Clinical Systems Neuroscience Takashi Hanakawa, MD, PhD (
[email protected]) Department of Advanced Neuroimaging, Integrative Brain Imaging Center, National Center of Neurology and Psychiatry, Kodaira, Tokyo 187-8551, Japan
Word count (9595 words)
2
Abstract Here ‘clinical systems neuroscience’ is advocated as an academic discipline for clinicians and scientists who are, or will be, involved in research related to neurological, neurosurgical, and psychiatric disorders. Clinical systems neuroscience concerns a neural function as an operation of a system, not of a simple collection of elements. It deals with issues situated in the intersection between clinical practice and neuroscience at the system, cognitive and behavioral levels. A technical background of clinical systems neuroscience is recent emergence of non-invasive brain mapping, brain stimulation, and brain-machine interface techniques. Clinical systems neuroscience is about to depart from a method for understanding pathophysiology of neuropsychiatric disorders to a clinical tool supporting diagnosis as well as treatment.
Keywords Neurology, Neurosurgery, Psychiatry, Systems Neuroscience, Cognitive Neuroscience, Behavioral Neuroscience, Neuroimaging, Neurophysiology, Brain Stimulation, Brain-machine interface
3
1. Principles of clinical systems neuroscience Here the author proposes ‘clinical systems neuroscience’ as a discipline situated at the intersection between clinical medicine and neuroscience at the levels of systems, cognition and behavior. Clinical systems neuroscience concerns malfunctions resulting from disease of the neural systems related to movement, sensory perception, imagery, memory, attention, language, learning, cognition, attention, emotion, social interaction, and decision making. In neuroscience, high-level brain functions such as cognition, language, and decision making had been the subject of cognitive/behavioral neuroscience. In my opinion, however, accumulating knowledge has made the distinction between systems neuroscience and cognitive/behavioral neuroscience unclear in terms of neural correlates 1, computational principles 2, and, probably, the level of organization. More importantly for clinician-scientists, an isolated set of knowledge specific to each neuroscience subdiscipline is not quite useful in clinical settings because problems existing in a neurological patient often span from systems neuroscience to cognitive/behavioral neuroscience. Therefore, I propose a discipline covering basic knowledge and methodology bridging over systems, cognitive and behavioral neuroscience for clinician-scientists (e.g. neurologists, neurosurgeons and psychiatrist, neuroradiologists, and rehabilitation therapists who have scientific and inquisitive minds). Clinical systems neuroscience would provide a lingua franca for clinicians who are interested in cutting-edge research, neuroscientists who are interested in malfunctions of the nervous system, and also engineers who want to develop innovative systems for patients with neuropsychiatric disorders. Remarkable development of computer science, informatics, and information technology (IT) in the past few decades supports non-invasive measurement and analysis of brain activity offline and now online, and greatly contributed to the emergence of clinical systems neuroscience. Integration of multimodal neuroimaging in both space and time, or “multidimensional” neuroimaging 3, which combines methodology in clinical neurophysiology, neuroimaging, neurostimulation and neuro-IT, would become an even more powerful tool for clinical systems neuroscience.
2. Clinical Systems Neuroscience: Past Several streams of clinical and academic disciplines join together to form clinical systems neuroscience. Starting from the clinical side, knowledge from clinical experiences was a foundation of clinical neuroscience. The discovery of association between loss of speech production (motor aphasia) and left inferior frontal lesion by Paul Broca 4 was the starting block for the understanding of localization of cognitive functions to specific convolutions of the brain. This epoch-making discovery was followed by a variety of lesion studies associating cognitive functions
4
with a specific region of the brain: for instance, memory studies inspired by the case of patient HM suffering from bilateral hippocampal lesions 5 and behavioral studies inspired by the case of Phineas Gage suffering from a medial prefrontal lesion 6. A lineage of these lesion studies in neurology and cognitive neuropsychology is an ancestor of clinical systems neuroscience. Of note, the brains or skulls of those historical patients have often been preserved and later examined with modern neuroimaging techniques 6,7. Functional neurosurgery has also provided an important basis of clinical systems neuroscience. Invaluable knowledge about functional organization of the cerebral cortex comes from Wilder Penfield’s electrical stimulation to the brain of a patient awake during neurosurgical interventions. Recordings from, and stimulations to, the human brain using surgically implanted electrodes are important opportunities for both understanding of systems in human brains and developing new therapeutic methods. From a technical perspective, clinical neurophysiology is a medical specialty that has used electrophysiological methods such as electroencephalography (EEG) and electromyography (EMG) for recording spontaneous and evoked responses in both peripheral and central nervous systems. Using these recording techniques, clinical neurophysiology aims at diagnosing patients with nervous system disorders and also at clarifying their pathophysiology. The concept of clinical neurophysiology partially overlaps with that of clinical systems neuroscience. Another technical background is provided by neuroradiology, which is a subspecialty of radiology, focusing on diagnosing abnormalities in the central and peripheral nervous systems, using magnetic resonance imaging (MRI) and computed tomography (CT). MRI is one of the most powerful tools subserving clinical systems neuroscience. Also, another medical sub-discipline, nuclear medicine provides useful technology, such as positron emission tomography (PET) for clinical systems neuroscience. As a sub-discipline of basic neuroscience, systems neuroscience studies the brain and spinal cord as systems, not as elements, in anatomical, functional and computational viewpoints often at the level of neural circuits (Figure 1). Systems neuroscience typically, but not exclusively, studies how an organism perceives stimuli from external environment, makes a decision, and translates the decision into behaviors. Neuroanatomical and neurophysiological techniques, and more recently neuroimaging, are typical technology applied to systems neuroscience. Cognitive neuroscience is based upon both psychology and neuroscience, and concerns neurobiological substrates of cognitive abilities and mental processes. Neuroimaging techniques that enabled non-invasive measurements of brain activity during various mental processes in humans have greatly contributed to the advance of cognitive neuroscience. Behavioral neuroscience weights more on behavioral aspects that can be also studied in non-human animal species. These subdisciplines of neuroscience partially overlap with each other. Moreover, as mentioned earlier, problems often span from systems neuroscience to cognitive/behavioral neuroscience levels from clinical points of view. In my opinion, therefore, the concept of systems neuroscience should be extended to cover issues handled typically by cognitive/behavioral basic neuroscience. Clinical systems neuroscience gets started when such an extended version of systems neuroscience
5
is applied to clinically oriented questions, using non-invasive physiology, imaging, stimulation, and neuro-IT methodology.
3. Clinical Systems Neuroscience: Present Clinical systems neuroscience examines functional neuroanatomy of the nervous systems in living humans, especially that of patients with neurological, neurosurgical, psychiatric disease or other systemic disease involving nervous system, with non-invasive methods. An exception would be functional neurosurgery. Currently available tools are based upon electrophysiological recordings, neuroimaging, brain stimulation, and neuro-IT such as neuroprostheses. Explanation of technical details of each method and its application to each functional system is beyond the scope of this article. Hence, I will only briefly explain summarize technical backgrounds and characteristics of each tool that have supported emergence of clinical systems neuroscience. Likewise, the space is too limited to discuss a complete list of application to all the neuropsychiatric disorders. Therefore, I only describe a few typical diseases as typical applications of each tool for solving problems in clinical systems neuroscience.
3.1. Electrophysiology 3.1.1. Electroencephalography (EEG) Since its discovery by Hans Berger in 1929, EEG has been widely used as a clinical or scientific method to monitor ongoing brain activity. Signals acquired with scalp electroencephalogram (also alleviated as EEG) are believed to reflect summation of local field potentials (LFPs). LFPs recorded by EEG are considered to reflect mostly excitatory postsynaptic potentials (EPSP), partly inhibitory postsynaptic potentials (IPSP), generated in the apical dendrites of pyramidal cells. Hence, what we see in EEG is mostly inputs into a large population of cortical neurons; when synchronized, these inputs may produce oscillatory EEG patterns that can be recorded from surface EEG. A scalp EEG is recorded with an instrument (electroencephalograph) and electrodes attached onto the scalp, often arranged according to the International 10-20 system (Fp1, Fp2, F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T3, T4, T5, T6, Fz, Cz, Pz, A1 and A2). The letters ‘F’, ‘T’, ‘P’ and ‘O’ represent frontal, temporal, parietal and occipital lobes, respectively. The letter ‘C’ refers to electrodes nearby the central sulcus, corresponding to the primary motor and somatosensory cortices.
