either direction, which matches the standards logic in section 5.

2.3 Headache and migraine: no direct evidence, one pathway

No controlled study of display polarity or contrast level has headache or migraine as its primary outcome. The only polarity experiments that measured headache at all found nothing [10] or folded it into a fatigue score without reporting it separately [14]. Searches for "dark mode migraine" return only vendor and patient-education pages, none citing a study. The one evidence-backed pathway from screen appearance to headache is pattern glare, which depends on stroke and line periodicity, contrast and flicker rather than polarity; section 4 covers it.

No controlled study compares #FFFFFF text on #000000 against an off-white-on-dark-grey palette for discomfort or headache. The nearest evidence is from gratings. In 95 young adults viewed double-masked, discomfort thresholds for 2.3 cycle-per-degree gratings were reached at about 70% contrast in people who reported benefit from coloured overlays against about 87% in the others (p = 0.015), and symptom count correlated with the threshold [17]. Pattern contrast therefore does drive discomfort in susceptible individuals, and the same paper notes that text "resembles stripes because of horizontal lines of words and the periodic vertical strokes of letters". But the relationship is not a simple "more contrast is worse" law. In images built from natural statistics, "discomfort increased with increasing color contrast or decreasing luminance contrast" [18]. That result has a direct implication for syntax highlighting: a theme whose tokens differ mainly in hue while sitting at similar, low luminance against the background is the pattern those stimuli flagged as aversive. Each highlight colour should carry luminance contrast, not just chromatic contrast.

Formal guidance does not help choose between 10:1 and 21:1. WCAG 2.2 sets minimums only (4.5:1 at level AA, 3:1 for large text, 7:1 at AAA); its Understanding document "does not say that contrast above the requirement is a problem, and it does not set a maximum contrast", though it acknowledges that "some people with cognitive disabilities need low-contrast color combinations" [19]. A moderated palette such as #CCCCCC on #1E1E1E is roughly 10:1 and passes everything; the choice between that and 21:1 white on black is unconstrained by standards and unstudied for comfort. The common advice to soften the palette is plausible under the pattern-glare mechanism but is not evidence-based.

4.1 Rows of text form stripes near the most provocative spatial frequency

Wilkins and Nimmo-Smith showed in 1987 that successive lines of printed text form a horizontal stripe pattern resembling the gratings known to provoke discomfort, illusions and, in photosensitive patients, seizures; judged clarity depended most on inter-line separation, and they recommended slightly tighter letter spacing with wider line spacing [20]. The full text is paywalled, so the paper's own cycles-per-degree figures could not be checked, and the often repeated claim that double or triple line spacing removed most symptoms could only be found second-hand, so it is left out here. The underlying grating parameters are well established: "patterns of high contrast having a striped configuration with spatial frequency around 3 cycles/degree, and with stripes of equal width and spacing (duty cycle of ~50%), tend to produce maximum effect" [21], and the clinical Pattern Glare Test uses 0.5, 3 and 12 cycle-per-degree gratings, with migraine patients reporting the most symptoms on the 3 cpd plate [22].

The modern version of the argument is quantitative. In a 2020 study, 15 participants adjusted font, size and background in an e-reader to their preference; a Fourier-based discomfort predictor fell in 29 of 30 settings, and across 72 texts the x-height-to-line-spacing ratio alone accounted for 79% of the variance in that predictor, with preferred ratios of 31-46% and the authors suggesting the typical 35-45% "might be reduced by increasing line spacing" [23]. Reading-speed experiments give a more modest picture: in central vision, speed rose with vertical spacing up to about 1.2-1.5 times standard and then plateaued [24], and no recent empirical work on fine-grained leading exists on modern displays [25]. So the case for extra leading beyond 1.5x rests on comfort modelling, not speed, and no study has measured headache as a function of line spacing on any screen.

Two pieces of arithmetic, this doc's own derivation and not published figures, show why this matters for terminals. A 3 cpd grating corresponds to a line pitch of 20 arcminutes, which at 60 cm is about 3.5 mm. On a 14-inch MacBook Pro at its default scaled resolution (about 127 logical pixels per inch), a 14 px font at line-height 1.2 gives a 16.8 px pitch, about 3.4 mm, which at 60 cm is very close to 3 cycles per degree, the peak of the pattern-glare curve. Raising the same font to 20 px at line-height 1.5 gives a 6 mm pitch, about 1.7 cpd, well off the peak. Separately, because monospaced fonts have x-height fractions near 0.5, a terminal at line-height 1.0-1.2 has an x-height-to-line-pitch ratio of roughly 43-52%, above the 35-45% that print typography uses and that Wilkins already considers too tight; line-height 1.4-1.6 brings it to 32-37%. These derivations assume an x-height fraction of about 0.52 and a 60 cm viewing distance; other fonts and distances shift the numbers, not the conclusion that stock terminal settings sit in the wrong part of the curve.

4.2 Font size: reading speed collapses below 0.2 degrees of x-height

Legge and Bigelow's review concludes that reading speed rises with print size up to a critical print size of about 0.2 degrees of x-height for normally sighted readers and is then flat across a tenfold "fluent range" of roughly 0.2-2 degrees; x-height, not nominal point size, is the vision-relevant measure, because the x-height fraction varies from 0.28 to 0.58 across typefaces (mean about 0.46, with screen-designed faces near 0.52) [26]. Newspaper running text averages 0.23 degrees and online news text 0.21 degrees, just inside that range. The same review reports Sheedy's threshold-legibility finding that Verdana and Arial were the most legible screen faces and Times New Roman the least, a result about thresholds rather than comfort [26]. On screen, 10 pt fonts were read significantly more slowly than 12 pt, with no 12 against 14 pt difference [25]. The size literature is entirely reading-speed and threshold based; no study measures hours-long discomfort as a function of x-height in adults.

