SemOpenAlex |

SemOpenAlex

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2313595323> ?p ?o ?g. }

Showing items 1 to 100 of ±105 with 100 items per page.

W2313595323 endingPage "109" @default.
W2313595323 startingPage "102" @default.
W2313595323 abstract "An Integrationist Approach to Studying the Brainstem and Spinal Cord Investigating the function of the human brainstem and spinal cord is an inherently difficult task, with its proximal region surrounded by the skull base and its longitudinal extent circumscribed by the bony spinal column, cartilaginous disks, and ligaments. Invasive animal and tissue-based experiments have formed our understanding of the electrophysiological basis of spinal cord function, and the discipline of psychophysics has extended this knowledge to include the perceptual effect of physical stimuli we experience from the environment. After all, in broad terms, the spinal cord can be considered the first stop for incoming sensory signals from the periphery and the final stop within the central nervous system (CNS) for motor signals originating from higher cortical and subcortical centers in the brain. In this narrative review, we consider the use of spinal functional magnetic resonance imaging (fMRI) to investigate the function of the human brainstem and spinal cord. Take, for example, the experience of holding a mobile phone while it is vibrating in response to an incoming call. In the neuroscience community, we distinguish between the word flutter, used to describe the sensation of an oscillating stimulus in the range of 5 to 40 Hz, and the word vibration, used to describe the sensation of an oscillating stimulus at frequencies > 60 Hz.1 Through an elegant psychophysics experiment comparing human perception with mechanoreceptive afferent response patterns of the monkey hand, it was established that flutter can be accurately localized to an area of skin, whereas vibration is only poorly localized to deep tissues; through a series of experimental inquiries, the sensation of flutter was found to be transmitted via rapidly adapting sensitive mechanoreceptors, including fast-adapting (FA) type I receptors (the Meissner corpuscle) in the glabrous skin and hair follicle and field receptors in the hairy skin. Vibration, on the other hand, is transmitted via the FA II receptors (Pacinian corpuscle). Afferent fibers of FA I and hair follicle receptors terminate in lamina III and IV of the dorsal horn and form a continuous rostrocaudal column in the spinal cord.2 Primary afferent fibers from FA II endings terminate in the medial aspect of lamina III-IV and form a series of elliptical zones along the rostral-caudal extent of the dorsal horn.3 Several parallel ascending pathways transmit flutter-vibration information to the brain. These include direct projections through the fasciculi gracilis and cuneatus to the dorsal column nuclei; hair follicle and FA I afferents activate the spinocervical tract and the spinothalamic tract, whereas FA II receptors do not.4-6 As one can glean from these electrophysiological and tracer studies, the flow of afferent information from a vibrating mobile phone to the spinal cord and on to higher centers in the brain is quite complex. An extension of these detailed investigations to include all regions of the spinal cord and brainstem is simply not possible with invasive point-based measurements. If we desire to understand the brainstem and spinal cord as a functional unit, rather than the function of specific cell types or anatomic regions such as the dorsal horn or dorsal column nuclei, we have to turn to methodologies with an appropriate spatial-temporal profile that allows simultaneous data collection over the entire brainstem-spinal cord axis. In doing so, the aim will be to build on our existing knowledge of the structure and function of the human spinal cord with an emphasis on integrative function (or dysfunction). Imaging science has decisively established functional segregation as a principle of CNS organization and in recent years has moved toward improving our understanding of functional integration within the CNS.7 To state this another way, the primary motor cortex is both functionally and spatially discrete from the facial nucleus in the brainstem, yet both work in a concerted effort when an individual recognizes an old friend and reaches out with a smile to embrace him or her. How spatially distinct regions are integrated to perform specific functions remains somewhat elusive. Aside from physical axons that might join 2 distinct nuclei or brain regions, a concept that dates back to invasive tracer methods and is now studied with modern diffusion tensor imaging methods,8 understanding the integration of distant regions of the CNS to perform a common task offers complex challenges. Modern imaging techniques in neuroscience aim to take advantage of the unique spatial-temporal