6
The letter ‘A’ identifies the earlobes. The letter ‘z’ (zero) refers to electrodes placed on the midline. Even numbers and odd numbers refer to electrode positions on the right and left hemispheres, respectively. For research purpose, electrode caps with more dense arrays of electrodes (e.g., 64, 128 and 256 electrodes) are available. EEG often reveals conspicuous oscillatory activity. Oscillatory EEG activities are classified into α (8-15 Hz), β (16-31 Hz), γ (over 32 Hz), δ (below 4 Hz) and θ (4-7 Hz) ranges or bands according to their frequency. Resting-state EEG activity is dominated by α-range oscillations over the parieto (P)-occipital (O) electrodes (“posterior dominant rhythm”) in healthy subjects. The posterior dominant rhythm is suppressed by eye opening (visual inputs) and by reduction of arousal levels. Although how the posterior dominant rhythms are generated and modulated is not clear, fMRI combined with EEG has started to cast lights on modulation mechanisms of the posterior dominant rhythm at rest 8. Meanwhile, the sensorimotor rhythms (SMRs), or mu rhythms, are α-band oscillations recorded over the central (C) electrodes. SMRs are considered to reflect idling state of the motor system since movement suppresses SMRs. However, since SMRs are also suppressed by motor imagery or motor intention, amplitude of SMR is now used as a feature for detecting motor intention of patients who cannot actually move (i.e. brain-machine interface). Combination of normal oscillatory activities observed in EEG recording is collectively called background EEG activity. Diffuse slowing and/or reduction of background EEG activity indicate diffuse impairment of brain functions (diffuse encephalopathy), especially when accompanied with disturbance of consciousness. The complete loss of EEG background activity is termed electrocerebral inactivity or silence, which is seen in case of the complete and irreversible of brain functions (brain death). Local slowing or reduction of EEG background activity suggests focal abnormality of brain functions resulting from, for example, space occupying lesions and ischemic lesions. Detecting epileptiform discharges with EEG is an indispensable process in the diagnosis of epilepsy. Epileptiform discharges are distinctive waves or complexes lasting for several tens of milliseconds to a few hundred milliseconds, often called “spikes” or “sharp waves”, distinct from background activity. Physiologically, an epileptiform discharge corresponds to a paroxysmal depolarization shift, which is abnormal and large EPSP occurring in a group of neurons in epileptic foci. Advantages of scalp EEG as a tool of clinical systems neuroscience include noninvasiveness, fine temporal resolution, and prevalence among research and clinical institutes all over the world. Conversely, shortcomings of scalp EEG may be poor spatial resolution/localization and inaccessibility to deep brain structure. To supplement the poor localization, source localization techniques have been developed with the aid of anatomical/functional MRI 9 and modeling methods 10. With the development of these techniques along with other visualization technology, EEG has now become an even more powerful tool in basic and clinical systems neuro-
7
science than before. An example would be brain-machine interface using SMRs or evoked potentials such as P300 as will be discussed later. Furthermore, EEG combined with other methodology is an emerging method for taking advantage of high temporal resolution of EEG and supplementing spatial localization with other imaging methods. For example, simultaneous EEG and neuroimaging, especially fMRI, is an emerging method for localizing epileptic foci 11.
3.1.2. Intracranial recording There are a few invasive electrophysiological techniques established for clinical application. Electrocorticography (ECoG) is an invasive technique by which EEG can be recorded from array of electrodes placed over the surface of the cerebral cortex. Typically, arrays of platinum electrodes are placed subdurally by means of surgery. ECoG has been the most reliable and powerful tool for localizing epileptic foci in patients with intractable epilepsy who are candidates for surgical removal of epileptic foci. Another advantage of ECoG is that it can map functional areas of the cerebral cortex beneath electrode arrays so that surgical removal of the tissue does not damage functionally important (“eloquent”) brain regions. This mapping can be done with electrical stimulation to the brain through the electrodes or recording of evoked potentials with cognitive, motor and sensory tasks. Brain mapping with ECoG provide complimentary information with other brain mapping tools such as fMRI 12. Moreover, ECoG may clinically serve realization of high-performance BMI 13 since ECoG has much better signal-to-noise ratio than scalp EEG. Another type of intracranial electrophysiological recordings of clinical importance can be obtained from electrode implanted for deep brain stimulation (DBS) to treat movement disorders. Recordings from DBS electrodes placed in the subthalamic nucleus or the globus pallidus have provided invaluable information about pathophysiology of Parkinson’s disease. In particular, it has been shown that β-range oscillation is enhanced in the subthalamic nucleus and high-frequency stimulation suppresses the oscillation, thereby inducing clinical benefit in Parkinson’s disease 14.
3.1.3. Evoked potentials and event-related potentials Triggered averaging EEG signals time-locked to repetitive sensory stimuli allow us to record evoked potentials by means of both scalp and intracranial recordings. Similarly, it is possible to record averaged EEG potentials time-locked to particular cognitive or motor events (event-related potentials or ERPs). Sensory evoked potentials (SEPs) are evoked responses in the peripheral nerves, spinal cord and brain following electrical stimulation to peripheral nerves such as the median nerve and the tibial nerve. SEPs are useful in detecting damage to the
8
posterior column-medial lemniscus sensory system, which conveys touch sense, vibration sense, and proprioception. In patients with multiple sclerosis, sensoryevoked responses can be delayed or diminished. Conversely, patients with cortical myoclonus often show “giant SEP” reflecting cortical hyperexcitability 15. Visual evoked potentials (VEPs) are responses to flash stimuli or visual pattern stimuli, and are recorded over occipital EEG electrodes. Abnormal VEPs may result from dysfunctions of the visual system including the visual nerves, optic tracts, lateral geniculate, optic radiation and primary visual cortex. Short-latency auditory evoked potentials (AEPs) are often called brainstem auditory evoked potentials (BAEP). BAEPs typically show seven positive peaks within 10 milliseconds after click sounds. BAEPs are useful in testing functions of the brainstem. P300 is a positive ERP recorded at approximately 300 milliseconds after presentation of rare stimuli in an “oddball” paradigm in which frequent and rare stimuli are randomly presented and a participant is asked to count the number of rare stimuli. P300 is applied to a type of BMI (“P300 speller”) detecting brain activity related to selection of stimuli such as letters 16. Also based on an oddball paradigm, mismatch negativity is an ERP following rare stimuli regardless of whether a participant pays attention to the stimuli or not. Mismatch negativity is reduced in patients with chronic schizophrenia 17. Movement-related cortical potentials (MRCP) or Bereitschaftpotentials (BPs) are slow cortical potentials preceding voluntary movements. MRCP/BPs are thought to originate from the supplementary motor areas and primary motor cortex. Contingent negative variations (CNV) are slow negative EEG potentials observed when a participant is anticipating an event of interest 18. CNVs are influenced by attention to the event.
3.1.4. Magnetoencephalography (MEG) Similar to EEG, the main source of the MEG signals is synchronous postsynaptic currents in the pyramidal neurons of the cerebral cortex 19. These electric currents accompany an electromagnetic field, which is theoretically detectable with coils outside of the body. In contrast to distorted electrical signals measured with scalp EEG because of the poorly conducting skull bones, the magnetic field measured with MEG is not affected by the skull. MEG therefore has a great potential to localize the source of synaptic activity quite precisely. Although an electromagnetic field in the brain is very small (10–100 fT), development of superconducting quantum interference devices (SQUIDs) has made detection of electromagnetic field in the human brain possible 20. Now, a helmet-type, whole-scalp covering MEG system with more than 300 SQUID sensors is available. It should be noted that MEG and EEG complement each other. MEG has different sensitivity to electromagnetic fields, depending upon the direction of the currents.
9
MEG is most sensitive to currents tangential to the skull, namely post-synaptic activity occurring in the banks of cerebral gyri buried within the sulci/fissures, whereas EEG has sensitivity to signals from the depth of the brain and from the cerebral convexity. With these limitations in mind, electromagnetic source modeling that can provide information on the timing and direction of current flow is of great help in interpreting MEG data. The most straightforward approach has been to use an equivalent current dipole (ECD) model. A single ECD or multiple ECDs can be modeled. In minimum-norm or minimum-current estimates (MNE/MCE), many small dipoles are placed throughout the cortex and their strengths are estimated as a function of time 21. In beamforming procedure 22, the brain volume is scanned by a sequential application of spatial filters to pass activity from a specific brain area with maximum gain, while suppressing activity from other areas. Despite these progresses, the currently available source analysis procedure cannot circumvent the ‘inverse problem’; that is, the source of electromagnetic field cannot be uniquely solved. MEG methodology has been applied to solving questions in basic systems/cognitive/behavioral neuroscience including sensory perception, movement, language, social interaction and development/learning. MEG can pinpoint cortical generators of various evoked and event-related potentials that had been studied with EEG. MEG can also be applied to studying oscillatory brain activity and inter-areal connectivity. In the clinical domain, MEG plays an important role in epilepsy and preoperative mapping 23. MEG has also been applied to pathophysiological studies in movement disorders 24 and schizophrenia (e.g. mismatch field).
