REVIEWS NATUREIVol 453112 June 2008 and in the (see 'Neural signals in Supplem d b ker microe trodes.The acellular field potentia captures at leas nce of NBR and de ns n minant neuronal n an electrode tip, on bet eld potentia Multiple ur activity and local fiel conclusions ot be yo tial data indicating that such ormation beta integrative soma nt hat all to th 8u时 hange in the powe ENy ban r tha excitatory input area o indi co sumption depending on th ontribution of the afor this case i Last but not inte ntribute alter CBF in respons may occasionally act as 'shunts'for the ry int sm nined se eparately,local spiking of the pes ofe to n erim ding toad ElctiophysiologcadlstudisCaminingtcndinidualcontr corti al output through the a ngof individual 6 other ty ct the the d it tothe direc f the lat d micro in the visual g rather than m dal cell ou of this fact,the nture of the EIN that s action and it han n was not the degre Icorrelates of the BOLD signa the the bution of any type any given time are conidered to mic resp current sink,w C2008 Macmillan Publishers Limited.All rights reserved
In contrast, human fMRI studies reported haemodynamic and metabolic downregulation accompanying neuronal inhibition in motor39 and visual cortices40, suggesting that the sustained negative BOLD response (NBR) is a marker of neuronal deactivation. Similarly, combined fMRI and electrophysiological experiments showed a clear correspondence of NBR and decreased population spiking in haemodynamically ‘negative’ areas in the monkey primary visual cortex41. Decreases in blood oxygenation and volume were also found to be co-localized with predominant neuronal inhibition and arteriolar vasoconstriction during somatosensory stimulation in rats42. Thus, without understanding the intrinsic correlation between direct or indirect inhibitory activity and concomitant changes in energy metabolism in a given situation, conclusions cannot be drawn. Unfortunately, the few published theoretical estimates of energy budget have not considered the metabolic costs of spikes in interneurons and of the inhibitory postsynaptic potentials (IPSPs) they produce43. Modelling of inhibition is unlikely to be straightforward. On the one hand, the density of cortical inhibitory neurons is 10–15 times lower than excitatory neurons16, and for each one of them the electrochemical gradient, down which Cl2 moves postsynaptically at inhibitory synapses, is weaker than that of Na1 at excitatory synapses, requiring less energy to pump Cl2 back. In fact, the transport cycles of the cation–chloride co-transporters, which have a key role in intracellular Cl2 regulation, are driven without the direct hydrolysis of ATP, by using the energy from the cation gradients generated by the Na,K-ATPase44. On the other hand, inhibitory interneurons are fast spiking45,46. For example, the firing of pyramidal cells in hippocampus is 1.4 Hz, whereas that of interneurons in the strata pyramidale and oriens is 15 Hz and 10 Hz, respectively. Similarly, cortical inhibitory interneurons may discharge 2–3 times faster than pyramidal cells47. In principle, inhibition may increase or decrease energy consumption depending on the contribution of the aforementioned factors (for a recent comprehensive review on inhibitory neurons and brain metabolism, see ref. 48). Last but not least, neurons directly affect microvessels. Pericytes, the flat, contractile connective-tissue cells, often attached to the abluminal surface of the capillary endothelial cells, might directly alter CBF in response to changes in neural activity49. Moreover, a body of evidence suggests that increased activity of single inhibitory interneurons results in precise vasomotor responses in neighbouring brain microvessels, and these contractile or dilatory responses were attributed to arteriole smooth muscle50. The diversity of the haemodynamic responses to neural inhibition obtained in different types of experiments is therefore hardly surprising: it is primarily due to the fact that regional inhibition itself might have a number of different causes, including early shunting of the weak cortical input, leading to a reduction of recurrent excitation rather than an increase in summed inhibition; increased synaptic inhibition; shunting of the cortical output through the axo-axonic connections of the chandelier cells; or any combination thereof. In the first case inhibition might result in a clear NBR; in the other two it might reflect the local metabolism increases induced by the unaffected input and its ongoing processing, resulting in fMRI activations. The fMRI responses might further blur the origin of inhibition owing to the direct effects of the latter on the arterioles and microvessels. Evidently much research is needed to characterize the actual state of an area and its participation in behaviour, but quite independent of this fact, the nature of the EIN suggests that mass action and its surrogate haemodynamics are ambiguous signals, the interpretation of which must be constrained by the concurrent use of other methodologies. Neurophysiological correlates of the BOLD signal EIN and mesoscopic neural signals. The active regions