In this study, we observed that a child's age and preference for ingesting carotenoid-containing foods, particularly green leaves, can influence the perception of an ambiguous image like the dress. As hypothesized, these results suggest a potential relationship with the gradual accumulation of MPs throughout a child's lifetime.
Lutein, zeaxanthin, and meso-zeaxanthin, the constituents of MPs, accumulate at high concentrations in the fovea, constituting the majority of retinal carotenoids (
Landrum & Bone, 2001). Various tissues, including the retina, rely on the dietary intake of these carotenoids for their accumulation. Their concentrations have been identified in maternal blood and umbilical cord blood samples, with lutein and zeaxanthin prevailing in these tissues despite not being the most abundant carotenoids in serum or diet (
Thoene et al., 2019). Additionally, further research has highlighted these two carotenoids as the primary ones in human milk (
Canfield et al., 2003;
Lipkie, Morrow, Jouni, McMahon, & Ferruzzi, 2015).
As individuals age, the levels of MPs tend to increase. Notably, in newborn and fetal human eyes, a low ratio of meso-zeaxanthin to lutein/zeaxanthin has been observed, suggesting potential insufficiency and/or inefficiency in isomerization (
Krinsky et al., 2003). The earliest measurements of MPs in newborns were reported by Sasano et al., who recorded an MPOD of 0.05 at 33 weeks of postmenstrual age (PMA) and a mean first measurable MPOD at 0.076. Their study demonstrated a significant correlation between MPOD levels and PMA, indicating that MPOD increases with age in premature infants (
R² = 0.91,
p = 0.0001). By 40 weeks of PMA, the mean MPOD reached 0.1033 (
Sasano et al., 2018). Bernstein et al. reported a steady rise in MPOD over the first 7 years of life, ranging from 0.05 to nearly 0.4, with linear regression analysis predicting an MPOD of 0.0835 at birth. The MPOD levels eventually reached those found in adult populations (
Bernstein et al., 2013). These observations support our hypothesis that older children are more likely to perceive #TheDress as WG, potentially due to a higher accumulation of MPs. This aligns with findings from Lafer-Sousa et al., who reported that older individuals and women were more likely to perceive the image as WG (
Lafer-Sousa, Hermann, & Conway, 2015). It is noteworthy that MPs continue to increase during adulthood, reaching maximal levels between 30 years (
Hong, Jung, Lee, & Chang, 2020) and 55 years (
van der Veen et al., 2009), followed by a gradual decline in older ages. Van der Veen et al. reported higher MPODs in women compared to men across all age groups, except for those aged 80 and above. This observation parallels the tendency of women to perceive more WG, aligning with the decline in WG perceptions through age reported by Gonzalez Martin-Miron and Wallisch, as further discussed in the following sections.
Research indicates that MP carotenoids contribute to improved visual and cognitive performance across various age groups. A meta-analysis of 9 clinical trials demonstrated that carotenoid supplementation enhances cognitive performance in adults aged 45 to 78 years (
Davinelli, Ali, Solfrizzi, Scapagnini, & Corbi, 2021). Studies on young adults showed improved visual task performance with glare disability and photostress recovery after 6 months of carotenoid supplementation (
Stringham & Hammond, 2008). Furthermore, a randomized, double-masked, placebo-controlled trial revealed that lutein and zeaxanthin supplementation in healthy young adults led to enhancements in visual memory, complex attention, and reasoning ability. Notably, visual memory exhibited the most significant improvement, suggesting the influence of these carotenoids on functions predominantly mediated by the occipital cortex, a region accumulating lutein and zeaxanthin (
Renzi-Hammond et al., 2017). In children, MPOD positively correlated with cognitive control performance (
Walk et al., 2017). In a prospective cohort study, higher maternal intake of lutein and zeaxanthin during the first and second trimesters of pregnancy was associated with better verbal intelligence, behavioral regulation ability in mid-childhood, and improved social-emotional development (
Mahmassani et al., 2021). Additionally, higher maternal zeaxanthin concentrations were linked to a lower risk of poor visual acuity in children (
Lai et al., 2020).
The perception of #TheDress appears to be influenced by early-stage optical, retinal, and neural factors, as proposed by
Rabin et al. (2016). Adult subjects with higher MP levels tended to perceive the image as white and gold, demonstrating differences in processing times, possibly attributed to varying MP concentrations. Visual evoked potentials further indicated distinct neural responses for those perceiving the image as white and gold, suggesting a correlation with MP levels (
Rabin et al., 2016). fMRI studies on #TheDress revealed increased brain activation in regions such as the middle frontal gyrus, inferior and superior parietal lobule, middle temporal gyrus, and inferior frontal gyrus among individuals perceiving the dress as WG (
Schlaffke et al., 2015). The occipital and frontal cortex, which contain lutein and zeaxanthin, exhibited enhanced activation in these studies, emphasizing the association between carotenoids and specific brain regions crucial for visual processing and higher cognitive function (
Vishwanathan, Neuringer, Snodderly, Schalch, & Johnson, 2013). This correlation suggests that lutein and zeaxanthin may play a role in promoting brain health and cognition by supporting neural structures and enhancing neuronal efficiency in regions involved in visual perception and decision-making (
Mewborn et al., 2018).
