brain imaging

brain imaging: altered states of consciousness.
brain imaging: the methods.
brain imaging: the neural correlates of consciousness.
brain imaging: the neural correlates of will.

brain imaging: altered states of consciousness.

In trying to understand consciousness, it is useful to distinguish between the level and the content of consciousness. The level of consciousness is closely related to arousal and determines whether we are awake or asleep. Arousal is essentially maintained by activating structures in the brain stem, the diencephalon, and the basal forebrain. Arousal is a prerequisite for the emergence of any conscious experience. The content of consciousness (what we are conscious of) is thought to rely on the function of thalamocortical and corticocortical networks. Some cortical areas are more closely related to conscious experience than are others. In conscious, resting subjects, the brain is most active in several regions of polymodal association cortex: the prefrontal cortex, inferior parietal cortex, and precuneus/posterior cingulate cortex (Fig. 1).

A number of altered states of consciousness (slow wave sleep, anaesthesia, coma) are associated with low levels of arousal, implying modification in the activity of the subcortical structures. Slow wave sleep is an example of decreased level and contents of consciousness. This is characterized by a global decrease in blood flow in the brain, which is particularly prominent in the subcortical activating areas as well as in those cortical areas that are the most active during wakefulness.

More rarely, the level of consciousness is preserved but the content of consciousness is abolished. For instance, patients in vegetative state look wide awake — they may move, grimace, cry, or laugh — but they are unconscious: as far as we can tell, they have no content of consciousness. In this case, the brain stem is still very active but the cortex is not, especially in the association areas mentioned above.

Of all the altered states of consciousness, rapid eye movement sleep (REM sleep) is probably the most intriguing. During REM sleep we are not fully aware of the environment. On the other hand, our content of consciousness is far from empty. If woken up from REM sleep, we most often report a dream. The distribution of the brain activity in REM sleep is very special. While polymodal association cortex is relatively quiescent (as in the other state of altered consciousness), there is a marked activation of two types of cortical regions:

(i) the limbic areas, which are involved in emotion and memory (amygdala, hippocampal formation, anterior cingulate cortex), and
(ii) the cortices of the occipital and temporal lobes, concerned with vision and hearing and language.

It is tempting to relate this activation to the dream experience. During REM sleep, it is hard to tell whether the level of consciousness is decreased: the brain is as active as during wakefulness. However, the structures that maintain this activation are limited to a small portion of the brain stem, the cholinergic pontine nuclei.

Altered states of consciousness are not only characterized by a change in the distribution of cerebral activity. Interactions between brain areas (functional and effective connectivity) are also relevant. The available data stresses the importance of corticocortical and thalamocortical networks involving parietal and prefrontal polymodal association cortex. In the vegetative state, interactions between different regions of polymodal association cortex are reduced. The brain's responses to sound and touch do not go beyond primary sensory cortex. It seems that the primary cortex is disconnected from downstream uni-and polymodal association cortex. It could be argued these alterations in connectivity are the direct result of extensive brain damage. However, this would not explain the resumption of normal interactions between brain regions that has been observed in patients who recover consciousness after a vegetative state.

Finally, although more difficult to assess by functional brain imaging, the neurochemical modulation of brain activity by subcortical structures profoundly influences the state of consciousness. Whereas many neurochemical systems are active during wakefulness (histamine, serotonine, noradrenaline (norepinephrine), acetylcholine, orexin, etc.), slow wave sleep is characterized by a decreased firing in most if not all of these systems and the hallmark of REM sleep is a prominent cholinergic drive in the absence of serotonergic and noradrenergic modulation.


Fig. 1. Upper panel, lateral view of the brain. Lower panel, medial view of the brain. In black, the regions critical for maintaining arousal (A: brain stem, B: basal forebrain). In grey, the cortical areas most active in the resting conscious state (C: lateral and medial prefrontal cortex, D: inferior parietal cortex, E: precuneus/posterior cingulate). In conditions of decreased contents of consciousness (sleep, coma, vegetative state), these areas are relatively quiescent. Note that the thalamus (F) is a crucial interface between structures maintaining arousal and networks supporting the contents of consciousness.


