Department of Psychology, Princeton University, Princeton, NJ 08544
Michael Graziano, Phone : (609) 258-7555, Fax : (609)258-1113, Email : firstname.lastname@example.org
Topics Studied in the Graziano Lab...
In the 1990s I studied neuronal coding of peripersonal space. This research focused on multisensory neurons in the monkey brain that encode the locations of objects near the body.
In the 2000's my lab studied the organization of the motor cortex. This research focused on a map of complex, ethologically relevant movements in the motor cortex of monkeys.
Since 2010 my lab has begun to study the brain basis of consciousness. This research has focused on the relationship between awareness, attention, and social perception in the human brain.
These three topics are summarized below, the most recent first.
1. Consciousness and the social brain: a new direction for the lab.
Recently the focus of my lab turned to the brain basis of awareness. How does an information-processing machine produce subjective awareness?
Consciousness is a large, complex, and difficult topic. My conscious mind encompasses a great deal of information, including memories, knowledge about myself (self awareness), and information about the outside world. But the problem can be simplified at least somewhat by focusing on the more specific property of awareness, rather than on the bewildering range of information about which I can be aware. Some process allows me to be aware of information. What is the process of awareness? Why does it apply to some information in the brain but not all?
In my lab we are pursuing a specific theory of awareness, the "attention schema" theory (Graziano and Kastner, 2011). The theory is described in detail in a book (Graziano, Oxford UP, in press). We proposed that specialized machinery in the brain computes the feature of awareness and attributes it to other people in a social context. The same machinery, in that hypothesis, also attributes the feature of awareness to oneself. Damage to that machinery disrupts one's own awareness.
The attention schema theory was motivated by two sets of previous findings.
First, certain regions of the cortex are recruited during social perception as people construct models of other people's minds. (e.g. Brunet et al., 2000; Ciaramidaro et al., 2007; Fletcher et al., 1995; Gallagher et al., 2000; Goel et al., 1995; Saxe and Kanwisher, 2003; Saxe and Wexler, 2005; Vogeley et al., 2001). These regions include, among other areas, the superior temporal sulcus (STS) and the temporoparietal junction (TPJ) bilaterally but with a strong emphasis on the right hemisphere.
Second, when these same regions of cortex are damaged, people suffer from a catastrophic disruption of their own awareness of events and objects around them. The clinical syndrome of hemispatial neglect, or loss of awareness of one side of space, is particularly profound after damage to the TPJ or STS in the right hemisphere (Karnath et al., 2001; Valler and Perani, 1986).
The conjunction of these two previous findings suggests that awareness is a computed feature constructed by an expert system in the brain. The feature of awareness can be attributed to other people in the context of social perception. It can also be attributed to oneself, in effect creating one's own awareness.
Why construct the feature of awareness and attribute it to other people?
In order to understand and predict the behavior of other people, it is useful to monitor other people's attentional state. Attention is a data handling method by which some signals in the brain are enhanced at the expense of others. According to the attention schema theory, when the brain computes that person X is aware of thing Y, it is in effect modeling the state in which person X is applying an attentional enhancement to signal Y. Awareness is an attention schema. In that theory, the same process can be applied to oneself. One's own awareness is a schematized model of one's own attention.
In the attention schema theory, awareness is an internal model of attention. But internal models, or schemas, are rarely entirely accurate. They are simplifications. Sometimes they are in error and diverge from reality. Usually, awareness and attention track each other. If you attend to something, you are usually aware of it. But numerous experiments suggest that it is sometimes possible to attend to something and yet be unaware of it. (e.g. Kentridge et al., 2004; Koch and Tsuchiya, 2007; Lamme, 2004). The attention schema theory predicts this dissociation between awareness and attention. In the same manner, the brain contains a body schema, an internal model of the physical body. The body schema usually tracks the body, but is not always accurate, and in laboratory conditions can be dissociated from the actual state of the body.
Human cultures are replete with myths and illusions regarding consciousness: the out-of-body experience, the belief that the center of awareness can leave the body and float through space as a spirit or ghost, the feeling of being stared at, the belief that an invisible force comes out of the eyes when we look at something. Why are such physically impossible notions so prevalent? The attention schema theory may provide one explanation. These myths may reflect intrsinsic inaccuracies in the brain's own internal model of attention. The brain models attention schematically, not as a competitive interaction between neurons that enhances information processing, but as an ethereal, plasma-like substance that is privy to experience.
