Chapter 17 Chapter 15 Some Experiments

"We cannot acquire any knowledge of the empirical world merely by purely logical thinking." --Albert Einstein A particular neuron in a monkey's brain might be sensitive to the color of a particular area of ​​the visual field.But how can we be sure that it is directly involved in the perception of that color?Maybe it's just part of the brain's system for directing attention to that area of ​​the visual field, for example.If so, a person who lost the neurons that perceive true color due to brain damage would see the world only in black and white, but his attention might still be drawn to a patch of color.

This is not just an abstract possibility.Alan Cowey of Oxford and colleagues studied in detail a man who had lost color perception (in layman's terms, he could see only black, white and different shades) due to brain damage. grey).They pointed out that as long as two small color squares (adjusted to equal brightness) are placed next to each other in the experiment, the subjects can tell whether the two squares have the same color.In fact, the subject categorically denied that he could perceive the colors of the two squares.If the two squares are not next to each other, he cannot complete the task, and his judgment is completely a guess.This shows quite clearly that the brain can still use some information about color when it is not perceiving color.

In order to find out whether the response of certain neurons in the monkey's brain was related to what it saw, William Newsome of Stanford University performed a series of remarkable experiments.The cortical region chosen for the experiments was the MT area (sometimes referred to as "V5").The neurons here respond well to motion, but not directly to color, or not at all (see Chapter 11).Experiments have shown that damage to this area makes it difficult for monkeys to respond to visual movements.But the impairment often fades after a few weeks, perhaps because the brain has learned to use other pathways.

Following earlier work by others, Newsom and colleagues first studied how individual neurons in the MT respond to selected motor signals.These signals consist of rapidly changing patterns of random dots displayed on a television screen.An extreme case is where all of these transient points are moving in one direction.This movement is easily recognizable.The other extreme is to have zero average movement of the points, which is like sometimes seeing "snowflakes" on the screen when changing channels on a TV.The observer must report whether the motion is in a given direction or in the opposite direction, and when the mean motion is zero, the outcome is random.

Newsom and colleagues used various combinations of these flashing patterns.If all the motion is in one direction, the monkey (or human) will always correctly signal the direction of motion. If only some points are moving in one direction and the other points are moving randomly, the observer will sometimes make mistakes.The smaller the proportion of points moving in that particular direction, the more mistakes are made.By varying this ratio, it is possible to plot the accuracy of the observer as a function of the percentage of points with the same direction of motion. ①Using a special mathematical method, find out those who are judging the direction of motion in the most effective way.

In all, they studied more than two hundred different neurons.About a third of these neurons made judgments with the same accuracy as the monkeys.Some judged poorly, but others judged movement much better than the monkeys.So, since the monkey has these cortical neurons in its brain, why isn't it more successful at making judgments?The most likely answer is that the monkey cannot just choose one neuron (that is, the one that judges most effectively) to control its response.Its brain must have used a population of neurons.It's unclear how it does this. This experiment does show that the visual information required to make a choice is present in the behavior of neurons in the MT area, so we cannot say that those neurons are incapable of this task.Unfortunately, this does not prove that they actually performed this task.

Newsom's next experiment went a step further.He and his colleagues asked the question: Could the monkeys' judgment be improved if we could appropriately stimulate neurons in the MT area and make them fire when they were performing difficult discrimination tasks? It is technically not easy to stimulate just one neuron.Fortunately, in the MT area of ​​the cortex, neurons with similar response patterns (that is, responses to a specific direction of motion at a specific part of the visual field) usually form clusters with each other.In this way, electrically stimulating a small area close to the target neuron will likely cause these neurons with similar characteristics to be stimulated together.

They performed a total of 62 experiments.The electrical stimulation significantly improved the monkeys' discrimination of motion in about half of the cases, a rather surprising result.It means that by firing neurons in the appropriate places in the visual cortex, we can change the way monkeys respond to specific visual stimuli.Current must be applied at this specific location.If the current stimulates other parts of the MT area of ​​the cortex, it has little effect on the monkey's completion of this special task. Does this mean that a small area of ​​the MT area is involved in the neural connections that recognize that movement?This is certainly possible, but there are many difficulties in affirming this conclusion.

One possible objection is that although the monkey displayed the appropriate (discrimination) behavior, it didn't actually see anything.It merely responds like an automaton, without visual perception, and a definitive answer to this objection would require a complete knowledge of the monkey and human visual systems; therefore, at present, we can only assume that monkeys have visual perception , until evidence shows otherwise. One could also argue that even though the monkey had visual awareness, it did not develop visual awareness for this particular task.This seems unlikely, since monkeys and humans make similar choices in this task, that is, their psychometric curves are fairly consistent.Monkeys did not perform much worse than humans.It is likely that both brains employ similar mechanisms; however, there is a difficulty.