(Published 2004)

— Chris Frith/Pierre Macquet

Bibliography
• Frith, C., Perry, R., and Lumer, E. (1999). 'The neural correlates of conscious experience: an experimental framework'. Trends in Cognitive Sciences, 3.
• Laureys, S., et al. (2000a). 'Auditory processing in vegetative state'. Brain, 123.
• — —  et al. (2000b). 'Restoration of thalamo-cortical connectivity after recovery from persistent vegetative state'. Lancet, 355.
• Maquet, P. (2000). 'Functional neuroimaging of normal human sleep by positron emission tomography'. Journal of Sleep Research, 9.
• Plum, F., and Posner, M., (1972). The Diagnosis of Stupor and Coma.
• Steriade, M., and McCarley, R. W. (1990). Brainstem Control of Wakefulness and Sleep.

brain imaging: the methods.

It has long been recognized that the brain is the origin of our mental life, but only recently has it become possible to study the relationship between brain and mind in humans in any detail. The change has come about largely through the development of non-invasive brain-imaging techniques in the last quarter of the 20th century. Before these developments our knowledge was either indirect, from experiments in animals, or came from the study of patients who had sustained damage to circumscribed brain regions as result of tumours, strokes, or accidents. Even in these cases it was not easy to know the precise location of this damage prior to examination of the brain after death. The development of computed tomography (CT) scanners in 1972 and magnetic resonance (MR) scanners in 1977 made it possible to locate the damage in exquisite detail (see Fig. 1) (Mansfield 1977). CT scanners use a series of X-rays projected through the skull to reconstruct an image of the brain. MR scanners use a powerful magnet that causes billions of hydrogen atoms (part of molecules throughout the brain) to spin in the same direction. A radio wave is passed through the brain, perturbing these spins, and causing them to emit signals that can be decoded by a computer to produce detailed structural images of the brain.

The brain contains 10 billion nerve cells, which are constantly electrically active. The electrical activity of a single cell, or a small group of cells, can be measured directly by inserting microelectrodes into the brain. However, this technique can only be used in humans if there is a medical reason for the procedure. For about 100 years we have known that electrical activity generated by large populations of nerve cells can also be detected at the surface of the scalp by using electrodes attached to very sensitive amplifiers (see electroencephalography). Such measurements reveal rapid fluctuations in activity known as the electroencephalogram (EEG). Recently, devices have been developed that can measure magnetic activity at the surface of the scalp induced by these changes in electrical activity (MEG). Both techniques give detailed information about the timing of events happening in the brain, but are poor at locating the source of this activity unless it is close to the surface of the brain.

Although the brain accounts for less than 2 per cent of a person's weight, the continuous electrical activity of nerve cells consumes 20 per cent of the body's energy. Energy is supplied, in the form of oxygen and glucose, by a network of blood vessels throughout the brain. It has been known for over 100 years that a local increase in neural activity causes a local increase in blood flow (Roy and Sherrington 1890). It is therefore possible indirectly to measure neural activity by measuring local changes in blood flow (Fulton 1928). Positron emission tomography (PET) was the first technique used to measure blood flow changes by injecting trace amounts of radioactivity into the bloodstream (see Figs. 2 and 3) (Ingvar 1975). More recently it has proved possible to use MRI (magnetic resonance imaging) to detect changes in blood flow non-invasively, without the need for radioactive injections (Kwong et al. 1992). Oxygen is transported to nerve cells in the bloodstream bound to haemoglobin as oxyhaemoglobin. When the oxygen is released, oxyhaemoglobin becomes deoxyhaemoglobin. Fortuitously, these two molecules differ slightly in their magnetic properties, and so any change in the proportion of oxyhaemoglobin to deoxyhaemoglobin can be detected by MRI. For reasons that are still not clear the local increase in blood flow that is elicited by an increase in neural activity is greater than is needed to replace the oxygen that has been used. As a result there is an increase in the ratio of oxyhaemoglobin and deoxyhaemoglobin, which is detected by MRI as blood oxygenation level dependent (BOLD) contrast (Ogawa et al. 1990). This technique is known as functional MRI (fMRI) and has slightly higher spatial and temporal resolution than PET.

