The effect of "red," "blue-green," and "white" light on electroencephalograms of autistic children



Journal Title

Journal ISSN

Volume Title



This study was conducted to investigate differences between autistic and normal children's electroencephalograms under several visual stimuli. There were 15 autistic children, 11 boys and 4 girls with a mean age of 10 years. The control group consisted of 15 normal children, 11 boys and 4 girls with a mean age of 10 years. Previous experiments have shown that autistic children respond more favorably to a red light visual stimulus than a white light (Ertl, Elliott, 1972). Bergman and Escalona (1949) reported that autistic children demonstrated unusual sensitivities to most stimuli, therefore, based upon the previous experiments, it was decided to use color as the primary independent variable. Each experimental trial was divided into five bands of EEG (Delta, Theta, Alpha, Beta, and Gamma rhythm) which created five variables for each group. Therefore a 2 x 3 x 3 research design was created from the two groups of children, three colors, and three configuration stimuli. The independent variables (color and configuration) consisted of forty-five 35 mm slides. The color was adopted from Munsell's (1966) color code for hue and chroma (red, blue-green, and white). The configurations were a nondescript geometric form, an alphabetical character, and a simple picture. There were 15 slides per color and five slides for each configuration. Therefore, each color set of 15 slides consisted of three subsets of five slides with one of three configurations on each set of five. Both groups viewed the nene sets of slides for five seconds per slide. All EEG recordings were completed inside an anechoic, shielded room to eliminate external light, sound, temperature, and vibration. While each slide was projected for viewing by each child on a screen, the EEG was simultaneously amplified and then recorded on 35 mm film by light emitting diodes. The film was developed and the recordings were converted into digital tapes for computer analysis after Fourier transformation and Hybrid analog to digital operations. A multivariate test. (MANOVA) of equality of mean vectors (bands 1-5) shows significant differences in EEGs between groups (df 5/158, F 14.8, P .0001). Also,Bartlett's chi square test of successive canonical variates produced a chi square of 61; df 5, and a P of less than .0001 for group differences. A Principle Components analysis of the five bands scores showed two components. The first three bands clustered on one component and the last two on the other. Tables of R square and Beta values revealed the greatest R square value of .30 occurred in the normal vs autistic test. The greatest Beta value consistently occurred within the Theta band. A Q Type factor analysis identified clusters of autistic children in the red color and blue-green color. Normal children had greater (significantly different) EEGs than autistic children. This might indicate that autistic children generate less brain energy than normal children when presented with the same visual stimuli. If this is true, then possibly autistic children need more intense visual stimuli to produce the same level of brain energy as do normal children. The tendencies for autistic children to cluster within the red and blue-green colors seem to indicate that color may perform a vital function in increasing the level of brain energy in autistic children. The results of this dissertation seem to concur with Bergman and Escalona (1969), and Bernal and Miller 1970) that autistic children have varying degrees of responsiveness to visual stimuli. The results also present data, demonstrating EEG tracings significantly different between autistic and normal children, not reported in the literature.