Quantification of image colorization with visual evoked potentials and functional magnetic resonance imaging

Visual Evoked Potentials (VEPs) result from the response of the visual system to a flashing or alternating visual stimulus. This study examines the effect of image colorization upon the VEP waveform using multiple analytical methods including Hjorth parameters, entropy, frequency power spectrum and time-frequency analysis.; A colorization method was developed and implemented that maximized the ability of the subject to detect all color levels within the stimulus image. This colorization was used to enhance VEP stimuli of circular checkerboards and mammogram ROIs. A novel VEP recording method was developed which allowed for the presentation of up to 8 different stimuli during one recording session. The statistics of the grayscale stimuli were compared to the colorized stimuli. The effect of the colorization enhancement mechanism on the VEP were analyzed and presented.; The VEP of the CCB stimuli showed differences in the time-frequency analysis, Hjorth activity and entropy of the Oz recordings. The mammogram ROI VEPs had decreases in the Hjorth activity and energy. VEP recordings of T5 and T6 were also analyzed. The results of a power spectrum analysis showed that there were frequency peaks which occurred only at T5 and T6, but not the Oz electrode. The peaks were in the alpha and beta frequency bandwidths from 5--15 Hz. This indicated that the visual signal had activated additional regions of the visual cortex due to colorization.; Functional Magnetic Resonance Imaging (fMRI) was used to verify the results of the VEP experiments. The first protocol compared the CCB and CALC mammogram ROI VEP stimuli to a static grayscale background. The second protocol directly compared the grayscale to the colorized stimuli.; The fMRI results verified the VEP T5 and T6 findings which indicated additional cortical activity due to the colorization. A comparison of the grayscale to the colorized stimuli indicated that the additional activity was found in Brodmann areas 18 and 19, and specifically areas V2, V3, V4 and V7. Colorization was determined to add a significant component to the visual signal that elicits separate and distinct responses from grayscale stimuli and can be used to further enhance mammogram image interpretation.