The changes measured by PET and fMRI can be detected throughout the brain, and just as accurately in deep structures as at the surface with a spatial resolution of a few millimetres. However, blood flow measures are far less sensitive to the timing of the neural activity than EEG. Changes in blood flow associated with a brief increase in neural activity do not occur for about five seconds after the onset of the neural activity and carry on for several seconds after the neural activity has ceased. Furthermore the changes in blood flow do not indicate whether the neural activity is excitatory or inhibitory. Nevertheless, blood flow measures can provide important information about neural activity which is entirely consistent with previous work using recordings from single cells in the brains of animals (Posner and Raichle 1994; Frackowiak et al. 1997).


Fig. 1. Structural MRI showing a discrete lesion.



Fig. 2. A PET scanner in use.



Fig. 3. PET image of auditory stimulation.


(Published 2004)

— Chris Frith/Geraint Rees

Bibliography
• Frackowiak, R. S. J., Friston, K. J., Frith, C. D., Dolan, R. J., and Mazziotta, J. C. (1997). Human Brain Function.
• Fulton, J. F. (1928). 'Observations on the vascularity of the human occipital lobe during visual activity'. Brain, 51.
• Ingvar, D. H. (1975). 'Patterns of brain activity revealed by measurements of regional cerebral blood flow'. In Ingvar, D. H., and Lassen, N. A. (eds.), Brain Work.
• Kwong, K. K., Belliveau, J. W., Chesler, D. A., et al. (1992). 'Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation'. Proceedings of the National Academy of Sciences of the USA, 89/12.
• Mansfield, P. (1977). 'Multiplanar image formation using NMR spin echoes'. Journal of Physiology, 10.
• Ogawa, S., Lee, T. M., Kay, A. R., and Tank, D. W. (1990). 'Brain magnetic resonance imaging with contrast dependent on blood oxygenation'. Proceedings of the National Academy of Sciences of the USA, 87/24.
• Posner, M. I., and Raichle, M. E. (1994). Images of Mind.
• Roy, C. S., and Sherrington, C. S. (1890). 'On the regulation of the blood supply to the brain'. Journal of Physiology, 11.

brain imaging: the neural correlates of consciousness.

The environment of a brain scanner is extremely restricted. The volunteers must lie flat on their back and move as little as possible. As a result in many early studies volunteers were required to do nothing more than think (e.g. 'Imagine walking through your front door and then turning left at every intersection'). The brain activity associated with such thoughts can easily be detected. More recent studies have shown that the location of the activity relates directly to the content of the thought. If you imagine moving your finger then activity is observed in the brain's motor system. If you imagine seeing a face or a house then activity is seen in the appropriate part of the inferior temporal cortex where objects are represented (see localization of brain function). Patients with visual disorders who suffer from visual hallucinations of particular objects show similar patterns of activity. These studies show that it is possible to find a localized physiological marker for purely mental activity, that does not result in any observable behaviour. Before the advent of brain imaging we could only know about this mental activity from the subjective report of our volunteer.

The ability to study the relation between mental activity and brain activity has led to a major scientific enterprise of identifying the neural correlates of consciousness (NCC). The ultimate, but far-off aim of this enterprise is to understand how mental activity can emerge from a physical entity such as a brain. So far much of this endeavour has been concerned with identifying neural activity associated with the contents of visual consciousness. In order to identify activity specific to consciousness, situations have been contrasted in which there is a difference in the contents of consciousness, but no difference in the visual signals striking the eye. Such a situation occurs when we view an ambiguous figure such as Rubin's face–vase illusion. Although the visual input does not change the contents of our consciousness spontaneously switches from a face to a vase and back again. Brain-imaging studies have shown that activity in visual areas of the temporal cortex closely follows what we consciously perceive, while activity at earlier stages of the visual system is less closely correlated. This is consistent with recordings from single neurons in monkeys exposed to similar situations. In addition, brain-imaging studies in humans have shown that there is also activity in frontal and parietal cortex that is time-locked to the occurrence of switches. This activity may represent a mechanism that actively causes the switches between alternative interpretations of the outside world.