Again in this doc's own arithmetic: at 60 cm, 0.2 degrees of x-height is 2.1 mm and 0.25-0.3 degrees is 2.6-3.1 mm. On the 14-inch MacBook Pro at default scaling, a 12 px terminal font has an x-height near 1.25 mm, about 0.12 degrees at 60 cm, well below the critical print size; reaching 0.2 degrees needs roughly 20 logical px, and 0.25 degrees about 25 px. On a 27-inch 4K monitor at 2x scaling (about 82 logical px per inch) a 14 px font gives about 0.2 degrees at 65 cm, right at the threshold.

4.3 Typeface: stroke periodicity matters, named coding fonts are untested

Wilkins' group showed across eight studies that the first peak of a word image's horizontal autocorrelation, a measure of how evenly spaced the vertical strokes are, predicts how striped words look and how fast they are read; high-peak words were read about 10% more slowly [23]. In the same group's 20-font analysis, serif fonts had significantly higher peaks than sans (t(18) = 3.45, p = 0.003), Times was among the most striped, Arial, Verdana and Open Sans among the least, and the single monospaced "typewriter" font had the lowest peak of all, apparently because fixed advance widths break stroke regularity across letters of different widths; the font is not named, and whether this generalizes to modern coding fonts is untested [23]. Font effects are also strongly individual: in crowdsourced tests with hundreds of adults, speed differed by 35% between each person's fastest and slowest font, and the best font varied widely across people [27].

The monospaced-against-proportional literature is small, old and mixed: proportional pitch read faster at medium and large sizes while fixed pitch was better near the acuity limit, and a later study could not replicate the size-dependent difference [26].

For the fonts developers actually argue about, the answer is simple: no peer-reviewed controlled study of JetBrains Mono, Fira Code, Menlo, SF Mono or Atkinson Hyperlegible for eye strain, headache or reading performance was found. JetBrains' description of its font's fatigue-reducing letterforms is a design rationale, not a measurement [28], and a listicle's "78% of coders reported less eye strain" has no source and is unusable [29].

4.4 Letter spacing, line length and viewing distance

Letter spacing should stay at the font's default for skilled adult readers: RSVP reading speed peaks at standard spacing and declines for both tighter and wider tracking, with visual span following the same curve [30]. Line length is not a demonstrated strain factor: 100-character lines read as fast as or faster than 25-character lines, though readers judged about 55 characters "easiest" [25]. Viewing distance interacts with size: in 24 adults doing two hours of demanding work at 52, 73 and 85 cm with text size held constant, the nearest distance produced less blurred vision, fewer dry or irritated eyes, less headache and better convergence recovery, so the authors recommend 52-73 cm when characters are near the acuity limit [31]. Conversely, people with a distant resting vergence reported more strain at 50 cm than at 100 cm even with doubled characters, and freely chosen distances averaged 74 cm [32]. The rule that fits both is to fix the angular size first, then let the user choose the distance they find restful.

5.1 Blue-light lenses and software warm tints

Blue-light filtering has the strongest evidence in this doc, and it is negative. A double-masked RCT randomized 120 symptomatic computer users to clear or blue-blocking spectacles for a two-hour computer task: no effect on the prespecified primary outcomes of symptom score (p = .394) or critical flicker fusion (p = .304), and the conclusion that "blue-blocking lenses did not alter signs or symptoms of eye strain with computer use relative to standard clear lenses" [33]. The 2023 Cochrane review of 17 RCTs (619 participants) could run no meta-analyses, found no trial at low risk of bias in all domains, and concluded that such lenses "may not attenuate symptoms of eye strain with computer use", with indeterminate effects on sleep [34]. TFOS says the same: "blue-light blocking interventions do not appear to be an effective management strategy" [4].

For Night Shift, f.lux and similar warm tints the evidence is not negative so much as absent. The only RCT of Night Shift (167 adults, actigraphy) found "no significant differences in sleep outcomes across the three experimental groups" and did not measure eye strain or headache [35]; no eye-strain trial of any software warm tint was found. The one laboratory study that varied screen colour temperature in 36 people found that the warmest setting, 2800 K, "resulted in the highest visual fatigue" whether in daytime or night-time ambient light, with lower fatigue at 4500 and 6500 K; the full text was not read and the statistics are unverified [15]. So there is no positive evidence for warm tints, and a weak signal that very warm settings increase subjective fatigue.

5.2 Ambient brightness and luminance ratios come from standards, not trials

The public preview of ISO 9241-303 (2011 edition) states that task areas viewed in sequence should have luminances "between 0,1 L and 10 L" of the screen average, that a task-to-surround ratio "up to 1:10, has no adverse effect" on a stationary visual field, and that for a 500 lux office "it is often recommended that the display luminance be in the range of 100 cd/m2 to 150 cd/m2" [36]. These clauses are quoted from the preview, not the purchased standard.

The experimental literature is thin. In 50 people reading dark-mode text at 28 cd/m2, raising ambient light from 0 to 10 lux improved objective pupil and blink measures (F(2,147) = 17.08, eta-squared 0.20) without changing subjective fatigue, though the paper's results contain internal inconsistencies [37]. The defensible practice is standards-backed rather than trial-backed: keep display luminance around 100-150 cd/m2 in an ordinary office, do not run a bright screen in a dark room, and match theme polarity to the room. A dark theme in a dim room and a light or bright theme in a bright room both stay inside the 1:10 envelope.

