Cerebral Palsy (CP) designates a large group of motor and sensory impairments caused by a non-progressive lesion of the brain that occurs early in life [1]. These defects are permanent but may exhibit some plasticity. The most frequent and severe motor impairment in such children is spastic cerebral palsy (SCP) [2, 3]. Cognitive alterations, mental retardation, epilepsy, and hearing loss are frequently associated with SCP [4]. Additionally, ophthalmological disturbances, such as oculomotor abnormalities and visual acuity loss, are often observed [5]. Higher prevalence of refractive error than in normal children has consistently been found in SCP children [6, 7]. Hypoxia/anoxia insults have been associated with this higher prevalence. Cortical visual impairment is also found in children with SCP. The most often cortical damage in these children is found in retrochiasmatic visual pathways [8–10].

In children with SCP, the degree of motor impairment has been shown to be positively correlated with the loss of visual acuity, in which more severely affected children have worse visual acuity than less-affected children, measured by the Visual Acuity Card Test [11]. We confirmed these previous results by correlating visual acuity measured by sweep visual evoked potential (sweepVEP) with Gross Motor Function Classification System (GMFCS) [12] score [13–15]. Recently, visual functions related to higher-order cortical processing were also found to be impaired in SCP children. The most frequent impairments were in visually guided object recognition [16], eye-hand coordination [17], language comprehension, visuospatial abilities, and visual memory [18, 19].

Although these previous studies found retrochiasmatic visual pathway impairments and visual acuity loss, no relationship was empirically observed between visual pathway impairments and the reduction of visual acuity [8]. Structural impairment has been suggested in frontal, parietal, and temporal cortical areas based on functional damage in visual tasks, including visuospatial abilities [19–21], ocular posture [22], and eye movement control during fixation [23, 24]. As mentioned above, we found that grating acuity is impaired in SCP children using a sweepVEP paradigm [13–15] and behavioral methods [25].

Color perception has been previously investigated in SCP children in the works of Sakuma [26] and Kozeis et al. [27]. In both studies, color discriminations were analyzed based on simple ordering or figure-found tasks and did not found impairments in SCP children.

In the present study, we measured chromaticity discrimination in spastic CP children for the protan, deutan, and tritan color confusion axes. Our hypothesis was that chromatic vision, similar to visual acuity, is generally impaired and positively correlated with the degree of motor impairment since visual pathway damage would probably affect other functions than spatial vision. Although SCP impairment is attributable to cortical and/or subcortical damage, only post-receptoral visual pathways were expected to be affected and, in this case, a red-green, blue-yellow or general impairment.

A small negative correlation was found between age and chromaticity thresholds in the control group, suggesting an increase in color discrimination performance with age. The opposite result was found for the SCP children, in which a positive correlation was found for all of the visual conditions. The hemiplegic group had almost no correlation. These data are summarized in Table 3.Positive correlations were also found between chromaticity thresholds and GMFCS scores within groups, indicating that worse motor impairment was associated with worse color discrimination (see Figure 2 for binocular data and Figure 3 for monocular data).The magnitude of color discrimination loss was obtained by normalizing the data from SCP children. Tetraplegic children had higher color discrimination loss, with an average of five-times worse discrimination than control children. Diplegic and hemiplegic children had similar losses, with an average of two-times worse discrimination (Figure 4).

	Binbocular	Monocular OD	Monocular OS
Correlation Coefficient*
Controls
Protan axis	r = -0.11 (0.041)	r = -0.16 (0.023)	r = -0.06 (0.066)
Deutan axis	r = -.023 (0.008)	r = -0.11 (0.040)	r = -0.17 (0.021)
Tritan axis	r = -0.14 (0.031)	r = -0.04 (0.074)	r = -0.18 (0.019)
Tetraplegic
Protan axis	r = 0.70 (0.041)	r = 0.69 (0.033)	r = 0.90 (0.033)
Deutan axis	r = 0.30 (0.038)	r = 0.10 (0.045)	r = 0.10 (0.056)
Tritan axis	r = 0.00 (0.036)	r = 0.06 (0.079)	r = 0.01 (0.109)
Diplegic
Protan axis	r = 0.48 (0.047)	r = 0.30 (0.053)	r = 0.34 (0.039)
Deutan axis	r = 0.42 (0.049)	r = 0.49 (0.042)	r = 0.26 (0.045)
Tritan axis	r = 0.43 (0.038)	r = 0.36 (0.043)	r = 0.35 (0.041)
Hemiplegic
Protan axis	r = 0.36 (0.066)	r = 0.25 (0.079)	r = 0.04 (0.049)
Deutan axis	r = 0.33 (0.052)	r = -0.07 (0.067)	r = 0.05 (0.064)
Tritan axis	r = 0.11 (0.061)	r = 0.21 (0.059)	r = 0.10 (0.061)

*Correlation coefficient (p-value).

