Below you'll find a review of the topics we've covered for Attention Effects and Attentional Control (corresponding to Chapters 6 and 7 in Purves - Introduction to Cognitive Neuroscience and the accompanying lectures in NE202/PS339).
In class we discussed this definition for attention, provided by William James:
Attention … is the taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible objects or trains of thought, localization, concentration, of consciousness are of its essence. It implies withdrawal from some things in order to deal effectively with others, and is a condition which has a real opposite in the confused, dazed, scatter brained state which in French is called distraction, and Zerstreutheit in German.A lovely blog post expounding on James's ideas on attention can be found here.
As poetic and intuitive this definition is, it doesn't tell us anything about the brain basis for attention. In fact, attention doesn't seem to be just one function in the brain at all; attention seems to be the name that we give to a collection of phenomena that allow the brain to selectively process a subset of the information that is readily available to it. Instead of asking, "What is attention?" (asking about one thing - attention), maybe we can ask, "How does the brain select and prioritize information?" (this question allows there to be many different answers, and no one answer has to be completely right on its own).
When I think about trying to understand attention, I'm reminded of the parable of the blind men and the elephant. You can read about it here or, if you're in the mood for some sick recorder music, you can listen to a version here (scroll down and listen to the episode segment, "The Rahjah's Question"). Basically, when we study attention we are trying to understand what it is and how it works from many different angles, because we don't yet have a complete picture.
Often when study attention, we are really studying the effect attention has on the processing of information in the brain. We said in class that to study the effects of attention, we need to "keep the stimuli the same" as well as control for eye position and other external factors. This is because we do not want to measure differences in the brain between different stimuli, we want to measure differences that occur only because the person in the experiment has shifted their attention in some way.
There are several methods used in the attention experiments that we've discussed:
Check out this informative video featuring two people with hemispatial neglect:
At 1:44, the neurologist tests the patient on which finger he is moving. The patient does not see the doctor's left finger, until he brings it in closer to the center of the patient's visual field. Even though the patient can now see the left finger move, when the doctor moves both fingers at the same time, the patient only sees the right finger move. This loss of awareness of the left hand side in the presence of right hand stimulus is an example of extinction. Extinction is separate but related to hemispatial neglect in that both describe deficits of attention that are confined to half of the environment. Here is another example of extinction, now in the somatosensory domain:
Neglect can be ego-centered or object-centered. These are the coordinate systems for neglect. Ego-centered means that the part of space that is ignored is to the left with respect to the patient's point of view. Object-centered means that the part of space that is ignored is to the left of each individual object. A paper comparing ego-centered and object-centered neglect can be found here. Check out pages 15-17 for figures showing results of a picture copying task performed by two patients with different types of neglect.
We discussed two theories for how hemispatial neglect works, and why it usually only affects the left side of things. The two models were the Representation Model and the Attentional Bias Model. Both are described in this article but here's a quote (with notes from me):
Perhaps the most widely accepted, ‘standard’ theory of neglect (a.k.a. the Representational Model) postulates that the right hemisphere controls shifts of attention to both sides of space while the left hemisphere only controls attention to the right side (Mesulam 1981). Damage to the right hemisphere impairs attention to the left hemifield while damage to the left hemisphere can be compensated. A second, ‘opponent-process’ theory (a.k.a. the Attentional Bias Model) proposes that each hemisphere promotes orienting in a contralateral direction, but the strength of this bias is stronger in the left than right hemispheres (Kinsbourne 1987). Left hemisphere lesions cause only mild right spatial neglect because the unopposed orienting bias generated by the right hemisphere is relatively weak.
Bálint's syndrome is a condition that can result from extensive bilateral lesions to the posterior parietal lobes. Patients with Bálint's syndrome exhibit simultagnosia, difficulty directing attention, and difficulty disengaging attention. A great description of the three main deficits associated with Bálint's syndrome are described on Wikipedia.
Here's a link to the video of a patient with Bálint's syndrome that we watched in class.
Here's a link to a short article describing two patients' experiences with Bálint's syndrome.
When we choose to concentrate on something we are voluntarily selecting what information we want to process, and we call this endogenous attention. It is under your conscious control. We also selectively process information that we didn't decide to concentrate on, but is important to us. Exogenous attention is reflexive and alerts us to important stimuli. When something "grabs" your attention, this is exogenous attention.
The term "bottom-up" generally refers to signals in the brain that travel from the sensory systems "up" to cortex. "Top-down" means that a signal is being sent "down" from higher-order cortices in order to affect the sensory systems. Bottom-up attention is the same as exogenous attention, and top-down attention is the same as endogenous attention. (You were right, Jacob)
Recall that our goal in studying attention is to answer the question, "How does the brain select and prioritize information?" We learned above that this selection can be voluntary (endogenous) or involuntary (exogenous). The directed attentional network is hypothesized to support endogenous attention (by consciously directing or focusing) and the orienting system supports exogenous attention (it disengages your attention and then re-orients your attention to a new stimulus).
