How Technologies of Imaging are Shaping Clinical Research and Practice in Neurology
- First Online:
- Cite this article as:
- Kopp, N. Medicine Studies (2009) 1: 315. doi:10.1007/s12376-010-0037-1
- 57 Views
The brain is our most complex organ, in terms of gross and microscopic structure. It is heterogeneous, with many areas and networks differing from one another in function. And, what is more, the brain is a ‘hidden entity’, embedded in an envelope made of bones, the skull. Brain imaging really came to age in medicine 40 years ago, thanks to computers. The technologies of structural anatomy like computerized tomography and magnetic resonance imaging have brought about a revolution in neurology by showing the lesion and its topography. With most recent developments, neuro-imaging continues to shape practices in neurology and clinical research.
Important Steps in Development of Imaging Technologies in Neurology
Several brain imaging technologies are presently available in neurology. One is computerized tomography (CT) scan based on X-ray technology and computers. It gives discrimination between densities, e.g., CT scan can, in urgency, establish a diagnosis and topography of a life-threatening extra-dural hematoma, in a patient with cranial trauma.
Two other technologies should be emphasized and justify a short explanation of their physical principles: positron emission tomography (PET scan) and magnetic resonance imaging (MRI).
The patient submitted to PET scan is given a safe dose of a radioactive compound. In order to study and evaluate brain activity, fluorodeoxyglucose (FDG), a modified glucose molecule, is chosen. Glucose is the main source of energy for the brain. The injected or inhaled FDG will enter the patient’s bloodstream and will travel to the brain. If an area of the brain is more active, more glucose will be needed there. The more glucose is consumed, the more FDG is absorbed. To measure the amount of radioactive material absorbed by the brain, the patient lies on a movable bed that slides into the tunnel-like opening of the device.
The PET scan measures the energy emitted when positively charged particles (positrons) from the radioactive molecule, the FDG, collide with electrons (negatively charged particles) in the brain of a patient. Between 30 min and 2 h are needed for the scan to be completed. A computer turns the measurements into multicolored two- or three-dimensional images. Nowadays, the public is familiar with these images shown in the media. The intensity of the energy that is recorded (i.e., the activity of a group of nervous cells) is indicated by a series of tiny dots (“voxels”) with color ranging from deep blue to flashy red color. Commonly, the parts of brain that show high activity are colored in red, others that show low activity in deep blue. A person ignorant of the complexity of different technological steps feels he/she sees the brain in activity, and therefore getting direct access to the mental processes of a patient in its quality (i.e., content of “thinking”). Actually this is, in a way, an illusion. What one actually sees is the quantitative assessment of energy consumption through a construction performed by the computer. A few experimental PET scans, used mainly for research, have extremely powerful computers (e.g., computers of the CRAY type) that provide high resolution, that is to say give out better pictures.
The MRI is in fact, stricto sensu, a “nuclear magnetic resonance imaging”. But there is no radioactivity used or produced by the procedure. In order not to induce anxiety, “nuclear” is dropped from the name of the device. MRI is the most useful and widely used imaging machine in modern medical centers.
What is the physical principle of MRI? MRI machines are giant magnets. The machine looks very much like a PET scan, but it has the added feature of an invisible magnetic field. The nuclei of some atoms of matter (such as brain tissue) can be in “resonance” according to the type of magnetic field. In order to record this physical phenomenon, a short radio frequency wave is used. It transiently modifies the orientation of protons that turn around the activated nuclei. When protons go back to their initial state, they give back the recordable energy under the form of a signal. This signal is picked up by a receiver antenna and analyzed by powerful computers that construct a three-dimensional image, presented in several successive sections.
When the resonance of hydrogen nuclei, very abundant in water (approximately 80% of human body consists of water) and fat, is monitored under an intense magnetic field, one can visualize the anatomical structure. Since blood contains large amounts of water molecules and, thus, lots of hydrogen atoms, the hydrogen atoms will produce pulses of energy when a person is immersed in a magnetic field. The energy emitted reflects increases in blood flow and, therefore, is seen to correspond to brain activity.
Brain imaging was developed in the field of medicine during the 20th century, inducing several successive revolutions.
Roentgen’s discovery of X-ray in 1895 allowed the birth of radiology and set forth a revolution in traumatology and oncology. Radiography became progressively less dangerous for staff and patients, and more efficient with the development of CT.
Hans Berger, in 1929, obtained the first functional images of the brain, the electro-encephalogram (EEG). This rather old technique gives, if not a “real brain image”, a crude topographical information of the brain’s electric activity which can, presently, be increased by augmenting the number of electrodes on the scalp. It can allow ambulatory examination. In the sixties, the field of expertise of EEG was expanded with the stimulus-evoked EEG, the “scalp-recorded event- related potential”. This was the first technique to unveil fundamental knowledge about the working of the human brain in near real time.