scale offered, for example, by simultaneous fMRI of the brain, brainstem, and spinal cord. In doing so, these methods have the ability to capture spatially isolated changes in different regions of the CNS. Using advanced postprocessing methods, we are able to learn how different regions of the CNS function as a unit to perform the functions of life such as respiration, movement, and the perception of sensation. With a spatial scale on the order of cubic millimeters and a temporal scale on the order of seconds, fMRI experiments should not be thought of as providing low-resolution evidence of neurophysiological events. For example, we know from electrophysiology studies that when the skin of an animal is warmed, cells contained in the dorsal horn of the spinal cord at the appropriate segmental level respond by firing action potentials to regions in the brainstem, thalamus, and eventually the somatosensory cortex. Imaging experiments are not required to build on this body of evidence. Rather, such imaging tools should take advantage of their unique spatial-temporal scale to provide novel insights into how different regions of the brainstem and spinal cord function as a unit. In a recent systematic review, we summarized the first 20 years of spinal fMRI research by carefully examining the elements of experimental design in healthy control subjects.9 Specifically, we examined the different tasks or stimulation protocols used to elicit spinal cord function and the contrast mechanisms used to capture functional changes in the spinal cord (signal enhancement by extravascular protons vs blood oxygen-level dependent [BOLD] contrast). Such investigations have provided a platform by which to launch studies aimed at uncovering physiological changes associated with pathological processes such as spinal cord injury (SCI). For example, we have recently highlighted the ability of spinal fMRI to uncover evidence of spinal cord plasticity in incomplete SCI patients who went on to recover from their injury.10 In this review, we begin by outlining the neurobiological processes that underlie the hemodynamic response function, the ultimate source of the fluctuating fMRI signal during experiments designed either at rest or in response to a particular task or stimulus. We then discuss how this time course data can be used to reach conclusions about the function of underlying regions of the brainstem and spinal cord. One way to do this is by considering the brainstem and spinal cord as a network that can be visualized or modeled with 1 of 2 broad approaches: functional connectivity, a tool that can be used to generate phenotypes of neural response patterns, and effective connectivity, a tool that can be used to study hypotheses about how the brainstem and spinal cord interact with each other. To conclude, we discuss the limitations of each of these types of connectivity analyses and what it means for the study of the brainstem--spinal cord as a functional network. WHAT GENERATES THE FMRI SIGNAL? Neurovascular Coupling and Functional Hyperemia Current fMRI technique is based on a local vascular response to an increase in neuronal activity. Given that the deoxygenated blood has paramagnetic properties, its transient concentration change produces an MRI signal increase. This technique is called BOLD.11 How the CNS maintains optimal perfusion of the brain and spinal cord in different metabolic states remains to be determined. As a case in point, imagine the resting spinal cord, perfused with a baseline level of blood flow, having to suddenly respond to increased afferent information. Using the example illustrated above, imagine holding a mobile phone that suddenly vibrates in response to an incoming call. When this information is received, in the form of action potentials and increased synaptic activity at the segmental dorsal horn, the metabolic requirements of these cells increase and require a concomitant increase in oxygen and glucose to maintain function. As others have pointed out, the ratio of fractional changes in blood flow to the local metabolic rate of oxygen consumption is not only different in different regions of the CNS but also of critical importance to understanding the magnitude of fMRI signal changes in response to specific stimuli.12 Given that the brain and spinal cord have little capacity to store energy and that they must meet metabolic requirements for complex processing in a matter of seconds, certain mechanisms are in place to ensure the continual processing of neural signals. Although the details of these mechanisms are under intense study on specific regions of the brain (with no studies completed on the spinal cord), 2 broad terms are generally agreed on in the literature: Neurovascular coupling describes the signaling events between neurons and the surrounding milieu of astrocytes, pericytes, and microvascular networks, and functional