3.1.5. Assessment of peripheral nervous systems and muscles Conventional electrophysiology techniques for studying the functions of peripheral nervous system and muscles are of importance alone or in combination for assessing peripheral functions of motor, sensory and autonomic systems. It is also important to use those in combination with modern neuroimaging and neurostimulation techniques. Needle EMG is an important technique to detect motor unit potentials. Active denervation can be diagnosed by detecting fibrillation potentials and positive sharp waves during muscle relaxation. Surface EMG is useful for assessing gross muscle activities. For quantification, integrated EMG has often been used since it is proportional to isometric tension. Surface MEG is also used for recording muscle activities evoked by electrical or magnetic stimulation of peripheral nerves and motor cortex (nerve conduction studies, F-waves, H-reflex, etc.). It is possible to record surface EMG during neuroimaging 25,26, and simultaneous brain stimulation and neuroimaging 27-29. Surface EMG help assess gross pattern of involuntary movement such as tremor, myoclonus, chorea and dystonia. Therefore, for neuroimaging studies on pathophysiology of involuntary movement, it is essential to
10
monitor and record surface EMG simultaneously and use the data as a covariate in the imaging analysis for proper interpretation of imaging findings.
3.2 Neuroimaging The concept of the clinical systems neuroscience substantially relies upon the development of neuroimaging techniques that have emerged in the last two decades of the 20th century. Positron emission tomography (PET), single-photon emission computed tomography (SPECT), near-infrared spectroscopy (NIRS) and MRI exemplify functional neuroimaging techniques. These techniques allow us to obtain functional, neurochemical and/or anatomical information from healthy and diseased human brains non-invasively. These methods are occasionally called brainmapping techniques since these are useful in identifying which parts of the brain represent a function of interest and then to create functional “maps” of the brain. A philosophy underlying such acts is an assumption that a spatially segregated part of the brain code one or more distinct brain functions (functional segregation). As coordinates of brain maps, researchers in the filed most often use Talairach and Tournoux’s coordinate system 30 developed originally for assisting neurosurgeons for planning and executing stereotaxic surgery. A surface-based coordinate system is also proposed for mapping the cerebral cortex 31,32. However, the concept of functional segregation may not account for all the control mechanisms of cognition and behaviors. In this regard, it is also important to test a hypothesis that functions of the brain may primarily arise from interactions across different brain regions (functional integration). It should be noted that the functional segregation and functional integration are not mutually exclusive. Hence, it may be worthwhile considering both types of control mechanisms always to explain a particular ability of the brain before jumping into conclusion.
3.2.1. Magnetic resonance imaging (MRI) MRI is a medical imaging technique utilizing a nuclear magnetic resonance (NMR) phenomenon in which nuclei in a magnetic field absorb and emit electromagnetic radiation. Nuclei with an intrinsic magnetic moment and angular momentum (such as 1H and 13C) can be tested with NMR, but currently available medical MRI systems almost exclusively use 1H because of high prevalence of 1H (proton) in the body. Protons behave like microscopic magnets. Within a strong static magnetic field, some protons are aligned parallel to the magnetic field and some are aligned anti-parallel to the field. The number of parallel protons and that of anti-parallel protons are not equal; the number of parallel protons is slightly larger than that of anti-parallel protons. This difference (approximately 2 x 1015 protons in a 1 x 1 x 5-mm3 voxel) produces a magnetization vector, which is a vectorial summation of all the magnetic protons in the voxel. Here, a ‘voxel’ refers to a unit in a regular grid in three-dimensional space, and is a unit of MRI ac-
11
quisition. Now, protons have a resonance frequency of 128 MHz at 3 Tesla. When a 128-MHz electromagnetic pulse is emitted from a transmit coil, protons absorb it and start to resonate in a 3-Tesla MRI system. By creating a magnetic gradient and modulating electromagnetic pulses, MRI can tag the position of a voxel where resonated protons would emit electromagnetic radiation. And, when the electromagnetic pulse disappears, the resonated protons start to emit electromagnetic radiation during a process toward the original condition. This is called a relaxation process during which the electromagnetic radiation from the protons can be detected with receiver coils. There are a few tissue factors that influence the relaxation process: the longitudinal relaxation time (T1) and the transverse relaxation time (T2). T2 assumes a homogenous magnetic field in the environment, but a biological environment has magnetic inhomogeneity. For example, magnetic inhomogeneity may result from deoxy-hemoglobin (deoxy-Hb), which is a paramagnetic substance in red blood cells. Transverse relaxation time of the tissue in the presence of magnetic inhomogeneity is called T2*, which is shorter than T2. Conventional brain structural MRIs, such as T1-weighted, T2-weighted, proton density, and fluid attenuated inversion recovery (FLIAR) images, have been of great importance in neuroradiological diagnosis of disorders. These images can detect atrophy, ischemic lesions, hemorrhages, space occupying lesions, inflammation, metal deposition, and so forth. For both clinical and scientific studies, three-dimensional T1-weighted MRIs with fast spoiled gradient echo (FSPGR) or magnetization prepared rapid gradient echo (MPRAGE) sequence (Figure. 2) have been often used for volumetry, voxel-based morphometry (VBM), tensorbased morphometry (TBM) and cortical thickness measurement. Gray matter volume or cortical thickness as studied with these methods has been shown to correlate with many aspects of human abilities 33. Recently, VBM has also been used as a method to elucidate neuroplasticity associated with development as well as use-dependent neuroplasticity associated with various kinds of learning and training in adults 33,34. GM/cortical thickness is also informative in clinical studies in dementia, movement disorders, and psychiatric disorders 35-38. An emerging MRI methodology is diffusion-weighted MRI (DWI) techniques (Figure. 2) such as diffusion tensor imaging (DTI), diffusion kurtosis imaging (DKI), diffusion spectral imaging (DSI), and q-space imaging (QSI). DWI is a technique to visualize diffusion of protons, especially those of water molecules. DWI has placed a great impact on clinical practice in detecting early stroke 39. DTI-derived images, such as fractional anisotropy (FA) and mean diffusivity (MD)/apparent diffusion coefficient (ADC) images allow quantitative analysis of changes in water diffusion associated with morphological changes and damages of the brain tissue. Moreover, diffusion-based fiber tractography has proven to be useful in assessing macroanatomical fiber connections in both healthy and disease people. For example, DWI-based tractography is useful in determining the dominant hemisphere 40 and in guiding neurosurgical procedure 41. It is also possible to
12
assess longitudinal changes in DWI-derived parameters resulting from neuroplastic changes associated with learning 33. A functional neuroimaging method using deoxy-Hb as an internal contrast agent is called blood-oxygenation level-dependent (BOLD) fMRI 42. To date, BOLD fMRI has been mostly achieved with T2*-weighted MRI such as gradient-echo echo planar imaging (GE-EPI), and uses T2* signals as a surrogate measure of synaptic/neuronal activations (see Neurovascular-metabolic-vascular coupling below). For more than two decades, BOLD fMRI has been the most widely and intensively used neuroimaging/brain-mapping tool, despite the cost of establishing and maintaining an MRI facility. There are several reasons for this. First of all, BOLD fMRI has a finest spatial resolution (typically, a few mm) among the currently available non-invasive brain mapping tools, and, moreover, accurate localization in the whole brain from the cortical surface to the deep nuclei and cerebellum. Nowadays, spinal cord fMRI is also feasible. FMRI has a time resolution (typically, a few seconds) fair enough to capture BOLD signal time course after a brief stimulus. Second, along with fMRI, MRI can obtain a variety of structural information with high spatial resolution as described above (such as 3D T1-weighted image). Third, BOLD fMRI can not only study functional segregation but also functional integration (functional connectivity and effective connectivity) by using either psychophysiological interactions 43-45 or dynamic causal modelling 46. Task fMRI has been applied to an enormous number of issues in systems/cognitive/behavioral neuroscience: sensory perception 47-50, movement 26,51-53, language 54-57, imagery 26,58, attention 59, cognition 25,44,60-62, , decision making 63, and social interaction 64,65. Similarly, task fMRI has been applied to a variety of disorders involving the central nervous system 66-68. Note, however, that an interpretation of task fMRI results is often difficult because task difficulty and task performance are often significantly different between a healthy group and a patient group. To circumvent this problem, recent attention has been drawn to a resting-state functional connectivity MRI (rsfcMRI) in which participants undergo no particular task 69,70. This variation of fMRI method depends upon ultra-slow fluctuations (0.01-0.1 Hz) of BOLD signals, which perhaps reflect changes in synaptic/neural activities occurring spontaneously across nodes of a functional network. Thus, without any tasks, rsfcMRI can identify motor, visual, attention, auditory, as well as default mode network (DMN) 71 (Figure. 2). For example, abnormalities in rsfcMRI have been identified in dementia, schizophrenia, mood disorders, epilepsy and movement disorders 72. Because of its strongly magnetic environment, it was previously difficult to acquire fMRI data simultaneously with other recording. However, because of technical advances, we can now combine fMRI with many other techniques such as TMS 27-29 and EEG 8. In particular, simultaneous EEG-fMRI may become an important tool in clinical research in epilepsy 11.