In the 1930s, pneumo-encephalography, with spinal tap and air injection, gave images of cerebral ventricles and subarachnoid spaces. This fairly invasive technique was not without danger (especially in patients with elevation of intracranial pressure). It usually was the cause of violent secondary headaches. This technique disappeared in the 70s, at least in developed countries.
At the same period, in the 1930s, angiography was discovered. At present, it allows nonsurgical endo-vascular intervention on some intracranial vascular lesions, noticeably aneurysms.
The CT Scan Revolution
In the 1970s, computerized axial tomography (CAT Scan) produced the first in vivo images of the global content of the skull (subarachnoid spaces, ventricles, and brain). CT scan is limited to structural anatomy. Its use marked the beginning of the regression of the traditional “clinico-pathological” examinations with correlations of post mortem brain examinations and clinical data. It represented a considerable progress in diagnosis by showing brain lesions (of etiological value) and its precise topography. This technology had a side effect: the beginning of the regression of time consuming—but irreplaceable—neurological clinical and neuropathological examination. And shorter examination means less time spent in contact with the patient. But this extensive use of CT scan meant a hugely positive step. Its inventors were laureates of the Nobel Prize of Medicine.
Today, one may even report on further technological innovations with regard to CT scan consisting of either multiple contiguous scans increasing spatial resolution or contrast methods allowing detailed reconstructing of intracranial vessels.
Other New Technologies and the Revolution of Functional MRI
Within a few years, other brain imaging technologies appeared. These technologies provided information of the highest value, in terms of structural anatomy and functional activity. There is an exponential increase of methods and possibilities to examine, analyze, and represent brain activities related to psychological and cognitive dynamics involving multiple electrical and chemical phenomena: “At least for some neuroscientists (..), it is the declared goal to explore all classical philosophical and psychological questions related to cognitive functioning with the methods and concepts of cellular biology” (Kollek 2004).
Monophotonic emission tomography (MET) can give neurologists functional information on the metabolic profile.
Functional MRI (fMRI) appeared at the beginning of the 1990s, another noninvasive and not noxious technology of relatively low cost. This was another revolution. fMRI has been used in association with different imagery technologies. MRI gives more detailed structural images than CT scans. It can also provide information on the physico-chemical state of tissue, their vascularization and perfusion. fMRI is, presently, the most widely used technique in cognitive neuroscience research–raising a bundle of epistemological and ethical questions (i.e., technological boundaries and unexpected clinical anomalies in research). In picking up issues of localization, fMRI could be regarded as a “new” phrenology. It is indeed more grounded than the one of the phrenologists of the early 19th century. But it still, at best, provides only a description and not an explanation of mental processes (cognition, emotion etc.).
Magnetoencephalography (MEG), though expensive, is used to get information on the electrophysiological activity of the whole brain with very high temporal performance. MEG also gives spatial information, but mainly on the surface of the brain convolutions. It can be used in association with other techniques, especially fMRI.
Near infrared spectroscopy (NIRS) is an optical method, which provides a noninvasive measure of regional brain activity based on the absorption of different wavelengths of light as it passes through the head. It is interesting in the sense that it can be used in ambulatory natural conditions, and not only under the constraints of a research laboratory setting.
What Do Activation Maps Represent?
Does activation of a region mean that it is actually involved in the task being performed by the subject? This question implies that we have a decent understanding of how neural activity in a region unequivocally shows its participation in the studied mental state or behavior. But we do not. The traditional cortical input-elaboration-output scheme, commonly presented, as an instantiation of the perception-cognition-action model is probably an over-simplification (Logothetis 2008).
In association, genetic studies potentially can increase dramatically the power of fMRI as a scientific means. At another level, proteomics is likely to provide cartography entirely different from functional imaging. At present, we do not know, how both, proteomics and imaging relate to each other; nor, how they change over time, not to mention when associated with data on life experience and on personality traits. However, the combined use of imagery, biological data (such as hormone levels), clinical data, and genetics data may help in the understanding of some traits and acquired psychological findings.
Electrode Recording (On the Basis of Implants)
Electric recording is not imagery in itself but registers neuronal events in situ in topographically distributed electrodes. Besides it is performed in association with imagery such as fMRI, intracerebral electrodes’ use is another way to boost the power of fMRI. Given a craniotomy must be done, it is indeed very invasive. In fact once electrodes are implanted in the gray matter, for instance, they allow to validate epileptic foci in the brain of patients suffering from epilepsy. Another important neurosurgical treatment, deep brain stimulation, is a standard procedure in Parkinson’s disease.
In the literature, direct electric recordings (i.e. “single neurons recording”) are discussed as powerful “mind-reading”-techniques in the future (Levy 2007). Given their invasiveness and other methodological preconditions (e.g., requirement of training and cooperation of the patient), it is questionable if they will ever be applied to nonmedical settings. Even if single neuron recordings some day could be used as “mind-reading”-techniques, they will provide us with a very reduced understanding of an individual’s mind (Baertschi 2009).