of the membrane of a discharging neuron at any given time are considered to act as a current sink, whereas the inactive ones act as a current source for the active regions (see ‘Neural signals’ in Supplementary Information). The linear superposition of currents from all sinks and sources forms the extracellular field potential measured by microelectrodes. The extracellular field potential captures at least three different types of EIN activity: single-unit activity representing the action potentials of well isolated neurons next to the electrode tip, multiple unit activity reflecting the spiking of small neural populations in a sphere of 100–300 mm radius, and perisynaptic activity of a neural population within 0.5–3 mm of the electrode tip, which is reflected in the variation of the low-frequency components of the extracellular field potential. Multiple unit activity and local field potentials (LFPs) can be reliably segregated by frequency band separation. A high-pass filter cutoff in the range of 500–1,000 Hz is used in most recordings to obtain the multiple unit activity, and a low-pass filter cutoff of approximately 250 Hz to obtain LFP. A large number of experiments have presented data indicating that such a band separation does indeed underlie different neural events (see ‘Neural signals’ in Supplementary Information). LFP signals and their different band-limited components (alpha, beta, gamma, and so on) are invaluable for understanding cortical processing, as they are the only signs of integrative EIN processes. In fact, LFPs do not, as initially thought, solely reflect population postsynaptic potentials, but also integrative soma–dendritic processes— including voltage-dependent membrane oscillations and afterpotentials following soma–dendritic spikes—that all together represent the local (perisynaptic) activity in a region (see ‘Neural signals’ in Supplementary Information). A shortcoming of the LFP is its ambiguity. A change in the power of LFP in a particular frequency band most likely occurs for any mode of operations of the EIN. As most of the excitatory input into an area is local, LFPs will also indirectly reflect some of the postsynaptic effects of pyramidal cell activity. In addition, LFPs have a certain neural-class bias, which in this case is determined by geometry and regional architecture. The arrangement of the pyramidal and Purkinje cells will give rise to large LFP modulations; in contrast, interneurons will contribute only weakly because of their star-shaped dendrites and their geometrical disorder. Finally, inhibitory synapses may occasionally act as ‘shunts’ for the excitatory currents through low-resistance channels, in which case large synaptic conductance changes may produce little effect in the membrane potential, and result in weak and hard-to-measure multiple unit activity and LFPs. When individual LFP bands are examined separately, local spiking activity is occasionally found to affect certain frequency bands, whereas that of neuromodulation affects others51–53. It is evident that the most useful information will not be derived by one type of signal alone, but rather by the study of relative changes in one signal or the other. Electrophysiological studies examining the individual contributions of different LFP frequency bands, multiple unit activity, and spiking of individual neurons are probably our only realistic chance of gaining insights into the neural mechanisms of haemodynamic responses and their meaning in the context of different cognitive tasks. Mesoscopic signals and the BOLD signal. The relationship of neocortical LFPs and spiking activity to the BOLD signal itself was examined directly in concurrent electrophysiology and fMRI experiments in the visual system of anaesthetized54 and alert55 monkeys. These studies found that the BOLD responses reflect input and intracortical processing rather than pyramidal cell output activity. Initially, both LFPs and spiking seemed to be correlated with the BOLD response, although quantitative analysis indicated that LFPs are better predictors of the BOLD response than multiple-unit or single-unit spiking. The decisive finding leading to the papers’ conclusion, however, was not the degree of correlation between the neural and the fMRI responses or the differential contribution of any type of signal into the BOLD responses55, but rather the striking, undiminished haemodynamic responses in cases where spiking was entirely absent despite a clear and strong stimulus-induced modulation of the field REVIEWS NATUREjVol 453j12 June 2008 874 ©2008 Macmillan Publishers Limited. All rights reserved