If a person performs this task repeatedly, his behavior becomes almost mechanical, and he will report that he barely glimpses the movement, but nevertheless chooses much better than by chance.Since the task cannot be described to monkeys in words, it is more difficult to train than humans.Newsom's monkeys may have been overtrained so that their behavior became more or less mechanical, with little visual perception involved. I doubt that this objection matters.Because when all the flashing points of light are moving in one direction, we see this movement very clearly, and almost certainly the monkey sees it too.Unfortunately, the stimulation current caused little difference in this case, since the monkeys were already performing the task almost as well.Perhaps an experiment could be conducted in which the monkey first learns to identify the direction of movement of another motor stimulus (such as a directed rod) and is tested with this moving blip before it is overtrained.This kind of experiment is not easy to do because of its risks, but it may be worth a try.

A more vehement objection is that although the behavior of neurons in the MT area of ​​the cortex appears to be associated with discrimination in monkeys, and thus may also be associated with visual perception, this does not mean that these specific neurons are responsible for generating perception. known places.They may affect other neurons (perhaps elsewhere in the visual hierarchy) by firing.And those neurons are really related to awareness. The only way to answer this question is to study other cortical regions.If we fail to find neurons with similar discriminative abilities elsewhere, the likelihood that neurons in the MT area are associated with awareness increases.In the long run, until we know more about all visual areas, especially how they are interconnected, we cannot hope to narrow down the areas of visual perception.Regardless, some of Newsom's experiments are a very important first step in this direction of research. If certain stimuli in the visual field cause the neurons in question to fire, we naturally suspect that this neuron might be the neural counterpart to those stimuli.However, as just explained, this conclusion is not inevitable.Are there some more efficient ways to narrow down the search for perceptual neurons?Can we find a situation in which the visual input is kept constant, but the perception is changing?Then we could try to find out which neurons in the monkey's brain fire as a function of input; more importantly, which neurons change as a function of perception. A notable case is the observation of Neck cubes (see Figure 4).At this point the figure remains the same, but when we see it as three-dimensional, the perception begins as one form, then another, and so on.It is not clear what part of the brain has the perception of a three-dimensional cube.We should investigate certain situations that are easy to localize in the monkey visual system. One notable possibility is based on the known phenomenon of binocular rivalry.This condition occurs when the two eyes receive different visual input related to the same part of the visual field.The primary visual system on the left side of the head receives input from the field of view to the right of the gaze point of the eyes (and vice versa on the right).Two conflicting inputs are said to be "competing" if the inputs on both sides cannot be fused, but instead see one input and then the other, alternating and so forth. You can see a rather dramatic example of binocular rivalry at the San Francisco exposition.It was designed by Sally Duensing and Bob Miller.In the exposition demonstration, the observer places his head in a fixed position and keeps his gaze fixed.Through a properly placed mirror, one eye of the observer sees the face of another person in front of him, while the other eye sees a blank screen to the side.If the observer shakes his hand in front of this screen, the face is erased from its original position in his vision!Hand movements are visually salient, in a way that grabs the brain's attention.If you don't pay attention, you can't see the face.If the observer moves his eyes, the face reappears. In some cases only part of the face disappeared.For example, one or both eyes are sometimes left behind.If the observer is looking at a smile on a person's face, there will be a situation where the face disappears and only the smile remains.For this reason, this effect is known as the "Cheshire Cat effect" (Cbeshire Cateffect, named after the cat in the cat).You can experiment with a simple pocket mirror yourself.It turned out to be very interesting.The experiment works better if there is a uniform white background behind both the observed and the observer's hands. No such experiments have been performed on monkeys so far.A much simpler experiment was performed at MIT.Nikos Logothetis and Jeffrey Schall trained rhesus monkeys to make judgments about seeing horizontal gratings moving up or down.To create binocular rivalry, an upward-moving grating was projected into one eye of the monkey, while a downward-moving grating was projected into the other eye, causing the two images to overlap in the monkey's field of view.It turned out that the monkey alternately signaled that it saw upward and downward motion, the same way we would respond.Note that the motor stimulus that reaches the monkey's eyes is always the same, while the monkey's perception changes approximately every second. ① The MT area of ​​the cortex mainly detects motion and is not interested in color.What happens to the behavior of neurons in the MT area for short periods of time when the monkey's sense