hyperemia describes the resultant increase in blood flow to meet the increased metabolic requirements of regional neural processing. The reason for the distinction between neurovascular coupling and functional hyperemia rests in the notion that different regions of the CNS may use unique signaling mechanisms to call on increased blood flow to meet the metabolic demands of that specific region. For example, although it is agreed on that both the cerebral cortex and the cerebellum have the capacity to recruit blood flow to meet increased metabolic requirements (functional hyperemia), there is evidence that these regions have unique neurovascular coupling mechanisms whereby the cerebellum, a phylogenetically older portion of the CNS, uses nitric oxide as a direct mediator between neurons and blood vessels, and the newer cerebral cortex uses nitric oxide as a mediator of neuron-to-vessel signaling by modulating astrocyte signaling pathways that dilate and constrict blood vessels.13 Furthermore, although neurovascular coupling within the cortex and cerebellum involves nitric oxide, there is evidence that brainstem nuclei use adenosine.14 To date, there have been no studies into the mechanisms of either neurovascular coupling or functional hyperemia on the human spinal cord. Communicating Energy Requirements A number of specific signaling molecules have been proposed to communicate the metabolic requirements of neural tissue and the need for increased or decreased blood flow to meet local oxygen and glucose requirements. Whether such signaling pathways act through a negative feedback loop (also known as the metabolic hypothesis), in which changes in neuronal activity drive changes in metabolism, which in turn drive vasoconstriction/vasodilation, or through a feed-forward loop (also known as the neurogenic hypothesis), whereby neuronal activity directly drives vasoconstriction/vasodilation, is open for debate.15-17 Regardless of where the balance lies, and there is ample evidence for both arguments in the literature, a number of signaling mechanisms have been established. Acting through far-reaching projections are neurons of the nucleus basalis magnocellularis (vasodilation effect of acetylcholine), dorsal raphe nucleus (vasoconstriction effect of serotonin), and rostral ventrolateral medulla in which oxygen-sensitive neurons influence cortical blood flow through thalamic intermediates and the modulatory actions of acetylcholine, serotonin, and noradrenergic input from the locus coeruleus.15,16,18,19 On a finite spatiotemporal scale, subclasses of cortical inhibitory interneurons have been found to act directly on smooth muscle of vascular arterioles whereby certain transmitters relax smooth muscle to dilate vessels (nitric oxide and vasoactive intestinal peptide) and others constrict smooth muscle (somatostatin and neuropeptide Y) in a concerted effort to regulate blood flow in response to activity within these networks.20 There is a rather large body of evidence linking the action of excitatory neurons acting through an astrocyte intermediate to the regulation of blood flow. Commonly referred to as the neurovascular unit, the neuron-astrocyte-arteriole complex acts in the following fashion: Excitatory glutamate leads to an increase in intracellular astrocytic calcium levels, initiating the conversion of arachidonic acid to either prostaglandin E and epoxyeicosatrienoic acid or 20-hydroxyeicosatetraenoic acid, resulting in a net vessel dilation or constriction, respectively.21,22 The identification and characteristics of these signaling molecules provide a strong foundation to establish how these pathways vary in different regions of the CNS and ultimately how they function to ensure adequate blood supply during times of need. The concepts outlined here are summarized in Figure 1.FIGURE 1: Energy supply, energy use, and blood flow regulation in the brain. A, adenosine triphosphate (ATP) is generated from glycolysis and mitochondrial oxidative phosphorylation in neurons and glia. ATP is consumed (red arrows) mainly by ion pumping in neurons to maintain the ion gradients underlying synaptic and action potentials after Na+ entry (blue arrows) through ionotropic glutamate receptors (iGluR) and voltage-gated Na+ channels (NaV). It is also used in glia for Na+-coupled neurotransmitter uptake by excitatory amino acid transporters (EAATs) and for metabolic processing (shown for conversion of glutamate to glutamine) and on maintaining the resting potentials of the cells. B, the negative-feedback control hypothesis for vascular energy supply in which a decrease in energy level induces an increased cerebral blood flow (CBF). C, the feed-forward regulation hypothesis for vascular energy supply. Reproduced with permission from Atwell et al.13FUNCTIONAL CONNECTIVITY Using Correlations to Define Phenotypes With a better understanding of how the fMRI