No differences were found between the visual conditions (monocular and binocular) in control and SCP children. Also no differences were found within SCP group considering etiology and prematurity.

Color vision is impaired in SCP children. We evaluated chromaticity discrimination along the protan, deutan, and tritan color confusion axes. In all of these axes, discrimination thresholds were reduced compared with control children. The tetraplegic group was the most affected, exhibiting higher thresholds compared with the diplegic and hemiplegic groups. To our knowledge, no studies have evaluated color vision in SCP children. Our study has important implications because cerebral palsy has been studied for a long time, with much data generated about the visual changes associated with ophthalmologic strabismus and lens accommodation paresis findings [6, 8, 9, 11, 31–37].

Visual acuity is classically reduced in SCP children. Recent studies have shown that it is positively correlated with motor function measured by the GMFCS [13–15, 25, 32, 34, 37]. Similar to previous visual acuity studies, we found a positive correlation between chromaticity discrimination thresholds and GMFCS scores within each motor group condition. We argue that the visual acuity reduction in our sample did not interfere with color vision because the worst visual acuity of our sample (20/50) means a discrimination threshold of 2.5 cycles per degree of visual angle. The stimulus had 1.25 degrees (or 46 minutes of arc), which was sufficiently wide to be seen by our children.

The color vision impairment observed in the present study was also related to the motor condition. Tetraplegic children had worse function compared with diplegic and hemiplegic children. Hemiplegic children had similar color vision as control children. Better visual and cognitive functions in hemiplegic children have been described in previous studies [9, 11, 15, 32, 35]. This finding corroborates studies of motor function that reported less neurological impairment compared with diplegic and tetraplegic children [2, 4, 18].

The developmental characteristic of chromaticity discrimination was assessed in the present study. We found that chromaticity discrimination is under development because the children in the control group exhibited a negative correlation between age and chromaticity thresholds obtained in each color confusion axis. This means that older children had lower thresholds. This is consistent with the findings of other studies that suggested that chromaticity discrimination is fully developed at the end of adolescence [30, 31, 38, 39].

One way to evaluate the development of chromaticity discrimination in SCP is to assess whether the threshold values decrease in older children. Considering the age of the children evaluated in the present study, our hypothesis is that, even with overall deficits in chromatic discrimination, the development of that function was occurring. However, we found a completely opposite result. As age increased, chromaticity thresholds increased, suggesting a progressive reduction of color vision discrimination. Thus, in children with SCP, color vision is both reduced compared with normal children and developmentally impaired. One possible interpretation could be related to sensorial impairment that affects the increase in perceptual complexity. Studies have demonstrated that multisensorial and associative cerebral areas are the most sensitive to external stimulation [40, 41]. Some associative areas, such as the dorsolateral prefrontal cortex, have been considered to subserve critical cognitive abilities during early infancy, and improvements in these abilities are evident over roughly the next 10 years. These authors found a reduction of visual contrast sensitivity function in children with diseases that affect the dorsolateral prefrontal cortex. Thus, damage to sensory and perceptual systems early in life may interfere with chromatic discrimination function in a complex way and interfere with the course of their development. We also demonstrated that visual acuity is also impaired in SCP children [13–15, 25]. The same cortical mechanism may be involved in both functions.

We evaluated the reliability of the test by computing the correct responses to the catch stimulus. If effects on cognitive function were affecting our measurement of chromaticity discrimination, then the reliability of the responses in the catch trial could be important. Because we had 100% correct responses in the catch trials for all of the subjects under all of the conditions, we can consider that the reduction of chromaticity discrimination reflected impairment in sensorial pathways.

Etiology and prematurity were not variables that affected chromaticity discrimination. Perinatal anoxia was the most frequent etiology, consistent with other studies [4, 42, 43]. However, no differences were found between the children with anoxia and those with another etiology, thus corroborating previous studies in which etiology was unrelated to visual impairments in children with SCP [8–10, 43–45].

Similar to the control group, the SCP children showed no difference in chromaticity discrimination thresholds in the monocular and binocular viewing conditions. These absences of a binocular summation had been reported to chromaticity discrimination function in normal subjects using the commercial version of the CCT [46].

There are some limitations that we consider critical for the interpretation of our findings. The need for a minimum level of cooperation from the patient may be underestimating the defect in color vision associated with the SCP. Psychophysical measures by involving behavioral, verbal or motor responses could certainly impair threshold measurements in different populations such as that we studied. The impaired motor behavior in the SCP children certainly adds a level of uncertainty that could not be precisely estimated. Although all controls performed to ensure the validity of the measures as measure reliability for the catch trials, the uncertainty in the threshold level may have been influenced by the lack of a clear and accurate motor behavior.

In conclusion, color perception based on chromaticity discrimination is generally reduced in SCP children. Tetraplegic children had worse color discrimination than the other two SCP groups (diplegic and hemiplegic). The impairment in chromaticity discrimination was proportional to GMFCS score for all SCP motor impairment subgroups.