A similar strategy can be used to study brain activity associated with unconscious processes. In some situations there can be a change of visual input of which we are not aware, but which still influences our behaviour and therefore must have caused a change in brain activity (see blindsight). We can identify this activity by studying situations in which there is a change of visual input, but no change in the contents of consciousness. For example, patients with visual extinction following parietal damage will sometimes fail to report a visual stimulus in the left visual field when it is paired with a second stimulus on the right. Brain imaging has shown that this unseen stimulus in the left visual field nevertheless activates areas of visual cortex in the right hemisphere. Even in normal observers without brain damage, changes in the visual environment that they do not notice can nevertheless produce activity in visual cortex. Similarly, when a word is flashed briefly and immediately followed by a mask it is not seen, but can produce brain activity related to its identity (see subliminal perception). These studies show that the mere presence of activity in the visual cortex is not sufficient to evoke awareness. Instead, it may be that the level or type of activity is important, or its association with activity in other areas of the brain.

Visual stimuli of which we are not aware can induce emotional responses. Aversive conditioning can be used to elicit an emotional response from a subject when an angry face is presented. If awareness of the angry face is prevented by masking with a neutral, expressionless face, then brain imaging can still detect activity in the amygdala that reflects the learned emotional response to that face. This shows that emotional associations of visual stimuli can be extracted unconsciously by the brain and influence behaviour.

The use of brain imaging to study the NCC is still in its infancy, but has already produced interesting findings. Activity measured with brain imaging in areas of visual cortex may be necessary but not sufficient to result in awareness. In addition, areas outside striate and extrastriate visual cortex, in the parietal and frontal lobes, are consistently associated with visual awareness.

(Published 2004)

— Chris Frith/Geraint Rees

Bibliography
• Dehaene, S., Naccache, L., Cohen, L., (2001). 'Cerebral mechanisms of word masking and unconscious repetition priming'. Nature and Neuroscience, 4.
• Ffytche, D. H., Howard, R. J., Brammer, M. J., David, A., Woodruff, P., and Williams, S. (1998). 'The anatomy of conscious vision: an fMRI study of visual hallucinations'. Nature and Neuroscience, 1.
• Frith, C., Perry, R., Lumer, E. (1999), 'The neural correlates of conscious experience: an experimental framework'. Trends in Cognitive Science, 3.
• Metzinger, T. (ed.) (1999). Neural Correlates of Consciousness: Empirical and Conceptual Questions.
• Morris, J. S., Ohman, A., and Dolan, R. J. (1998). 'Conscious and unconscious emotional learning in the human amygdala'. Nature, 393.
• Rees, G. (2001). 'Neuroimaging of visual awareness in patients and normal observers'. Current Opinion in Neurobiology, 11.

brain imaging: the neural correlates of will.

My voluntary actions, as opposed to my reflexes, are free in the sense that I could have done otherwise. However, in most experimental studies of action the volunteer temporarily gives up this freedom and performs exactly as instructed by the experimenter (e.g. 'lift your finger whenever you hear the tone', 'press the left button when the red light goes on'). A degree of freedom can be introduced into such experiments by giving the volunteer a choice (e.g. 'lift your finger whenever you feel like it', 'whenever you hear the tone lift either your left or your right index finger'). Actions associated with free choices are sometimes referred to as willed actions. Brain regions where activation is associated with willed actions have been identified in a number of neuroimaging studies. These regions are activated whatever the modality of the action (finger movements, arm movements, speech) and whether the choice concerns the timing or the nature of the action, i.e. when to move or what to move.