Other ocular variables, such as hyperopia, cataract density, and age, have been previously associated with the perception of #TheDress (
González-Martín-Moro et al., 2021). However, our observations differ from those reported by Gonzalez Martin-Moro et al., where all three variables were associated with perceiving the image as BB. In their study involving a sample of 1,092 subjects aged 2 to 99 years, the authors noted that in the first decade of life, perceived colors were approximately 60% WG and 40% BB (we reported 50.6% WG and 49.3% BB). The authors were unable to explain the association of age with BB perception and concluded that ocular examination did not contribute to understanding the visual phenomenon. Here, we propose the gradual accumulation of MPs during childhood as a potential factor associated with perception. This hypothesis aligns with the concept of stability in perception as children reach MP optical density levels comparable to adults, as reported by
Bernstein et al. (2013). The life cycle analysis of #TheDress perception supports this idea, demonstrating variations through age. In a cohort of 7,868 subjects, those perceiving the dress as WG reached a proportion of around 60% after adolescence, remaining stable for approximately 20 years. Subsequently, the proportion dropped to 30% between the ages of 65 and 75 (
Wallisch, 2017), a decline also observed after the sixth decade of life in Gonzalez-Martin-Moro's report of 1,025 subjects (
Gonzalez-Martin-Moro et al., 2021). Notably, our study suggests an apparent increase in WG perceptions as age advances from 2 to 10 years, prompting interest in understanding the trend beyond the age of 10.
In examining the relationship between GLP and the perception of #TheDress as WG, it is crucial to note that green-leafy vegetables serve as the primary source of lutein and zeaxanthin. These carotenoids, crucial for MPs, are found in abundance in green leaves, making them significant contributors to visual processes (
Böhm et al., 2021;
Estévez-Santiago, Beltrán-de-Miguel, & Olmedilla-Alonso, 2016;
Franke, Fröhlich, Werner, Böhm, & Schöne, 2010;
Sommerburg, Keunen, Bird, & van Kuijk, 1998). While oils, eggs, and fish contribute to carotenoid intake, they appear to be less significant in comparison (
Nolan et al., 2014). In our study, GLP was chosen as the primary outcome for analysis, given the crucial role of green-leafy vegetables as principal contributors to MPs. Even though TCP did not yield a statistically significant difference, GLP showed a significant correlation with perceiving #TheDress as WG. This aligns with previous research indicating positive correlations between dietary lutein and zeaxanthin, serum levels, and MPOD, particularly driven by vegetable and fruit consumption (
Burke, Curran-Celentano, & Wenzel, 2005). Burke et al. reported that individuals with lower fruit and vegetable consumption had lower MPOD, while those with higher consumption exhibited higher MPOD levels (
r = 0.35,
p = 0.001). Additionally, their study highlighted a decrease in MPOD with age, reinforcing the link between age and macular health (
Burke et al., 2005). Observing GLP provides valuable insights into macular health and, in this context, correlates with the perception of #TheDress, offering a glimpse into the potential influence of dietary factors on visual experiences.
Our study has several limitations. First, it employs a cross-sectional analysis, and second, the sample size is relatively small when compared to larger Internet surveys (
Lafer-Sousa et al., 2015;
Wallisch, 2017). These issues may lead to a cohort effect, potentially affecting our sample of children, particularly in the age analysis. Third, while we believe this study provides crucial insights into visual perception and potential mechanisms related to MPs, it may be considered a “low-level” effect influencing the colors of the dress (
Wallisch, 2017). It has been suggested that almost 70% of factors influencing the perception of the dress could be environmental rather than genetic (
Mahroo et al., 2017). Some of the “low-level” factors influencing the sensory end of perception, as proposed by Wallisch, could be the MPOD, as suggested by
Rabin et al. (2016). Our study aligns with these conclusions. However, results from these low-level factors neither explain nor demonstrate a high correlation with the perception. In contrast, the perceived direction of the light source illuminating the dress has been linked to the precept (
Chetverikov & Ivanchei, 2016;
Wallisch, 2017), where an assumed shadow cast on the dress would make it appear “yellower” through color constancy mechanisms (
Wallisch, 2017;
Witzel, Racey, & O'Regan, 2017). Although we did not inquire about the children's assumptions regarding the light direction on the dress in our study, analyzing this in a cohort of children could be challenging, especially with 2-year-olds.
Another factor found by
Wallisch (2017) to be related “in a dose-dependent fashion” is the circadian type of each subject, classified as either “owls” or “larks.” This observation also relates to the perception of the illumination on the dress. Exploring how the circadian rhythm of children changes and transitions from “larks” toward “owls” (
Randler, Faßl, & Kalb, 2017) and its relation to their perception could shed light on the observed shift toward WG in our sample. These findings may also be related to the effects achieved by action potentials emerging from intrinsically photosensitive retinal ganglion cells (ipRGCs). Through their absorption of blue wavelengths and using melanopsin, ipRGCs are known to influence conscious visual perception, circadian rhythm, and photo-entrainment (
Dacey et al., 2005;
Do, 2019;
Rabin et al., 2023).
Finally, could some of the mechanisms proposed in our study affect perception in older ages? As observed, a decrease in WG perceptions happens around 60 years of age (
Wallisch, 2017). Older age and the grade of nuclear opacity were associated with BB perception. Additionally, BB perceptions were correlated with unhealthy ocular conditions, such as glaucoma, AMD, and other retinal diseases (Person's correlation coefficient 0.64) (
Gonzalez-Martin-Moro et al., 2021). Could a decrease in MPs at an older age partially influence BB perceptions? This question warrants investigation through direct MP measurements, as observed in cohorts, such as patients with mild to moderate Alzheimer's disease, displaying lower levels of MPs, poorer visual function, and an increased risk of AMD (
Nolan, Loskutova et al., 2014).