Revolutions of Imaging in Neurology Induce a Revolution in Cognitive Neuroscience
In sum, these technological developments may bring clinicians basis for diagnosis of cognitive or emotive disorders. In the field of clinical neurology, functional imagery already has practical uses. At a preoperative step of a case of pharmaco-resistant temporal epilepsy, fMRI can show the precise location and the lateralization of speech areas that the operating surgeon will have to spare. Another example is the recent use of fMRI in describing emotional disorders in autism, schizophrenia, or fronto-temporal dementias.
The development of a general technology, able to read the “thoughts” of people, even in the absence of preliminary training and the establishment of a base line would be possible only if there is a great deal of commonality in the neural correlates of mental states among people. It would indeed be useful for neurologists and psychiatrists to build up a kind of “translation manual” that allows drawing correlations between particular brain states and cognitive or emotive components of mental activity in question. Given that my thought that “elephants are gray” has specific neural correlates, which are different from the correlates of your thought that “elephants are gray”, the generation of a manual will be very difficult or even impossible.
Methodological Issues of Imaging Technologies in Neurology
CT scan and MRI brain imaging technologies allow us to see the structural anatomy of the brain, whereas functional MRI visualizes specific aspects of the brain’s metabolism correlated, supposedly, to the activity of neurons. The following section aims at showing the necessity to acquire more knowledge on what is a “normal” brain before tackling the problems of identifying abnormalities/pathologies of the brain.
Normal Structural Brain Anatomy
Imaging technologies progressively gain a higher and higher resolution, therefore allowing more and more precise and detailed analysis of different brain areas. This immediately has impact on medical diagnosis and treatment, for instance, in showing very small structures that were not visible before, such as the entorhinal cortex, which is a discrete temporal area where atrophy begins in patients with Alzheimer’s disease. The possibility to have a highly performing method of structural anatomy at disposal is a prerequisite to good fMRI studies of the brain aimed at diagnosis.
The basic question is the vague limits separating the “normal” and the “pathological”, “normality” and “abnormality”. For example, neuropathological examinations of unexpected and unexplained infant death syndromes (UUIDS) frequently disclose variations of brain structures (i.e., the cerebellar olives). But these reports are based on the results of comparison of cases of UUIDS with “controls” dead because of a brutal and massive obstruction of breath airways or fatal trauma. In short, theses anatomical studies are based on a weak definition of the “pathology”. As a matter of fact, here, death is implicitly considered as a syndrome whereas it is a process, a very complex physiopathological process. The common name of the disorder is SIDS, which stands for sudden infant death syndrome. The so-called syndrome is still considered as a cause, an explanation for death. However, reliable data of what are “normal” brain structures and anatomy during development are lacking. In vivo studies of the anatomy of brain development and plasticity are possible fields of research, also in patients at risk of UUIDS, such as apneic infants.
Information on functional brain anatomy can be obtained with new technologies of imaging. For example, one can mention the significant correlations between mnesic performances and metabolic values in groups of patients with Alzheimer’s disease: Brain areas that have to be scrutinized vary with the variety of memory considered, for example, working memory (bilateral posterior associative cortex), semantic memory (left posterior and anterior associative cortex), or verbal episodic memory (hippocampal region and posterior cingulate gyrus, predominantly on the left hemisphere).
New Brain Atlases
Imaging studies of many subjects in the same program allow the realization of population-based brain atlases. They offer a powerful framework to synthesize results from disparate imaging studies. These atlases use new analytical tools to fuse data across subjects, modalities, and time. They detect group-specific features, which are not visible on individual patients’ scans. These atlases, when built, can be stratified in subpopulations to reflect a particular clinical group, such as persons at genetic risk for Alzheimer’s disease or patients with mild cognitive impairment or different dementia types (fronto-temporal dementia, Lewy body disease, etc.). Some brain atlases incorporate developmental dynamic data that may in the near future assist research in paediatric disorders, the above-mentioned UUIDS syndrome is one of them.
But a pharaonic task in normative research is needed here. One can only mention disease-specific atlases, statistical brain templates, and cortical patterns (Toga et al. 1997). This type of work not only implies the need of large series; furthermore, it should be coupled with life narration, psychological tests (i.e., IQ), or other assessment of personality traits, clinical, and biological data, (i.e., genetics). Who will pay? Which journal will accept to publish such long lasting longitudinal studies? The scientific authorities may favor innovation, experimental strategies of research, more than pharaonic descriptive studies.
Normal Functional Brain Anatomy
Not only do interindividual variations exist in MRI “normal brain” structure but variations also exist in normal brains analyzed with fMRI. Again we emphasize that enormous works have to be performed on brain structural normal anatomy, prior to possibilities of diagnosis of some neurological disorders. In the future, studies designed in terms of functional normal anatomy–prior to application to functional imagery diagnosis in patients with cognitive deficits–will be needed. Indeed, the validation of activation in areas associated with language processing (e.g., Wernicke’s and Broca’s areas) really makes sense to the neurosurgeon: Prior to surgical intervention, this means to ascertain if the brain is left or right lateralized, respectively, if tumor growth is affecting brain areas associated with cognitive functions that have to be spared during operation: However, as Logothetis recently has pointed out, “the limitations of fMRI are not related to physics or poor engineering (..) they are instead due to the circuitry and functional organization of the brain, as well as to inappropriate experimental protocols that ignore this organization” (Logothetis 2008).