The dorsal attentional network is a group of brain regions thought to correspond to the directed attention network. The ventral attentional network is thought to correspond to the orienting system. We started out with an observation - that attention can be voluntary or involuntary and defined endogenous vs. exogenous attention. Then we hypothesize that these two types of attention are supported by different systems (directed attention network vs. orienting system). Finally, we use fMRI to look in the brain to try and find the brain networks that support these systems (dorsal and ventral attention networks).
Change blindness and inattentional blindness are both examples of "looking without seeing." They are both phenomena that expose the fact that even though we think we are seeing everything in a scene, we are actually only processing a subset of the information at a time. This exposes an important property of attention; that it employs limited resources in the brain.
An example of inattentional blindness is here:
Inattentional blindness occurs because we are busy focusing our attention elsewhere, so we are "blind" to other details.
Change blindness occurs when we don't notice a change. This shows us that even though we thought we saw everything in a scene, we were actually only perceiving a fraction of the information. This is yet another way our attention is limited. In case you missed the demo in lecture, here's a fun video showing people experiencing change blindness:
The most common definition of a "cue" is given as "a thing said or done that serves as a signal to an actor or other performer to enter or to begin their speech or performance." Here, an attentional cue is a stimulus that serves as a signal to direct your attention at something. Spatial attentional cueing is a signal to direct your attention to a certain spatial location, and improves detection of a subsequent stimulus at that location. One famous example of a behavioral experiment that uses spatial cues is the Posner cueing task. This task can be used to study both endogenous and exogenous attention.
Here's a demo of the Posner cueing task that uses both valid and invalid cues. A valid cue is a cue that correctly directs your attention to a target, and an invalid cue directs your attention away from where a target will appear. Invalid cues will normally make you slower to detect a target.
Film and television make use of attentional cues to draw the viewer's eye to things, or in some cases, draw the eye away from certain elements in a frame. An article in Wired magazine quotes Jon Favreau (the director of Iron Man 2) as saying,
"We’re constantly calculating where we think the audience’s eye is going to be, and how to attract it to that area and prioritize within a shot what you can fake," Favreau said. "The best visual effects tool is the brains of the audience," he said. "They will stitch things together so they make sense."
You don't have to be looking right at something in order to pay attention to it. Overt attention is when you are looking right at the thing you want to pay attention to. Covert attention is when you direct your eyes to one thing, but pay attention to something you are not looking at.
The abstract of this article puts it succinctly:
The superior colliculus (SC) has long been known to be part of the network of brain areas involved in spatial attention, but recent findings have dramatically refined our understanding of its functional role. The SC both implements the motor consequences of attention and plays a crucial role in the process of target selection that precedes movement. Moreover, even in the absence of overt orienting movements, SC activity is related to shifts of covert attention and is necessary for the normal control of spatial attention during perceptual judgments. The neuronal circuits that link the SC to spatial attention may include attention-related areas of the cerebral cortex, but recent results show that the SC's contribution involves mechanisms that operate independently of the established signatures of attention in visual cortex. These findings raise new issues and suggest novel possibilities for understanding the brain mechanisms that enable spatial attention.We also discussed the superior colliculus as a possible location in the brain where the "bottom-up" saliency map is represented.
The superior colliculus is famous for the Sprague effect.
The saliency and priority maps (or models) are two theories for explaining how we choose where to look next.
The saliency map is an example of a "bottom-up" theory: where we direct our attention depends on the physical saliency of the scene in front of us. This means that we don't choose where to look, our gaze is "grabbed" by salient stimuli (exogenous attention, involuntary). BTW, "salient" just means "important" or "noticeable." From lecture:
Saliency at a given location is determined primarily by how different this location is from its surround in color, orientation, motion, depth etc.Possible brain regions supporting a saliency map are the superior colliculus or pulvinar (part of the thalamus).
The priority map is a theoretical "top-down" map telling your eyes where to look next. It combines information about physical saliency, where we've already looked in a scene, and our current goals. With this amount of integration of different kinds of information, a priority map would have to be found in cortex - possibly the intrapariental sulcus.
When we are in a crowded room with lots of people talking, we can use our selective attention to focus on just the conversation we are interested in. This is the cocktail party effect. Studies of selective attention will often simulate competing auditory streams by having a person listen to different auditory streams in each ear (dichotic listening) or a mixture of two streams at the same time in both ears (diotic listening; usually done with speech stimuli). By comparing responses to the attended vs. unattended auditory stream, we can measure the effect that paying attention has on auditory processing.
I found some examples of cocktail party effect stimuli here.
This article reviews the cocktail party effect in more detail.
Early vs. late attentional selection refers to how quickly attentional effects can be measured. This timing has great theoretical importance. We want to know, does attention act like a filter that only lets some stimuli get through to our brain, or does it act more like a volume knob for cortical processing of stimuli we pay attention to? The "filter" account would mean that attention effects can be measured a very short time after the stimulus presentation (early). Late attentional selection means that attentional effects are measured after more time has passed since the initial stimulus was presented. This late selection would suggest that the stimulus signal has had time to reach higher-level stages of cortical processing and attention affects that high-level processing. Both early and late effects of attention have been observed in different experiments.
More info on these two models of attention can be found here.