signal is generated, we turn to the utility of capturing this signal: How can we use this information to understand the function of the human spinal cord in a way that ultimately improves our care of individuals afflicted with disease or traumatic injury to the spinal cord? The word phenotype is classically used to describe an observable characteristic of an organism. When the observable characteristic is viewed through the lens of an fMRI experimental design and data analysis tools, the word loses some of its inherent meaning. The acquired spinal cord and brainstem fMRI signal is surely dependent on the genetic makeup of the individual and perhaps even the environmental upbringing of that individual, but interpretation of this signal is heavily weighted toward the methods used for data acquisition and analysis. Therefore, a spinal imaging phenotype must be defined in terms of the signal acquisition parameters, the stimulus or resting-state design, and the data analysis methods and can be visualized in terms of correlations between different regions of this distributed system (Figure 2).FIGURE 2: The classic definition of phenotype describes the physical appearance of a trait based on both the genetic makeup and the environmental conditions. Left, a photograph of the garden pea plant, Pisum sativum, used by Mendel in his classic genetic experiments. The color of the pea plant flower (phenotype) is a result of the genetic makeup of the parent plants. Right, a correlation analysis of spinal functional magnetic resonance imaging (fMRI) data in which the left C5 dermatome was heated to 44°C and a correlation analysis was conducted between the C5 spinal cord dorsal horn segment and all other regions of the brainstem and spinal cord in this healthy 31-year-old woman. This spinal fMRI phenotype, although distantly related to the genetic makeup of the individual, is more dependent on the methods used to acquire and analyze the spinal fMRI time-series data. We gratefully acknowledge Dr Patrick Stroman for assistance in creating the spinal fMRI phenotype figure illustrated on the right.Defining the Phenotype Functional connectivity is, in general terms, defined as the temporal correlation between spatially remote neurophysiological events. This is expressed as a deviation from statistical independence across such events. Functional connectivity can refer to correlations across subjects, different trials, or time points. With the broad range of potential targets for a correlation analysis, the definition of functional connectivity can be confusing.23,24 Functional connectivity in the resting state has taken on a meaning of its own and is reviewed in a very accessible manner here.25 Functional connectivity can also be applied to stimulus- or task-based studies in which inter-regional correlations can be linked to events that may be associated with the task such as underlying genetics or personality.26 Inter-regional correlations have been applied in the setting of SCI in which the number of interspinal connections (correlations between a seed region and other regions of the spinal cord; in this reference, the seed region is the dorsal quadrant of the spinal segment of the dermatome stimulated) was associated with the underlying degree of injury (American Spinal Injury Association grade).10 Defining a spinal fMRI phenotype on the basis of a functional connectivity analysis requires consideration and declaration of 4 particular domains of the experimental design: the stimulation (or resting) paradigm, fMRI data acquisition, and preprocessing and processing steps (see the Table).TABLE: Characteristics of a Spinal Functional Magnetic Resonance Imaging Phenotype Based on a Functional Connectivity AnalysisaFirst and foremost, the objectives of the study must be clearly defined and in line with the experimental paradigm (stimulus or task based or resting state). From a spinal cord neuroscience point of view, this is perhaps the most challenging portion of defining a phenotype. In a task-based paradigm, consideration must be given to the type of stimulus and its intensity, duration, and interstimulus interval. In addition, one must consider the effect of the peripheral nervous system in transmitting the stimulus to the spinal cord and whether effects such as habituation may play a role.27,28 Task-based experimental designs should consider the muscle groups involved in the task and the ventral spinal roots that subserve those muscles. As pointed out in an earlier review, the effect of sensory stimulation during a motor task should be considered.9 In a stimulus-based, task-based, or resting-state design, the number of data points collected should be considered, and predetermined scan times should be adhered to on the basis of sample size and power calculations. One previously used approach estimates the number of time points necessary in an individual fMRI data