CH 2 PDF QUIZ PDF

SOME QUESTIONS WE WILL CONSIDER ◗ What is cognitive neuroscience, and why is it necessary? (26) A s we discussed in Chapter 1, research on the mind has been on a roller-coaster ride that began with a promising start in the 1800s with the early research of Donders and Ebbinghaus, only to be derail...

SOME QUESTIONS WE WILL CONSIDER ◗ What is cognitive neuroscience, and why is it necessary? (26) A s we discussed in Chapter 1, research on the mind has been on a roller-coaster ride that began with a promising start in the 1800s with the early research of Donders and Ebbinghaus, only to be derailed by Watson’s behaviorism in the early 1900s and Skinner’s operant conditioning in the 1930s. Finally, in the 1950s and 1960s, clearer minds decided ◗ How is information transmitted from one place to another in the nervous system? (29) that it was important to return to the study of the mind and began doing experiments based ◗ How are things in the environment, such as faces and places, represented in the brain? (42) would have a huge impact on our understanding of the mind. In the 1950s, a number of re- ◗ What are neural networks, and what is their role in cognition? (45) began long before the 1950s, but technological advances led to a large increase in physio- on the information-processing model that was inspired by digital computers. But just as this cognitive revolution was beginning, something else was happening that search papers began appearing that involved recording nerve impulses from single neurons. As we will see, research studying the relationship between neural responding and cognition logical research beginning just about the same time the cognitive revolution was happening. In this chapter, we take up the story of cognitive neuroscience—the study of the physiological basis of cognition. We begin by discussing the idea of “levels of analysis,” which is our rationale behind studying the physiology of the mind, and we then go back in time to the 19th and early 20th century to look at the early research that set the stage for amazing discoveries that were to be made beginning in the 1950s. Levels of Analysis Levels of analysis refers to the idea that a topic can be studied in a number of different ways, with each approach contributing its own dimension to our understanding. To understand what this means, let’s consider a topic outside the realm of cognitive psychology: understanding the automobile. Our starting point for this problem might be to take a car out for a test drive. We could determine its acceleration, its braking, how well it corners, and its gas mileage. When we have measured these things, which come under the heading of “performance,” we will know a lot about the particular car we are testing. But to learn more, we can consider another level of analysis: what is going on under the hood. This would involve looking at the mechanisms responsible for the car’s performance: the motor and the braking and steering systems. For example, we can describe the car as being powered by a four-cylinder 250 HP internal combustion engine and having independent suspension and disc brakes. But we can look even deeper into the operation of the car by considering another level of analysis designed to help us understand how the car’s engine works. One approach would be to look at what happens inside a cylinder. When we do this, we see that when vaporized gas enters the cylinder and is ignited by the spark plug, an explosion occurs that pushes the cylinder down and sends power to the crankshaft and then to the wheels. Clearly, considering the automobile from the different levels of driving the car, describing the motor, and observing what happens inside a cylinder provides more information about cars than simply measuring the car’s performance. Applying this idea of levels of analysis to cognition, we can consider measuring behavior to be analogous to measuring the car’s performance, and measuring the physiological processes behind the behavior as analogous to what we learned by looking under the hood. And just as we can study what is happening under a car’s hood at different levels, we can 26 Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 26 4/18/18 4:22 PM 27 Neurons : Basic Principles study the physiology of cognition at levels ranging from the whole brain, to structures within the brain, to chemicals that create electrical signals within these structures. Consider, for example, a situation in which Gil is talking with Mary in the park (Figure 2.1a), and then a few days later he passes the park and remembers what she was wearing and what they talked about (Figure 2.1b). This is a simple behavioral description of having an experience and later having a memory of that experience. But what is going on at the physiological level? During the initial experience, in which Gil perceives Mary as he is talking with her, chemical processes occur in Gil’s eyes and ears, which create electrical signals in neurons (which we will describe shortly); individual brain structures are activated, then multiple brain structures are activated, all leading to Gil’s perception of Mary and what is happening as they talk (Figure 2.1a). Meanwhile, other things are happening, both during Gil’s conversation with Mary and after it is over. The electrical signals generated as Gil is talking with Mary trigger chemical and electrical processes that result in the storage of Gil’s experiences in his brain. Then, when Gil passes the park a few days later, another sequence of physiological events is triggered that retrieves the information that was stored earlier, which enables him to remember his conversation with Mary (Figure 2.2b). We have gone a long way to make a point, but it is an important one. To fully understand any phenomenon, whether it is how a car operates or how people remember past experiences, it needs to be studied at different levels of analysis. In this book, we will be describing research in cognition at both the behavioral and physiological levels. We begin our description of physiology by considering one of the basic building blocks of the nervous system: the neuron. PERCEPTION Groups of brain structures activated Brain structures activated Neurons activated (a) Chemical processes MEMORY Storage activated Brain storage Neurons activated Chemical processes (b) ➤ Figure 2.1 Physiological levels of analysis. (a) Gil perceives Mary and their surroundings as he talks with her. The physiological processes involved in Gil’s perception can be described at levels ranging from chemical reactions to single neurons, to structures in the brain, to groups of structures in the brain. (b) Later, Gil remembers his meeting with Mary. The physiological processes involved in remembering can also be described at different levels of analysis. Neurons: Basic Principles How is it possible that the 3.5-pound structure called the brain could be the seat of the mind? The brain appears to be static tissue. Unlike the heart, it has no moving parts. Unlike the lungs, it doesn’t expand or contract. And when observed with the naked eye, the brain looks almost solid. As it turns out, to understand the relation between the brain and the mind—and specifically