NATUREVol 45312 June 2008 REVIEWS potentials>Similar dissociations between spikes and CBF had thickness",the overall length of neuronal processes remains rela been demonstrated earlier and very recently in a number of studies tively constant,with axonal length being approximately 4km mm using other techniques6-ss and dendrite length 0.4 km mm.Overall,synaptic density and the The findings are in close agreement with a number of older auto- ratio of excitatory to inhibitory synapses also remain constant. radiography studies,also showing that regional glucose utilization is Given these neuro-statistical data,what are the actual contents of a directly related to neuronal synaptic activity3.For example,the neuroimaging voxel?An examination of the 300 top-cited cognitive greatest 2-DG uptake occurs in the neuropil (that is,in areas rich fMRI studies suggests that the commonly used in-plane resolution is in synapses,dendrites and axons,rather than in cell bodies).During 9-16 mm2,for slice thicknesses of 5-7mm.The average voxel size orthodromic and antidromic electrical microstimulation,only before any pre-processing of the data is thus 55 ul(or 55 mm3).Often orthodromic microstimulation,which involves presynaptic term- the effective size is 2-3 times larger due to the spatial filtering that inals,increases glucose consumption.Similarly,the highest density most investigators apply to improve the functional SNR.Less than of cytochrome oxidase(an enzyme of the respiratory chain)is found 3%of this volume is occupied by vessels and the rest by neural in somato-dendritic regions that are adjacent to axon terminals. elements(see Fig.3)A typical unfiltered fMRI voxel of 55 ul in size Finally,as mentioned earlier,presynaptic activity increases metabol- thus contains 5.5 million neurons,2.2-5.5 X 1010 synapses,22 km of ism even if the output is inhibited (that is,the spiking activity is dendrites and 220km of axons. abolished). This 'large population view'is in contrast to the scope of the Despite all this evidence,some discussion still concentrates on the traditional microelectrode recordings.It would be nice if we could importance of the firing rate of action potentials of projection neu- monitor every relevant neuron in the cortex during intracortical rons in the generation of the haemodynamic responses,perhaps microelectrode recordings,but this is practically impossible stemming from the fact that important early studies of neural corre- Instead,the typical electrophysiological measurements in behaving lates of behaviour took the mean spiking rate to be the gold standard animals report only on the properties of most active large neurons for quantifying neuronal activation.These discussions,however, that constitute a minority.The strong selection bias during extracel- often suffer from a certain amount of contention seeking where none lular recordings is partly due to practical limitations (for example, is warranted.In many cases,spikes do indeed correlate with LFPs, injury or simply size bias2)and partly to the physiological properties and they will also correlate with the BOLD signal.In addition,unusu- of neurons and/or the organizational principles of neural networks. ally high correlations between multiple unit activity and BOLD signal In fact,many different types of electrical and optical measurements (or LFP and multiple unit activity)may result from excessive signal- provide evidence that a substantial proportion of neurons,including smoothing owing to sampling rates of several seconds rather than a the cortical pyramidal cells,might be silent.Their silence might fraction of a second,as well as inter-subject averaging when simultan- reflect unusually high input selectivity or the existence of decoding eous physiology and fMRI measurements are not possible (see ref.55 for discussion). schemes relying on infrequent co-spiking of neuronal subsets.Most Predicting neural activity from the fMRI signals.Functional MRI important for the comparison of neuroimaging and electrophysiol- signals are presumed to result from changes in the activity of the ogy results is the fact that lack of measurable neuronal spiking may neuronal populations responsible for the functions in question(for not necessarily imply lack of input and subthreshold processing. example,stimulus-or task-selective neurons).This assumption is A direct analogy between neuronal spiking as measured in animal mainly based on decades of electrophysiology research with record- experiments and the fMRI signal obtained in human recording is thus simply unrealistic and might often lead to incorrect conclusions. ings from isolated single neurons in experimental animals,in which particular sensory stimuli that the animal perceives or tasks that it It is hardly surprising that most studies so far relying purely on BOLD performs were found to increase the firing rate of certain cells but not fMRI have failed to reveal the actual neural properties of the studied of others.The psychologist or cognitive neuroscientist who finds area,at least those properties (for example,selectivity to various cortical area X to be activated by the task at hand implicitly or expli- visual features)that were previously established in electrophysio- citly assumes that-if an electrode were placed in the subject's logical studies. brain-an increase in the spiking rate of those specialized neurons An example is cortical area V5(or MT)that has been extensively underlying the subject's behaviour would be observed.This might studied in the context of motion processing and perception465 well be true in some cases,but not in all.When attempting to inter- Electrophysiology has shown that the vast majority of the V5 neurons