3.2.2 Positron emission tomography (PET)
13
Positron emission tomography (PET) is a nuclear medicine imaging technique that uses a positron-emitting radionuclide-labeled tracer, and can create a threedimensional functional image of the body including the brain. The labeled tracer is typically injected into the blood stream, and the tracer distributed in the body according to its biochemical characteristics. The positron-emitting radionuclide emits a positron, which travels for a short distance (~1 mm) in tissue and interacts with an electron. When the two particles meet, they annihilate each other and emit annihilation radiation (a pair of 511-keV gamma photons emitted to the opposite directions). This pair of gamma photons simultaneously reaches scintillators of the scanner, and produces a burst of light, which is detected by photomultiplier tubes. By detection of this ‘coincidence event’, PET can accurately localize the distribution of the tracer in the brain. After collecting many coincidence events, a tomographic reconstruction method is applied to yield a three-dimensional image dataset. A shortcoming of the PET is radiation exposure. Moreover, a PET facility, which is typically equipped with a PET scanner(s), a cyclotron, and a laboratory for tracer synthesis, is quite expensive. This is because many positron-emitting radionuclides used for brain PET have short lives and therefore tracer must be synthesized just before the use. Still, PET is a powerful and indispensible functional imaging tool for both basic and clinical domains of systems neuroscience. The strength of PET over other imaging modalities includes high signal-to-noise ratio. Moreover, PET can obtain and specific molecular information depending upon the tracer in use (thereby called ‘molecular imaging’). Brain PET images using 18Flabeled fluorodeoxyglucose (18F-FDG) reflect glucose metabolism associated with levels of regional neural/synaptic activities. Reduction of 18F-FDG indicates pathological changes of brain metabolism associated with neuropsychiatric disorders such as epilepsy 73 and Alzheimer’s disease 74. Moreover, the development of tracers binding to β-amyloid, such as Pittsburgh compound B (PIB), has dramatically changed PET diagnosis of Alzheimer’s disease, preclinical Alzheimer’s disease, and mild cognitive impairment (MCI) 74. Development of new amyloid tracers is following. Other tracers detecting neuroaggregate (e.g., tau) in the human brain are also on the way toward clinical applications. PET also allows for many types of neurotransmitter imaging: dopamine receptors (11C-raclopride and 18F-fallypride), enzymes related to dopamine synthesis (18F-dopa), acetylcoline synthesis (11C-methylpiperidin propionate), and serotonin transporters (11C-DASB). 15 O-labeled water distributes in the brain in proportional to regional cerebral blood flow (rCBF). Changes in rCBF are coupled with changes in neural/synaptic activities as will be discussed later (Neuro-metabolic-hemodynamic coupling). Thus, brain PET images of H215O distribution during a particular task can identify brain regions with neural activation during the task (“PET activation study”) in healthy 25,75 and diseased people 76-78. An activation study is also possible to test dopamine release during a particular task. Internally released dopamine competes with externally administered 11C-raclopride for dopamine receptor D2 binding. Therefore, reduction of 11C-raclopride uptake during an experimental task com-
14
pared with a control task indicates increased dopamine release in the experimental task 79. In recent PET systems, silicon avalanche photodiodes are used in the place of conventional photomultiplier tubes. Since silicon avalanche photodiodes function within a strong magnet, it is now possible to combine PET and MRI as a single hardware device (hybrid PET/MRI system).
3.2.3. Single-photon emission tomography (SPECT) Single-photon emission computed tomography (SPECT) is another nuclear medicine imaging technique using a gamma-emitting radioisotope. A SPECT machine is usually equipped with multiple gamma cameras to acquire multiple two dimensional images from multiple directions. After image acquisition, a tomographic reconstruction method is applied to yield a three-dimensional dataset. SPECT has less spatial resolution, but is less expensive, than PET. SPECT has been used as a tool to measure brain perfusion with 99mTc-HMPAO (hexamethylpropylene amine oxime) or 99mTc-ECD (ethyl cysteinate dimer). Perfusion SPECT can detect diffuse perfusion changes in dementia. Focal changes can be detected in cerebral vascular disorders. Perfusion SPECT has a particular application in functional neuroimaging. PAO and ECD are fixed within the brain within several minutes after injection (“snapshot method”). Therefore, SPECT provides a unique opportunity to study mechanisms underlying walking difficulty in neurological disorders. Patients with Parkinson’s disease show underactivation of the supplementary motor areas during Parkinsonian walking 80 whereas they show overactivity in the lateral premotor cortex during paradoxical improvement of walking under appropriate visual guidance 81. A follow-up study in vascular Parkinsonism replicated dysfunctions of supplementary motor areas and compensatory activity in the lateral motor cortex during Parkinsonian gait 82. SPECT with 123I-ioflupane (FP-CIT) has become available for dopamine transporter (DAT) imaging to visualize nigrostriatal dopaminergic nerve terminals. Parkinson’s disease and dementia with Lewy bodies show reduced DAT signals in the striatum (putamen and caudate nucleus).
3.2.4. Near-infrared spectroscopy (NIRS) NIRS uses the near-infrared region of the electromagnetic spectrum (wavelength, 700-1400 nm) and measures absorbance at specific wavelengths corresponding to oxyhemoglobin (oxy-Hb) and deoxy-Hb. An NIRS device is instrumented with pairs of an emitter and a detector of probes attached over the scalp. The near infrared light in the spectrum of 700-900 nm penetrates the scalp and skull, and then reaches the surface of the brain. In this range of spectrum, oxy-Hb and deoxy-Hb
15
are strong absorbers of the light. Therefore, NIRS can measure relative concentration of oxy-Hb and deoxy-Hb from the brain surface. The detector typically measure reflected (back-scattered) lights in which photons may take a variety of paths, often modeled as an ellipsoid in shape. The distance between the emitter and the detector (typically 2-3 cm) decides the size of the ellipsoid. The modified BeerLambert Law is used to calculate micromolar-level changes in tissue oxy-Hb and deoxy-Hb. Task-induced changes in oxy-Hb and deoxy-Hb are considered to reflect synaptic/neuronal activity according to the principle of neuro-metabolichemodynamic coupling. Therefore, similar to fMRI, functional NIRS (fNIRS) is possible, by measuring oxy-Hb and deoxy-Hb as surrogate markers of synaptic/neuronal activity changes correlated with a task. Although this assumption has been confirmed by a few simultaneous fMRI-fNIRS studies 83, it is also known that fNIRS is strongly influenced by hemodynamic changes in the scalp 84,85. Hence, best efforts should be made to remove the effects of scalp blood flow changes for fair interpretation of fNIRS data. Because of dependence upon hemodynamic signals, temporal resolution of fNIRS is an order of second, although sampling of fNIRS can be carried out at an order of millisecond. Other limitations of fNIRS are difficulty in measuring deep brain structures and low spatial information. On the other hand, fNIRS has obvious advantages over other neuroimaging methods. fNIRS is completely safe and non-invasive, and its setup is relatively easy. Moreover, there is no need for rigid restraint of head and body. Such advantages of fNIRS are best exemplified by application to infants 86. Another example is application to studying cortical motor area activity during walking 87. As a tool of clinical systems neuroscience, applications of fNIRS to developmental neurology and pediatrics are particularly warranted, considering thin skulls in pediatric population and thereby easy penetration of near-infrared lights to the brain. Assessment of brain activity during walking has been employed to examine pathophysiology of disturbances in patients with walking difficulty 88. Other important application could be application to restless patients as well as critically ill patients. Lastly, conveniences to use fNIRS may bolster wide clinical applications in daily practice. Of note, psychiatrists in Japan have started to apply fNIRS to everyday clinical diagnosis of psychiatric disorders 89.
3.3 Neuro-metabolic-hemodynamic coupling Many neuroimaging methods measure signals derived of hemodynamic changes as a surrogate maker of neural/synaptic activity. It is quite well established that hemodynamic signals faithfully reflect both neural activity and synaptic activity, putting an emphasis on local field potentials (LFP) closely related to synaptic activity 90. For the interpretation of results obtained with the neuroimaging method described above, however, it is important to understand the relationship among
16
neuronal/synaptic activity, metabolic changes and hemodynamic changes (neurometabolic-hemodynamic coupling) occurring at molecular and cellular levels. Neurons maintain a negative membrane potential (-70 mV) at rest. A neuron can generate an action potential on demand, when depolarized above a threshold by EPSP. Both action and synaptic potentials are generated by inflow of cations, primarily sodium. Once depolarized, a neuron needs pump sodium out of the cell to restore the resting membrane potential. Na+/K+ ATPase, depending its energy onto adenosine triphosphate (ATP), achieves this burden. For the generation of ATP, it is believed that the brain almost exclusively depends upon metabolism of glucose. In fact, the brain consumes ~20% of glucose required for the total body. Considering these cascades of events, it is quite reasonable that regional glucose consumption is tightly coupled with local neural and synaptic activities 91. Local neural/synaptic activity also increases cerebral blood flow nearby (Figure 3). Regional cerebral blood flow (rCBF) increases proportionally to local glucose consumption. Although it remains unclear how cerebral blood flow is regulated in response to neural/synaptic activity, at least two possible mechanisms have been proposed to date. A conventional view is that a messenger such as nitric oxide (NO) first relaxes smooth muscles of arterioles and enlarges their diameters, leading to blood flow increases. Capillaries and venules are passively enlarged in response to blood flow increases like a balloon 92. Very recently, however, Hall et al. 93 proposed a different view. According to them, synaptic/neural activitydependent glutamate release first results in relaxation of pericytes of capillaries through the actions of a messenger. A key messenger of this process is assumed to be prostaglandin E2 (PGE2), which has been known as a pivotal mediator in the neuro-immune system and neuro-endocrine system of the brain. The pericytemediated blood flow increase also requires NO to inhibit synthesis of a vasoconstrictor substance (20-hydroxyeicosatetraenoic acid). Capillaries dilate faster than arterioles do, and account for 84% of the blood flow increase. This capillary-first theory may bring a paradigm shift in the understanding of neural/synaptic activity and hemodynamic changes. Apart from precise regulation mechanisms of neuro-vascular coupling, it is established that the blood flow increase in response to neural/synaptic activity exceeds demands for oxygen consumption. This leads to dilution of deoxy-Hb, which is an intrinsic MRI contrast medium representing a level of blood oxygenation. Deoxy-Hb is paramagnetic, and its increases thus decrease homogeneity of local magnetic fields. Therefore, deoxy-Hb can be used as an internal contrast agent reflecting neural activity-dependent blood flow changes, and is called bloodoxygenation level-dependent (BOLD) signals. Synaptic/neuronal activations can be indirectly detected with T2*-weighted MRI, and is often called BOLD fMRI 42.