to understand the physiological basis for everything we perceive, remember, and think—it is necessary to look within the brain and observe the small units called neurons that create and transmit information about what we experience and know. Early Conceptions of Neurons For many years, the nature of the brain’s tissue was a mystery. Looking at the interior of the brain with the unaided eye gives no indication that it is made up of billions of smaller units. But in the 19th century, anatomists applied special stains to brain tissue, which increased Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 27 4/18/18 4:22 PM CHAPTER 2 Cognitive Neuroscience (a) (b) Clouds Hill Imaging Ltd./Corbis Documentary/Getty Images 28 ➤ Figure 2.2 (a) Nerve net theory proposed that signals could be transmitted throughout the net in all directions. (b) A portion of the brain that has been treated with Golgi stain shows the shapes of a few neurons. The arrow points to a neuron’s cell body. The thin lines are dendrites or axons (see Figure 2.3). the contrast between different types of tissue within the brain. When they viewed this stained tissue under a microscope, they saw a network they called a nerve net (Figure 2.2a). This network was believed to be continuous, like a highway system in which one street connects directly to another, but without stop signs or traffic lights. When visualized in this way, the nerve net provided a complex pathway for conducting signals uninterrupted through the network. One reason for describing the microstructure of the brain as a continuously interconnected network was that the staining techniques and microscopes used during that period could not resolve small details, and without these details the nerve net appeared to be continuous. However, in the 1870s, the Italian anatomist Camillo Golgi (1843–1926) developed a staining technique in which a thin slice of brain tissue was immersed in a solution of silver nitrate. This technique created pictures like the one in Figure 2.2b, in which fewer than 1 percent of the cells were stained, so they stood out from the rest of the tissue. (If all of the cells had been stained, it would be difficult to distinguish one cell from another because the cells are so tightly packed.) Also, the cells that were stained were stained completely, so it was possible to see their structure. Meanwhile, the Spanish physiologist Ramon y Cajal (1852–1934) was using two techniques to investigate the nature of the nerve net. First, he used the Golgi stain, which stained only some of the cells in a slice of brain tissue. Second, he decided to study tissue from the brains of newborn animals, because the density of cells in the newborn brain is small compared with the density in the adult brain. This property of the newborn brain, combined with the fact that the Golgi stain affects less than 1 percent of the neurons, made it possible for Cajal to clearly see that the nerve net was not continuous but was instead made up of individual units connected together (Kandel, 2006). Cajal’s discovery that individual units called neurons were the basic building blocks of the brain was the centerpiece of neuron doctrine—the idea that individual cells transmit signals in the nervous system, and that these cells are not continuous with other cells as proposed by nerve net theory. Figure 2.3a shows the basic parts of a neuron. The cell body is the metabolic center of the neuron; it contains mechanisms to keep the cell alive. The function of dendrites that branch out from the cell body is to receive signals from other neurons. Axons (also called nerve fibers) are usually long processes that transmit signals to other neurons. Figure 2.3b shows a neuron with a receptor that receives stimuli from the environment—pressure, in this example. Thus, the neuron has a receiving end and a transmitting end, and its role, as visualized by Cajal, was to transmit signals. Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 28 4/18/18 4:22 PM Neurons : Basic Principles 29 Dendrite Axon or nerve fiber Cell body (a) Touch receptor Stimulus from environment (b) Nerve fiber Electrical signal ➤ Figure 2.3 (a) Basic components of a neuron in the cortex. (b) A neuron with a specialized receptor in place of the cell body. This receptor responds to pressure on the skin. Cajal also came to some other conclusions about neurons: (1) There is a small gap between the end of a neuron’s axon and the dendrites or cell body of another neuron. This gap is called a synapse (Figure 2.4). (2) Neurons are not connected indiscriminately to other neurons but form connections only to specific neurons. This forms groups of interconnected neurons, which together form neural circuits. (3) In addition to neurons in the brain, there are also neurons that are specialized to pick up information from the environment, such as the neurons in the eye, ear, and skin. These neurons, called receptors (Figure 2.3b), are similar to brain neurons in that they have an axon, but they have specialized receptors that pick up information from the environment. Cajal’s idea of individual neurons that communicate with other neurons to form neural circuits was an enormous leap forward in the understanding of how the nervous system operates. The concepts introduced by Cajal—individual neurons, synapses, and neural circuits—are basic principles that today are used to explain how the brain creates cognitions. These discoveries earned Cajal the Nobel Prize in 1906, and today he is recognized as “the person who made this cellular study of mental life possible” (Kandel, 2006, p. 61). The Signals That Travel in Neurons Cajal succeeded in describing the structure of individual neurons and how they are related to other neurons, and he knew that these neurons transmitted signals. However, determining the Cell body (a) Axon Nerve impulse Neurotransmitter molecules (b) Synapse Neurotransmitter being released ➤ Figure 2.4 (a) Neuron synapsing on the cell body of another neuron. (b) Close-up of the synapse showing the space between the end of one neuron and the cell body of the next neuron, and neurotransmitter being released. Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 29 4/18/18 4:22 PM 30 CHAPTER 2 Cognitive Neuroscience exact nature of these signals had to await the development of electronic amplifiers that were powerful enough to make the extremely small electrical signals generated by the neuron visible. In the 1920s, Edgar Adrian was able to record electrical signals from single sensory neurons, an achievement for which he was awarded the Nobel Prize in 1932 (Adrian, 1928, 1932). METHOD Recording from a Neuron Adrian recorded electrical signals from single neurons using microelectrodes—small shafts of hollow glass filled with a conductive salt solution that can pick up electrical signals at the electrode tip and conduct these signals back to a recording device. Modern physiologists use metal microelectrodes. Figure 2.5 shows a typical setup used for recording from a single neuron. There are two electrodes: a recording electrode, shown with its recording tip inside the neuron,1 and a reference electrode, located some distance away so it is not affected by the electrical signals. The difference in charge between the recording and reference electrodes is fed into a computer and displayed on the computer’s screen. ➤ Figure 2.5 Recording an action potential as it travels down an axon. (a) When the nerve is at rest, there is a difference in charge, called the resting potential, of 270 millivolts (mV) between the inside and outside of the axon. The difference in charge between the recording and reference electrodes is fed into a computer and displayed on a computer monitor. This difference in charge is displayed on the right. (b) As the nerve impulse, indicated by the red band, passes the electrode, the inside of the fiber near the electrode becomes more positive. (c) As the nerve impulse moves past the electrode, the charge in the fiber becomes more negative. (d) Eventually the neuron returns to its resting state. Recording electrode (inside axon) Push Monitor Reference electrode (outside axon) Resting potential –70 Time Nerve impulse (b) Pressure-sensitive receptor Charge inside fiber relative to outside (mV) (a) +40 –70 –70 (c) Back at resting level –70 (d) In practice, most recordings are achieved with the tip of the electrode positioned just outside the neuron because it is technically difficult to insert electrodes into the neuron, especially if it is small. However, if the electrode tip is close enough to the neuron, the electrode can pick up the signals generated by the neuron. 1 Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 30 4/18/18 4:22 PM Neurons : Basic Principles 31 When the axon, or nerve fiber, is at rest, the meter records a difference in 1/ sec. 10 potential between the tips of the two electrodes of 270 millivolts (a millivolt is 1/1000 of a volt), as shown on the right in Figure 2.5a. This value, which stays the same as long as there are no signals in the neuron, is called the rest10 action potentials ing potential. In other words, the inside of the neuron has a charge that is 70 mV more negative than the outside, and this difference continues as long as the neuron is at rest. Time Figure 2.5b shows what happens when the neuron’s receptor is stimu(a) lated so that a nerve impulse is transmitted down the axon. As the impulse 1 action passes the recording electrode, the charge inside the axon rises to 140 millipotential volts, compared to the outside. As the impulse continues past the electrode, (expanded) the charge inside the fiber reverses course and starts becoming negative again (Figure 2.5c), until it returns to the resting potential (Figure 2.5d). This imTime 1/ pulse, which is called the action potential, lasts about 1 millisecond (1/1000 1000 sec. of a second). (b) Figure 2.6a shows action potentials on a compressed time scale. Each vertical line represents an action potential, and the series of lines indicates that a ➤ Figure 2.6 (a) A series of action potentials number of action potentials are traveling past the electrode. Figure 2.6b shows displayed on a time scale that makes each action one of the action potentials on an expanded time scale, as in Figure 2.5. There potential appear as a thin line. (b) Changing the are other electrical signals in the nervous system, but we will focus here on the time scale reveals the shape of one of the action potentials. action potential because it is the mechanism by which information is transmitted throughout the nervous system. In addition to recording action potentials from single neurons, Adrian made other discoveries as well. He found that each action potential travels all the way down the axon without changing its height or shape. This property makes action potentials ideal for sending signals over a distance, because it means that once an action potential is started at one end of an axon, the signal will still be the same size when it reaches the other end. At about the same time Adrian was recording from single neurons, other researchers were showing that when the signals reach the synapse at the end of the axon, a chemical called a neurotransmitter is released. This neurotransmitter makes it possible for the signal to be transmitted across the gap that separates the end of the axon from the dendrite or cell body of another neuron (see Figure 2.4b). (a) Although all of these discoveries about the nature of neurons and the signals that travel in them were extremely important (and garnered a number of Nobel Prizes for their discoverers), our main interest is not in how axons transmit signals, but in how these signals contribute to the operation of the mind. So far, our description of how signals are transmitted is analogous to describing how the Internet transmits electrical signals, without describ(b) ing how the signals are transformed into words and pictures that people can understand. Adrian was acutely aware that it was important to go beyond simply describing nerve signals, so he did a series of experiments to relate nerve signals to stimuli in the environment and therefore to people’s experience. Adrian studied the relation between nerve firing and sensory experience by measuring how the firing of a neuron from a receptor in the skin changed as he applied more pressure Time (c) to the skin. What he found was that the shape and height of the action potential remained the same as he increased the pressure, but the rate of nerve firing—that is, the number of ➤ Figure 2.7 Action potentials action potentials that traveled down the axon per second—increased (Figure 2.7). From recorded from an axon in this result, Adrian drew a connection between nerve firing and experience. He describes response to three levels of pressure this connection in his book The Basis of Sensation (1928) by stating that if nerve impulses stimulation on the skin: (a) light, “are crowded closely together the sensation is intense, if they are separated by long intervals (b) medium, and (c) strong. the sensation is correspondingly feeble” (p. 7). Increasing stimulus intensity What Adrian is saying is that electrical signals are representing the intensity of the stimcauses an increase in the rate of ulus, so pressure that generates “crowded” electrical signals feels stronger than pressure that nerve firing. Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 31 4/19/18 6:18 PM 32 CHAPTER 2 Cognitive Neuroscience generates signals separated by long intervals. Later experiments demonstrated similar results for vision. Presenting high-intensity light generates a high rate of nerve firing and the light appears bright; presenting lower intensity light generates a lower rate of nerve firing and the light appears dimmer. Thus, the rate of neural firing is related to the intensity of stimulation, which, in turn, is related to the magnitude of an experience, such as feeling pressure on the skin or experiencing the brightness of a light. Going beyond Adrian’s idea that the magnitude of experience is related to the rate of nerve firing, we can ask, how is the quality of experience represented in neural firing? For the senses, quality across the senses refers to the different experience associated with each of the senses—perceiving light for vision, sound for hearing, smells for olfaction, and so on. We can also ask about quality within a particular sense, such as for vision: color, movement, an object’s shape, or the identity of a person’s face. One way to answer the question of how action potentials determine different qualities is to propose that the action potentials for each quality might look different. However, Adrian ruled out that possibility by determining that all action potentials have basically the same height and shape. If all nerve impulses are basically the same whether they are caused by seeing a red fire engine or remembering what you did last week, how can these impulses stand for different qualities? The short answer to this question is that different qualities of stimuli, and also different aspects of experience, activate different neurons and areas in the brain. We begin the long answer to this question in the next section by taking up the idea of representation, which we introduced in Chapter 1. Representation by Neural Firing In Chapter 1, we defined the mind as a system that creates representations of the world so that we can act within it to achieve our goals (page 6). The key word in this definition is representations, because what it means is that everything we experience is the result of something that stands for that experience. Putting this in neural terms, the principle of neural representation states that everything a person experiences is based on representations in the person’s nervous system. Adrian’s pioneering research on how nerve impulses represent the intensity of a stimulus, in which he related high nerve firing to feeling greater pressure, marks the beginning of research on neural representation. We now move ahead to the 1960s to describe early research that involved recording from single neurons in the brain. The Story of Neural Representation and Cognition: A Preview In the 1960s, researchers began focusing on recording from single neurons in the primary visual receiving area, the place where signals from the eye first reach the cortex (Figure 2.8a). The question being asked in these experiments was “what makes this neuron fire?” Vision dominated early research because stimuli could be easily controlled by creating patterns of light and dark on a screen and because a lot was already known about vision. But as research progressed, researchers began recording from neurons in areas outside the primary visual area and discovered two key facts: (1) Many neurons at higher levels of the visual system fire to complex stimuli like geometrical patterns and faces; and (2) a specific stimulus causes neural firing that is distributed across many areas of the cortex (Figure 2.8b). Vision, it turns out, isn’t created only in the primary visual receiving area, but in many different areas. Later research, extending beyond vision, found similar results for other cognitions. For example, it was discovered that memory is not determined by a single “memory area,” because there are a number of areas involved in creating memories and remembering them later. In short, it became obvious that large areas of the brain are involved in creating cognition. Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 32 4/18/18 4:22 PM Representation by Neural Firing (a) Visual cortex (b) Multiple visual areas 33 (c) More areas and network communication ➤ Figure 2.8 (a) Early work on neural representation and cognition focused on recording from single neurons in the visual cortex, where signals first arrive at the cortex. (b) Researchers then began to explore other places in the brain and found that visual stimulation causes activity that is distributed across many areas of the cortex. (c) Recent work has focused on looking at how these distributed areas are connected by neural networks and how activity flows in these networks. Note that, with the exception of the visual area in (a), the locations of the areas in this figure do not represent the locations of actual areas. They are for illustrative purposes only. As it became clear that understanding neural representation involves casting a wide net across the brain, many researchers began considering the way different areas are connected to one another. The idea of neural signals transmitted between many destinations in an interconnected brain has led to today’s conception of the brain as containing a vast Recording highway system that can be described in terms of “neural electrode networks” (Figure 2.8c). We will now fill in the details, beginning with the discovery of neural feature detectors. Feature Detectors One possible answer to the question “how can nerve impulses stand for different qualities?” is that perhaps there are neurons that fire only to specific qualities of stimuli. Early research found some evidence for this (Hartline, 1940; Kuffler, 1953), but the idea of neurons that respond to specific qualities was brought to the forefront by a series of papers by David Hubel and Thorsten Wiesel, which would win them the Nobel Prize in 1981. In the 1960s, Hubel and Wiesel started a series of experiments in which they presented visual stimuli to cats, as shown in Figure 2.9a, and determined which stimuli caused specific neurons to fire. They found that each neuron in the visual area of the cortex responded to a specific type of stimulation presented to a small area of the retina. Figure 2.9b shows some of the stimuli that caused neurons in and near the visual cortex to fire (Hubel, 1982; Hubel & Wiesel, 1959, 1961, 1965). They called these neurons feature detectors because they responded to specific stimulus features such as orientation, movement, and length. The idea that feature detectors are linked to perception was supported by many different experiments. One of these experiments involved a phenomenon called (a) Oriented bar Oriented moving bar Short moving bar (b) ➤ Figure 2.9 (a) An experiment in which electrical signals are recorded from the visual system of an anesthetized cat that is viewing stimuli presented on the screen. The lens in front of the cat’s eye ensures that the images on the screen are focused on the cat’s retina. The recording electrode is not shown. (b) A few of the types of stimuli that cause neurons in the cat’s visual cortex to fire. Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 33 4/19/18 6:20 PM 34 CHAPTER 2 Cognitive Neuroscience experience-dependent plasticity, in which the structure of the brain is changed by experience. For example, when a kitten is born, its visual cortex contains feature detectors that respond to oriented bars (see Figure 2.9). Normally, the kitten’s visual cortex contains neurons that respond to all orientations, ranging from horizontal to slanted to vertical, and when the kitten grows up into a cat, the cat has neurons that can respond to all orientations. But what would happen if kittens were reared in an environment consisting only of verticals? Colin Blakemore and Graham Cooper (1970) answered this question by rearing kittens in a space in which they saw only vertical black and white stripes on the walls (Figure 2.10a). After being reared in this vertical environment, kittens batted at a moving vertical stick but ignored horizontal objects. The basis of this lack of response to horizontals became clear when recording from neurons in the kittens’ brains revealed that the visual cortex had been reshaped so it contained neurons that responded mainly to verticals and had no neurons that responded to horizontals (Figure 2.10b). Similarly, kittens reared in an environment consisting only of horizontals ended up with a visual cortex that contained neurons that responded mainly to horizontals (Figure 2.10c). Thus, the kittens’ brains had been shaped to respond best to the environment to which they had been exposed. Blakemore and Cooper’s experiment is important because it is an early demonstration of experience-dependent plasticity. Their result also has an important message about neural representation: When a kitten’s cortex contained mainly vertically sensitive neurons, the kitten perceived only verticals, and a similar result occurred for horizontals. This result supports the idea that perception is determined by neurons that fire to specific qualities of a stimulus (orientation, in this case). This knowledge that neurons in the visual system fire to specific types of stimuli led to the idea that each of the thousands of neurons that fire when we look at a tree fire to different features of the tree. Some neurons fire to the vertically oriented trunk, others to Vertically reared cat Horizontally reared cat Vertical Vertical Horizontal Horizontal Vertical (a) (b) Vertical (c) ➤ Figure 2.10 (a) Striped tube used in Blakemore and Cooper’s (1970) selective rearing experiments. (b) Distribution of orientations that caused maximum firing for 72 cells from a cat reared in an environment of vertical stripes and (c) for 52 cells from a cat reared in an environment of horizontal stripes. Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 34 4/18/18 4:22 PM Representation by Neural Firing the variously oriented branches, and some to more complex combinations of a number of features. The idea that the tree is represented by the combined response of many feature detectors is similar to building objects by combining building blocks like Legos. But it is important to realize that the visual cortex is an early stage of visual processing, and that vision depends on signals that are sent from the visual cortex to other areas of the brain. Figure 2.11 indicates the location of the visual cortex in the human brain, as well as additional areas that are involved in vision, and some other areas we will be discussing later. The vision areas are part of a vast network of areas that make up about 30 percent of the cortex (Felleman & Van Essen, 1991). Some of these visual areas receive signals directly from the visual cortex. Others are part of a sequence of interconnected neurons, some of which are far down the line from the visual cortex. Following Hubel and Wiesel’s pioneering research, other researchers who began exploring these “higher” levels of the visual pathway discovered neurons that respond to stimuli more complex than oriented lines. Motor cortex 35 Parietal lobe Frontal lobe Occipital lobe Visual cortex Extrastriate body area Temporal lobe Parahippocampal place area (underside of brain) Fusiform face area (underside of brain) ➤ Figure 2.11 Some of the structures of the human brain that we will be referring to in this chapter. Pointers indicate the locations of these areas, each of which extends over an area of the cortex. Neurons That Respond to Complex Stimuli How are complex stimuli represented by the firing of neurons in the brain? One answer to this question began to emerge in the laboratory of Charles Gross. Gross’s experiments, in which he recorded from single neurons in the monkey’s temporal lobe (Figure 2.11), required a great deal of endurance by the researchers, because the experiments typically lasted 3 or 4 days. In these experiments, the results of which were reported in now classic papers in 1969 and 1972 (Gross et al., 1969, 1972), Gross’s research team presented a variety of different stimuli to anesthetized monkeys. On a projection screen like the one shown in Figure 2.9a, they presented lines, squares, and circles. Some stimuli were light and some dark. The discovery that neurons in the temporal lobe respond to complex stimuli came a few days into one of their experiments, when they had found a neuron that refused to respond to any of the standard stimuli, like oriented lines or circles or squares. Nothing worked, until one of the experimenters pointed at something in the room, casting a shadow of his hand on the screen. When this hand shadow caused a burst of firing, the experimenters knew they were on to something and began testing the neuron with a variety of stimuli, including cutouts of a monkey’s hand. After a great deal of testing, they determined that this neuron responded best to a handlike shape with fingers pointing up (far-right stimuli in Figure 2.12) (Rocha-Miranda, 2011; also see Gross, 2002). After expanding the types of stimuli presented, they also found some neurons that responded best to faces. Later researchers extended these results and provided many examples of neurons that respond to faces but don’t respond to other 1 1 1 2 3 3 4 4 5 6 types of stimuli (Perrett et al., 1982; Rolls, 1981) (Figure 2.13). ➤ Figure 2.12 Some of the shapes used by Gross et al. (1972) to study the Let’s stop for a moment and consider the reresponses of neurons in the temporal lobe of the monkey’s cortex. The shapes sults we have presented so far. We saw that neuare arranged in order of their ability to cause the neuron to fire, from none (1) rons in the visual cortex respond to simple stimuli to little (2 and 3) to maximum (6). like oriented bars, neurons in the temporal lobe (Source: Based on Gross et al., 1972.) Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 35 4/18/18 4:22 PM CHAPTER 2 Cognitive Neuroscience respond to complex geometrical stimuli, and neurons in another area of the temporal lobe respond to faces. What is happening is that neurons in the visual cortex that respond to relatively simple stimuli send their axons to higher levels of the visual system, where signals from many neurons combine and interact; neurons at this higher level, which respond to more complex stimuli such as geometrical objects, then send signals to even higher areas, combining and interacting further and creating neurons that respond to even more complex stimuli such as faces. This progression from lower to higher areas of the brain is called hierarchical processing. Does hierarchical processing solve the problem of neural representation? Could it be that higher areas of the visual system contain neurons that are specialized to respond only to a specific object, so that object would be represented by the firing of that one type of specialized neuron? As we will see, this is probably not the case, because neural representation most likely involves a number of neurons working together. Bruce Goldstein 36 Firing rate 20 10 Sensory Coding 0 The problem of neural representation for the senses has been called the problem of sensory coding, where the sensory code refers to how neurons represent various characteristics of ➤ Figure 2.13 Firing rate, in nerve impulses per second, of a the environment. The idea that an object could be represented neuron in the monkey’s temporal lobe that responds to