pret the fMRI signal by modelling,or when comparing the results of in monkeys are direction and speed selective.Neuroimaging loca- human neuroimaging to those obtained in monkey physiology lized the homologue of area V5 in humans as an area responding experiments,it is useful to take the following facts into consideration. stronger to moving than to stationary stimuli.Later studies suggested In humans,there are about 90,000-100,000 neurons under 1 mm that human V5 is sensitive to motion direction,and that it may be of cortical surface.This number is relatively constant for all structur- thought of as containing large populations of directionally selective ally and functionally distinct areas,including the somatosensory, units,just like its monkey homologue.The studies of directional temporal,parietal,frontal and motor cortical areas3.An exception specificity exploited the phenomenon of motion after-effect induced is the primary visual cortex of certain primates,including monkey by motion adaptation.After prolonged exposure to a stimulus mov- and human,which has approximately twice as many neurons.The ing in one direction,subjects perceive a subsequent static stimulus to number of cortical neurons under unitary cortical surface is also move in the opposite direction.It is assumed that motion after-effect similar across many species,including mouse,rat,cat,monkey and is due to the fact that the balance of mutual inhibition (opponency) human.Its small variability is the result of a trade-off between cor- between detectors for opposite directions of movement is distorted tical thickness and neural density.The former varies from area to area after adaptation.The sensitivity of the detectors selective for the and from species to species (for example,from mouse to human the adapting direction is reduced,which in turn releases from inhibition cortex becomes approximately three times thicker).Neural density the neurons selective for the opposite direction.Using this phe- varies inversely to cortical thickness.On average,density is 20,000 to nomenon,human studies demonstrated that the fMRI response to 30,000 neurons per mm';it peaks in the primary visual cortex by a a stationary stimulus was greater when the stimulus was preceded by factor of 4,and it is minimal in the motor cortex.Synaptic density a motion-after-effect-inducing,unidirectional adaptation,than ranges from 0.4 to 1 X 10per mm'.Depending on the thickness of when preceded by bidirectional adaptation Given the existing the cortex(2-4 mm),the number of synapses beneath I mm2surface physiology data in the monkey V5,these findings were interpreted is around 10(0.8-4X 10).Although the number of synapses as demonstrating that the BOLD signal directly reflects direction- and the axonal length per neuron increases with increasing cortical selective spiking activity of the area. 875 2008 Macmillan Publishers Limited.All rights reserved
potentials54,55. Similar dissociations between spikes and CBF had been demonstrated earlier and very recently in a number of studies using other techniques56–58. The findings are in close agreement with a number of older autoradiography studies, also showing that regional glucose utilization is directly related to neuronal synaptic activity35. For example, the greatest 2-DG uptake occurs in the neuropil (that is, in areas rich in synapses, dendrites and axons, rather than in cell bodies). During orthodromic and antidromic electrical microstimulation, only orthodromic microstimulation, which involves presynaptic terminals, increases glucose consumption. Similarly, the highest density of cytochrome oxidase (an enzyme of the respiratory chain) is found in somato-dendritic regions that are adjacent to axon terminals. Finally, as mentioned earlier, presynaptic activity increases metabolism even if the output is inhibited (that is, the spiking activity is abolished). Despite all this evidence, some discussion still concentrates on the importance of the firing rate of action potentials of projection neurons in the generation of the haemodynamic responses, perhaps stemming from the fact that important early studies of neural correlates of behaviour took the mean spiking rate to be the gold standard for quantifying neuronal activation. These discussions, however, often suffer from a certain amount of contention seeking where none is warranted. In many cases, spikes do indeed correlate with LFPs, and they will also correlate with the BOLD signal. In addition, unusually high correlations between multiple unit activity and BOLD signal (or LFP and multiple unit activity) may result from excessive signalsmoothing owing to sampling rates of several seconds rather than a fraction of a second, as well as inter-subject averaging when simultaneous physiology and fMRI measurements are not possible (see ref. 55 for discussion). Predicting neural activity from the fMRI signals. Functional MRI signals are presumed to result from changes in the activity of the neuronal populations responsible for the functions in question (for example, stimulus- or task-selective neurons). This assumption is mainly based on decades of electrophysiology research with