New Brain Imagery Combined with Stereotaxy
Against this background, the combination of different imaging techniques is very important for stereotactic neurosurgery, hence, the localization and protection of healthy brain tissue surrounding the lesion (see earlier). The same is true for the implantation of intracerebral electrodes for the treatment of epilepsy, in deep brain stimulation (Parkinson’s disease, obsessive-compulsive disorders etc.).
Diagnostic and Therapeutic Benefits Versus Challenges in Neurology
Interview of the patient: Identification of symptoms and complaints, past medical history;
Identification of objective signs at detailed clinical examinations;
Grouping of symptoms and signs in syndrome(s): Clinical examination of a patient may disclose a loss of muscle tone (hypotonia), an incoordination (ataxia) of volitional (voluntary) movements, and disorders of equilibrium and gait. Taken together, these signs constitute a cerebellar syndrome. Clinical examination may show a corticospinal (also called pyramidal) syndrome with progressive paralysis, muscle hypertonia, and jerky deep tendon reflexes;
Topographical diagnosis: A purely cerebellar syndrome points to the cerebellum. In an infant, a fairly rapid onset may indicate a malignant tumor and the most frequently—in an infant—it is a medulloblastoma. In case of a slow onset cerebellar syndrome associated with a pyramidal syndrome, neurodegenerative disorder is a possibility;
Synthetic diagnosis: At our brain imaging era, the topographical diagnosis is ascertained and associated with structural (lesion) information. This often leads to a probable diagnosis such as neurodegeneration or medulloblastoma. In the latter, only biopsy will allow certainty.
What the pathologic process is: inflammatory? immunological? vascular? proliferative? degenerative? due to malformation?
Where the process is located in the brain?
When the neurologist has clear imaging responses, retrospectively, some refinements of thorough neurological examination may-unduly-seem to appear obsolete. Vice versa, hitherto undescribed morphologic or functional details disclosed in imaging may induce the discovery of undescribed clinical signs.
Neuroimaging has also contributed, together with biology, to blur the frontiers between neurology and psychiatry. This is testified by the fact that in patients with motor conversion symptoms (hysteria), where traditionally no brain lesion is described, at fMRI examination, an abnormal functioning has been found. The overall pattern suggests more complex mental activity in patients with conversion disorder than in normal controls (Stone et al. 2007). Another good illustration of this blurring is obsessive compulsive disorders. Forty years ago, this nosological entity was called “obsessive neurosis”. It was considered as having a psychogenic origin. Since then anatomical anomalies have been described and some cases are successfully treated by electrical stimulation through deep brain electrodes, which does not mean that psychological (behavioral) therapies are abandoned. The same disease (the same patient) has a “neurological” side and a “psychiatric” side… In the following section, we will discuss some disorders that are regarded to be neurological in their nature. In recent years, more and more disorders are conceptualized as “neuropsychiatric” ones.
In Sketching a Few Relevant Diseases....
we will see that neuroimaging is a powerful tool for diagnosis and physiopathological assessment of a growing number of diseases, whereas definitive cure, in many of them, is not possible at present. A noticeable exception is the endovascular cure of intracranial aneurysms (see below). The presentation of these issues is very far from being exhaustive and limited to some of the most frequent and/or demanding situations:
Repeated MRI neuroimagery is also able to monitor, in patients with Alzheimer’s disease, the extension of atrophy from hippocampus and the inferior aspect of temporal lobe, to temporal pole, the external face of temporal lobe, to parietal lobe, the associative parieto-temporo-occipital crossroad (frontal one). But this would in most situations be unnecessary and extravagant. Conversely, PET scan can disclose the abnormal amyloid deposits [see supplementary Figure] pulling diagnosis toward certitude. PET scan can also disclose and precise an involvement of one or more dysfunctions of neurotransmitters. Ultimately, according to Klitzman (2006) Alzheimer’s disease might eventually be seen as composed of several physiopathological disorders, with the help of new technologies, particularly brain imaging. But these disorders, presently, cannot be cured; at best, their worsening can be slowed down.
A major ethical issue has cropped up and is rapidly growing. It is the one of early diagnosis or diagnosing of preclinical Alzheimer in subjects known to be at genetic risk (Basset et al. 2006). A study has shown a possibility to differentiate mild cognitive impairment from mild Alzheimer’s disease by the in vivo study of serotonin receptors 5-HT1A (Truchot et al. 2007). Incidental findings should be fairly frequently encountered, given the high incidence of the disease. Imagery is liable to detect a vascular alteration which is a major causal factor enhancing the degenerative process. The highly valuable efficiency of diagnostic imaging is used to identify other cortical degenerative disorders such as frontal dementia or Lewy body disease.
