Visual Extinction - Visual Extinction and Image Preprocessing

Visual Extinction and Image Preprocessing

Visual extinction has been involved in several studies regarding preprocessing of images in the brain. A study performed by Dr Vuilleumier examined the effect of using faces as the stimulus in studying an extinction event. Subjects were presented with two simultaneous stimuli – some combination of a schematic face, a shape, a word, and a scrambled face. Faces in the contralesional field were less likely to be extinguished than other stimuli, but faces in the ipsilesional field appeared to induce more extinction events. The appearance of a face was prioritized in attention, despite the handicap presented of extinction damage.

Visual extinction has also been used to demonstrate brain bias towards gestalt processing. When presented with a figure containing illusory contours, patients were able to correctly report the presence of stimuli in both contralesional and ipsilesional hemispheres, due to their unconscious processing of the whole field to produce the illusion. This experiment implied that the attention center prioritizes the visualization of surfaces over other stimuli – therefore, although under race model the ipsilesional stimuli should extinguish the contralesional, the creation of the gestalt takes priority over detection of both. Further, a study using Gabor signals (alternating blurred and noisy black and white bars, commonly used by opticians in diagnostic tests) investigated how the orientation of these signals affected their extinction rate. Bilateral stimuli were least extinguished when both stimuli were oriented horizontally, although both stimuli being oriented vertically also showed a reduction in extinguishing rate when compared to one stimulus vertical and one horizontal – in what could be assumed by the brain to represent two different surfaces.

Combining the brain bias towards facial processing and surface discovery, a followup study by Dr. Vuilleumier exposed subjects to paired visual stimuli in three phases. In the initial phase, stimuli were simply bilateral pairs, and extinction frequency was measured. In the next stage, the stimuli were surrounded with an oval, so that they appeared to represent eyes in a face. The extinction rate for this appearance was greatly reduced. Finally, the patients were shown the stimuli minus the contextual oval – however, the reduced extinction rate persisted, as though the patients had learned to group these stimuli. The effect did not occur when the stimuli were surrounded by a nonoval shape, pointing towards facial processing as being the key to this phenomenon.

Visual feedback has also been studied in relation to visual extinction. Patients were asked to touch a known target in a darkened room. A light attached to the patient’s hand was sometimes briefly illuminated, to provide information about where the hand was in relation to the target. In some of these trials, a distracting light was also lit, which induced an extinction event in the patient. Although the patient reported in such cases that he had not seen the indicator light on his hand, their performance was correspondingly better, similar to the results when visual feedback had been available. Although the patient was not cognizant of having received the information, they were able to correctly act upon it, in a manner similar to blindsight.

Visual extinction has also been researched with regard to the effect of repetition on visual detection rate. Patients were shown a colored (red or green) letter (O or E), one to each visual field, and then asked variably to report the color or shape of one letter or the other. Extinction was found to be increased in the contralesional field when the patient was asked to report on a repeated characteristic – if both stimuli had been the same shape, or same color – regardless of whether the other characteristic had also been changed. This is another example of repetition blindness.