of motion is sometimes up and sometimes down?The answer was that some neurons fired in response to sensation, while the average firing rate of the remaining neurons remained relatively constant regardless of the direction in which the monkey was seeing movement. (The actual data is much messier than this simple description.) This result suggests that it is unlikely that all of the neurons in the visual cortex firing at a certain moment are related to our visual perception.Of course, it would be better if there were more such examples.Unfortunately, this does not precisely define the location of the perception neurons.As explained for Newsome's experiments, the real link may be the firing of neurons elsewhere in the visual hierarchy that are affected (at least in part) by the firing of those in the MT.Rama Canjun once suggested that this kind of competition may not be a competition of real movement, but a competition of shape, and its real location may lie in the lower level of the visual hierarchy system, perhaps in the V1 or V2 area of ​​the cortex.Also, even if some perception neurons are indeed located in the MT area, the present results do not confirm which neurons they are.In which cortical areas are they located? Which classes ① Time intervals follow the r distribution.What type of neurons tend to change in response to perceptual rather than visual input?As with the discussion of Newsom's results, there is also the possibility that the monkeys were overtrained.While this is unlikely, as training has little impact on competition, it still raises concerns.Again, even with reservations, these are important experiments.Further research will lead us to explain visual perception in neural terms. Under other conditions, is there a situation where the visual input is constant but the perception changes for some reason?Of course, sometimes the observer will suddenly "see" an object that was not discovered before, as in the hidden Dalmatian dog in Figure 9.But it's not easy to do this kind of research in monkeys.People say, "Look, I see a dog now. I haven't seen a dog before." It's much harder to get a monkey to tell us that.Furthermore, once the observer identified the dog from the picture, he was usually able to recognize it directly in subsequent experiments.It is therefore difficult to repeat the same experiment many times.And this repetition is exactly what is necessary to obtain scientifically reliable results. One possibility is to study the effects in the brain of images that fade from awareness.These images are stabilized on the retina. (Recall that we usually prevent this fading by making all sorts of small eye movements.) Initially, a small device was placed on the eyeball to stabilize the image on the retina, which was very uncomfortable for the eye.It projects selected optical patterns onto the retina.No matter how the eye moves, the pattern remains in the same place on the retina and thus gradually fades away. Several experiments of this kind were carried out in the 1950s, but since then, although the devices for producing stable images have become more sophisticated and comfortable, they do not seem to have been carried out since then. One might think that this regression process occurs mainly in the retina and is therefore of little interest to us.But that seems unlikely to be true.These earlier studies showed that complex images do not always fade as a whole.A straight line is often taken as a whole, but several sides that make up a square or triangle may disappear independently.Zigzag shapes are less stable than arcs.The activities of what Gestalt psychologists call "good graphics" are more holistic than those of "bad graphics".If there is a pattern that is a capital B with rough curved lines running through it, the curved lines usually fade earlier than the letter day.This suggests that extinction mainly occurs in the brain, rather than in the eyes.So it's worth trying, for example, to train a macaque to signal what it sees while awake.Fix various patterns on its retina and see which neurons are affected when parts of the image fade from awareness. Another possibility is further study of Rama's compelling experimental reports (see Figure 19).Artificially damaging a small portion of the monkey's cortex V1 creates a localized blind spot (called the "blind spot").The experiment involved the apparent movement of two unaligned parallel line segments at rest when they touched this blind zone.This research could be possible in monkeys if we could train them to report with signals, to distinguish between motion and inactivity, alignment and misalignment, interruption and continuation, and so on.As far as I know, no one has tried it so far. A simple experiment has been done in monkeys' true blind spots. (See Chapter 3 for a psychological description of our blind spots.) There is an area in V1 that corresponds to the blind spot, where the cortex only receives direct input from one eye, while the other eye's photoreceptors do not cover the field of view this part of the . (Recall that most of the neurons in area V1 on one side of the brain receive input from both eyes, although it only processes information from the opposite half of the field of view.) One might think that the neurons in the area of ​​the point of view only respond to inputs from one eye. Eye signals respond.Surprisingly, this is not true.Rlcardo Gattass of the Federal University of Rio de Janeiro and colleagues have shown that some neurons in macaque monkeys' blind spots do respond to input from both eyes.This unexpected input from an eye that is partially blind in this area may come directly or indirectly from adjacent