6.1 100 Hz lighting halves headaches; a 14.9 kHz backlight does not

Invisible flicker is the environmental factor with the strongest headache-specific evidence. In a double-blind crossover in office workers, conventional fluorescent ballasts (100 Hz modulation at 43-49% depth) were compared with electronic ballasts at about 32 kHz: "the average incidence of headaches and eyestrain was more than halved under high-frequency lighting", and headaches also fell with office height above ground, that is, with more daylight [38]. IEEE Std 1789-2015, listed as Inactive-Reserved since March 2026, set "low-risk" limits of modulation percentage below 0.08 times frequency up to 1250 Hz and "no-effect" limits of 0.0333 times frequency up to 3000 Hz, with no limit above either; the standard is paywalled, so these figures are quoted from the US Department of Energy's summary [39], [40].

Sensitivity extends far above the flicker-fusion threshold through the phantom-array effect, in which modulated light smears into a visible trail during saccades. In 25 migraineurs and 30 non-migraineurs, migraineurs saw more stimuli (odds ratio 3.31), 64% against 19% reported discomfort within four hours and 41% against 8% reported headache; peak sensitivity was 400-800 Hz, stimuli up to 7000 Hz were more visible than steady light, and at 10,000 Hz visibility no longer differed from steady light, leading the authors to suggest about 10 kHz as a reasonable minimum PWM frequency [41].

Against those thresholds the measured Apple figures are reassuring. Notebookcheck's bench measurement of the 2024 M4 Pro MacBook Pro 14 found PWM at 14,880 Hz at every brightness level with about 90% amplitude, and noted that "most users sensitive to PWM should not notice any flickering"; for context, 52% of the devices it has tested use no PWM, and the average frequency among those that do is about 7605 Hz [42]. That is a reviewer measurement, not peer-reviewed data, and no peer-reviewed study has measured headache against display PWM frequency. But 14.9 kHz is above the IEEE no-effect cutoff and above Miller's 10 kHz suggestion, so there is no physical basis for attributing an M4 MacBook Pro headache to backlight PWM. OLED panels are the suspect category: Notebookcheck measured a Samsung OLED laptop panel at 240 Hz PWM "with a comparatively large amplitude", squarely in the phantom-array band [43]. The more plausible invisible-flicker source in a developer's room is cheap LED or fluorescent lighting with 100 or 120 Hz ripple, the exact condition Wilkins tested.

6.2 Refresh rate: no evidence either way for static text

No controlled, peer-reviewed study compares 60 Hz with 120 Hz for comfort or eye strain during static text work on a modern sample-and-hold panel. The historical refresh-rate literature concerns CRT phosphor flicker, which LCD and OLED panels do not reproduce, and Sheppard and Wolffsohn's review does not discuss refresh rate at all [2]. Claims that 120 Hz "reduces eye strain" for text are not evidence-based; its plausible benefit is motion clarity when scrolling.

7.1 Blink rate, blink completeness and dry eye

The blink-rate reduction during screen work is one of the most replicated findings in this field: Sheppard and Wolffsohn collect reductions from 18.4 to 3.6, 22 to 7 and 15.5 to 5.3 blinks per minute, and a pooled dry-eye prevalence among computer users of 49.5% against 5-33% in the general population [2]. In 21 subjects reading on a computer, the symptom score correlated with the percentage of incomplete blinks (p = 0.002), which ranged from 0.9% to 56.5%, and forcing the blink rate up to 23.5 per minute with an audible tone "did not produce a significant change in the symptom score"; the authors concluded that "actions to achieve complete corneal coverage during blinking may be more helpful" [44]. A monitor in primary gaze increases corneal exposure, which is why a slightly downward gaze is recommended [3].

The link from dry eye to headache is associational. A meta-analysis of 11 observational studies covering 3.6 million people found dry eye associated with all-cause headache (OR 1.59, 95% CI 1.41-1.79), migraine (OR 1.50) and tension-type headache (OR 1.61) [45]. A "Celik et al. 2022" dry-eye and migraine paper that circulated during the research does not exist under that name and is not cited. No trial tests whether treating screen-related dry eye reduces headache. Dry eye fits the external symptom cluster: if the headache comes with burning, grittiness or tearing, ocular-surface management is a reasonable adjunct; if not, it is unlikely to be the main driver.

7.2 Break schedules: the 20-second part of 20-20-20 fails its only controlled test

The 20-20-20 rule rests on one small uncontrolled study: 29 symptomatic users received 20-20-20 reminders for two weeks, dry-eye and strain symptoms decreased (p <= 0.045), but "this improvement was not maintained one week after discontinuation", and no ocular-surface sign changed [46]. The only controlled test of the 20-second component, a 30-person crossover with 20-second breaks every 5, 10, 20 or 40 minutes of demanding tablet reading, found symptoms rose in every condition with no effect of the breaks (p = .70); the authors note "there is relatively little peer-reviewed evidence to support this rule" [47]. The stronger evidence is for longer breaks. In a NIOSH field crossover with 42 data-entry operators, adding a 5-minute break in every hour that otherwise had none significantly lowered eyestrain and musculoskeletal discomfort "without reductions in data-entry performance" [48], and a 10-person lab comparison found a 30-minutes-work, 5-minutes-rest schedule gave the lowest eyestrain [49]. The active ingredient looks like a minutes-long break every hour or half-hour, not a 20-second glance every 20 minutes. No break-schedule trial uses headache as a primary outcome.