Other Cortical and Sub-Cortical Neurodegenerative Disorders
Secondly, we can consider degenerative motor disorders. In Parkinson’s disease, for instance, detailed structural imaging is necessary before performing implantation of deep cerebral electrodes. In addition, functional analysis of dopaminergic neuronal systems has largely confirmed the classical neuropathological and post mortem chemical studies of different neurotransmitters (Javoy-Agid and Agid 1990). Both types of data-imaging and post mortem chemical/histological study complete each other. These confirmed sets of data will be helpful, for instance in therapeutic strategies based on dopaminergic stem cells grafts. Functional imagery will permit a monitoring of the graft.
Attention Deficit and Hyperkinetic Disorder children (ADHD)
According to Schneiweiss (2006), the cause of this disorder is not precisely known yet. Conventional structural imaging studies (CT, MRI) disclose no obvious cerebral lesion in ADHD. But they point to probable developmental anomalies. Some symptoms, especially the tendency to be extremely inattentive, suggest the implication of a frontal dysfunction. Imaging studies have shown a decrease in cerebral blood flow in frontal lobes of ADHD patients and structures connecting them with limbic structures. But these studies could not, to our knowledge, be reproduced. Quantitative structural MRI techniques have shown the absence of the normal hemispheric asymmetry (frontal lobe, caudate nucleus, and globus pallidus). Anomalies have been described in inferior aspect of prefrontal region, bilaterally, an area controlling attention and impulse control. Paloyelis et al. (2007), in a systematic review of literature, discuss the role that fMRI could eventually play as a diagnostic tool or in treatment outcome predictions. Overmeyer et al. (2001) and Amat et al. (2006) describe lesional images that seem to involve in and possibly disrupt at large scale the neurocognitive network for attention.
Coma, Qualities of Consciousness
The brain imaging of severely damaged patients has the potential to tell researchers and clinicians a great deal about the patient’s state of consciousness. But there is a catch: Before they can make a good use of this brain imaging information, researchers need to know more precisely the relation between brain activity and consciousness. Brain activity that indicates information processing does not necessarily indicate consciousness.
According to Ackerman (2006), neuroimaging, at present, cannot predict reliably, which patients with severe disorders of consciousness are going to improve, become more conscious, and recover a bearable life. An estimation in the United States gives 15.000 patients in persistent vegetative state (PVS) and 100,000 more in a new category the minimally conscious state (MCS). In fact, these figures cannot be confirmed in the current incomplete state of medical knowledge of PVS and MCS. Both are usually diagnosed simply with the observation of behavior. Fortunately, these disorders are no more orphan topics (there was, until recently, apparently no good reasons to study these patients because they were simply too few), and they are now emerging as a legitimate and productive area of research, especially with fMRI. A recent study used fMRI in a unique way to assess the neurocircuitry that underlies receptive language in MCS. The researchers made a personalized, narrative tape recording of siblings or other relatives for each patient research subject. They recorded stories, sometimes very moving narratives (e.g., “remember when we went riding our bicycle together on main street…”) about situations or circumstances they had shared with the patient. The narratives evoked activity in areas of the brain (Broca’s area and Wernicke’s area) that are thought to house networks or language function in healthy people. A patient had, in addition a considerable amount of visual activity, suggesting that the patient was also visualizing the narrative he was hearing.
Up till recently, MCS and related states had been tentatively explained as the product of a cortical network that disease or injury had made incomplete. With these new data, one may speculate, instead, that a very high threshold of activation might characterize some disorders of consciousness. Compelling ethical questions is raised about the quality of cognition and perception by patients with MCS at different stages of emergence. That quality is sometimes underestimated. Consequently, physicians and nurses feel free to talk overtly of the patient and of his/her gloomy prognosis, at his/her bedside. Some of these cases may be in fact cases of Locked-In-Syndrome in which the patient has a normal consciousness but typically has lost all of motricity, except for eye movements, which are preserved. It seems to us noteworthy to mention here the case of Karen Ann Quinlan, who, in the mid seventies, became famous through the American media: This young woman was one of the first cases of long lasting coma. Her situation is one of cases that had great impact on medical ethics in the USA (Pence 2004).
Imagery is, here, indeed crucial: first, in emergency to ascertain and localize an intracranial hematoma or to diagnose and visualize traumatic lesions of the spinal column and/or of the spinal cord, and later in assessing the consequences. The evaluation of the accident toll of a cranial trauma must be urgently considered. It usually starts by a skull X-ray exam. But CT scan or, above all, structural MRI of the skull and-in many instances-of cervical vertebrae, limbs, etc. are needed. In case of an intracranial hematoma, (extradural hematoma, notably) surgery may be an emergency. Weeks or months after the accident, the after effect will be assessed with help of imagery also.