set by considering the temporal signal-to-noise ratio, the desired significance, and the effect size.10,29 Referenced here is an example of a power calculation for a spinal cord fMRI investigation on the BOLD signal responses to controlled hypercapnia in humans.30 Correlations of fMRI time-series data between spatially distinct regions in the human brainstem and spinal cord can also be influenced by the fMRI data acquisition and the preprocessing and processing steps. Although a detailed review of each of these steps is beyond the scope of this work, we highlight a few of the most important concepts here and refer the interested reader to other more in-depth references.9,30-34 Data acquisition can take the form of a gradient echo planar imaging sequence (most commonly used for the brain), implying a reliance on the BOLD hemodynamic response function; a spin echo planar imaging sequence, implying a reliance on the BOLD effect; or a regular spin echo sequence, implying the signal enhancement by extravascular protons contrast mechanism.27,33,35,36 It is important to recall that all imaging acquisition methods aim to capture the signal generated by a neurophysiological event. Advantages of gradient echo planar imaging sequences include faster acquisition of the brainstem--spinal cord volume and high sensitivity to the BOLD contrast. However, they suffer from signal dropout and image distortions caused by susceptibility artifacts, which are particularly prominent in the spinal region as a result of the presence of multiple tissue structures with different magnetic susceptibilities (eg, spinal cord, cerebrospinal fluid, bones, cartilaginous disks, fat, muscles). Spin echo planar imaging sequences reduce signal drop, whereas regular spin echo sequences reduce both dropout and distortions. Data preprocessing (identification of segmental spinal cord levels, motion correction, temporal filtering) and processing (statistical analysis) steps have the potential to greatly influence the observed phenotype. Emphasis should therefore be placed on carefully defining each of the outlined elements of a spinal fMRI phenotype with the explicit understanding that even seemingly subtle variations in these elements may result in a different expression of the phenotype. Visualizing and Comparing Phenotypes With the spinal fMRI phenotype very clearly defined, we are left with a rather complex 4-dimensional data matrix in which voxels represent specific 3-dimensional locations with the brainstem--spinal cord axis and the intensity of the voxel changes throughout the time domain. Transforming a spinal fMRI time-series data set into a tool that can be used to classify subjects on the basis of either disease state or response to treatment can be done with a number of techniques. Here, we review 2 of the most common approaches: advanced functional connectivity analyses and multivoxel pattern analysis. Functional connectivity analyses use correlation to generate subject-specific functional connectivity maps.37,38 The correlations are performed between regions, which can be either anatomically defined (regions of interest) or functionally defined using parcellation methods (eg, k-means clustering, independent component analysis). These functional connectivity maps can then be subjected to second-level analysis using univariate statistics to compare features of the map between disease states. Null hypothesis testing is used to determine which features of the map might be different between healthy individuals and those with neurological impairment. We recently used this type of analysis to study differences in the spinal cord processing of thermal heat in healthy control subjects and patients with chronic cervical SCI.10 Specifically, we demonstrated that patients with chronic, incomplete SCI process thermal sensory information differently from healthy control subjects and that such differences in spinal cord functioning persist in patients who fully recover from their injury. To the best of our knowledge, this is the first evidence of spinal cord plasticity obtained with spinal fMRI techniques; we have summarized the pertinent findings in Figure 3 and refer the interested reader to the original paper for more information.10FIGURE 3: A spinal functional connectivity analysis demonstrating that patients with chronic spinal cord injury (SCI) process thermal sensory stimuli different from healthy control subjects. To the best of our knowledge, this is the first evidence of human spinal cord plasticity obtained from spinal functional magnetic resonance imaging methodologies. From left to right, intraspinal connectivity analysis in a healthy control subject (projected image), statistical comparison of control subjects and incomplete SCI patients (bar graph), intraspinal connectivity analysis in a patient with incomplete SCI (projected image), and statistical comparison of