The goal of this study was to determine whether reading in color could be used to acquire letter-color associations and whether those learned associations are expressed in behavior similar to that found in grapheme-color synesthetes. Specifically, this study tested certain synesthetic traits, such as automaticity, perceptual experiences, and letter-color pair recollection. Fifteen participants read specially prepared colored books and were tested with a modified Stroop task, a perceptual crowding task, and were asked to report about their experiences of reading in color. In addition, we contacted the participants months later to test whether they remembered their letter-color pairs. The results show, for the first time to our knowledge, that letter-color associations can be formed by reading in color. Furthermore, the level of learning these associations is influenced by the frequency of exposure as can be seen by an increase in the size of the Stroop effect for those participants who continued to read, as well as the difference in the size of the Stroop effects between upper- and lower-case letter conditions. The degree to which an individual reports internalizing these letter-color associations was reflected in behavior on the Stroop task. Unlike synesthetes, who are defined by their high consistency of grapheme-color pairs and performance advantage during a perceptual crowding task, readers do not remember their letter-color pairs several months later nor do they show a significant improvement on the crowding task compared to controls.
One must keep in mind that these results alone are not enough to conclude the presence of synesthesia, since over-learned associations produce a Stroop effect. Learned associations might not be conscious or perceptual in nature as is the case in synesthesia. Interestingly, the questionnaire data from participants correlated positively with the size of this Stroop effect in such a way that people who reported being more likely to internally experience the colored letters tend to show a larger Stroop effect. The demand characteristics of the study as well as the question itself may have influenced the participants to think about the letters in color. However, a significant correlation with the reaction time data is evidence to suggest that the self-report ratings of internal experiences are indeed related to learning letter-color associations. Individual differences are immediately apparent in the data. Some individuals who read relatively more did not show as large an effect as other individuals who read relatively less. Furthermore, word count did not correlate with the Stroop effect. We can infer that some individuals are inherently more or less sensitive to forming these associations than others. During the course of this research, we began to wonder if perhaps an individual’s visual mental imagery is a key factor in understanding the differences in sensitivity to forming letter-color associations. We are currently investigating this hypothesis. Individual differences are a major theme in synesthesia research; the term ‘synesthesia’ encompasses many varieties of the phenomenon. Within grapheme-color synesthesia, synesthetes are commonly classified as ‘projectors’ or ‘associators’ depending on the spatial localization of the synesthetic experience 
. In addition to projector and associator classifications, grapheme-color synesthetes can be categorized into ‘higher’ or ‘lower’ synesthetes depending on what induces the synesthetic experiences (e.g. ‘A’ may or may not induce a different experience than ‘a’) 
. Even within these categorizations of grapheme-color synesthesia, much variation in the specific qualities of the experiences exists. The best example of this is the fact that the specific colors (and also textures) of the graphemes vary across people. In addition, some individuals experience their grapheme-color synesthesia from visual stimuli only or from auditory stimuli only, while it is also possible that both modalities are inducers, and that the mental images of the graphemes are inducers. Despite the wide range of differences found between synesthetes, there are some general patterns present within larger samples of synesthetes. For example, frequent graphemes tend to induce common colors 
and correlate with color saturation 
and luminance 
. For this reason, we choose four high-frequency graphemes and paired them with four high-frequency colors. We expected that the frequency of viewing letters plays a role in acquiring letter-color associations through reading in color.
In line with our hypotheses about frequency effects, differences were apparent in the data between upper- and lower-case letters. The Stroop effect was present in both upper- and lower-case versions of the same letter. However, the effect was significantly larger for lower-case letters, which are more common in novels, suggesting that the frequency of exposure plays a role. An additional explanation, which is not necessarily mutually exclusive, is that the letter-color association for the lower-case version of a letter is transferred via its semantic relationship to the upper-case version of the same letter. We could not test this hypothesis directly in the current study. In addition, we hypothesize that differences in learning color associations for certain letters may be explained by the differences in letter-position frequency of those letters. Upper-case letters are typically found at the beginning of words. In contrast, lower-case letters are found in all the letter positions within words. The frequencies of letter position may interact with the relative letter frequencies of the individual letters. For example, ‘a’ is less frequent than ‘e’, ‘s’, or ‘t’, yet appears most often as the first letter in the English language. There is evidence to suggest that visual attention while reading differs for letter positions 
. The current data suggest that this interaction may have an effect on the formation of these letter-color associations, but further experiments are necessary in order to fully understand the nature of such an effect.