Brain tumors probably represent the most evident application of neuroimaging since it shows the proliferating process, tumor location (most brain tumors show predilection for one several locations: pinelomas in pineal gland, medulloblastoma in cerebellar hemisphere, glioblatoma in fronto callosal area, etc.), tumor size and structure (necrosis, heterogeneity, hemorrhage, neovascularisation indicate malignancy). Histopathology and/or cytopathological examination is necessary to establish the diagnosis. Biopsy or the puncture of the tumor is unavoidable. They are guided by imaging also. The microscopic diagnosis begins intra-operatively (“frozen sections”) and may often guide the surgeon. In rare instances (in frail, weak patients with a large and cortical expanding process fairly easily accessible to the needle), a puncture, performed under CT scan, can suffice for a cytological diagnosis. This procedure, relatively inexpensive, is widely used in poorly equipped hospitals of developing countries.
Pharmaco-resistant, intractable epilepsy patients, are sometimes treated by neurosurgery. In such cases, a detailed structural imagery is needed prior to implanting electrodes and topographical monitoring. Then follows surgical resection, usually a partial temporectomy. The conjunction of structural imagery (disclosing discrete malformative dysplasias), cortical electroencephalogram in different areas, and per-operative histopathology allow a targeted resection of the epileptic foci, whereas normal brain is spared. Functional MRI permits to determine, prior to operation, the localisation of speech areas, to be spared by the neurosurgeon.
Drevets et al. (1997) found in people with history of depression a 48% loss of gray matter in their left subgenual prefrontal cortex, a region involved in the mediation of emotional and autonomic response to socially significant or provocative stimuli and the modulation of neurotransmitter systems targeted at by anti-depressants. Nugent et al. (2006) found, in bipolar disorder, morphometric abnormalities in the posterior cingulate/retrosplenial cortex and superior temporal and lateral orbital cortices. They consider that this supports the hypothesis that the extended network of neuroanatomical structures sub serving visceromotor regulation, contains structural alterations in bipolar disorder. In addition, abnormalities of areas known to exhibit increased metabolism in depression supports the hypothesis that repeated stress and elevated glucocorticoids secretion may result in neuroplastic changes in bipolar disorder. According to Friedman et al. (2006), the data available from literature, based on functional brain imagery, offer support that depressive states in bipolar disorder are likely to be neurologically different from depressive states in major depressive disorder, though both share clinical similarities. For Pollock and Kuo (2004), “inconsistencies in clinical imaging studies, patient selection and medication effects, disease heterogeneity and comorbidity and imaging technology should be considered. To sum it up, the potential of brain imaging investigations in bipolar psychopathology is great, and the role of neuroimaging in psychiatry has only begun”.
Quite different clinical symptoms correspond to this notion: loss of contact with reality, delusions, hallucinations, and withdrawal from outside world, emotional indifference. These symptoms can be grouped in clinical entities, which are almost innumerable, without consensus and depend on schools of thought. An example is the progressive scheme: psychopathological description, association of symptoms, clinical entity grouping, and nosology according to etiology. Post mortem examinations showed nonspecific lesions (conclusion of the First international congress of neuropathology, Rome 1950). As schizophrenic patients age, they often present dementia; cerebral atrophy may be found at brain imaging.
Difficulties in diagnosis in the field of schizophrenia are due to conceptual limitations. The DSM IV American manual has fragmented the nosology of schizophrenias. Schizophrenia entity has a multifactorial causality. In research, the diseased individuals have to be matched with “normal controls”, thus are dependant on concepts and definitions that vary notably with culture and normative sample. Long-term neuroleptic treatment may induce deep metabolic and even anatomical changes. Modern structural neuroanatomy discloses presently abnormalities, evocative of developmental origin (Goldman et al. 2009). Genetic predisposition is not clear enough, at present, to be predictive. fMRI studies are performed however, in schizophrenia. For instance, in the study of self-consciousness (Vogely 2003), it is suggested that in the pathophysiology of schizophrenia, the clinical subsyndromes like cognitive disorganization and derealization syndromes reflect disorders of partial features of self-consciousness. Abnormal activity of some brain structures like the amygdala linked to emotion (like abnormal fear) can be seen in schizophrenics [see supplementary Figure].
Studies of neuroimaging in schizophrenia, as well as in several other disorders, need to be included in heavy, long lasting, multidisciplinary protocols, including genetics (Potkin et al. 2009). These studies must be longitudinal over years, decades, and based on many patients and controls. In these protocols, both structural and functional imaging have a major role to play. Pharaonic is, here again, a suited adjective.
Risks for Patients Linked to the Neuroimaging Technologies
Many medical technologies impose certain risks for patient’s health. This is also true for some of the brain imaging technologies. Harm ranges from the iatrogenicity of imaging technologies to incidental findings of life-threatening processes.
Patient’s Risks of Exposure to Ionizing Radiations
Due to the specific technological setup, risks associated with ionizing radiations are absent in performing fMRI. But, they have to be considered with techniques like PET scan and CT scan, especially if these examinations are repeated on the same patient: A chest X-ray gives out a 10 millirems radiation; a head CT 150 millirems.