The brain regions most commonly implicated are the dorsolateral prefrontal cortex (DLPFC) and anterior cingulate cortex (ACC), regions on the outer and inner surface of the frontal lobes, respectively. Though these regions appear necessary for willed action they are not (in themselves) sufficient; their 'output' must be relayed through 'lower' brain regions such as the basal ganglia, motor cortex, and thalamus. However, although these lower centres may be sufficient to support stereotypical or learned behaviours, they are supplemented by activity in DLPFC and ACC when choice occurs (see Spence and Frith 1999).

Diseases and disorders that impair the execution of wilful actions in humans have been well described: Parkinson's disease, schizophrenia, severe depression, each may be associated with disturbance of DLPFC and ACC function. While patients with these conditions may have difficulty initiating actions, there are also other conditions in which patients are unable to prevent their limbs from making superficially 'purposeful' movements. Such patients, exhibiting 'utilization behaviour' or 'alien limbs' often have lesions in orbitofrontal, medial premotor, or callosal regions, giving rise to a failure of response inhibition. People with utilization syndrome, consequent upon bilateral orbitofrontal lesions, seem to be 'driven' by their environments, responding even when they have no desire to do so. Their hands may pick up and grasp objects 'against their will' (Shallice et al. 1989). Hence, there appears to be a tension between the appropriate initiation of an action which the agent wishes to perform, and the inhibition of alternative performances which are inappropriate in the current context. Different neural circuits may be implicated in these complementary functions.

Using EEG it is possible also to look at the timing of the brain activity associated with willed actions. Prior to willed action, changes in brain activity (the readiness potential) can be detected up to a second before the action is initiated. These changes probably occur in ACC and in the premotor cortex. In contrast, when we respond as instructed to a stimulus, changes in brain activity are only observed a few hundred milliseconds prior to the action. In a controversial experiment Benjamin Libet asked volunteers to indicate at what time they had the urge to act. This time was typically about 300 milliseconds after the first detectable changes in brain activity associated with the ensuing act (Libet et al. 1983). Patrick Haggard has replicated and extended this observation, showing that the timing of the urge to act is correlated with the timing of the onset of the lateralized component of the readiness potential (Haggard and Eimer 1999). Some have argued that these results show that human choices are determined rather than free because changes in brain activity can be observed before the choice has been made. However, we only know about the time of the awareness of the choice. It is still possible that the choice is freely made before we are aware of it (Spence 1996).

Whether or not our actions are determined, we have a strong subjective experience of wilfully choosing courses of action and then initiating them (the experience of agency). In some psychiatric and neurological disorders this experience of agency is compromised. Schizophrenic patients with delusions of control experience their own actions as being made by some alien force. This experience resembles similar experiences reported by people with parietal lobe epilepsy or tumours and is associated with hyperactivity of the right inferior parietal lobe (Spence et al. 1997). Such hyperactivity has been shown to remit as patients' symptoms resolve.

(Published 2004)

— Chris Frith/Sean Spence

Bibliography
• Haggard, P., and Eimer, M. (1999). 'On the relation between brain potentials and awareness of voluntary movements'. Experimental Brain Research, 126.
• Libet, B., Gleason, C. A., Wright, E. W., et al. (1983). 'Time of conscious intention to act in relation to onset of cerebral activity'. Brain, 106.
• Shallice, T., Burgess, P. W., Schon, F., and Baxter, D. M. (1989). 'The origins of utilisation behaviour'. Brain, 112.
• Spence, S.A. (1996). 'Free will in the light of neuropsychiatry'. Philosophy, Psychiatry and Psychology, 3.
• — —  and Frith, C. D. (1999). 'Towards a functional anatomy of volition'. Journal of Consciousness Studies, 6.
• — —  Brooks, D. J., Hirsch, S. R., et al. (1997). 'A PET study of voluntary movement in schizophrenic patients experiencing passivity phenomena (delusions of alien control)'. Brain, 120.

---