face by the firing of a specialized neuron that responds only to stimuli but not to nonface stimuli. that object is called specificity coding. This is illustrated in (Source: Based on E. T. Rolls & M. J. Tovee, 1995.) Figure 2.14a, which shows how a number of neurons respond to three different faces. Only neuron 4 responds to Bill’s face, only neuron 9 responds to Mary’s face, and only neuron 6 responds to Raphael’s face. Also note that the neuron specialized to respond only to Bill, which we can call a “Bill neuron,” does not respond to Mary or Raphael. In addition, other faces or types of objects would not affect this neuron. It fires only to Bill’s face. Although the idea of specificity coding is straightforward, it is unlikely to be correct. Even though there are neurons that respond to faces, these neurons usually respond to a number of different faces (not just Bill’s). There are just too many different faces and other objects (and colors, tastes, smells, and sounds) in the world to have a separate neuron dedicated to each object. An alternative to the idea of specificity coding is that a number of neurons are involved in representing an object. Population coding is the representation of a particular object by the pattern of firing of a large number of neurons (Figure 2.14b). According to this idea, Bill’s, Mary’s and Raphael’s faces are each represented by a different pattern. An advantage of population coding is that a large number of stimuli can be represented, because large groups of neurons can create a huge number of different patterns. There is good evidence for population coding in the senses and for other cognitive functions as well. But for some functions, a large number of neurons aren’t necessary. Sparse coding occurs when a particular object is represented by a pattern of firing of only a small group of neurons, with the majority of neurons remaining silent. As shown in Figure 2.14c, sparse coding would represent Bill’s face by the pattern of firing of a few neurons (neurons 2, 3, 4, and 7). Mary’s face would be signaled by the pattern of firing of a few different neurons (neurons 4, 6, and 7), but possibly with some overlap with the neurons representing Bill, and Raphael’s face would have yet another pattern (neurons 1, 2, and 4). Faces Nonfaces Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 36 4/18/18 4:22 PM Representation by Neural Firing Population Coding Sparse Coding 1 2 3 4 5 6 7 8 9 10 Neuron number 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Neuron number 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Neuron number 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Firing rate Specificity Coding 37 Firing rate Bill Firing rate Mary Raphael (a) (b) (c) ➤ Figure 2.14 Three types of coding: (a) Specificity coding. The response of 10 different neurons to each face on the left is shown. Each face causes a different neuron to fire. (b) Population coding. The face’s identity is indicated by the pattern of firing of a large number of neurons. (c) Sparse coding. The face’s identity is indicated by the pattern of firing of a small group of neurons. Notice that a particular neuron can respond to more than one stimulus. For example, neuron 4 responds to all three faces, although most strongly to Mary’s. Recently, neurons were discovered when recording from the temporal lobe of patients undergoing brain surgery for epilepsy. (Stimulating and recording from neurons is a common procedure before and during brain surgery, because it makes it possible to determine the exact layout of a particular person’s brain.) These neurons responded to very specific stimuli. Figure 2.15 shows the records for a neuron that responded to pictures of the actor Steve Carell and not to other people’s faces (Quiroga et al., 2007). However, the researchers who discovered this neuron (as well as other neurons that responded to other people) point out that they had only 30 minutes to record from these neurons, and that if more time were available, it is likely that they would have found other faces that would cause this neuron to fire. Given the likelihood that even these special neurons are likely to fire to more than one stimulus, Quiroga and coworkers (2008) suggested that their neurons are probably an example of sparse coding. There is also other evidence that the code for representing objects in the visual system, tones in the auditory system, and odors in the olfactory system may involve the pattern of ➤ Figure 2.15 Records from a neuron in the temporal lobe that responded to different views of Steve Carell (top records) but did not respond to pictures of other well-known people (bottom records). (Source: Quiroga et al., 2008) Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 37 4/18/18 4:22 PM 38 CHAPTER 2 Cognitive Neuroscience activity across a relatively small number of neurons, as sparse coding suggests (Olshausen & Field, 2004). Memories are also represented by the firing of neurons, but there is a difference between representation of perceptions and representation of memories. The neural firing associated with experiencing a perception is associated with what is happening as a stimulus is present. Firing associated with memory is associated with information about the past that has been stored in the brain. We know less about the actual form of this stored information for memory, but it is likely that the basic principles of population and sparse coding also operate for memory, with specific memories being represented by particular patterns of stored information that result in a particular pattern of nerve firing when we experience the memory. Saying that individual neurons and groups of neurons contain information for perception, memory, and other cognitive functions is the first step toward understanding representation. The next step involves looking at organization: how different types of neurons and functions are organized within the brain. T E ST YOUR SELF 2.1 1. Describe the idea of levels of analysis. 2. How did early brain researchers describe the brain in terms of a nerve net? How does the idea of individual neurons differ from the idea of a nerve net? 3. Describe the research that led Cajal to propose the neuron doctrine. 4. Describe the structure of a neuron. Describe the synapse and neural circuits. 5. How are action potentials recorded from a neuron? What do these signals look like, and what is the relation between action potentials and stimulus intensity? 6. How has the question of how different perceptions can be represented by neurons been answered? Consider both research involving recording from single neurons and ideas about sensory coding. 