The charts of approximately 30,000 patients were reviewed who underwent 190,000 CT scans to estimate cumulative radiation exposure and lifetime attributable risk of radiation-induced cancer. The results of this study suggest that cumulative CT radiation exposure added incrementally to baseline cancer risk in the cohort. A subgroup of seven percent had an estimated lifetime attributable risk greater than one percent (Sodickson et al. 2009).
The Risks of Magnetic Field
The presence of metal inside the cranium, such as metallic surgical devices, can prohibit the use of some imaging technologies. With brain MRI, there is the possibility of displacement of the metal, which may be the cause of a lesion of nervous tissue.
Incidental findings are frequent on images of the brain performed, for instance, in traumatology or for unusual headaches. Thus, they represent the major issue of risk-assessment. Fortunately, the large majority of these findings are no big threat to health. Many of them can be called “anatomical variants” and are more easily detected by experienced neuroradiologists. Their detection should become easier with progress in neuroanatomy, which will be brought by normative studies in large series representative of populations in different stages of development, plasticity, and aging through life. The need is great for such studies.
Only one to two percent of cases of incidental findings may have serious outcomes. Eskandary et al. (2009) studied the frequency of incidental findings of 3,000 CT scans of trauma patients. Most frequent findings were not threatening, e.g., enlargement of cisterna magna (11), arachnoid cyst (7), and hydrocephaly (3). But also, fourteen tumors were found, such as meningiomas (3), osteomas (3), lipomas of mid-line (3), craniopharyngiomas (2), low-grade astrocytoma (1), oligodendroglioma (1), and medulloblastoma (1).
The most demanding finding, in terms of emergency, is the incidental disclosure of an intracranial aneurysm on brain imaging performed after a trauma, or, in a few cases, of (“healthy”) subjects participating in a neurocognitive research program. Mamourian (2007), director of such a program, encountered such a situation personally. The decision to operate, taken by a neurosurgeon (surgical intervention with craniotomy) or a neuroradiologist (operation with endovascular catheter), will depend on several factors such as age, gender, blood pressure, and anatomical location of the aneurysm.
Ruptured intracranial aneurysms are almost all treated (except extreme old age or conditions approaching brain death) in order to avoid disastrous rebleeding. Concerning unruptured intracranial aneurysms, Brisman et al. (2006) summarizes the situation as a dilemma: To treat or not to treat?, Brisma and co-workers think that with the increased ability to diagnose aneurysms comes the need to identify, which incidentally revealed aneurysms should be treated, and with what modality. Physicians offering both modalities of treatment will take the decision: neurosurgeon for “clipping”, interventional radiologist for “coiling”. Many factors are to be considered: the location of the aneurysm, as much as the age, co-morbidities, and the anxiety of the patient (Damocles’ sword). The most important factor is the size of the aneurysm. The authors consider that the natural history risk of rupture of an incidentally discovered aneurysm is not known. They mention the retrospective work of the International Study of Unruptured Intracranial Aneurysms (ISUIA) (1998). Among 2,600 subjects from 53 centers, small aneurysms had a risk of rupture of 0.05% per year. For posterior communicating artery aneurysms of 13–25 mm, the risk at 5 years was 18.4%.
According to Raymond (2009), his critical review of available literature for treating incidental intracranial aneurysms did not provide reliable numerical data of sufficient precision and quality to feed mathematical models of prediction of potential treatment benefits: “Some clinical dilemmas survive decades of controversy “should incidental intracranial aneurysms be treated or left alone?” is one of them. Eminent registries have claimed that the preventive treatment of incidental intracranial aneurysms is rarely indicated (Wiebers et al. 2003)”. According to the author, different approach using randomized trials is now in order.
Benign tumors like meningiomas can be discovered incidentally. In case of gliomas, imagery can disclose heterogeneity, necrosis, neovascularization, and ill-defined limits–all findings in favor of malignancy.
In an asymptomatic aged or middle aged person, the finding of a cerebral atrophy may raise the question of a preclinical, predementia state, and of a possible Alzheimer’s disease to come: should this finding be announced to the patient?
These incidental findings raise ethical questions and fully justify the procedure of informed consent to imaging.
Patients’ Psychological Reactions When an MRI is Performed
Some patients can hardly bear to have their head motionless and confined inside a small space. Indeed, a discomfort can be a manifestation of respiratory insufficiency where lying down is unpleasant. It can also be a psychological reaction such as claustrophobia. A paranoid patient may resist undergoing an MRI, and suspiciously misinterpreting the motives and intentions of physicians. He may develop an “influence” delirium. A depressed patient may feel more distress and confusion when knowing that his brain shows anomalies and, thus, may feel he cannot get better, relieved from his symptoms. Anxious people may be overly concerned with infra-clinical findings, such as a cerebral atrophy (e.g., elderly with regard to the validation of dementia).
Finally, one must mention the–at present only theoretical–possibility of problems with confidentiality or manipulation of images (data misuse) by some individuals highly performant in computer science (remember the image is a construction by computers).