cortical tissue that receives input from both eyes.Regardless of where it comes from, experiments have shown that the neurons in the blind spot of V1 respond to the signal by firing pulses in the manner described in Chapter 3, and fill in the external graphics.At the same time, this decisively rejects Dennett's argument (outlined in Chapter 4).Such a neat example illustrates a general principle: Whenever you clearly see a feature of a visual scene, there must be neurons firing whose activity clearly symbolizes that feature. (Another example of this principle is the neural response to subjective contours described in Chapter 11.) Compared with the usual example of neural responses elicited by visual input, this particular blind spot phenomenon provides us with little information about the localization of perceptual neurons.It would be helpful in our quest if it could be extended, as suggested earlier, to study how perception varies for constant visual input (Fig. 19). Another avenue is to study the conditions under which different visual inputs produce the same perception, or at least some of the building blocks of that perception.An example is an experiment by Tom Albright of the Salk Institute and collaborators in the MT area of ​​macaque cortex.It turns out that the firing of certain neurons in the MT area has remarkably consistent properties even when the studied moving objects vary considerably.For example, when a piece of ripples moves across the visual field, they cause certain neurons to fire in the MT area in much the same way as a straight rod moving in the same direction at the same location.Although the patterns are different, their movements are similar. (They call this "shape-cue invariance".) So far they have not shown whether there is anything special about the type, location or firing behavior of this neuron.If they were perception neurons, we might expect that their firing (or some of their properties) would always correlate with visual perception, whatever the input signal. Since the evidence is so far weak, it is reasonable to ask the question: Can one precisely study the behavior of the same neuron when the animal is alert and when it is unconscious?For technical reasons, it is difficult to perform such experiments when the animals are anesthetized and unconscious.However, there are experiments comparing cats in a state of alertness and slow-wave sleep. ① Neuroscientists Margaret Livingstone and David in 1981.Huber published such an experiment.Most of the neurons they studied were in the V1 area of ​​the cortex. ②The animal's eyes are open, so the neurons in V1 respond to the visual signals generated by the computer on the screen placed in front of the animal even during slow-wave sleep.When they recorded a response from a particular neuron, they woke the animal up and tested it again with the same stimulus it had just seen. When the animal was awake, each neuron they studied responded in roughly the same pattern as when it was asleep, that is, if it was sensitive to a line with a certain orientation at a certain location in the visual field, it did so regardless of whether the animal was awake or sleeping. state, its optimal stimulus is the same, but usually the signal-to-noise ratio is better when awake. ① In any case, the firing rate of a considerable number of cells is higher when animals are awake than when they are asleep.Perhaps this is not surprising, but the interesting result is that the changes in responses are more pronounced in the lower layers of the cortex (layers 5 and 6) than in the higher layers. They used a chemical substance (radioactive 2-deoxyglucose) to confirm this general result.This substance can reveal the average behavior generated by visual stimuli at these cortical levels.These behaviors are averaged over a period of about half an hour.In one case the animals were awake, and a different radioactive isotope was used as a comparison when the animals were sleeping.The results are roughly the same.When the animal was conscious, there was a dramatic change in behavior in the lower layers of the cortex, but little change in the higher layers. This prompts such a broad inference that goes far beyond the current evidence.That is, activity in the higher layers of the cortex is largely unconscious, while at least some lower layer neurons are associated with consciousness.I must admit that I love this assumption too much.If so, it would be wonderful.But I can't take it wholeheartedly, maybe there is something else that makes the lower levels of slow wave sleep less active. Can we learn anything about awareness by studying the mechanisms of attention?Experimental studies of the neural mechanisms of attention have been underway for some time.Some experiments were done on conscious monkeys.They have recorded the firing of neurons in various parts of the brain when monkeys perform specific visual tasks, and some experiments have used PET scans as described in Chapter 8 in humans.I do not intend to repeat all of these experiments; instead, I will only briefly describe one of them and its results. Robert Desimone of the National Institute of Mental Health in Bethesda, Maryland, and colleagues trained monkeys to fixate on a point on one side of a visual display and attend (fixedly) to a feature of the display. predict.Various signals then flashed.The experimenters studied the response of a specific neuron in the V4 area of ​​the cortex, which is more sensitive to color, to the visual display at that location.Suppose the neuron under study responds to a red rod with a certain orientation, while a green rod has no effect on it. (Of course, some of the other neurons