recordings from isolated single neurons in experimental animals, in which particular sensory stimuli that the animal perceives or tasks that it performs were found to increase the firing rate of certain cells but not of others. The psychologist or cognitive neuroscientist who finds cortical area X to be activated by the task at hand implicitly or explicitly assumes that—if an electrode were placed in the subject’s brain—an increase in the spiking rate of those specialized neurons underlying the subject’s behaviour would be observed. This might well be true in some cases, but not in all. When attempting to interpret the fMRI signal by modelling, or when comparing the results of human neuroimaging to those obtained in monkey physiology experiments, it is useful to take the following facts into consideration. In humans, there are about 90,000–100,000 neurons under 1 mm2 of cortical surface. This number is relatively constant for all structurally and functionally distinct areas, including the somatosensory, temporal, parietal, frontal and motor cortical areas16,59. An exception is the primary visual cortex of certain primates, including monkey and human, which has approximately twice as many neurons. The number of cortical neurons under unitary cortical surface is also similar across many species, including mouse, rat, cat, monkey and human. Its small variability is the result of a trade-off between cortical thickness and neural density. The former varies from area to area and from species to species (for example, from mouse to human the cortex becomes approximately three times thicker). Neural density varies inversely to cortical thickness. On average, density is 20,000 to 30,000 neurons per mm3 ; it peaks in the primary visual cortex by a factor of 4, and it is minimal in the motor cortex59,60. Synaptic density ranges from 0.4 to 1 3 109 per mm3 . Depending on the thickness of the cortex (2–4 mm), the number of synapses beneath 1 mm2 surface is around 109 (0.8–4 3 109 ). Although the number of synapses and the axonal length per neuron increases with increasing cortical thickness61, the overall length of neuronal processes remains relatively constant, with axonal length being approximately 4 km mm23 and dendrite length 0.4 km mm23 . Overall, synaptic density and the ratio of excitatory to inhibitory synapses also remain constant. Given these neuro-statistical data, what are the actual contents of a neuroimaging voxel? An examination of the 300 top-cited cognitive fMRI studies suggests that the commonly used in-plane resolution is 9–16 mm2 , for slice thicknesses of 5–7 mm. The average voxel size before any pre-processing of the data is thus 55 ml (or 55 mm3 ). Often the effective size is 2–3 times larger due to the spatial filtering that most investigators apply to improve the functional SNR. Less than 3% of this volume is occupied by vessels and the rest by neural elements (see Fig. 3) A typical unfiltered fMRI voxel of 55 ml in size thus contains 5.5 million neurons, 2.2–5.5 3 1010 synapses, 22 km of dendrites and 220 km of axons. This ‘large population view’ is in contrast to the scope of the traditional microelectrode recordings. It would be nice if we could monitor every relevant neuron in the cortex during intracortical microelectrode recordings, but this is practically impossible. Instead, the typical electrophysiological measurements in behaving animals report only on the properties of most active large neurons that constitute a minority. The strong selection bias during extracellular recordings is partly due to practical limitations (for example, injury or simply size bias62) and partly to the physiological properties of neurons and/or the organizational principles of neural networks. In fact, many different types of electrical and optical measurements provide evidence that a substantial proportion of neurons, including the cortical pyramidal cells, might be silent63. Their silence might reflect unusually high input selectivity or the existence of decoding schemes relying on infrequent co-spiking of neuronal subsets. Most important for the comparison of neuroimaging and electrophysiology results is the fact that lack of measurable neuronal spiking may not necessarily imply lack of input and subthreshold processing. A direct analogy between neuronal spiking as measured in animal experiments and the fMRI signal obtained in human recording is thus simply unrealistic and might often lead to incorrect conclusions. It is hardly surprising that most studies so far relying purely on BOLD fMRI have failed to reveal the actual neural properties of the studied area, at least those properties (for example, selectivity to various visual features) that were previously established in electrophysiological studies. An example is cortical area V5 (or MT) that has been extensively studied in the context of motion processing and perception64,65. Electrophysiology has shown that the vast majority of the V5 neurons in monkeys are direction and speed selective. Neuroimaging localized the homologue of area V5 in humans as an area responding stronger to moving than to stationary stimuli. Later studies suggested that human