2. The deep principle of organization of the motor map.
More than 140 years ago, Fritsch and Hitzig (1870) discovered that electrical stimulation of specific sites in the dog brain caused specific parts of the dog's body to twitch. Fritsch and Hitzig did not describe a map of the body as it was later understood, but they found a functional topography. The exact nature of that topography was debated over the next century and a half. If it was a map of the body, it was overlapping, fractured, and in some cases re-represented. If it was composed of many maps, the reason for so many and their different contributions was not entirely clear.
In my lab, in 2002, we reported an unexpected discovery (Graziano et al., 2002). That discovery led to a series of experiments over ten years. Those experiments led to the formulation of an underlying principle that we believe can finally account for the map in motor cortex. The normal movement repertoire of an animal is a complex, diverse, highly dimensional collection of actions with many similarity relationships among different actions. This movement repertoire appears to be arranged on the cortical surface. The complexity and overlap in the movement repertoire is reflected in the complexity and overlap in the map.
The initial discovery that led to this action-map view involved electrical stimulation in the motor cortex of monkeys. (For an integrated review of this body of work, see Graziano, 2008.) In most previous experiments, electrical stimulation was applied to the motor cortex in brief bursts, generally shorter than 50 ms. We extended the stimulation to a behaviorally relevant time scale. We stimulated for half a second because that matched the typical duration of a monkey's reaching and grasping. The longer stimulation train evoked complex movements that were coordinated among many joints and that resembled common movements from the animal's normal behavioral repertoire. Different actions could be reliably evoked from different locations in cortex. Every monkey tested has the same map-like arrangement of actions.
These movements included ethologically relevant behaviors such as closing the hand in a grip while bringing the hand to the mouth and opening the mouth; extending the hand away from the body with the palm facing away from the body and the grip opened as if in preparation to grasp an object; bringing the hand inward to a region just in front of the chest while shaping the fingers, as if to manipulate an object; squinting the facial muscles while turning the head sharply to one side and flinging up the arm, as if to protect the face from an impending impact; and moving all four limbs as if leaping or climbing. The behavioral repertoire of the animal seemed to be rendered onto the cortical sheet.
This initial work became controversial because of the use of a method, stimulation on a behavioral time scale, that was common enough and well accepted in many subfields but was simply not traditional in the study of motor cortex (Strick, 2002). Unfortunately, that controversy over the initial work has tended to cloud the deeper issues and to cloud the many other methods that have been brought to bear.
In my view the most important method applied to the issue was a mathematical and theoretical one. If the map in motor cortex is a two-dimensional rendering of a highly complex movement repertoire, then it should be possible to measure the typical movement repertoire of a monkey, mathematically flatten it on a sheet, and thereby reproduce in great detail the complicated, fractured, overlapping map that is known from the data. We did the study and found just that (Aflalo and Graziano, 2006; Graziano and Aflalo, 2007). Indeed the method worked far better than we could possibly have anticipated. To our astonishment, a mathematically flattened version of a monkey's normal movement repertoire accounted for the overarching organization of the primary motor cortex, the supplementary motor cortex, the dorsal premotor cortex, the ventral premotor cortex, the frontal eye field, and the supplementary eye field. A large swath of cortex, about 20% of the cortical mantle, could be understood as a mapping of the animal's behavioral repertoire.
Other methods that we used to study this map of the behavioral repertoire include: studying muscle activations using short, traditional-style stimulation; single neuron recording during untrained, complex, ethologically meaningful movements; chemical inhibition and disinhibition of cortical regions, to down-regulate or up-regulate a specific behavior; and examination of the natural movement repertoires of monkeys. All of these studies are detailed in a book (Graziano, 2008) and in our many journal publications on the topic.
Some answers to common questions about this line of research are provided next.
i. How do we account for the division between premotor cortex and primary motor cortex?
How can a single, encompassing map of behavioral repertoire be reconciled with the large body of evidence indicating that these cortical areas are distinct in structure, connectivity, and function, and moreover that the SMA and premotor cortex can be further subdivided into many smaller functional areas?
A mapping of behavioral repertoire across the cortical surface does not argue against functional cortical subdivisions. On the contrary, it provides a deep underlying explanation for the subdivisions. In this hypothesis, the separation among areas is driven by the statistical clustering within the movement repertoire. The physiological differences among areas, and the differences in anatomical connectivity, may be partly because different cortical regions emphasize different segments of the animal's movement repertoire that require different control strategies and different patterns of sensory input and motor output. The view of a map of behavioral repertoire therefore seems to be more supportive of, rather than in conflict with, more traditional views of functional divisions in the motor cortex.
ii. How do we account for direction tuning?