8.1 Refractive error and astigmatism: less common than believed, but potent when present

The International Classification of Headache Disorders recognizes headache attributed to refractive error (11.3.2) as "generally symptomatic after prolonged visual tasks", and adds that "refractive error is much less commonly a cause of headache than is generally believed" [50]. The one adult series describing the phenotype matches the developer's complaint uncomfortably well: in 90 Tunisian patients, headaches were daily in 90% and in the second half of the day in 82%, and multivariate analysis retained prolonged screen work (p = 0.013) and astigmatism (p = 0.03) among the risk factors, with "improvement in headache in 100% of cases" after correction; that is a small retrospective series with no control for regression to the mean, and the 100% figure is not a credible effect size [51]. The sober counterpoint is an interview study of 105 people with uncorrected refractive error and 71 controls: headache frequency was similar in both, refractive-error headache was identified in only 6.7%, but of those 72.5% improved with correction and 38% remitted [52]. No RCT of refractive correction against placebo lenses with headache as the outcome exists in adults.

Why an exam still comes first is that sub-acuity amounts of astigmatism are a proven symptom driver even when 20/20 acuity is intact. Inducing just +0.50 D of astigmatism in eight adults for 25 minutes of VDT reading produced greater eyestrain (p = 0.01) "even though visual acuity is relatively unaffected" [53]. Most striking, induced oblique astigmatism of 1 D and 2 D raised median post-task symptom scores from 2.0 to 6.5 and 40.0 (p < 0.0001) without changing reading rate [54]. A developer whose last prescription left a quarter or half dioptre of cylinder uncorrected because "it does not affect acuity" has a documented cause of afternoon eye-ache.

8.2 Convergence insufficiency: often untested, measurable, treatable in signs more than symptoms

Convergence insufficiency (CI) is the most common non-strabismic binocular anomaly in symptomatic adults: 18.2% in a hospital series of 231 young adults with 6/6 acuity, where headache was the most frequent presenting symptom [55], and 7.7% in 65 university students with no refractive error [56]. It can be screened with the Convergence Insufficiency Symptom Survey (CISS), where a cut-off of 21 in adults gives 97.8% sensitivity and 87% specificity [57]. The treatment evidence is excellent in children and weaker in adults. In 221 children, office-based therapy reduced symptom scores to 15.1 against 21.3-24.7 for home therapy, pencil push-ups and placebo (p < .001) [58], and the Cochrane network meta-analysis rates office therapy's superiority over placebo for clinical signs as high-certainty (RR 3.04, 95% CI 2.32-3.98) [59]. But in the one adult RCT, office therapy normalized convergence (near point 12.8 to 5.3 cm) yet 58% remained symptomatic and placebo therapy produced a 31% symptom response [60], and across six adult trials Cochrane found office therapy improved fusional vergence but "no evidence of a difference for NPC or the CISS score" against placebo [59]. No trial reports headache as a separate outcome.

These tests have to be asked for because they are discretionary: the American Optometric Association's adult-examination guideline lists near point of convergence, phorias, vergence ranges and facility, and accommodative amplitude and facility as "potential components", with "professional judgment and patient symptoms" determining which are performed [61]. Every CI prevalence series above required normal acuity, so an adult can pass several refraction visits with undetected CI unless the complaint is described and binocular testing requested.

8.3 Accommodative problems and "computer glasses": plausible, unproven

Accommodative anomalies are common in near-work populations: accommodative excess in 10.8% and insufficiency in 6.2% of students [56], and poor accommodative facility the most common diagnosis, at 20.3%, in 153 symptomatic computer users [2]. Yet over 30 minutes of laptop reading at 50 cm, "no significant changes in accommodation or vergence were observed", and discomfort tracked fixation disparity instead [62]. A 2009 systematic review found that beyond CI "there is a lack of published randomized, clinical trials that support the evidence for the efficacy of each treatment" [63]. Low-plus "computer" lenses for a pre-presbyopic adult are therefore an empirical trial with no controlled headache evidence.

8.4 Migraine: stripes and flicker are real triggers, screen time is a weak one

If the headache is pulsating, unilateral, worsened by activity and accompanied by nausea or light and sound sensitivity, it meets migraine criteria and the relevant evidence changes [50]. Stripe sensitivity discriminates sharply: 82% of migraineurs against 6.2% of non-migraineurs were stripe-sensitive in a 102-person sample [64], and a review of optometric correlates calls the migraine and pattern-glare link "striking" while rating the evidence for refractive and binocular causes of migraine as poor [65]. For a migraineur, a terminal is a high-contrast striped stimulus at a few cycles per degree, so the line-spacing, size and contrast measures above are the direct countermeasures, and flicker-free room lighting has the strongest environmental evidence [38]. Screen time itself is a weak and uncertain trigger: in 4927 French students the highest screen-time quintile had an odds ratio of 1.37 for migraine and no association with non-migraine headache, cross-sectionally [66]. Tinted lenses rest on thin evidence: a double-masked crossover of precision tints in 17 adults found effects "suggestive rather than conclusive" [67]. A daily, bilateral, pressing headache without photophobia or nausea fits tension-type or refractive headache better than migraine.