control subjects and recovered SCI patients. The projected images represent a single subject. The blue lines represent intraspinal connections; the orange-yellow lines, spinal cord--caudal brainstem connections; and the red spectrum, spinal cord--rostral brainstem connections. The number of intraspinal connections to the prime cluster is shown as a projected image for both uninjured control subjects (A) and patients with chronic incomplete SCI (B) and is shown graphically for SCI patients who fully recovered (C). A 1-way analysis of variance comparing the mean number of intraspinal connections and the prime cluster across healthy control subjects (blue and red bar graph), patients with incomplete SCI (green and purple bar graph), and recovered SCI participants (yellow and orange bar graph) was significant (P < .001). Significant post hoc Tukey tests included the difference between control participants stimulated in the C5 dermatome and patients with incomplete SCI stimulated above the level of injury (*P = 0.045, blue vs green bar graph) and recovered SCI participants stimulated above the level of injury (#P = .03, blue vs yellow bar graph). The interested reader is referred to the original manuscript for further details.10Two shortcomings of using a second-level analysis of functional connectivity data include the notion that secondary univariate analysis ignores spatially distributed patterns of functional connectivity and that, although null hypothesis testing may be used to categorize results, it does not provide a means for evaluating the predictive power of the results. Multivoxel pattern analysis methods, sensitive to spatially distributed information, aim to overcome the limitations of standard functional connectivity analysis.39,40 Feature selection, defined as “the process of selecting a subset of features that are useful for predication,” is an integral component of multivoxel pattern analysis algorithms whereby patterns are extracted from time-series data sets and used to differentiate observations.41 In this way, the performance of the learned pattern is quantified by the prediction error, say between a healthy individual and a person who sustained an SCI. Prediction error therefore allows a means of measuring how well the observed data fit a model rather than how poorly the data match the null hypothesis. Such analyses have been successfully applied in brain fMRI data whereby clinically depressed individuals could be distinguished from healthy counterparts.41 Although it remains to be determined whether multivoxel pattern analysis methods can be successfully applied to spinal fMRI data, such methods would certainly be advantageous for elucidating different clinical states of patients with diseases or injury to the spinal cord. The hope in this realm lies in the notion that spinal fMRI phenotypes (either resting or task based) would provide a more sensitive measure of spinal cord function than is possible with the current gold standard of clinical examination. If this does indeed pan out, it will be possible to monitor an individual patient’s response to novel therapeutic trials. It may also be possible to quantify a patient’s subjective experience of pain. If this is indeed the case, such novel methodologies may also aid in our understanding of spinal mechanisms of pain and how individuals respond to treatments such as dorsal column stimulation. EFFECTIVE CONNECTIVITY Estimating Causality to Understand Spinal Function Effective connectivity refers explicitly to the influence that 1 neural system exerts over another either at a synaptic or population level.7 Thus, effective connectivity analyses of spinal fMRI data aim to estimate causality within the functioning human brainstem and spinal cord. At its core, effective connectivity is rooted in model comparison in which models are usually generated on the basis of neuronal coupling architectures that have been experimentally determined. This is in stark contrast to functional connectivity approaches, which are purely descriptive in nature. Causality in neural function is not a straightforward concept. Besides unidirectional information flow (eg, a vibrating mobile phone causing neuronal activity in the dorsal horn of the spinal cord), the CNS also takes advantage of feedback circuitry and, of particular importance to the spinal cord, descending modulation. These few modes of information transfer can quickly add up to complicated neuronal circuitry in which estimating the causality of 1 discrete neurophysiological event on another becomes an arduous task. Nonetheless, a few proposals have been put forward to estimate causality in fMRI data. Although none of these has been applied to spinal fMRI data sets, we briefly review a few examples applied to brain fMRI studies. Spatially distributed discrete neurophysiological events have the potential to influence one another. Conditional