How We Filter Out Unwanted Visual Stimuli:
Clues from Functional Brain Imaging

The find-it puzzle, “Where’s Waldo,” works because Waldo’s coke-bottle glasses and his red and white cap help us identify him (with some effort) in a busy scene of people and objects.

Princeton neuroscientist Sabine Kastner studies the brain mechanisms behind this kind of selective visual attention. In her paper, “Towards a Neural Basis of Human Visual Attention: Evidence from Functional Brain Imaging,” Kastner proposes that selective attention operates in the brain at multiple processing levels in the visual system and beyond.

In studies using functional magnetic resonance imaging (fMRI), a checkerboard stimulus evoked a higher signal response while participants were attending to the stimulus than when they attended away from the stimulus and focused on reading a set of letters. These attention effects were found in visual cortical areas and also in a region of the thalamus called the lateral geniculate nucleus. Kastner identifies this area as the first stage where attention modulates the processing of visual stimuli.

There is also evidence that at the next level, multiple visual objects may compete for neural representation within the visual cortex by essentially suppressing each other’s neural activity. A push-pull mechanism in this area may facilitate the processing of the visual stimuli attended to, while suppressing the stimuli not attended to. This may be a fundamental mechanism by which attention filters out unwanted information from cluttered visual scenes.

The researchers also suggest that a network of higher-order areas generates top-down signals that mediate the attention effects observed in the visual system. These signals may bias visual processing in favor of the attended to location. This was evidenced in a study where baseline activity increased when participants’ attention was directed to a location in the visual field in the anticipation of a visual stimulus to appear. Kastner concludes that attention can be best described as a multi-level selection process.
< Back