8.5 Posture and tension-type headache: associated, direction unclear

Chronic tension-type headache is "daily or very frequent... typically bilateral, pressing or tightening in quality and of mild to moderate intensity", not worsened by routine activity, and ICHD-3 warns that "in many uncertain cases there is overuse of medication" [50]. A meta-analysis of 12 observational studies found chronic primary headache patients have greater forward head posture than asymptomatic people (Hedges g 0.68, 95% CI 0.25-1.1), but the studies are cross-sectional, so the posture may be a consequence of the headache as much as a cause [68]. The American Optometric Association's screen geometry (top of screen near eye level, screen 15-20 degrees below the line of sight, 50-70 cm away) is consistent with the posture literature but is professional guidance, not trial evidence [69]; a laptop on a desk forces a bent neck or a cramped keyboard, so a riser plus external keyboard, or an external monitor, is an evidence-consistent step.

8.6 When to see a physician rather than an optometrist

The SNNOOP10 list gives the red and orange flags for secondary headache: systemic symptoms, neoplasm history, neurologic deficit, sudden onset, onset after 65, pattern change or a new headache, positional headache, precipitation by sneezing, coughing or exercise, papilledema, progressive course, pregnancy, painful eye with autonomic features, post-traumatic onset, immune pathology, and painkiller overuse or a new drug at onset [70]. A daily headache that is new, progressive or accompanied by daily analgesic use belongs with a physician first regardless of eye findings. The remainder, with an afternoon pressure pattern that clears off-screen, belongs with a comprehensive eye examination that covers refraction (cycloplegic where latent hyperopia or accommodative spasm is suspected), binocular and accommodative testing, and the ocular surface [61]. A two-week diary of time of day, side, quality, nausea or photophobia, and whether the pain clears off-screen lets the clinician classify the headache under ICHD-3 before anyone changes a theme.

"Headache measured" means a study measured headache or migraine; most did not.

The ranking weighs evidence strength, expected effect size and cost. The first three items for the developer come before any theme debate.

10.1 For the developer: habits and medical follow-up, in order

10.2 For a terminal-focused IDE: default theme and layout, in order

The useful reframing is that "dark mode versus light mode" is the wrong axis. The variables that move measured outcomes are luminance reaching the eye, angular text size, line pitch relative to x-height, and whether anything in view modulates at 100 Hz; polarity is a proxy for the first and nothing else. The strongest results here are negative (blue-light lenses) or clinical (sub-dioptre astigmatism multiplying symptom scores, lighting flicker halving headaches), while the display-side folklore with the widest circulation (softened palettes, named coding fonts, 120 Hz, warm tints, 20-20-20) has either no controlled test or a failed one. The three things most likely to help a developer today are an exam that measures convergence and small cylinder, a font two sizes larger with 1.5 line height, and a five-minute walk every hour; none of them is a theme.