probabilities, for example, the notion that A causes B, can be estimated with the Patel conditional pairwise probability (if only 2 nodes are considered) or more complex methods if > 2 regions are involved in the flow of information (such as Bayes nets, structural equation modeling).42-44 A prerequisite of estimating such causality rests in the notion of imposing structural constraints on possible models. For example, if one were attempting to study the effect of the pre-Botzinger complex on respiratory mechanisms of the human brainstem, it would be necessary to define this region (along with other brainstem and spinal cord respiratory centers). With spatial regions of interest clearly outlined, it becomes possible to generate models of information flow. As a generic example, one may probe whether information flows from A to B to C, from A to C to B, or from A to B and C (Figure 4). In turn, the strength of the model, when compared with the obtained spinal fMRI data, can provide an estimate of causal influence.FIGURE 4: Effective connectivity to estimate causality in the human central nervous system (CNS). It is known that spatially distributed discrete neurophysiological events have the potential to influence one another. By generating models of information flow through the CNS, spinal functional magnetic resonance imaging (fMRI) data can be used to estimate causality between discrete neuroanatomical regions. In this cartoon diagram, 3 causal models have been put forth to show to how region A influences regions B and C. By comparing these models to the obtained spinal fMRI data, one can determine which model best predicts the observed spinal fMRI data and hence can estimate causality within the functioning human brainstem and spinal cord.CONCLUSION In this narrative review, we have provided an overview of how spinal fMRI can be used to increase our understanding of the human spinal cord. First, we outlined how the human spinal cord generates periodic fluctuations in blood flow and summarized the definitions of neurovascular coupling and functional hyperemia. Next, we discussed the importance of defining a spinal fMRI phenotype according to 4 broad categories: the stimulation (or resting) paradigm, fMRI data acquisition, preprocessing steps, and processing steps. This phenotype can be visualized with a functional connectivity analysis to derive correlation-based metrics. The value of this approach lies in the classification of patients. Finally, we discussed the potential of effective connectivity to go beyond correlation-based classification of patients and outline the possibilities of spinal fMRI to estimate causality of function within the human brainstem and spinal cord. Given that most of the data analysis approaches discussed in this forward-looking review have not been applied to spinal fMRI data (rather, they have been used in brain fMRI analyses), we must end on a note of caution. The field of spinal fMRI has made exceptional gains in the last decade but faces equally extraordinary hurdles for the next decade. As alluded to in the opening paragraph, the spinal cord is one of the most difficult areas of the human body to image. Magnetic susceptibility artifacts (arising from surrounding airspaces, ligaments, and bones) and spinal cord motion are but 2 physiological hurdles limiting the robustness and reproducibility of single-subject studies. Full disclosure of methods and strategies to circumvent physiological obstacles will be paramount to realize the full potential of this technique to better aid patients with debilitating spinal disorders. Disclosure The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article. Acknowledgment We are grateful for the assistance of Dr Patrick Stroman, PhD, in creating the spinal fMRI image presented in Figure 2." @default.
W2313595323 created "2016-06-24" @default.
W2313595323 creator A5009765322 @default.
W2313595323 creator A5047977949 @default.
W2313595323 creator A5069282912 @default.
W2313595323 date "2013-08-01" @default.
W2313595323 modified "2023-10-15" @default.
W2313595323 title "Visualizing Integrative Functioning in the Human Brainstem and Spinal Cord With Spinal Functional Magnetic Resonance Imaging" @default.
W2313595323 cites W116098567 @default.
W2313595323 cites W1519239048 @default.
W2313595323 cites W184691178 @default.
W2313595323 cites W1970695058 @default.
W2313595323 cites W1973130592 @default.
W2313595323 cites W1976926042 @default.
W2313595323 cites W1996309271 @default.
W2313595323 cites W1998175349 @default.
W2313595323 cites W2002373576 @default.
W2313595323 cites W2002660165 @default.
W2313595323 cites W2006240253 @default.
W2313595323 cites W2010746653 @default.
W2313595323 cites W2013016035 @default.
W2313595323 cites W2016306388 @default.
W2313595323 cites W2022454153 @default.
W2313595323 cites W2027095655 @default.
W2313595323 cites W2036613205 @default.