17
3.4 Neurostimulation 3.4.1. Transcranial magnetic stimulation (TMS) Transcranial magnetic stimulation (TMS) is a widely applied non-invasive brain stimulation technique 94. Before the development of TMS, transcranial electrical stimulation (TES) had been invented to generate a high-voltage electric shock over the primary motor cortex 95. Although TES can produce a brief muscle response, the motor-evoked potential (MEP), TES is quite painful. Instead of directly applying electric current over the scalp, TMS generates a rapidly changing magnetic field by passing a brief electric current through a coil 96. Since a magnetic field easily penetrates through the scalp and skull without attenuation, it induces electric currents (“eddy currents”) in the brain in the direction to cancel out the TMS-generated magnetic filed. The eddy current in the brain thereby activates a group of neurons beneath the coil. When applied over the primary motor cortex, TMS can induce muscle twitching (MEP) with little or no pain. Different shapes of TMS coils differently stimulate the brain. A round coil induces strongest currents near the circumference of the coil, and there is no current at the center of coil. A figure-of-eight-shaped coil produces a more focal current at the intersection of the two round components. A cone-shaped coil increases current at the intersection of the two angled round coils. TMS to the primary motor cortex provides indispensible knowledge in the filed of basic and clinical motor neuroscience. Central conduction time is the time for neural signals to travel from the primary motor cortex to the motor neuron pool in the spinal cord. Subtraction of the peripheral conduction time from the MEP latency yields the central conduction time. Measurement of central conduction time may be useful in clinical diagnosis of movement disturbance. TMS to different parts of the primary motor cortex (“motor cortical homunculus”) induces MEPs in the different parts of the body. Such motor mapping has been proven to be useful in clinical neuroscience, for instance, to show that reorganization of motor somatotopy may underlie pathophysiology of “phantom limb pain” syndrome 97. Motor cortical excitability 94 can be measured through single-pulse TMS to the primary motor cortex: the resting and active motor thresholds, recruitment curve, and silent period. Paired-pulse TMS to the motor cortex allows us to measure short intracortical inhibition (SICI), short intracortical facilitation (ICF), and long intracortical inhibition (LICI). When two coils are placed over the bilateral motor cortices, we can measure transcallosal inhibition (TCI), which is inhibition in the primary motor cortex in one hemisphere by stimulating the primary motor cortex in the opposite hemisphere. TMS is an important tool to study non-motor systems as well. TMS to visual areas may cause visual experiences (phosphenes or moving phosphenes) or disrupt visual experiences (scotoma). For studies of other sensory and cognitive systems, interruption of ongoing neural activity with TMS and resulting changes of percep-
18
tion or cognition, if any, would prove causal relationship between neural activity and perception/cognition 98. TMS is a very important tool to induce short-term plasticity. There are several protocols proposed. Repetitive TMS (rTMS) at low frequencies (0.2-1 Hz) induces suppression of cortical excitability while rTMS at high frequencies (5 Hz or higher) causes an increase in cortical excitability. Theta burst stimulation (TBS) protocols deliver very short, high frequency trains of stimulation at a theta frequency (5 Hz). Intermittent TBS increases cortical excitability while continuous TBS decreases cortical excitability 99. Many studies have applied plasticity-inducing TMS protocols to therapy in Parkinson’s disease 100, dystonia 101, stroke 102, pain syndrome 103, and mood disorders 104. However, rTMS has a potential to cause seizure even in healthy humans although a single-pulse TMS to the brain is considered to be safe. Therefore, clinician and researchers should stick to safety guideline 105.
3.4.2. Transcranial current stimulation
Transcranial direct current stimulation (tDCS) is a non-invasive brain stimulation technique by which weak direct current is delivered to the brain. It has been long known from animal studies that weak direct current by intracerebral or epidural electrodes modulates cortical activity and may induce long-lasting changes in cortical excitability. Typically for tDCS, weak direct current (1-2 mA) is delivered for 10-20 minutes through electrodes placed on the scalp. Brain regions under the anodal electrode show increased cortical excitability (excitatory effects) while those under the cathodal electrode show decreased cortical excitability (inhibitory effects). How the tDCS induces change in cortical excitability is not completely understood; multiple factors such as changes in membrane potentials of neuronal population and release of neurotransmitters 106 may be involved. Since its reemergence 107, however, a number of studies have confirmed modulation of cortical excitability by tDCS and suggested its potential to induce behavioral changes 62,108. A potential to induce neuroplasticity indicates clinically useful plasticity may by induced in patients with disorders such as stroke 109. Other non-invasive electrical stimulation techniques such as transcranial alternating current stimulation (tACS) and transcranial random noise stimulation (tRNS) have been proposed to modulate motor, sensory and cognitive processes 110.
3.4.3 Deep brain stimulation (DBS) As briefly mentioned earlier, DBS to the subthalamic nucleus, globus pallidus or thalamus is a widely accepted for treating movement disorders (advanced Parkinson’s, dystonia and tremor) and, more recently, psychiatric disorders such as depression.
19
4. Clinical Systems Neuroscience: Future Traditionally, tools in clinical systems neuroscience have been used to clarify pathophysiology, or to assist diagnosis, of disorders involving nervous system. However, brain stimulation techniques have long been used to enhance functions of nervous system as a potentially therapeutic tool. One of the most successful cases would be DBS as an established functional therapy for movement disorders. We now see an even stronger and more continuous trend toward application of the systems neuroscience tools to restore, supplement or replace impaired functions in patients with dysfunctions of the nervous system. Such efforts include development of brain-machine interface (BMI), robot-assisted rehabilitation, and neurofeedback. In the near future, these innovative methods will not only furnish treatment options, but also provide completely new strategies, for treatments of patients with nervous system dysfunctions. Another trend would be further development of neuro-IT. Here I focus on two topics, BMI and neurofeedback.
4.1. Brain-machine interface (BMI) Outward BMIs, or brain-computer interfaces (BCIs), are emerging neuro-IT, which ‘decodes’ a behavioral decision from brain activity (“mind reading”) and feeds the decoded information to control IT devices. A typical usage of BMI is to decode patients’ motor intention to translate into control signals to move a robotic hand 111, or a computer cursor 112. Such outward BMI has originally been developed as a neuroprosthesis for e.g. tetraplegic patients with spinal cord injury or amyotrophic lateral sclerosis, or hemiplegic patients with stroke. Inward BMIs intend to replace lost sensation with neuro-IT. Inward BMIs include direct delivery of visual signals from engineered photo-sensors to the visual cortex. ECoG may help development of high-performance BMI, taking advantage of high signal-to-noise ratio 13. Invasiveness of the surgical procedure, long-term safety and stability of implanted ECoG electrodes may still be concerns for immediate clinical applications.
4.2. Neurofeedback Neurofeedback is an emerging area of interest in clinical systems neuroscience. Using EEG or real-time fMRI, information about brain activity is retrieved from the brain of a participant and shown to the participant online as sensory stimuli. Study participant try to change the property of the sensory stimuli linked to the information about his/her own brain activity. It has been shown that experimental
20
animals and humans can do this task. Interestingly, changing brain activity may result in changes in behavior 113. This suggests that neurofeedback may ne utilized as a therapeutic tool 114,115.
5. Concluding remarks Clinical systems neuroscience arises from the mutual interactions between neuroscience and clinical medicine, distinct from unidirectional “translation” from basic research to clinical medicine. The author foresees even greater advances in clinical systems neuroscience in the coming decade. Those changes may completely change how we deal with neuropsychiatric disorders. I therefore encourage young clinicians and researchers to join this exciting activity.