Outcomes Versus Expectations of Patients Being Attracted By the ‘Imagery’
Images transmit a great number of information to people looking at them. They are able to convey concepts, but also emotions. The image can be static like a CT scan; it can be dynamic like fMRI. For Ackerman (2006), the most powerful impression of a functional brain image is still the simplest one that we are actually seeing the brain in action. The effect will be particularly strong with people who are not familiar with such images, particularly people recently arrived from poor countries.
It would be interesting to think about possible narcissism of a subject with psycho-somatic disorders. Despite many steps preceding the final image (magnetic resonance, statistical operations, heavy computing), the final product seems to offer a direct view of the cerebral scene. The image of an easily recognizable shape lit up with attractive colors has a powerful visual impact. This image, Ackerman (2006) writes, “is as persuasive as the sonogram of a thumb sucking fetus, just as unforgettable”. Most people do not understand looking at the “picture” that they are looking at a mathematical abstraction. Images appear more and more in different medias, with less and less explanation of their contents. Looking at different magazines–general, scientific, health, medical ones–of these last 40 years, one sees the emergence of diagrams, photomontages, etc. based on the computer metaphor. At stage, the working of the brain commonly is understood as a computer immersed in a “soup” with transmitters, hormones, ions, catabolites, growth factors–if such a daring representation is permitted. The public is—vaguely—aware of the exaggerated claims of possibilities to “read minds” and regard imaging technologies as some “big brother” machine.
Of concern is the possibility that patients, when shown their pathologic brain image, feel encouraged. Like a patient with psychosomatic “fibromyalgia” saying: “I am not a malinger! I have a real disease!”
Two curious concepts appeared recently (Illes after Ackerman 2006): One concept may be called “neuro-realism”: A phenomenon must be real when studied with such a high technology. Also neuro-realism can quickly be adopted by judges in court or by the police. The other confusing concept is “neuro-essentialism”: Subjectivity and personal identity are not just housed in the brain but identical to it. Given this, we believe that “we are our brain”. This sense may be less important in nonwestern cultures (the Japanese, for instance, traditionally think that the self is in the chest). When we say that a person with Alzheimer’s disease is “no longer there” we mean that the person’s thoughts, feelings, and interactions are totally gone forever. The parents of their comatose daughter see in her all the things they loved in her: “She is still with us”. Both sides are true: A patient can still be here and not be here any more. A difficult ethical challenge!
The brain is heterogenous both structurally and in terms of function (motor, sensitive, visual, emotional, cognitive, sleep-wakefulness, etc.). The diseases of the brain vary in nature, frequency, and consequences with parts of the brain involved in the pathological process.
Neuroimaging has literally revolutionized and changed the practice of neurology, especially as a diagnostic tool showing anomalies and/or lesions. It also brings unbeatable topographical information: The limits of a lesion can be precisely defined and located, and the perilesional, nonspecific modifications (e.g. edema, gliotic reaction) can also be evaluated. Thus, the lesion can be eradicated whereas healthy brain matter can be spared. Furthermore, imaging techniques give information about the structure of the lesion, its density, its homogeneity or heterogeneity, etc., which by them can help the etiological diagnosis.
Functional imagery can detect changes, i.e. neuronal hyperactivity or hypoactivity that even a thorough post mortem neuropathological examination (markers of proliferation of a tumor, neurotransmitters, etc.) cannot disclose.
Imagery, in short, allows bringing the neurologist decisive information on the nature of the lesion and its topography, information that long was accessible only after craniotomy, if at all. Neurologists have at their disposal a growing panel of technological methods available for brain imaging for diagnosis. PET can disclose amyloid deposits, thus pulling the diagnosis of Alzheimers’ diseases toward certainty. Vascular imaging, i.e., angiography allows interventional radiology: it is, now, very frequently possible to cure an intracranial aneurism without opening the skull. Cognitive neuroscience data begin to play an important role in neurological diagnosis (e.g., abnormal emotions in schizophrenics via fMRI).
In sum, neuroimaging is an invaluable group of tools for neurological clinical research, especially in clinical cognitive neuroscience and in clinical neuropharmacology. Also, imaging technologies become more and more sensitive and keen. As illustration, we mention the fact that disclosing deposits of amyloid beta protein in the brain of a person having no symptom of Alzheimer’s disease raises difficult ethical questions. Is it fair to induce anxiety in the person whereas no really efficient treatment is available? Such person may in fact never develop clinical symptoms and signs of Alzheimer’s disease before they die of other diseases. The opinion of the very large majority of neurologists, psychiatrists, or gerontologists is that no screening for the disease should be done in people without any sign or symptom of Alzheimer’s disease.
The author wishes to thank Pierre Krolak-Salmon for his neurological expertise, Lara Huber for suggestions and editorial help, Elizabeth Valour for linguistic corrections, Emmanuel Jouanneau for information on neurosurgery, Jean Claude Froment and Francis Turjman for neuroradiology.