in V4 that were not studied at this time would also respond to green rods rather than red rods.) Each display included two colored rods, one red (the one for this neuron). effective stimulus), while the other is green (ineffective stimulus).Both are within the receptive field of the neuron.When the monkey paid attention to the position occupied by the red rod, the neurons fired the same as when the monkey did not pay attention, or higher. ①However, in experiments in which the monkeys focused on the green stick, the firing of this red-sensitive neuron was reduced.Attention, therefore, is not just a psychological concept.Its effects can be observed at the neuronal level.When the monkey pays attention to one place, the firing of neurons sensitive to the attended stimulus will increase, and when the monkey pays attention to other places, even though the eye position and the input visual information are exactly the same as last time, the firing of that neuron will also be weakened. . They describe the results thusly: Neurons in area V4...have such a large receptive field that many stimuli fall into it.One might expect that the behavior of such a cell would be characteristic of all stimuli in its receptive field.It has been found, however, that when a monkey restricts its attention to one location in the receptive field of a V4... "shrink" the same. Since understanding them is not easy, I will not describe their results in detail.They point out that the simple theory about the searchlight of attention doesn't seem to be true.To explain them requires a more complex mechanism, which has not yet been established.Is the thalamus involved in attention?As the "gateway to the cortex," the thalamus has a number of rather distinct regions, some of which are related to vision.The main pathway from the eye to the cortex needs to pass through the Laterai Geniculate Nucleus (abbreviated as LGN).The lateral geniculate body is part of the thalamus (described in Chapter 10). (Primate) Other thalamic visual areas are located in an area called the "postthalamic tubercle". ① It is a large thalamic nucleus, apparently much larger than the lateral geniculate body. David Lee Robinson of the National Eye Institute in Bethesda and colleagues performed extensive experiments on a portion of the posterior thalamic tubercle in monkeys.It appears that features eliciting responses in the posterior tubercle of the thalamus depend on their input from the visual cortex.rather than from the upper hill. ② If inhibition of a small area of ​​the posterior tubercle of the posterior thalamus was increased chemically, it would be more difficult for monkeys to shift attention; conversely, lowering inhibition would make shifting easier.Some experiments by others have suggested that the posterior thalamic tubercle plays a role in suppressing input from unrelated events.A study of three patients with thalamic lesions showed that they had difficulty forming attention.PET scans of normal individuals showed increased activity in the posterior tubercle of the thalamus when a visual task distracted them.These distractors made the participants try to pay more attention to complete the task.All these results (see ref. 13 for a review article) strongly suggest that these parts of the thalamus are closely related to multiple aspects of visual attention. ① There is still a wide area for further work here.Further more detailed studies are needed to investigate the exact connectivity of each of the above-mentioned PTT regions.For example, how are the corresponding regions of several retinal regions wired differently?Can we understand more precisely how each specific part of the posterior thalamic tubercle affects attention and how it interacts with neurons in the various cortical areas it is associated with?Further experimental work should answer these questions. (I discuss some speculative ideas about different regions of the posterior tubercle in Chapter 16.) How much do we learn about the neural mechanisms of visual perception from studies of the thalamus?Since attention is important to awareness, it would be foolish to ignore it.To unravel the mysteries of vision, we need to understand not only how the neocortex works, but also the lateral geniculate body and the posterior tubercle of the thalamus. Can relevant experiments be done on humans instead of monkeys?The beauty of this type of experiment is that the subjects can verbally report their experiences, which the monkeys cannot.However, inserting electrodes into a person's brain is unlikely for ethical reasons, although it is sometimes necessary for medical treatments. It is also possible to study brain waves from outside the skull, but the results are often more difficult to interpret. This method was originally developed by Benjamin Libet at the California State University, San Francisco.He likes to experiment on humans because he has reason to believe that other people are conscious. (He's less confident that monkeys are conscious, too.) In the past, not just psychologists and neuroscientists, but practicing doctors, have been seriously skeptical of any experimental work on consciousness.For surgeons and anesthesiologists, almost the only interest is how to anesthetize the patient during the operation so that he is not aware of what is going on.This is done, in part, to alleviate the suffering of patients and in part to prevent patients from suing them. (Ribet told me that he was wise not to experiment with consciousness in conscious people until he was tenured.) Ribet's main work concerns certain brain waves that precede voluntary movement, and how these events in the brain relate to the timing