21
Figure. 1. Clinical systems neuroscience. Neuroscience consists of multi-layered subdisciplines with different spatial scales: molecular neuroscience (~nm), cellar neuroscience (~μm), systems neuroscience and cognitive/behavioral neuroscience (~mm-cm). Advance of non-invasive techniques examining functions of the brain allows clinicians to ask meaningful scientific questions easily at systems, cognitive, and behavioral neuroscience levels. Neuroscientists, in turn, can explore a new venue for basic research according to unanswered questions in clinical medicine. Clinical systems neuroscience thus arises from the mutual interactions between neuroscience and clinical medicine, distinct from unidirectional “translation” from basic research to clinical medicine.
22
Figure 2. Three representative images indicating different information obtained with MRI from a healthy volunteer. A gray matter (GM) segmented image from a three-dimensional (3D) T1-weighted image (left). GM images have been used for a voxel-based morphometry (VBM) analysis testing GM volume correlated with a cognitive function or differences in GM volume across groups. Diffusion-weighted image (DWI) measures diffusion of water molecules. Assuming that diffusion of water molecules reflect orientation of fibers in the white matter (WM), DWI can yield, for example, a color map simulating a voxel-wise predominance of WM fiber orientations (red = right-left fiber orientation; green = anterior-posterior; blue = superior-inferior) (middle). A resting-state functional connectivity magnetic resonance imaging (rsfcMRI) shows the medial prefrontal area and posterior cingulate area (red) corresponding to parts of default mode network (DMN) (right). An independent component analysis (ICA) was used to separate different resting-state network (RSN) in an rsfcMRI data set without a priori hypothesis. DMN (red) is overlain onto a gradient-echo echo planar image used for acquisition of rsfcMRI data.
23
Figure 3. Neuro-vascular coupling. Compared with a no-stimulation baseline condition, increased synaptic and neural activities result in releases of neurotransmitter such as glutamate. Glutamate induces vaso-effective messengers including nitric oxide (NO) and prostaglandin E2 (PGE2). PGE2 dilates capillaries through actions on pericytes and NO dilates arterioles through actions on smooth muscles, resulting in increased cerebral blood flow (CBF). Although neuronal activation consumes oxygen, thereby converting oxy-hemoglobin (oxy-Hb) to deoxyhemoglobin (deoxy-Hb), excessive CBF rather dilutes deoxy-Hb. This series of process by which synaptic/neuronal activities give rise to hemodynamic changes is called neurovascular coupling, and is a basis of functional neuroimaging such as functional magnetic resonance imaging (fMRI) and functional near-infrared spectroscopy (fNIRS).
24
References
1
Hanakawa, T. Rostral premotor cortex as a gateway between motor and cognitive networks. Neuroscience research 70, 144-154 (2011).
2
Grush, R. The emulation theory of representation: motor control, imagery, and perception. Behav Brain Sci 27, 377-396; discussion 396-442 (2004).
3
Hanakawa, T. [Multidimensional neuroimaging approach for studying brain network]. Rinsho shinkeigaku = Clinical neurology 50, 901-902 (2010).
4
Broca, P. Nouvelle observation d’aphe´mie produite par une le´sion de la troisie`me circonvolution frontale. Bulletins de la Socie´te´ d’anatomie (Paris) 2e serie 6, 398–407 (1861).
5
Corkin, S. What's new with the amnesic patient H.M.? Nat Rev Neurosci 3, 153-160, doi:10.1038/nrn726 (2002).
6
Damasio, H., Grabowski, T., Frank, R., Galaburda, A. M. & Damasio, A. R. The return of Phineas Gage: clues about the brain from the skull of a famous patient. Science (New York, N.Y 264, 1102-1105 (1994).
7
Dronkers, N. F., Plaisant, O., Iba-Zizen, M. T. & Cabanis, E. A. Paul Broca's historic cases: high resolution MR imaging of the brains of Leborgne and Lelong. Brain 130, 1432-1441, doi:10.1093/brain/awm042 (2007).
8
Omata, K., Hanakawa, T., Morimoto, M. & Honda, M. Spontaneous Slow Fluctuation of EEG Alpha Rhythm Reflects Activity in Deep-Brain Structures: A Simultaneous EEGfMRI Study. PloS one 8, e66869, doi:10.1371/journal.pone.0066869 (2013).
9
Yoshimura, N., Dasalla, C. S., Hanakawa, T., Sato, M. A. & Koike, Y. Reconstruction of flexor and extensor muscle activities from electroencephalography cortical currents. NeuroImage 59, 1324-1337, doi:10.1016/j.neuroimage.2011.08.029 (2012).
10
Michel, C. M. et al. EEG source imaging. Clin Neurophysiol 115, 2195-2222, doi:10.1016/j.clinph.2004.06.001 (2004).
11
Chaudhary, U. J., Duncan, J. S. & Lemieux, L. Mapping hemodynamic correlates of seizures
using
fMRI:
A
review.
Human
brain
mapping
34,
447-466,
doi:10.1002/hbm.21448 (2013). 12
Hanakawa, T. et al. Functional mapping of human medial frontal motor areas. The combined use of functional magnetic resonance imaging and cortical stimulation. Exp Brain Res 138, 403-409 (2001).
13
Yanagisawa, T. et al. Electrocorticographic control of a prosthetic arm in paralyzed patients. Annals of neurology 71, 353-361, doi:10.1002/ana.22613 (2012).
25 14
Kuhn, A. A. et al. High-frequency stimulation of the subthalamic nucleus suppresses oscillatory beta activity in patients with Parkinson's disease in parallel with improvement in motor performance. J Neurosci 28, 6165-6173, doi:10.1523/JNEUROSCI.0282-08.2008 (2008).
15
Rothwell, J. C., Obeso, J. A. & Marsden, C. D. On the significance of giant somatosensory evoked potentials in cortical myoclonus. Journal of neurology, neurosurgery, and psychiatry 47, 33-42 (1984).
16
Farwell, L. A. & Donchin, E. Talking off the top of your head: toward a mental prosthesis utilizing
event-related
brain
potentials.
Electroencephalography
and
clinical
neurophysiology 70, 510-523 (1988). 17
Nagai, T. et al. Mismatch Negativity as a "Translatable" Brain Marker Toward Early Intervention
for
Psychosis:
A
Review.
Frontiers
in
psychiatry
4,
115,
doi:10.3389/fpsyt.2013.00115 (2013). 18
Walter, W. G., Cooper, R., Aldridge, V. J., McCallum, W. C. & Winter, A. L. Contingent Negative Variation: An Electric Sign of Sensorimotor Association and Expectancy in the Human Brain. Nature 203, 380-384 (1964).
19
Hari, R. & Salmelin, R. Magnetoencephalography: From SQUIDs to neuroscience. Neuroimage
20th
anniversary
special
edition.
NeuroImage
61,
386-396,
doi:10.1016/j.neuroimage.2011.11.074 (2012). 20
Cohen, D. Magnetoencephalography: detection of the brain's electrical activity with a superconducting magnetometer. Science (New York, N.Y 175, 664-666 (1972).
21
Lin, F. H., Belliveau, J. W., Dale, A. M. & Hamalainen, M. S. Distributed current estimates using
cortical
orientation
constraints.
Human
brain
mapping
27,
1-13,
doi:10.1002/hbm.20155 (2006). 22
Hillebrand, A., Singh, K. D., Holliday, I. E., Furlong, P. L. & Barnes, G. R. A new approach to neuroimaging with magnetoencephalography. Human brain mapping 25, 199211, doi:10.1002/hbm.20102 (2005).
23
Stufflebeam,
S.
M.,
Tanaka,
N.
&
Ahlfors,
S.
P.
Clinical
applications
of
magnetoencephalography. Human brain mapping 30, 1813-1823, doi:10.1002/hbm.20792 (2009). 24
Mima, T. et al. Pathogenesis of cortical myoclonus studied by magnetoencephalography. Annals of neurology 43, 598-607, doi:10.1002/ana.410430507 (1998).
25
Hanakawa, T. et al. The role of rostral Brodmann area 6 in mental-operation tasks: an integrative neuroimaging approach. Cereb Cortex 12, 1157-1170 (2002).
26
Hanakawa, T. et al. Functional properties of brain areas associated with motor execution and imagery. Journal of neurophysiology 89, 989-1002 (2003).
26 27
Hanakawa, T. et al. Stimulus-response profile during single-pulse transcranial magnetic stimulation
to
the
primary
motor
cortex.
Cereb
Cortex
19,
2605-2615,
doi:10.1093/cercor/bhp013 (2009). 28
Shitara, H., Shinozaki, T., Takagishi, K., Honda, M. & Hanakawa, T. Time course and spatial distribution of fMRI signal changes during single-pulse transcranial magnetic stimulation
to
the
primary
motor
cortex.
NeuroImage
56,
1469-1479,
doi:10.1016/j.neuroimage.2011.03.011 (2011). 29
Shitara, H., Shinozaki, T., Takagishi, K., Honda, M. & Hanakawa, T. Movement and afferent representations in human motor areas: a simultaneous neuroimaging and transcranial magnetic/peripheral nerve-stimulation study. Frontiers in human neuroscience 7, 554, doi:10.3389/fnhum.2013.00554 (2013).
30
Talairach, J. & Tournoux, P. Co-planar stereotaxic atlas of the human brain. (Thieme Medical Publishers, 1988).