One of the most important contributions to motor cortex physiology was the discovery of direction tuning by Georgopoulos et al. (1986, 1988). When a monkey reaches from a central location in a variety of directions, neurons in motor cortex are tuned to the direction of hand movement, broadly preferring one direction over the others. Yet on stimulating the motor cortex, we tended not to evoke a hand movement in a specific direction. Instead, the joints of the arm tended to seek a specific posture, regardless of the initial posture. As a result, the hand tended to move from any initial position toward a specific final position. The hand-to-mouth movements are an especially clear example of this apparent goal-directedness. How can a directional code for hand movement be reconciled with the stimulation-evoked movements to a goal posture?
We performed a single neuron experiment to track down the reason for this apparent discrepancy (Aflalo and Graziano, 2007). In summary, we found strong classical direction tuning in agreement with all previous studies, but only for movements confined to local regions of space and confined to a limited range of joint configurations. Once the entire workspace of the hand, and more importantly the entire 8-dimensional joint-space of the arm, were explored, direction tuning was present but minimal, and tuning to the goal configuration of the arm became dominant. Tuning to the configuration of the arm was too broad to be observed over local regions of the workspace.
Our suggestion was that direction tuning, and tuning to many other parameters such as speed or force or acceleration, must of course be important for motor control. However, most arm movements in the monkey's repertoire are variations on or adjustments around a relatively small set of common goal postures: for example, the hand at the mouth with the forearm supinated; the arm outstretched with the forearm pronated and the grip opened; the arm at the side with the shoulder adducted and the elbow bent, such that the hand is placed within the central space in front of the chest. The postures that we obtained on stimulation may be these canonical postures of the arm typical of general classes of behavior.
If zones within motor cortex relatively emphasize different categories of behavior, then why don't lesions in motor cortex result in loss of specific behavioral functions? Two studies address this question.
We found that chemical inhibition of the defense-related zone in motor cortex had a selective effect, reducing the monkey's ability to flinch (Cooke and Graziano, 2004).
Perhaps the best answer to the question comes from a study by Ramanathan et al. (2006). They found that electrical stimulation of the rat cortex evoked complex movements including reaching movements of the forepaw. When the reaching zone was lesioned, the rat's ability to reach was selectively impaired. Yet the impairment was temporary. The rats were able to re-learn the action. Once the rat re-learned, the motor cortex was found to have developed new regions from which reaching could be evoked. The extent of the rat's recovery of reaching ability correlated with the size of the new reaching representation in cortex. These results suggest that local lesions to specific zones in motor cortex do cause selective impairment of action types, but the deficit recovers rapidly as the functional map re-organizes.
Electrical stimulation must affect many connected structures. How does one know if an evoked movement is truly a function of the directly stimulated neurons around the electrode tip, or instead a function of the structures connected to those neurons? Perhaps the evoked movement is actually caused by spinal circuits, basal ganglia circuits, cerebellar circuits, or other motor circuits.
The simplest answer is: the most basic truth of the brain is that no neuron has a function by itself. Its function is defined by its connections with, and therefore its influence on, other neurons.
In the use of electrical stimulation, it is necessary to distinguish between two kinds of signal spread. Direct spread, sometimes called passive spread, is the spread of the electrical field around the electrode tip that activates neighboring neurons. Indirect or active spread is the spread of neuronal signals across synapses and through networks. Ideally the passive spread is minimized, or at least restricted to the experimentally targeted neurons. The active spread, the percolation of signal through connected networks, is the goal of the technique, allowing function to be probed. It is not an error or artifact to be avoided. The evoked movement is a function of the directly stimulated neurons precisely because of their effect on connected structures.
The motor cortex tends to be conceptualized in a simplistic manner: each spot in cortex "controls" a movement because it sends a command signal, via a relay in the spinal cord, to the muscles. This view, though firmly in the intuitions of many neuroscientists, is at odds with the way the system is put together. Motor cortex plugs into a complex network. The spinal cord itself is a sophisticated processor that is a part of that larger network. A spot in cortex initiates and influences movement presumably because of its influence on the motor network.
Are we proposing that the motor cortex can be divided into isolated areas, each of which controls a separate type of action, providing the animal with a limited number of action primitives?
No. The fundamental code in motor cortex is probably a population code. In our hypothesis, a hand-to-mouth movement for example involves neuronal activity across a wide extent of motor cortex with a hill or peak near the hand-to-mouth area. Activity in the hand-to-mouth area, such as through electrical stimulation, evokes a specific movement partly because of the manner in which it recruits the rest of motor cortex. Loss of the hand-to-mouth area should remove the most efficient, centralized representation of that movement and therefore reduce performance, but surrounding cortical regions can presumably combine their outputs to produce a less skilled hand-to-mouth movement. The view suggested here is therefore a combination of localized function and distributed population, an approach that was successful in explaining the maps and population codes present in eye movement areas such as the superior colliculus.