1. [1]
   J. K. Portello, M. Rosenfield, Y. Bababekova, J. M. Estrada, and A. Leon, "Computer-related visual symptoms in office workers," Ophthalmic Physiol. Opt., vol. 32, no. 5, pp. 375-382, Sep. 2012. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/22775070/
2. [2]
   A. L. Sheppard and J. S. Wolffsohn, "Digital eye strain: Prevalence, measurement and amelioration," BMJ Open Ophthalmol., vol. 3, no. 1, art. e000146, Apr. 2018. Accessed: Oct. 13, 2026. [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC6020759/
3. [3]
   M. Rosenfield, "Computer vision syndrome: A review of ocular causes and potential treatments," Ophthalmic Physiol. Opt., vol. 31, no. 5, pp. 502-515, Sep. 2011. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/21480937/
4. [4]
   J. S. Wolffsohn et al., "TFOS Lifestyle: Impact of the digital environment on the ocular surface," Ocul. Surf., vol. 28, pp. 213-252, Apr. 2023. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/37062428/
5. [5]
   J. E. Sheedy, J. N. Hayes, and J. Engle, "Is all asthenopia the same?," Optom. Vis. Sci., vol. 80, no. 11, pp. 732-739, Nov. 2003. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/14627938/
6. [6]
   C. Piepenbrock, S. Mayr, I. Mund, and A. Buchner, "Positive display polarity is advantageous for both younger and older adults," Ergonomics, vol. 56, no. 7, pp. 1116-1124, 2013. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1080/00140139.2013.790485
7. [7]
   C. Piepenbrock, S. Mayr, and A. Buchner, "Positive display polarity is particularly advantageous for small character sizes: Implications for display design," Hum. Factors, vol. 56, no. 5, pp. 942-951, 2014. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1177/0018720813515509
8. [8]
   C. Piepenbrock, S. Mayr, and A. Buchner, "Smaller pupil size and better proofreading performance with positive than with negative polarity displays," Ergonomics, vol. 57, no. 11, pp. 1670-1677, 2014. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1080/00140139.2014.948496
9. [9]
   A. Buchner, S. Mayr, and M. Brandt, "The advantage of positive text-background polarity is due to high display luminance," Ergonomics, vol. 52, no. 7, pp. 882-886, 2009. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1080/00140130802641635
10. [10]
    A. Buchner and N. Baumgartner, "Text-background polarity affects performance irrespective of ambient illumination and colour contrast," Ergonomics, vol. 50, no. 7, pp. 1036-1063, 2007. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1080/00140130701306413
11. [11]
    J. Dobres, N. Chahine, B. Reimer, D. Gould, B. Mehler, and J. F. Coughlin, "Utilising psychophysical techniques to investigate the effects of age, typeface design, size and display polarity on glance legibility," Ergonomics, vol. 59, no. 10, pp. 1377-1391, 2016. Accessed: Oct. 13, 2026. [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC5213401/
12. [12]
    Level Access, "Accessibility for people with astigmatism," Level Access blog, Jun. 30, 2020. Accessed: Oct. 13, 2026. [Online]. Available: https://www.levelaccess.com/blog/accessibility-for-people-with-astigmatism/
13. [13]
    P. Sengsoon and R. Intaruk, "Immediate effects of light mode and dark mode features on visual fatigue in tablet users," Int. J. Environ. Res. Public Health, vol. 22, no. 4, art. 609, 2025. Accessed: Oct. 13, 2026. [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC12027292/
14. [14]
    F. J. Pathari, Y. Nielsen, L. I. Andersen, and G. Marentakis, "Dark vs. light mode on smartphones: Effects on eye fatigue," in Proc. 17th Int. Conf. Advances in Computer-Human Interactions (ACHI), Barcelona, Spain, May 2024, pp. 150-154. Accessed: Oct. 13, 2026. [Online]. Available: https://personales.upv.es/thinkmind/dl/conferences/achi/achi_2024/achi_2024_3_150_20069.pdf
15. [15]
    X. Xie, S. Yu, and D. Chen, "Effects of screen color mode and color temperature on visual fatigue under different ambient illuminations," Int. J. Hum.-Comput. Interact., vol. 41, no. 2, pp. 821-833, 2025. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1080/10447318.2024.2305982
16. [16]
    T. Sethi and M. Ziat, "Dark mode vogue: Do light-on-dark displays have measurable benefits to users?," Ergonomics, vol. 66, no. 12, pp. 1814-1828, 2023. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1080/00140139.2022.2160879
17. [17]
    L. J. Monger, A. J. Wilkins, and P. M. Allen, "Pattern glare: The effects of contrast and color," Front. Psychol., vol. 6, art. 1651, 2015. Accessed: Oct. 13, 2026. [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC4621622/
18. [18]
    I. Juricevic, L. Land, A. Wilkins, and M. A. Webster, "Visual discomfort and natural image statistics," Perception, vol. 39, no. 7, pp. 884-899, 2010. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1068/p6656
19. [19]
    W3C, "Understanding Success Criterion 1.4.3: Contrast (Minimum)," Understanding WCAG 2.2, W3C Web Accessibility Initiative. Accessed: Oct. 13, 2026. [Online]. Available: https://www.w3.org/WAI/WCAG22/Understanding/contrast-minimum.html
20. [20]
    A. J. Wilkins and I. Nimmo-Smith, "The clarity and comfort of printed text," Ergonomics, vol. 30, no. 12, pp. 1705-1720, 1987. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1080/00140138708966059
21. [21]
    J. M. Gilchrist and P. M. Allen, "Lexical decisions in adults with low and high susceptibility to pattern-related visual stress: A preliminary investigation," Front. Psychol., vol. 6, art. 449, 2015. Accessed: Oct. 13, 2026. [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC4396132/
22. [22]
    B. J. W. Evans and S. J. Stevenson, "The Pattern Glare Test: A review and determination of normative values," Ophthalmic Physiol. Opt., vol. 28, no. 4, pp. 295-309, 2008. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/18565084/
23. [23]
    A. Wilkins, K. Smith, and O. Penacchio, "The influence of typography on algorithms that predict the speed and comfort of reading," Vision, vol. 4, no. 1, art. 18, 2020. Accessed: Oct. 13, 2026. [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC7175152/
24. [24]
    S. T. L. Chung, "Reading speed benefits from increased vertical word spacing in normal peripheral vision," Optom. Vis. Sci., vol. 81, no. 7, pp. 525-535, 2004. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/15252352/