W2313595323 cites W2038040839 @default.
W2313595323 cites W2048124855 @default.
W2313595323 cites W2060319077 @default.
W2313595323 cites W2061564920 @default.
W2313595323 cites W2063202713 @default.
W2313595323 cites W2063688349 @default.
W2313595323 cites W2064841852 @default.
W2313595323 cites W2065560362 @default.
W2313595323 cites W2075556735 @default.
W2313595323 cites W2095438393 @default.
W2313595323 cites W2095921032 @default.
W2313595323 cites W2105396029 @default.
W2313595323 cites W2108049514 @default.
W2313595323 cites W2110253210 @default.
W2313595323 cites W2117490209 @default.
W2313595323 cites W2123923307 @default.
W2313595323 cites W2138905229 @default.
W2313595323 cites W2143822459 @default.
W2313595323 cites W2156295356 @default.
W2313595323 cites W2156743186 @default.
W2313595323 cites W2160756801 @default.
W2313595323 cites W2161677827 @default.
W2313595323 cites W2162178181 @default.
W2313595323 cites W2163234334 @default.
W2313595323 cites W2234625571 @default.
W2313595323 cites W2409128968 @default.
W2313595323 doi "https://doi.org/10.1227/01.neu.0000430767.87725.95" @default.
W2313595323 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/23839360" @default.
W2313595323 hasPublicationYear "2013" @default.
W2313595323 type Work @default.
W2313595323 sameAs 2313595323 @default.
W2313595323 citedByCount "0" @default.
W2313595323 crossrefType "journal-article" @default.
W2313595323 hasAuthorship W2313595323A5009765322 @default.
W2313595323 hasAuthorship W2313595323A5047977949 @default.
W2313595323 hasAuthorship W2313595323A5069282912 @default.
W2313595323 hasBestOaLocation W23135953231 @default.
W2313595323 hasConcept C105702510 @default.
W2313595323 hasConcept C118552586 @default.
W2313595323 hasConcept C126838900 @default.
W2313595323 hasConcept C143409427 @default.
W2313595323 hasConcept C169760540 @default.
W2313595323 hasConcept C2779226451 @default.
W2313595323 hasConcept C2780775167 @default.
W2313595323 hasConcept C551621295 @default.
W2313595323 hasConcept C71924100 @default.
W2313595323 hasConcept C86803240 @default.
W2313595323 hasConceptScore W2313595323C105702510 @default.
W2313595323 hasConceptScore W2313595323C118552586 @default.
W2313595323 hasConceptScore W2313595323C126838900 @default.
W2313595323 hasConceptScore W2313595323C143409427 @default.
W2313595323 hasConceptScore W2313595323C169760540 @default.
W2313595323 hasConceptScore W2313595323C2779226451 @default.
W2313595323 hasConceptScore W2313595323C2780775167 @default.
W2313595323 hasConceptScore W2313595323C551621295 @default.
W2313595323 hasConceptScore W2313595323C71924100 @default.
W2313595323 hasConceptScore W2313595323C86803240 @default.
W2313595323 hasIssue "Supplement 1" @default.
W2313595323 hasLocation W23135953231 @default.
W2313595323 hasLocation W23135953232 @default.
W2313595323 hasLocation W23135953233 @default.
W2313595323 hasOpenAccess W2313595323 @default.
W2313595323 hasPrimaryLocation W23135953231 @default.
W2313595323 hasRelatedWork W1229094906 @default.
W2313595323 hasRelatedWork W1551508849 @default.
W2313595323 hasRelatedWork W1975548597 @default.
W2313595323 hasRelatedWork W2053596317 @default.
W2313595323 hasRelatedWork W2071486685 @default.
W2313595323 hasRelatedWork W2072932238 @default.
W2313595323 hasRelatedWork W2344611820 @default.
W2313595323 hasRelatedWork W2924093185 @default.
W2313595323 hasRelatedWork W3031450532 @default.
W2313595323 hasRelatedWork W4321107168 @default.