of the subject's awareness of attempting or wishing to move. ① His results showed that for this form of conscious awareness there must be some minimum duration (around 100 milliseconds) of neural activity.The exact value of this time may depend on the strength of the signal and the environment. Some of his other, more recent work concerns the effects of stimulating a part of the thalamus, the ventrobasal complex.The ventrobasal complex is primarily involved with sensations such as touch and pain, and this experiment has been done in some patients where electrodes placed in this part of the thalamus relieve their uncontrollable pain.Although these experiments did not involve vision, they may be relevant to the interpretation of blindsight (as discussed in Chapter 12).So I will describe them. The subjects' thalamus received a certain amount of stimulation.He (or she) then needs to judge (guess if necessary) when the stimulus was presented.More precisely, he said, determine whether the stimulus occurred within 1 second of a particular light being turned on, or within 1 second of a different light being turned on later.The subject indicated his choice by pressing one of the two buttons provided.If he doesn't know when the stimulus will appear, he has to guess, so on average with 50% accuracy, when the stimulus and response are over, he needs to press one of the three buttons to indicate whether he has ever noticed the stimulus .If the subject had perceived the stimulus in the usual position, even if briefly, he should press the first button.If he's unsure, or thinks he might feel something, press the second button.If he just feels nothing, press the third button. The experiments designed by Ribet and colleagues are so complex that I will only describe their general results.Stimulation consisted of 72 electrical pulses per second; different numbers of pulses were delivered in different experiments, and their amplitude was kept constant.The results showed that even when the pulse trains were too brief to induce awareness, the subjects performed better than random selection.And perceiving stimuli (even with uncertainty) requires quite long sequences. This implies, Ribet and colleagues explain, that pulsed stimuli of a certain duration are required to develop awareness, but unfortunately they did not systematically vary the intensity of stimuli in these experiments.However, these and earlier work have shown that increasing the intensity of a fixed-time sequence can change the response of the subjects, that is, from the unconscious state to the aware state.Briefly, in the somatosensory system, a weak or transient signal can affect behavior without eliciting awareness, whereas a stronger or prolonged stimulus of the same form can cause awareness.The precise neural behavior elicited by stronger or longer stimuli remains to be determined. This result implies that, when trying to explain blindsight, we cannot ignore a similar explanation that the pathway from the lateral geniculate body to areas such as V4 is too weak to produce visual awareness but sufficient for human behavior make an impact. Although the experiments described in this chapter cannot yet draw any strong conclusions about the precise neural correlates of visual perception, they do suggest that it is possible to study consciousness through experimental approaches.Such experiments must eventually lead to the solution of the problem, if we pursue them with enthusiasm and persistence. A parallel avenue is to try to guess the general nature of the answer and use it only as a guide for further experiments.Without this guidance, experiments cannot be performed, some of the speculative points of which are outlined in the next chapter.They have so far formed less of a coherent collection of ideas than of a hodgepodge of tentative suggestions.However, as we shall see, some of them can be grouped together reasonably. ① Such a curve is called a "psychometric curve". ① During rapid eye movement (REM) sleep, brain waves are similar to those of waking, suggesting that the brain is at least partially conscious at this time, just as we seem to be conscious when we dream.The brain waves of slow-wave sleep (non-rapid eye movement) are very different from those of alertness, and dreams are rarely seen at this time.It is therefore reasonable to assume that during slow wave sleep we are normally unconscious. ②They also tested some neurons in the lateral geniculate body. ① That is, the ratio of neuron firing rate to stimulus to background firing rate is higher. ① If the task is simple, then the distribution is about the same.If color discrimination becomes more difficult, attention will increase the firing rate. ①The posterior thalamic tubercle consists of three main parts and one smaller part.Two of these, the anterior and the lateral, correspond to areas of the retina, each with one or more projections to the visual field.They have bidirectional connections to most primary visual areas and receive strong, non-bidirectional connections from the superior colliculus.The third part, called the posterior tubercle of the midthalamus, does not have a correspondence with the retinal area, but has bidirectional connections mainly with the parietal and frontal lobes.It may respond to other senses, not just vision. It may be more involved in cognitive processes and less involved in forming vivid visual perceptions. ②回想一下，上丘与眼动控制有密切联系，而眼动控制是视觉注意的另一种形式。另一方面，从上丘到丘脑后结节的输入，看来更多的与视野不同部位中显著特征有关。 ①安德森（Jim Anderson）和范·埃森（Davidvan Essen）也提出了这种观点，作为他们的移动回路理论的一部分。 ①由于两个原因我将不描述这些实验：它们并不直接与视觉系统有关而这是我涉及的主要方面，而且很难解释并引起了争论。这样，如果要全面讨论它们，要用一定的篇幅来描述。这作为一个旁证来说大长了。它们更多的与自由意志问题有关，将在跋中作简要讨论。