7. How is neural representation for memory different from neural representation for perception? How is it similar? Localized Representation One of the basic principles of brain organization is localization of function—specific functions are served by specific areas of the brain. Many cognitive functions are served by the cerebral cortex, which is a layer of tissue about 3 mm thick that covers the brain (Fischl & Dale, 2000). The cortex is the wrinkled covering you see when you look at an intact brain (Figure 2.11). Other functions are served by subcortical areas that are located below the cortex. Early evidence for localization of function came from neuropsychology—the study of the behavior of people with brain damage. Localization Determined by Neuropsychology In the early 1800s, an accepted principle of brain function was cortical equipotentiality, the idea that the brain operated as an indivisible whole as opposed to specialized areas (Flourens, 1824; Pearce, 2009). But in 1861, Paul Broca published work based on his study of patients who had suffered brain damage due to strokes that caused disruption of the blood supply to the brain. These strokes caused damage to an area in the frontal lobe that came to be called Broca’s area (Figure 2.16). One of Broca’s patients was famously called Tan, because the word tan was the only word he could say. Other patients with frontal lobe damage could say more, but their speech Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 38 4/18/18 4:22 PM Localized Representation was slow and labored and often had jumbled sentence structure. Here is an example of the speech of a modern patient, who is attempting to describe when he had his stroke, which occurred when he was in a hot tub: 39 Wernicke’s area Alright. . . . Uh … stroke and un. . . . I . . . huh tawanna guy. . . . H . . . h . . . hot tub and. . . . And the. . . . Two days when uh. . . . Hos . . . uh. . . . Huh hospital and uh . . . amet . . . am . . . ambulance. (Dick et al., 2001, p. 760) Patients with this problem—slow, labored, ungrammatical speech caused by damage to Broca’s area—are diagnosed as having Broca’s aphasia. The fact that damage to a specific area of the brain caused a specific deficit of behavior was striking evidence against the idea of equipotentiality and for the idea of localization of function. Eighteen years after Broca reported on his frontal lobe patients, Carl Wernicke (1879) described a number of patients who had damage to an area in their temporal lobe that came to be called Wernicke’s area. Wernicke’s patients produced speech that was fluent and grammatically correct but tended to be incoherent. Here is a modern example of the speech of a patient with Wernicke’s aphasia: Broca’s area ➤ Figure 2.16 Broca’s area, in the frontal lobe, and Wernicke’s area, in the temporal lobe, were identified in early research as being specialized for language production and comprehension, respectively. It just suddenly had a feffort and all the feffort had gone with it. It even stepped my horn. They took them from earth you know. They make my favorite nine to severed and now I’m a been habed by the uh stam of fortment of my annulment which is now forever. (Dick et al., 2001, p. 761) Patients such as this not only produce meaningless speech but are unable to understand other people’s speech. Their primary problem is their inability to match words with their meanings, with the defining characteristic of Wernicke’s aphasia being the absence of normal grammar (Traxler, 2012). Broca’s and Wernicke’s observations showed that different aspects of language— production of language and comprehension of language—were served by different areas in the brain. As we will see later in this chapter, modern research has shown that the strict separation of language functions in different areas was an oversimplification. Nonetheless, Broca’s and Wernicke’s 19th-century observations set the stage for later research that confirmed the idea of localization of function. Further evidence for localization of function came from studies of the effect of brain injury in wartime. Studies of Japanese soldiers in the Russo-Japanese war of 1904–1905 and Allied soldiers in World War I showed that damage to the occipital lobe of the brain, where the visual cortex is located (Figure 2.11), resulted in blindness, and that there was a connection between the area of the occipital lobe that was damaged and the place in visual space where the person was blind (Glickstein & Whitteridge, 1987; Holmes & Lister, 1916; Lanska, 2009). For example, damage to the left part of the occipital lobe caused an area of blindness in the upper-right part of visual space. As noted earlier, other areas of the brain have also been associated with specific functions. The auditory cortex, which receives signals from the ears, is in the upper temporal lobe and is responsible for hearing. The somatosensory cortex, which receives signals from the skin, is in the parietal lobe and is responsible for perceptions of touch, pressure, and pain. The frontal lobe receives signals from all of the senses and is responsible for coordination of the senses, as well as higher cognitive functions like thinking and problem solving. Another effect of brain damage on visual functioning, reported in patients who have damage to the temporal lobe on the lower-right side of the brain, is prosopagnosia—an inability to recognize faces. People with prosopagnosia can tell that a face is a face, but they can’t recognize whose face it is, even for people they know well such as friends and family Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 39 4/18/18 4:22 PM 40 CHAPTER 2 Cognitive Neuroscience members. In some cases, people with prosopagnosia look into a mirror and, seeing their own image, wonder who the stranger is looking back at them (Burton et al., 1991; Hecaen & Angelergues, 1962; Parkin, 1996). One of the goals of the neuropsychology research we have been describing is to determine whether a particular area of the brain is specialized to serve a particular cognitive function. Although it might be tempting to conclude, based on a single case of prosopagnosia, that the damaged brain area in the lower temporal lobe is responsible for recognizing faces, modern researchers realized that to reach more definite conclusions about the function of a particular area, it is necessary to test a number of different patients with damage to different brain areas in order to demonstrate a double dissociation. METHOD Demonstrating a Double Dissociation A double dissociation occurs if damage to one area of the brain causes function A to be absent while function B is present, and damage to another area causes function B to be absent while function A is present. To demonstrate a double dissociation, it is necessary to find two people with brain damage that satisfy the above conditions. Double dissociations have been demonstrated for face recognition and object recognition, by finding patients with brain damage who can’t recognize faces (Function A) but who can recognize objects (Function B), and other patients, with brain damage in a different area, who can’t recognize objects (Function B) but who can recognize faces (Function A) (McNeal & Warrington, 1993; Moscovitch et al., 1997). The importance of demonstrating a double dissociation is that it enables us to conclude that functions A and B are served by different mechanisms, which operate independently of one another. The results of the neuropsychology studies described above indicate that face recognition is served by one area in the temporal lobe and that this function is separate from mechanisms associated with recognizing other types of objects, which is served by another area of the temporal lobe. Neuropsychological