A crowding task was administered to test whether reading in color could induce low-level visual effects. Although our results did not reach significance, there is evidence to suggest that reading in color may influence performance on the crowding task. Our crowding stimulus was presented for a very short duration (100 ms) and was not easily recognizable, although there are some limitations to this study. One limitation was that we were not able to monitor the participants’ eye movements in order to exclude trials where a saccade to the stimuli was made. Another limitation may be that the letters were presented in their upper-case forms; an upper-case font ensured equal visual angles in the horizontal and vertical axes in order to preserve the effect of crowding and minimize pop-out due to differences in the visual forms of the letters. The Stroop effect was present for upper-case letters, but was significantly less that in the lower-case condition, and this fact could have hindered the results of the crowding task. None of the participants reported any projector-type of experiences; therefore, a trend in the crowding effect is worth pursuing further to determine if such a perceptual effect could be induced by over-learned associations, for example with lower-case letters or with extended training. It is worth noting here that differences between types of synesthetes on the crowding task were also noted in the original research by Hubbard et al. 
. One synesthetic participant reported identifying the crowded grapheme on the basis of its color, even though he could not identify the grapheme. While another participant reported that she must first identify the grapheme to experience the synesthetic color. This same participant did not show any performance improvements on the perceptual crowding or embedded figures tasks compared to a group of controls. Again, the variation between synesthetes makes the task of defining ‘genuine synesthesia’ difficult, and one must rely on a set of measures in addition to perceptual tasks, including consistency, to determine if someone is synesthetic or not. In our sample, participants did not recall on average half of their four letter-color pairs correctly several months after testing, although they did recall the four colors used almost perfectly. This can immediately be contrasted with the high-level of consistency seen with genuine synesthesia, where an individual will describe for each grapheme in detail the exact same color each time he or she is asked, with many months to years in between.
What is very special about reading in color is the fact that the effects found are a ‘byproduct’ of the reading process. Although learning by reading in color is not implicit in the strict sense of the term (they are indeed aware of the letter-color pairs), the participants were never instructed to remember the letter-color pairs. It can be seen by the results of their letter-color pair recollection test that they do not remember these pairs well several months later. This begs the question: if people are able to ‘implicitly’ learn letter-color associations by reading in color, why do most synesthetic colors seem to be randomly assigned to their inducer? For example, given the prevalence of McDonald’s in today’s society, why do we not all (Americans especially) think of ‘M’ as yellow? A study on children’s books found little evidence to suggest that the synesthetic colors of a sample of Australian synesthetes came directly from exposure to the specific letter- and digit-color pairs contained in these books 
. One difference between a synesthete and a trained control, which is stressed in the study from Elias et al. 
, is that synesthetes generally do not know where their colors come from, while the control does. However, this is not always true. In one case, specific synesthetic colors were traced back to refrigerator magnets in the individual’s childhood 
. In another report, two twin boys have similar synesthetic colors as a shared jigsaw toy 
. A synesthete reported to one of the present authors that her colors were identical to her kindergarten alphabet. She was shocked to realize this years later as an adult when she went back into the classroom of her hometown and saw her colors as the alphabet on the classroom wall. It seems plausible that this happens in more cases than are reported or remembered, considering that grapheme-color synesthesia is present since early-childhood. If synesthetic colors stem directly from the environment, even in a small minority of cases, the phenomenon is crucial for understanding why synesthesia develops in some people and why some people seem to be more sensitive to certain stimuli than others. Furthermore, in order to understand how grapheme-color synesthesia may develop in children, we must take into account that the mechanisms underlying how a child learns to read and write are inherently different from those processes by which literate adults form letter-color associations through reading in color.
Reading in color has much potential for a long-term training program, probably more than directed training with a computer. Overall, the participants found it a pleasant and interesting experience and gladly took the opportunity to continue reading in color when asked. It is still unclear whether reading in color utilizes the same underlying neural mechanisms as the experience of grapheme-color synesthesia. Studies have shown differences in the structure of the brains of synesthetes, including difference in grey-matter density 
and white-matter connectivity 
. Structural differences are found in sensory cortices, parietal cortex, frontal cortex, and hippocampus (for a review, see 
). We are currently investigating to what extent changes in the structural and functional properties of the brain can be induced by reading in color. The human brain has surprising plasticity even in adulthood as seen by recent research on experience-induced neural plasticity from learning or practicing traits, such as changes in grey matter induced by learning how to juggle 
, studying for a medical school exam 
, meditation 
, and aerobic exercise in the elderly 
. Recently, it has been shown that learning new color names increases grey-matter volume in color areas V2/3 in left visual cortex 
. Changes in white matter have also been found due to experience-dependent plasticity 
. In light of these findings, it seems plausible that training synesthetic associations might induce changes in the structural and functional properties of the brain. It will be interesting to know if these changes will be similar or in line with the neurological differences between synesthetes and controls that have already been documented (e.g. 
This study, for the first time, tested reading in color as a method for acquiring associations between letters and color. Behavioral effects due to reading in color can be induced in adults, in a mostly implicit manner and in some cases synesthesia-like experiences are reported. Although these trainees are not synesthetic in the traditional sense, many similarities exist and we conclude that reading in color is a promising method that may be used to explore both the differences and similarities between synesthetes and people who have been trained on cross-modal associations.