31
Dale, A. M., Fischl, B. & Sereno, M. I. Cortical surface-based analysis. I. Segmentation and surface reconstruction. NeuroImage 9, 179-194, doi:10.1006/nimg.1998.0395 (1999).
32
Fischl, B., Sereno, M. I. & Dale, A. M. Cortical surface-based analysis. II: Inflation, flattening,
and
a
surface-based
coordinate
system.
NeuroImage
9,
195-207,
doi:10.1006/nimg.1998.0396 (1999). 33
Hosoda, C., Tanaka, K., Nariai, T., Honda, M. & Hanakawa, T. Dynamic Neural Network Reorganization Associated with Second Language Vocabulary Acquisition: A Multimodal Imaging Study. J Neurosci 33, 13663-13672, doi:10.1523/JNEUROSCI.0410-13.2013 (2013).
34
Tanaka, S. et al. Larger right posterior parietal volume in action video game experts: a behavioral and voxel-based morphometry (VBM) study. PloS one 8, e66998, doi:10.1371/journal.pone.0066998 (2013).
35
Namiki, C. et al. Small orbitofrontal traumatic lesions detected by high resolution MRI in a patient
with
major
behavioural
changes.
Neurocase
14,
474-479,
doi:10.1080/13554790802459494 (2008). 36
Yamada, M. et al. Social cognition and frontal lobe pathology in schizophrenia: a voxelbased
morphometric
study.
NeuroImage
35,
292-298,
doi:10.1016/j.neuroimage.2006.10.046 (2007). 37
Garraux, G. et al. Changes in brain anatomy in focal hand dystonia. Annals of neurology 55, 736-739 (2004).
38
Garraux, G. et al. Increased midbrain gray matter in Tourette's syndrome. Annals of neurology 59, 381-385 (2006).
39
Moseley, M. E. et al. Early detection of regional cerebral ischemia in cats: comparison of diffusion- and T2-weighted MRI and spectroscopy. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine 14, 330-346 (1990).
27 40
Matsumoto, R. et al. Hemispheric asymmetry of the arcuate fasciculus: a preliminary diffusion tensor tractography study in patients with unilateral language dominance defined by Wada test. Journal of neurology 255, 1703-1711, doi:10.1007/s00415-008-0005-9 (2008).
41
Okada, T. et al. Corticospinal tract localization: integration of diffusion-tensor tractography at 3-T MR imaging with intraoperative white matter stimulation mapping--preliminary results. Radiology 240, 849-857, doi:10.1148/radiol.2403050916 (2006).
42
Ogawa, S. et al. Intrinsic signal changes accompanying sensory stimulation: functional brain mapping with magnetic resonance imaging. Proc Natl Acad Sci U S A 89, 5951-5955 (1992).
43
Friston, K. J. et al. Psychophysiological and modulatory interactions in neuroimaging. NeuroImage 6, 218-229, doi:10.1006/nimg.1997.0291 (1997).
44
Abe, M. et al. Functional coupling of human prefrontal and premotor areas during cognitive manipulation. J Neurosci 27, 3429-3438 (2007).
45
Aso, K., Hanakawa, T., Aso, T. & Fukuyama, H. Cerebro-cerebellar interactions underlying temporal
information
processing.
J
Cogn
Neurosci
22,
2913-2925,
doi:10.1162/jocn.2010.21429 (2010). 46
Friston, K. J., Harrison, L. & Penny, W. Dynamic causal modelling. NeuroImage 19, 12731302 (2003).
47
Ban, H. et al. Topographic representation of an occluded object and the effects of spatiotemporal context in human early visual areas. J Neurosci 33, 16992-17007, doi:10.1523/JNEUROSCI.1455-12.2013 (2013).
48
Yamamoto, T. et al. Neural correlates of the stereokinetic effect revealed by functional magnetic resonance imaging. Journal of vision 8, 14 11-17, doi:10.1167/8.10.14 (2008).
49
Aso, T. et al. Subregions of human parietal cortex selectively encoding object orientation. Neuroscience letters 415, 225-230 (2007).
50
Sawamoto, N. et al. Expectation of pain enhances responses to nonpainful somatosensory stimulation in the anterior cingulate cortex and parietal Operculum/Posterior insula: an event-related functional magnetic resonance imaging study. J Neurosci 20, 7438-7445 (2000).
51
Toma, K. et al. Activities of the primary and supplementary motor areas increase in preparation and execution of voluntary muscle relaxation: An event- related fMRI study. J Neurosci 19, 3527-3534 (1999).
52
Hanakawa, T., Honda, M., Zito, G., Dimyan, M. A. & Hallett, M. Brain activity during visuomotor behavior triggered by arbitrary and spatially constrained cues: an fMRI study in humans. Exp Brain Res 172, 275-282 (2006).
28 53
Hanakawa, T., Dimyan, M. A. & Hallett, M. The representation of blinking movement in cingulate motor areas: a functional magnetic resonance imaging study. Cereb Cortex 18, 930-937 (2008).
54
Nakamura, K. et al. Modulation of the visual word retrieval system in writing: a functional MRI study on the Japanese orthographies. J Cogn Neurosci 14, 104-115 (2002).
55
Nakamura, K. et al. Participation of the left posterior inferior temporal cortex in writing and mental recall of kanji orthography: A functional MRI study. Brain 123, 954-967 (2000).
56
Crinion, J. et al. Language control in the bilingual brain. Science (New York, N.Y 312, 15371540 (2006).
57
Hosoda, C., Hanakawa, T., Nariai, T., Ohno, K. & Honda, M. Neural mechanism of language switch. J Neurolinguistics 25, 44-61 (2011).
58
Hanakawa, T., Dimyan, M. A. & Hallett, M. Motor planning, imagery, and execution in the distributed motor network: a time-course study with functional MRI. Cereb Cortex 18, 2775-2788 (2008).
59
Nakamura, K. et al. Attentional modulation of parieto-occipital cortical responses: implications for hemispatial neglect. J Neurol Sci 176, 136-143 (2000).
60
Hanakawa, T., Honda, M., Okada, T., Fukuyama, H. & Shibasaki, H. Differential activity in the premotor cortex subdivisions in humans during mental calculation and verbal rehearsal tasks: a functional magnetic resonance imaging study. Neuroscience letters 347, 199-201 (2003).
61
Hanakawa, T., Honda, M., Okada, T., Fukuyama, H. & Shibasaki, H. Neural correlates underlying mental calculation in abacus experts: a functional magnetic resonance imaging study. NeuroImage 19, 296-307 (2003).
62
Kasahara, K., Tanaka, S., Hanakawa, T., Senoo, A. & Honda, M. Lateralization of activity in the parietal cortex predicts the effectiveness of bilateral transcranial direct current stimulation on performance of a mental calculation task. Neuroscience letters 545, 86-90, doi:10.1016/j.neulet.2013.04.022 (2013).
63
Fukui, H., Murai, T., Fukuyama, H., Hayashi, T. & Hanakawa, T. Functional activity related to risk anticipation during performance of the Iowa Gambling Task. NeuroImage 24, 253-259, doi:10.1016/j.neuroimage.2004.08.028 (2005).
64
Shinozaki, J., Hanakawa, T. & Fukuyama, H. Heterospecific and conspecific social cognition in the anterior cingulate cortex. Neuroreport 18, 993-997 (2007).
65
Fukui, H. et al. The neural basis of social tactics: An fMRI study. NeuroImage 32, 913-920, doi:10.1016/j.neuroimage.2006.03.039 (2006).
66
Tanaka, S. et al. Abacus in the brain: a longitudinal functional MRI study of a skilled abacus user with a right hemispheric lesion. Frontiers in psychology 3, 315, doi:10.3389/fpsyg.2012.00315 (2012).
29 67
Oga, T. et al. Abnormal cortical mechanisms of voluntary muscle relaxation in patients with writer's cramp: an fMRI study. Brain 125, 895-903 (2002).
68
Bohlhalter, S. et al. Neural correlates of tic generation in Tourette syndrome: an eventrelated functional MRI study. Brain 129, 2029-2037 (2006).
69
Fox, M. D. & Raichle, M. E. Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging. Nat Rev Neurosci 8, 700-711 (2007).
70
Smith, S. M. et al. Functional connectomics from resting-state fMRI. Trends Cogn Sci 17, 666-682, doi:10.1016/j.tics.2013.09.016 (2013).
71
Raichle, M. E. et al. A default mode of brain function. Proc Natl Acad Sci U S A 98, 676682 (2001).
72
Greicius, M. Resting-state functional connectivity in neuropsychiatric disorders. Current opinion in neurology 21, 424-430 (2008).
73
Takaya, S. et al. Prefrontal hypofunction in patients with intractable mesial temporal lobe epilepsy. Neurology 67, 1674-1676 (2006).
74
Weiner, M. W. et al. The Alzheimer's disease neuroimaging initiative: progress report and future plans. Alzheimer's & dementia : the journal of the Alzheimer's Association 6, 202211 e207 (2010).