In the 1990s, I studied a category of neurons that encoded information about the space immediately surrounding the body. These studies focused on neurons in specific regions of the macaque monkey brain, including the basal ganglia, the posterior parietal lobe, and the ventral premotor cortex (Graziano and Gross, 1993; Graziano et al., 1994, 1997a). A subset of neurons in these brain areas respond to events in the space around the body. For example, a neuron might respond to a touch within a receptive field on the right cheek. The same neuron might respond to the sight of objects near or approaching the right cheek. The neuron might respond to the onset of sounds, with a greater response for sounds near the right cheek (Graziano et al., 1999). The same neuron might even respond in the dark, with no tactile stimulation of the cheek and no sound near the cheek, if the head is moved so that the cheek approaches the remembered location of an object (Graziano et al., 1997b). These multi-modal sensory cells therefore appear to encode the space immediately around the body.
The exact function of these sensory-like neurons in motor areas of the brain was unknown. Not until the more recent work in my lab did we develop a good sense of their main behavioral role. Electrical stimulation of pockets of these neurons in the cortex, in both the parietal lobe and the motor cortex, evoked a specific, complex, and highly characteristic action (Cooke et al., 2003; Cooke and Graziano 2004; Graziano et al., 2002). The animal would make the motions of defending the body surface from an impending impact. Our best interpretation is that these multisensory neurons are part of a specialized system that encodes the space near the body, computes margin of safety, and helps to coordinate movements in relation to nearby objects especially withdrawal or blocking movements (Graziano et al., 2006).
Aflalo TN, Graziano MSA (2006) Possible origins of the complex topographic organization of motor cortex: reduction of a multidimensional space onto a two-dimensional array. J Neurosci 26: 6288-6297.
Aflalo TN, Graziano MSA (2007) Relationship between unconstrained arm movements and single-neuron firing in the macaque motor cortex. J Neurosci 27: 2760-2780.
Astafiev SV, Shulman GL, Corbetta M (2006) Visuospatial reorienting signals in the human temporo-parietal junction are independent of response selection. Eur J Neurosci 23: 591-596.
Blanke O, Ortigue S, Landis T, Seeck M (2002) Stimulating illusory own-body perceptions. Nature 419: 269-270.
Brunet E, Sarfati Y, Hardy-Baylé MC, Decety J (2000) A PET investigation of the attribution of intentions with a nonverbal task. Neuroimage 11: 157-166.
Ciaramidaro A, Adenzato M, Enrici I, Erk S, Pia L, Bara BG, Walter H (2007) The intentional network: how the brain reads varieties of intentions. Neuropsychologia 45: 3105-3113.
Cooke DF, Graziano MSA (2004) Super-flinchers and nerves of steel: Defensive movements altered by chemical manipulation of a cortical motor area. Neuron 43: 585-593.
Cooke DF, Taylor CSR, Moore T, Graziano MSA (2003) Complex movements evoked by microstimulation of Area VIP. PNAS 100: 6163-6168.
Corbetta M, Kincade JM, Ollinger JM, McAvoy MP, Shulman GL (2000) Voluntary orienting is dissociated from target detection in human posterior parietal cortex. Nat Neurosci 3: 292-297.
Fletcher PC, Happé F, Frith U, Baker SC, Dolan RJ, Frackowiak RS, Frith CD (1995) Other minds in the brain: a functional imaging study of "theory of mind" in story comprehension. Cognition 57: 109-128.
Frith C (2002) Attention to action and awareness of other minds. Conscious Cogn 11: 481-487.
Fritsch G, Hitzig E (1870) Über die elektrishe Erregbarkeit des Grosshirns. Arch. f. Anat., Physiol und wissenchaftl. Mediz., Leipzig, 300-332. (On the electrical excitability of the cerebrum.) Translated by G. von Bonin. In: Some Papers on The Cerebral Cortex. WW Nowinski, Ed. 73-96. Springfield IL: Thomas.
Gallagher HL, Happé F, Brunswick N, Fletcher PC, Frith U, Frith CD (2000) Reading the mind in cartoons and stories: an fMRI study of 'theory of mind' in verbal and nonverbal tasks. Neuropsychologia 38: 11-21.
Georgopoulos AP, Kettner RE, Schwartz AB (1988) Primate motor cortex and free arm movements to visual targets in three-dimensional space. II. Coding of the direction of movement by a neuronal population. J Neurosci 8: 2928-2937.