V5 is sensitive to motion direction, and that it may be thought of as containing large populations of directionally selective units, just like its monkey homologue. The studies of directional specificity exploited the phenomenon of motion after-effect induced by motion adaptation. After prolonged exposure to a stimulus moving in one direction, subjects perceive a subsequent static stimulus to move in the opposite direction. It is assumed that motion after-effect is due to the fact that the balance of mutual inhibition (opponency) between detectors for opposite directions of movement is distorted after adaptation. The sensitivity of the detectors selective for the adapting direction is reduced, which in turn releases from inhibition the neurons selective for the opposite direction66. Using this phenomenon, human studies demonstrated that the fMRI response to a stationary stimulus was greater when the stimulus was preceded by a motion-after-effect-inducing, unidirectional adaptation, than when preceded by bidirectional adaptation67. Given the existing physiology data in the monkey V5, these findings were interpreted as demonstrating that the BOLD signal directly reflects directionselective spiking activity of the area. NATUREjVol 453j12 June 2008 REVIEWS 875 ©2008 Macmillan Publishers Limited. All rights reserved
REVIEWS NATUREVol 45312 June 2008 Yet,as I have indicated above,the BOLD signal is primarily affec- in monkeys'.These studies showed that only a small fraction of VI ted by changes in excitation-inhibition balance,and this balance may cells modulate their spiking during the perceptual changes;neuroi- be controlled by neuromodulation more than by the changes in maging,on the other hand,demonstrated fMRI-signal modulations spiking rate of a small set of neurons.In fact,the BOLD signal is that were nearly as large as those obtained during the physical strongly modulated by attention,and the results of the motion alternation of stimuli".The difference,once again,reflects the fact after-effect experiments could,in principle,be due to the fact that that neuromodulatory feedback from higher areas can be easily a stimulus with illusory motion automatically draws the attention of detected by means of fMRI,but not through the measurement of a subject more compared to a situation in which there is no motion single-unit activity.Interestingly,measurements of subthreshold after-effect.This hypothesis turned out to be correct,as a later activity in another study of perceptual multistability revealed per- study-in which balance in attentional load was accomplished by ception-related modulations in LFP,despite the unaltered spike having the subjects perform a concurrent visual task-found no ratesss.Such clear spiking and BOLD signal mismatches appear even signal differences between the motion after-effect and no motion in simple experiments probing sensory processing.Simple stimuli, after-effect conditions. such as those used in the aforementioned studies,are most likely to A similar example pertains to the differences in neurophysiological generate a proportional enhancement in both the afferent and effer- and fMRI responses in the primary visual cortex during different ent activity of any sensory area.The activation of high-level asso- perceptual states.It is known that physiological signals are in general ciation areas related to cognitive processing might be more sensitive stronger when stimuli are perceived as opposed to when they are not. or even dominated by feedback and neuromodulation,whose differ- Intriguingly,in some regions the BOLD response seems to reflect this ential effect on spiking and haemodynamic responses is utterly even more sensitively than physiological measures like spikes and unknown. multi-unit activity?.An example is the pattern of fMRI activation changes in V1 during binocular rivalry (that is,the perceptual alter- Conclusions and perspectives nations experienced when the two eyes view different stimuli).This The limitations of fMRI are not related to physics or poor engineer- phenomenon has been studied extensively psychophysically and ing,and are unlikely to be resolved by increasing the sophistication also over the last two decades in a series of electrophysiology studies and power of the scanners;they are instead due to the circuitry and Va Vb 100 Figure 3 Neural and vascular contents of a voxel.The left panel right;white spots are cross-sections of vessels).The average distance between demonstrates the relative density of vessels in the visual cortex of monkeys. the small vessels(capillaries)is about 50 um.This is approximately the The dense vascular mesh is displayed by perfusing the tissue with barium distance that oxygen molecules travel by diffusion within the limited transit sulphate and imaging it with synchrotron-based X-ray microtomography time of the blood.The dense population of neurons,synapses and glia (courtesy B.Weber,MPI for Biological Cybernetics).The vessel diameter is occupy the intervascular space,as depicted in the drawing at the top right-a colour coded.Cortical surface without pial vessels is displayed at the top; hypothetical distribution of vascular and neural elements in a small section white matter at the bottom.At the left of the panel is a Nissl slice from the (red rectangle).The drawing in the background shows some of the typical same area,showing the neural density for layers II through to the white neuronal types(for example,red,large pyramidal cell;dark blue,inhibitory matter (wm).Although the density of the vessels appears to be high in this basket cells;light blue,chandelier inhibitory neurons;and grey,stellate cells) three-dimensional representation,it is actually less the 3%(see section at the and their processes. 876 2008 Macmillan Publishers Limited.All rights reserved