25. [25]
    M. C. Dyson, "How physical text layout affects reading from screen," Behav. Inf. Technol., vol. 23, no. 6, pp. 377-393, 2004. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1080/01449290410001715714
26. [26]
    G. E. Legge and C. A. Bigelow, "Does print size matter for reading? A review of findings from vision science and typography," J. Vis., vol. 11, no. 5, art. 8, pp. 1-22, 2011. Accessed: Oct. 13, 2026. [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC3428264/
27. [27]
    S. Wallace et al., "Towards individuated reading experiences: Different fonts increase reading speed for different individuals," ACM Trans. Comput.-Hum. Interact., vol. 29, no. 4, art. 38, pp. 1-56, 2022. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1145/3502222
28. [28]
    JetBrains, "JetBrains Mono: A new font made for developers," JetBrains blog, Jan. 15, 2020. Accessed: Oct. 13, 2026. [Online]. Available: https://blog.jetbrains.com/blog/2020/01/15/jetbrains-mono-a-new-font-made-for-developers/
29. [29]
    Nestify, "Top 10 monospaced fonts for coders," Nestify blog. Accessed: Oct. 13, 2026. [Online]. Available: https://nestify.io/blog/top-10-monospaced-fonts-for-coders/
30. [30]
    D. Yu, S.-H. Cheung, G. E. Legge, and S. T. L. Chung, "Effect of letter spacing on visual span and reading speed," J. Vis., vol. 7, no. 2, art. 2, pp. 1-10, 2007. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1167/7.2.2
31. [31]
    D. Rempel, K. Willms, J. Anshel, W. Jaschinski, and J. Sheedy, "The effects of visual display distance on eye accommodation, head posture, and vision and neck symptoms," Hum. Factors, vol. 49, no. 5, pp. 830-838, Oct. 2007. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/17915601/
32. [32]
    W. Jaschinski-Kruza, "Eyestrain in VDU users: Viewing distance and the resting position of ocular muscles," Hum. Factors, vol. 33, no. 1, pp. 69-83, 1991. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1177/001872089103300106
33. [33]
    S. Singh, L. E. Downie, and A. J. Anderson, "Do blue-blocking lenses reduce eye strain from extended screen time? A double-masked randomized controlled trial," Am. J. Ophthalmol., vol. 226, pp. 243-251, Jun. 2021. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/33587901/
34. [34]
    S. Singh et al., "Blue-light filtering spectacle lenses for visual performance, sleep, and macular health in adults," Cochrane Database Syst. Rev., no. 8, art. CD013244, Aug. 2023. Accessed: Oct. 13, 2026. [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC10436683/
35. [35]
    K. M. Duraccio, K. K. Zaugg, R. C. Blackburn, and C. D. Jensen, "Does iPhone night shift mitigate negative effects of smartphone use on sleep outcomes in emerging adults?," Sleep Health, vol. 7, no. 4, pp. 478-484, Aug. 2021. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/33867308/
36. [36]
    International Organization for Standardization, "Ergonomics of human-system interaction, Part 303: Requirements for electronic visual displays," ISO 9241-303 (2011 edition), public preview, Nov. 2011. Accessed: Oct. 13, 2026. [Online]. Available: https://www.iso.org/standard/57992.html
37. [37]
    Q. Fan, J. Xie, Z. Dong, and Y. Wang, "The effect of ambient illumination and text color on visual fatigue under negative polarity," Sensors, vol. 24, no. 11, art. 3516, 2024. Accessed: Oct. 13, 2026. [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC11175232/
38. [38]
    A. J. Wilkins, I. Nimmo-Smith, A. I. Slater, and L. Bedocs, "Fluorescent lighting, headaches and eyestrain," Light. Res. Technol., vol. 21, no. 1, pp. 11-18, 1989. Accessed: Oct. 13, 2026. [Online]. Available: https://doi.org/10.1177/096032718902100102
39. [39]
    IEEE, "IEEE Recommended Practices for Modulating Current in High-Brightness LEDs for Mitigating Health Risks to Viewers," IEEE Std 1789-2015, IEEE Standards Association, Jun. 2015. Accessed: Oct. 13, 2026. [Online]. Available: https://standards.ieee.org/ieee/1789/4479/
40. [40]
    U.S. Department of Energy Solid-State Lighting Program, "A much-needed recommended practice for LED flicker," SSL Postings, Jun. 11, 2015. Accessed: Oct. 13, 2026. [Online]. Available: https://www.energy.gov/sites/prod/files/2015/06/f22/postings_06-11-15.pdf
41. [41]
    N. J. Miller, L. Irvin, M. Royer, and C. Strachan, "Visibility and annoyance of the phantom array effect varies with age and history of migraine," Light. Res. Technol., vol. 56, no. 7, pp. 676-706, 2024. Accessed: Oct. 13, 2026. [Online]. Available: https://www.pnnl.gov/publications/visibility-and-annoyance-phantom-array-effect-varies-age-and-history-migraine
42. [42]
    A. Osthoff, "Apple MacBook Pro 14 2024 review: The M4 Pro and matte display are massive upgrades," Notebookcheck, Nov. 8, 2024. Accessed: Oct. 13, 2026. [Online]. Available: https://www.notebookcheck.net/Apple-MacBook-Pro-14-2024-review-The-M4-Pro-and-matte-display-are-massive-upgrades.914597.0.html
43. [43]
    Notebookcheck, "Samsung's OLED panel in the Galaxy Book5 Pro 14 falls behind the competition," Notebookcheck, Oct. 2025. Accessed: Oct. 13, 2026. [Online]. Available: https://www.notebookcheck.net/Samsung-s-OLED-panel-in-the-Galaxy-Book5-Pro-14-falls-behind-the-competition.1133101.0.html
44. [44]
    J. K. Portello, M. Rosenfield, and C. A. Chu, "Blink rate, incomplete blinks and computer vision syndrome," Optom. Vis. Sci., vol. 90, no. 5, pp. 482-487, May 2013. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/23538437/
45. [45]
    S. Liu, H. Dong, S. Fang, and L. Zhang, "Risk of dry eye in headache patients: A systematic review and meta-analysis," Ann. Med., vol. 54, no. 1, pp. 2876-2885, 2022. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/36259538/
46. [46]
    C. Talens-Estarelles, A. Cervino, S. Garcia-Lazaro, A. Fogelton, A. Sheppard, and J. S. Wolffsohn, "The effects of breaks on digital eye strain, dry eye and binocular vision: Testing the 20-20-20 rule," Cont. Lens Anterior Eye, vol. 46, no. 2, art. 101744, Apr. 2023. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/35963776/
47. [47]
    S. Johnson and M. Rosenfield, "20-20-20 rule: Are these numbers justified?," Optom. Vis. Sci., vol. 100, no. 1, pp. 52-56, Jan. 2023. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/36473088/