75
Lerner, A. et al. Involvement of insula and cingulate cortices in control and suppression of natural urges. Cereb Cortex 19, 218-223, doi:10.1093/cercor/bhn074 (2009).
76
Lerner, A. et al. Neuroimaging of neuronal circuits involved in tic generation in patients with
Tourette
syndrome.
Neurology
68,
1979-1987,
doi:10.1212/01.wnl.0000264417.18604.12 (2007). 77
Lerner, A. et al. Regional cerebral blood flow correlates of the severity of writer's cramp symptoms. NeuroImage 21, 904-913 (2004).
78
Sawamoto, N. et al. Cognitive slowing in Parkinson disease is accompanied by hypofunctioning of the striatum. Neurology 68, 1062-1068 (2007).
79
Sawamoto, N. et al. Cognitive deficits and striato-frontal dopamine release in Parkinson's disease. Brain 131, 1294-1302 (2008).
80
Hanakawa, T. et al. Mechanisms underlying gait disturbance in Parkison's disease: A single photon emission computed tomography study. Brain 122, 1271-1282 (1999).
81
Hanakawa, T., Fukuyama, H., Katsumi, Y., Honda, M. & Shibasaki, H. Enhanced lateral premotor activity during paradoxical gait in Parkinson's disease. Annals of neurology 45, 329-336 (1999).
82
Iseki, K. et al. Gait disturbance associated with white matter changes: a gait analysis and blood flow study. NeuroImage 49, 1659-1666, doi:S1053-8119(09)01002-7 [pii]
10.1016/j.neuroimage.2009.09.023 (2010).
30 83
Sato, H. et al. A NIRS-fMRI investigation of prefrontal cortex activity during a working memory task. NeuroImage 83, 158-173, doi:10.1016/j.neuroimage.2013.06.043 (2013).
84
Takahashi, T. et al. Influence of skin blood flow on near-infrared spectroscopy signals measured on the forehead during a verbal fluency task. NeuroImage 57, 991-1002, doi:10.1016/j.neuroimage.2011.05.012 (2011).
85
Kirilina, E. et al. The physiological origin of task-evoked systemic artefacts in functional near infrared spectroscopy. NeuroImage 61, 70-81, doi:10.1016/j.neuroimage.2012.02.074 (2012).
86
Taga, G., Asakawa, K., Maki, A., Konishi, Y. & Koizumi, H. Brain imaging in awake infants by near-infrared optical topography. Proc Natl Acad Sci U S A 100, 10722-10727, doi:10.1073/pnas.1932552100 (2003).
87
Miyai, I. et al. Cortical mapping of gait in humans: a near-infrared spectroscopic topography study. NeuroImage 14, 1186-1192 (2001).
88
Miyai, I. et al. Premotor cortex is involved in restoration of gait in stroke. Annals of neurology 52, 188-194 (2002).
89
Takizawa, R. et al. Neuroimaging-aided differential diagnosis of the depressive state. NeuroImage 85 Pt 1, 498-507, doi:10.1016/j.neuroimage.2013.05.126 (2014).
90
Logothetis, N. K., Pauls, J., Augath, M., Trinath, T. & Oeltermann, A. Neurophysiological investigation of the basis of the fMRI signal. Nature 412, 150-157. (2001).
91
Heeger, D. J. & Ress, D. What does fMRI tell us about neuronal activity? Nat Rev Neurosci 3, 142-151, doi:10.1038/nrn730 (2002).
92
Buxton, R. B., Wong, E. C. & Frank, L. R. Dynamics of blood flow and oxygenation changes during brain activation: the balloon model. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine 39, 855-864 (1998).
93
Hall, C. N. et al. Capillary pericytes regulate cerebral blood flow in health and disease. Nature 508, 55-60, doi:10.1038/nature13165 (2014).
94
Hallett, M. Transcranial magnetic stimulation: a primer. Neuron 55, 187-199, doi:10.1016/j.neuron.2007.06.026 (2007).
95
Merton, P. A. & Morton, H. B. Stimulation of the cerebral cortex in the intact human subject. Nature 285, 227 (1980).
96
Barker, A. T., Jalinous, R. & Freeston, I. L. Non-invasive magnetic stimulation of human motor cortex. Lancet 1, 1106-1107 (1985).
97
Hanakawa, T. Neural mechanisms underlying deafferentation pain: a hypothesis from a neuroimaging perspective. Journal of orthopaedic science : official journal of the Japanese Orthopaedic Association 17, 331-335, doi:10.1007/s00776-012-0209-9 (2012).
31 98
Oshio, R. et al. Differential effect of double-pulse TMS applied to dorsal premotor cortex and precuneus during internal operation of visuospatial information. NeuroImage 49, 11081115, doi:S1053-8119(09)00806-4 [pii]
10.1016/j.neuroimage.2009.07.034 (2010). 99
Huang, Y. Z., Edwards, M. J., Rounis, E., Bhatia, K. P. & Rothwell, J. C. Theta burst stimulation of the human motor cortex. Neuron 45, 201-206 (2005).
100
Fregni, F., Simon, D. K., Wu, A. & Pascual-Leone, A. Non-invasive brain stimulation for Parkinson's disease: a systematic review and meta-analysis of the literature. Journal of neurology, neurosurgery, and psychiatry 76, 1614-1623, doi:10.1136/jnnp.2005.069849 (2005).
101
Murase, N. et al. Subthreshold low-frequency repetitive transcranial magnetic stimulation over
the
premotor
cortex
modulates
writer's
cramp.
Brain
128,
104-115,
doi:10.1093/brain/awh315 (2005). 102
Kim, Y. H. et al. Repetitive transcranial magnetic stimulation-induced corticomotor excitability and associated motor skill acquisition in chronic stroke. Stroke; a journal of cerebral circulation 37, 1471-1476, doi:10.1161/01.STR.0000221233.55497.51 (2006).
103
Lefaucheur, J. P., Drouot, X., Keravel, Y. & Nguyen, J. P. Pain relief induced by repetitive transcranial magnetic stimulation of precentral cortex. Neuroreport 12, 2963-2965 (2001).
104
Herrmann, L. L. & Ebmeier, K. P. Factors modifying the efficacy of transcranial magnetic stimulation in the treatment of depression: a review. The Journal of clinical psychiatry 67, 1870-1876 (2006).
105
Rossi, S., Hallett, M., Rossini, P. M., Pascual-Leone, A. & Safety of, T. M. S. C. G. Safety, ethical considerations, and application guidelines for the use of transcranial magnetic stimulation in clinical practice and research. Clin Neurophysiol 120, 2008-2039, doi:10.1016/j.clinph.2009.08.016 (2009).
106
Tanaka, T. et al. Transcranial direct-current stimulation increases extracellular dopamine levels
in
the
rat
striatum.
Frontiers
in
systems
neuroscience
7,
6,
doi:10.3389/fnsys.2013.00006 (2013). 107
Nitsche, M. A. & Paulus, W. Excitability changes induced in the human motor cortex by weak transcranial direct current stimulation. The Journal of physiology 527 Pt 3, 633-639 (2000).
108
Tanaka, S., Hanakawa, T., Honda, M. & Watanabe, K. Enhancement of pinch force in the lower leg by anodal transcranial direct current stimulation. Exp Brain Res 196, 459-465, doi:10.1007/s00221-009-1863-9 (2009).
109
Tanaka, S. et al. Single session of transcranial direct current stimulation transiently increases knee extensor force in patients with hemiparetic stroke. Neurorehabilitation and neural repair 25, 565-569, doi:10.1177/1545968311402091 (2011).
32 110
Guleyupoglu, B., Schestatsky, P., Edwards, D., Fregni, F. & Bikson, M. Classification of methods in transcranial electrical stimulation (tES) and evolving strategy from historical approaches
to
contemporary
innovations.
J
Neurosci
Methods
219,
297-311,
doi:10.1016/j.jneumeth.2013.07.016 (2013). 111
Hochberg, L. R. et al. Reach and grasp by people with tetraplegia using a neurally controlled robotic arm. Nature 485, 372-375, doi:10.1038/nature11076 (2012).
112
Wolpaw, J. R. & McFarland, D. J. Control of a two-dimensional movement signal by a noninvasive brain-computer interface in humans. Proc Natl Acad Sci U S A 101, 1784917854, doi:10.1073/pnas.0403504101 (2004).
113
Shibata, K., Watanabe, T., Sasaki, Y. & Kawato, M. Perceptual learning incepted by decoded fMRI neurofeedback without stimulus presentation. Science (New York, N.Y 334, 1413-1415, doi:10.1126/science.1212003 (2011).
114
deCharms, R. C. et al. Control over brain activation and pain learned by using real-time functional
MRI.
Proc
Natl
Acad
Sci
U
S
A
102,
18626-18631,
doi:10.1073/pnas.0505210102 (2005). 115
Subramanian, L. et al. Real-time functional magnetic resonance imaging neurofeedback for treatment
of
Parkinson's
disease.
doi:10.1523/JNEUROSCI.3498-11.2011 (2011).
J
Neurosci
31,
16309-16317,