research has also identified areas that are important for perceiving motion and for different functions of memory, thinking, and language, as we will see later in this book. Localization Determined by Recording from Neurons Another tool for demonstrating localization of function is recording from single neurons. Numerous studies, mostly on animals, used single-neuron recording to demonstrate localization of function. For example, Doris Tsao and coworkers (2006) found that 97 percent of neurons within a small area in the lower part of a monkey’s temporal lobe responded to pictures of faces but not to pictures of other types of objects. This “face area,” as it turns out, is located near the area in humans that is associated with prosopagnosia. The idea that our perception of faces is associated with a specific area of the brain is also supported by research using the technique of brain imaging (see Chapter 1, page 18), which makes it possible to determine which areas of the brains of humans are activated by different cognitions. Localization Demonstrated by Brain Imaging We noted in Chapter 1 that technological advances that cause a shift in the way science is done can be called a revolution. On that basis, it could be argued that the introduction of the brain-scanning techniques positron emission tomography (PET) in 1976 and functional magnetic resonance imaging (fMRI) in 1990 marked the beginning of the “imaging revolution.” Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 40 4/18/18 4:22 PM Localized Representation 41 As you will see throughout this book, brain scanning, and especially fMRI, has played an important role in understanding the physiological basis of cognition. Here we consider what fMRI research tells us about localization of function in the brain. We begin by describing the basic principle behind fMRI. METHOD Brain Imaging Functional magnetic resonance imaging (fMRI) takes advantage of the fact that neural activity causes the brain to bring in more oxygen, which binds to hemoglobin molecules in the blood. This added oxygen increases the magnetic properties of the hemoglobin, so when a magnetic field is presented to the brain, these more highly oxygenated hemoglobin molecules respond more strongly to the magnetic field and cause an increase in the fMRI signal. The setup for an fMRI experiment is shown in Figure 2.17a, with the person’s head in the scanner. As a person engages in a cognitive task such as perceiving an image, the activity of the brain is determined. Activity is recorded in voxels, which are small, cube-shaped areas of the brain about 2 or 3 mm on a side. Voxels are not brain structures but are simply small units of analysis created by the fMRI scanner. One way to think about voxels is that they are like the small, square pixels that make up digital photographs or the images on your computer screen, but because the brain is three-dimensional, voxels are small cubes rather than small squares. Figure 2.17b shows the result of an fMRI scan. Increases or decreases in brain activity associated with cognitive activity are indicated by colors, with specific colors indicating the amount of activation. It bears emphasizing that these colored areas do not appear as the brain is being scanned. They are determined by a procedure that involves taking into account how the brain is responding when the person is not engaged in a task and the change in activity triggered by the task. Complex statistical procedures are used to determine the task-related fMRI—the change in brain activity that can be linked specifically to the task. The results of these calculations for each voxel are then displayed as colorful activation patterns, like the one in Figure 2.17b. Many of the brain-imaging experiments that have provided evidence for localization of function have involved determining which brain areas were activated when people observed pictures of different objects. ➤ Figure 2.17 (a) Person in a brain scanner. (b) fMRI record. Colors indicate locations of increases and decreases in brain activity. Red and yellow indicate increases in brain activity; blue and green indicate decreases. Photodisc/Getty images (Source: Part b from Ishai et al., 2000) Percent Activation (a) –1 (b) 0 +1 +2 Copyright 2019 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. WCN 02-200-202 08271_ch02_ptg01.indd 41 4/18/18 4:22 PM CHAPTER 2 nonpreferred PPA Cognitive Neuroscience preferred 42 nonpreferred EBA preferred (a) (b) ➤ Figure 2.18 (a) The parahippocampal place area (PPA) is activated by places (top row) but not by other stimuli (bottom row). (b) The extrastriate body area (EBA) is activated by bodies (top) but not by other stimuli (bottom). (Source: Chalupa & Werner, 2003) ➤ Figure 2.19 Four frames from the movies viewed by participants in Huth et al.’s (2012) experiment. The words on the right indicate categories that appear in the frames (n 5 noun, v 5 verb). (Huth et al., 2012) Looking at Pictures We’ve already seen how neuropsychology research and single neuron recording identified areas that are involved in perceiving faces. A face area has also been identified by having people in a brain scanner look at pictures of faces. This area, which is called the fusiform face area (FFA) because it is in the fusiform gyrus on the underside of the temporal lobe (Kanwisher et al., 1997), is the same part of the brain that is damaged in cases of prosopagnosia (Figure 2.11). Further evidence for localization of function comes from fMRI experiments that have shown that perceiving pictures representing indoor and outdoor scenes like those shown in Figure 2.18a activates the parahippocampal place area (PPA) (Aguirre et al., 1998; Epstein et al., 1999). Apparently, what is important for this area is information about spatial layout, because increased activation occurs when viewing pictures both of empty rooms and of rooms that are completely furnished (Kanwisher, 2003). Another specialized area, the extrastriate body area (EBA), is activated by pictures of bodies and parts of bodies (but not by faces), as shown in Figure 2.18b (Downing et al., 2001). Looking at Movies Our usual experience, in everyday life, involves seeing scenes that contain many different objects, some of which are moving. Therefore, Alex Huth and coworkers (2012) conducted an fMRI experiment using stimuli similar to what we see in the environment, by having participants view film clips. Huth’s participants viewed 2 hours of film clips while in a brain scanner. To analyze how the voxels in these participants’ brains responded to different objects and actions in the films, Huth created a list of 1,705 different objects and action categories and determined which categories were present in each film scene. Figure 2.19 shows four scenes and the categories (labels) associated with them. By determining how each voxel responded to each scene and then analyzing his results using a complex statistical procedure, Huth was able to determine what kinds of stimuli each voxel Movie Clip Labels Movie Clip Labels butte.n desert.n sky.n cloud.n brush.n city.n expressway.n skyscraper.n traffic.n sky.n woman.n talk.v gesticulate.v book.n bison.n walk.v grass.n stream.n Copyright 2019 C