48. [48]
    T. L. Galinsky, N. G. Swanson, S. L. Sauter, J. J. Hurrell, and L. M. Schleifer, "A field study of supplementary rest breaks for data-entry operators," Ergonomics, vol. 43, no. 5, pp. 622-638, May 2000. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/10877480/
49. [49]
    R. Balci and F. Aghazadeh, "The effect of work-rest schedules and type of task on the discomfort and performance of VDT users," Ergonomics, vol. 46, no. 5, pp. 455-465, Apr. 2003. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/12745696/
50. [50]
    Headache Classification Committee of the International Headache Society, "The International Classification of Headache Disorders, 3rd edition (ICHD-3)," Cephalalgia, vol. 38, no. 1, pp. 1-211, Jan. 2018. Accessed: Oct. 13, 2026. [Online]. Available: https://ichd-3.org/
51. [51]
    H. Lajmi, R. Choura, B. Ben Achour, M. Doukh, Z. Amin, and W. Hmaied, "Headache associated with refractive errors: Characteristics and risk factors," Rev. Neurol. (Paris), vol. 177, no. 8, pp. 947-954, Oct. 2021. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/33483090/
52. [52]
    R. Gil-Gouveia and I. P. Martins, "Headaches associated with refractive errors: Myth or reality?," Headache, vol. 42, no. 4, pp. 256-262, Apr. 2002. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/12010381/
53. [53]
    N. P. Wiggins and K. M. Daum, "Visual discomfort and astigmatic refractive errors in VDT use," J. Am. Optom. Assoc., vol. 62, no. 9, pp. 680-684, Sep. 1991. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/1815002/
54. [54]
    M. Rosenfield, J. E. Hue, R. R. Huang, and Y. Bababekova, "The effects of induced oblique astigmatism on symptoms and reading performance while viewing a computer screen," Ophthalmic Physiol. Opt., vol. 32, no. 2, pp. 142-148, Mar. 2012. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/22150631/
55. [55]
    S. Chhetri et al., "Prevalence and clinical profile of non-strabismic binocular vision anomalies in the Nepalese population: A hospital-based study," J. Optom., vol. 18, no. 4, art. 100575, 2025. Accessed: Oct. 13, 2026. [Online]. Available: https://pmc.ncbi.nlm.nih.gov/articles/PMC12355990/
56. [56]
    E. Porcar and A. Martinez-Palomera, "Prevalence of general binocular dysfunctions in a population of university students," Optom. Vis. Sci., vol. 74, no. 2, pp. 111-113, Feb. 1997. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/9097328/
57. [57]
    M. W. Rouse et al., "Validity and reliability of the revised convergence insufficiency symptom survey in adults," Ophthalmic Physiol. Opt., vol. 24, no. 5, pp. 384-390, Sep. 2004. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/15315652/
58. [58]
    Convergence Insufficiency Treatment Trial Study Group, "Randomized clinical trial of treatments for symptomatic convergence insufficiency in children," Arch. Ophthalmol., vol. 126, no. 10, pp. 1336-1349, Oct. 2008. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/18852411/
59. [59]
    M. Scheiman, M. T. Kulp, S. A. Cotter, J. G. Lawrenson, L. Wang, and T. Li, "Interventions for convergence insufficiency: A network meta-analysis," Cochrane Database Syst. Rev., no. 12, art. CD006768, Dec. 2020. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/33263359/
60. [60]
    M. Scheiman et al., "A randomized clinical trial of vision therapy/orthoptics versus pencil pushups for the treatment of convergence insufficiency in young adults," Optom. Vis. Sci., vol. 82, no. 7, pp. 583-595, Jul. 2005. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/16044063/
61. [61]
    American Optometric Association, "Comprehensive adult eye and vision examination: Quick reference guide," 1st ed., AOA Evidence-Based Optometry, 2015. Accessed: Oct. 13, 2026. [Online]. Available: https://www.aoa.org/documents/EBO/Comprehensive_Adult_Eye_and_Vision%20QRG.pdf
62. [62]
    J. D. Collier and M. Rosenfield, "Accommodation and convergence during sustained computer work," Optometry, vol. 82, no. 7, pp. 434-440, Jul. 2011. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/21514899/
63. [63]
    P. Cacho Martinez, A. Garcia Munoz, and M. T. Ruiz-Cantero, "Treatment of accommodative and nonstrabismic binocular dysfunctions: A systematic review," Optometry, vol. 80, no. 12, pp. 702-716, Dec. 2009. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/19932444/
64. [64]
    D. A. Marcus and M. J. Soso, "Migraine and stripe-induced visual discomfort," Arch. Neurol., vol. 46, no. 10, pp. 1129-1132, Oct. 1989. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/2803073/
65. [65]
    D. E. Harle and B. J. W. Evans, "The optometric correlates of migraine," Ophthalmic Physiol. Opt., vol. 24, no. 5, pp. 369-383, Sep. 2004. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/15315651/
66. [66]
    I. Montagni, E. Guichard, C. Carpenet, C. Tzourio, and T. Kurth, "Screen time exposure and reporting of headaches in young adults: A cross-sectional study," Cephalalgia, vol. 36, no. 11, pp. 1020-1027, Oct. 2016. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/26634831/
67. [67]
    A. J. Wilkins, R. Patel, P. Adjamian, and B. J. W. Evans, "Tinted spectacles and visually sensitive migraine," Cephalalgia, vol. 22, no. 9, pp. 711-719, Nov. 2002. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/12421156/
68. [68]
    I. Elizagaray-Garcia, H. Beltran-Alacreu, S. Angulo-Diaz, M. Garrigos-Pedron, and A. Gil-Martinez, "Chronic primary headache subjects have greater forward head posture than asymptomatic and episodic primary headache sufferers: Systematic review and meta-analysis," Pain Med., vol. 21, no. 10, pp. 2465-2480, Oct. 2020. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/33118601/
69. [69]
    American Optometric Association, "Computer vision syndrome," aoa.org. Accessed: Oct. 13, 2026. [Online]. Available: https://www.aoa.org/healthy-eyes/eye-and-vision-conditions/computer-vision-syndrome
70. [70]
    T. P. Do et al., "Red and orange flags for secondary headaches in clinical practice: SNNOOP10 list," Neurology, vol. 92, no. 3, pp. 134-144, Jan. 2019. Accessed: Oct. 13, 2026. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/30587518/