Chapter 7 Three kinds of attentional selection
Because we have limited capacity, we need to attend to things to get them processed. This is assisted by three abilities that we have:
- Location selection
- Feature (e.g., color) selection
- Object selection and tracking
In this chapter we’ll talk about location and feature selection - a later chapter will discuss object selection and tracking.
7.1 Location selection
People typically move their eyes to look directly at where they are attending when they are selecting a single location.
But sometimes, people attend to multiple locations simultaneously. Then, their eyes point at one location but they may be attending to other locations simultaneously. In any previous readings or units of study you’ve taken, location selection may have been the only kind of attention discussed. For example, in standard cuing experiments (Posner, Snyder, and Davidson 1980), a location is cued, after which participants perform better at processing things in that location, because the participants attentionally selected that location.
The schematic illustrates a shape identification experiment. The arrow cues the participant to attend to the right, which results in them identifying the shape there (the star) more quickly than if the arrow had directed the participant’s attention somewhere else. We’ll revisit location cuing in some new contexts in 11.
7.2 Feature selection
People can select not only an individual location, but also an individual feature value, for certain features such as color.
In the below image, try just attending to the red objects. With some concentration, you may feel you can do it just by thinking of red.
Next, try concentrating on all the black objects. You can use your power of feature selection to do so. Similarly with the green - try it.
For any of these forms of selection to work as effectively as they do, your brain must be processing the color and the shape information simultaneously in parallel across the visual scene. If it didn’t, you’d be reduced to moving attention to each location to process its color and shape.
In the below, try concentrating on purple, green, or blue.
Feature selection also works pretty well for simple shapes. In the below, try thinking of squares, circles, triangles, or pluses.
7.3 Feature combination selection?
However, selecting a combination of features does not work as well. You can think “green circles”, but unless you’re so lucky that your attention semi-randomly lands on them right away, it will take longer to find the green circles.
Similarly, in the below image, in b it is easy to find the red object, in d it is easy to find the circle, but in f it takes longer to find the red circle. Selection of a combination of features is not effective.
Feature selection also is effective for direction of motion, such as to select upward or downward motion (Sàenz, Buraĉas, and Boynton 2003).
Color, shape, and motion direction are processed in parallel across the visual field, in some fashion that allows attention to be guided by them. Researchers don’t fully understand yet how this works, but as will be decribed in 9, Anne Treisman proposed that there is a separate map for each feature, which can guide attention.
Whether the mechanism turns out to be a feature map or something else, when we think of a particular color, shape, or motion direction, our brain can enhance activation of the associated neurons rapidly. But we can’t do so for a particular combination of color and shape. Instead, we end up enhancing the activation of both the objects that are that particular color and the objects that are that particular shape; we can’t confine our attention to those that have that combination of features. For example, if the color is red and the shape is circle, we’ll activate all the red objects (even if they’re squares) and all the circular objects (even if they’re blue), rather than just the red circular object.
7.5 Complex shapes
While shapes that differ dramatically from each other allows parallel selection, like of the circles versus squares above, we do not have the ability to have our visual cortex enhance all instances of a complex shape.
In the next search array, your task is to find the shape that does NOT have a squiggly tail.
The search array is below. Find the target - the object without a squiggly tail.
Finding the object without a squiggly tail takes, on average, much longer than in the simple search cases above.
How quickly can you find a lucky four-leaf clover?
Finding a four-leaf clover is hard! One reason is that feature attention is not effective for complex shapes (another reason, in the above display at least, is that the clover is small and crowded, so you have to look at each patch almost directly to perceive its leaves clearly).
Even fairly simple shapes like a ‘T’ can be beyond the powers of our featural shape selection. I wish we could just think to ourselves, ‘T!’, and enhance the activation of any T’s in the scene, making our attention go straight to their locations, but we can’t. So, on average it takes quite a while for attention to go to the location of the T.
In summary, while we have some ability to do featural selection for color, shape, and motion, we can only do so for very simple shapes. The reason for this is not fully understood. Something about how the brain is connected up allows us to think “red” and quickly attend to all the locations in a scene that contain red, but this ability extends to only a few features.
Answer these questions and relate them to the seventh learning outcome ( 2 ):
- What does “feature selection” mean?
- What kind of selection is Posner cuing an example of?
Posner, Michael I, Charles R R Snyder, and Brian J Davidson. 1980. “Attention and the Detection of Signals.” Journal of Experimental Psychology: General 109 (2): 160–74. https://doi.org/10.1037/0096-34126.96.36.199.
Sàenz, Melissa, Giedrius T Buraĉas, and Geoffrey M Boynton. 2003. “Global Feature-Based Attention for Motion and Color.” Vision Research 43 (6): 629–37. http://www.ncbi.nlm.nih.gov/pubmed/12604099.