Yet, as I have indicated above, the BOLD signal is primarily affected by changes in excitation–inhibition balance, and this balance may be controlled by neuromodulation more than by the changes in spiking rate of a small set of neurons. In fact, the BOLD signal is strongly modulated by attention68, and the results of the motion after-effect experiments could, in principle, be due to the fact that a stimulus with illusory motion automatically draws the attention of a subject more compared to a situation in which there is no motion after-effect. This hypothesis turned out to be correct, as a later study—in which balance in attentional load was accomplished by having the subjects perform a concurrent visual task—found no signal differences between the motion after-effect and no motion after-effect conditions69. A similar example pertains to the differences in neurophysiological and fMRI responses in the primary visual cortex during different perceptual states. It is known that physiological signals are in general stronger when stimuli are perceived as opposed to when they are not. Intriguingly, in some regions the BOLD response seems to reflect this even more sensitively than physiological measures like spikes and multi-unit activity70. An example is the pattern of fMRI activation changes in V1 during binocular rivalry (that is, the perceptual alternations experienced when the two eyes view different stimuli). This phenomenon has been studied extensively psychophysically and also over the last two decades in a series of electrophysiology studies in monkeys70. These studies showed that only a small fraction of V1 cells modulate their spiking during the perceptual changes; neuroimaging, on the other hand, demonstrated fMRI-signal modulations that were nearly as large as those obtained during the physical alternation of stimuli70. The difference, once again, reflects the fact that neuromodulatory feedback from higher areas can be easily detected by means of fMRI, but not through the measurement of single-unit activity. Interestingly, measurements of subthreshold activity in another study of perceptual multistability revealed perception-related modulations in LFP, despite the unaltered spike rates53. Such clear spiking and BOLD signal mismatches appear even in simple experiments probing sensory processing. Simple stimuli, such as those used in the aforementioned studies, are most likely to generate a proportional enhancement in both the afferent and efferent activity of any sensory area. The activation of high-level association areas related to cognitive processing might be more sensitive or even dominated by feedback and neuromodulation, whose differential effect on spiking and haemodynamic responses is utterly unknown. Conclusions and perspectives The limitations of fMRI are not related to physics or poor engineering, and are unlikely to be resolved by increasing the sophistication and power of the scanners; they are instead due to the circuitry and Figure 3 | Neural and vascular contents of a voxel. The left panel demonstrates the relative density of vessels in the visual cortex of monkeys. The dense vascular mesh is displayed by perfusing the tissue with barium sulphate and imaging it with synchrotron-based X-ray microtomography (courtesy B. Weber, MPI for Biological Cybernetics). The vessel diameter is colour coded. Cortical surface without pial vessels is displayed at the top; white matter at the bottom. At the left of the panel is a Nissl slice from the same area, showing the neural density for layers II through to the white matter (wm). Although the density of the vessels appears to be high in this three-dimensional representation, it is actually less the 3% (see section at the right; white spots are cross-sections of vessels). The average distance between the small vessels (capillaries) is about 50 mm. This is approximately the distance that oxygen molecules travel by diffusion within the limited transit time of the blood. The dense population of neurons, synapses and glia occupy the intervascular space, as depicted in the drawing at the top right—a hypothetical distribution of vascular and neural elements in a small section (red rectangle). The drawing in the background shows some of the typical neuronal types (for example, red, large pyramidal cell; dark blue, inhibitory basket cells; light blue, chandelier inhibitory neurons; and grey, stellate cells) and their processes. REVIEWS NATUREjVol 453j12 June 2008 876 ©2008 Macmillan Publishers Limited. All rights reserved