Georgopoulos AP, Schwartz AB, Kettner RE (1986) Neuronal population coding of movement direction. Science 233: 1416-1419.
Goel V, Grafman J, Sadato N, Hallett M (1995) Modeling other minds. Neuroreport 6: 1741-1746.
Graziano MSA (2008) The Intelligent Movement Machine: An Ethological Perspective on the Primate Motor System. Oxford University Press, Oxford, UK.
Graziano, MSA (in press) Consciousness and the Social Brain. Oxford University Press, Oxford UK.
Graziano MSA, Aflalo TN (2007) Mapping behavioral repertoire onto the cortex. Neuron 56: 239-251
Graziano MSA, Cooke DF (2006) Parieto-frontal interactions, personal space, and defensive behavior. Neuropsychologia 44: 845-859.
Graziano MSA, Gross CG (1993) A bimodal map of space: somatosensory receptive fields in the macaque putamen with corresponding visual receptive fields. Exp Brain Res 97: 96-109.
Graziano MSA, Hu XT, Gross CG (1997a) Visuo-spatial properties of ventral premotor cortex. J Neurophys 77: 2268-2292.
Graziano MSA, Hu XT, Gross CG (1997b) Coding the locations of objects in the dark. Science 277: 239-241.
Graziano MSA, Kastner S (2011) Human consciousness and its relationship to social neuroscience: A novel hypothesis. Cog Neurosci 2: 98-113.
Graziano MSA, Reiss LAJ, Gross CG (1999) A neuronal representation of the location of nearby sounds. Nature 397: 428-430.
Graziano MSA, Taylor CSR, Moore T (2002) Complex movements evoked by microstimulation of precentral cortex. Neuron 34: 841-851.
Graziano MSA, Yap GS, Gross CG (1994) Coding of visual space by pre-motor neurons. Science 266: 1054 -1057.
Karnath HO, Ferber S, Himmelbach M (2001) Spatial awareness is a function of the temporal not the posterior parietal lobe. Nature 411: 950-953.
Kentridge, R.W., Heywood, C.A., and Weiskrantz, L. (2004). Spatial attention speeds discrimination without awareness in blindsight. Neuropsychologia 42: 831-835.
Koch, C. and Tsuchiya, N. (2007). Attention and consciousness: two distinct brain processes. Trends Cogn. Sci. 11: 16-22.
Lamme, V.A. (2004). Separate neural definitions of visual consciousness and visual attention: a case for phenomenal awareness. Neural Networks 17: 861-872.
Ochsner KN, Knierim K, Ludlow DH, Hanelin J, Ramachandran T, Glover G, Mackey SC (2004) Reflecting upon feelings: an fMRI study of neural systems supporting the attribution of emotion to self and other. J Cogn Neurosci 16: 1746-1772.
Ramanathan D, Conner JM, Tuszynski MH (2006) A form of motor cortical plasticity that correlates with recovery of function after brain injury. PNAS 103: 11370-11375.
Saxe R, Kanwisher N (2003) People thinking about thinking people: fMRI investigations of theory of mind. NeuroImage 19: 1835-1842.
Saxe R, Moran JM, Scholz J, Gabrieli J (2006) Overlapping and non-overlapping brain regions for theory of mind and self reflection in individual subjects Soc Cogn Affect Neurosci 1: 229-234.
Saxe R, Wexler A (2005) Making sense of another mind: the role of the right temporo-parietal junction. Neuropsychologia 43: 1391-1399.
Shulman GL, Pope DL, Astafiev SV, McAvoy MP, Snyder AZ, Corbetta M (2010) Right hemisphere dominance during spatial selective attention and target detection occurs outside the dorsal frontoparietal network. J Neurosci 30: 3640-3651.
Strick PL (2002) Stimulating research on motor cortex. Nat Neurosci 5: 714-715.
Vallar G, Perani D (1986) The anatomy of unilateral neglect after right-hemisphere stroke lesions. A clinical/CT-scan correlation study in man. Neuropsychologia 24: 609-622.
Vogeley K, Bussfeld P, Newen A, Herrmann S, Happé F, Falkai P, Maier W, Shah NJ, Fink GR, Zilles K (2001) Mind reading: neural mechanisms of theory of mind and self-perspective. Neuroimage 14: 170-181.
Vogeley K, May M, Ritzl A, Falkai P, Zilles K, Fink GR (2004) Neural correlates of first-person perspective as one constituent of human self-consciousness J Cogn Neurosci 16: 817-827.