Our volitional self-initiated acts are preceded by the Bereitschaftspotential (BP) or readiness potential [34,35]. The BP has an early component (BP1) and a late component (BP2, see Fig. 1). The early component BP1 lasts from the very beginning of the BP (1-2 s or more prior to movement onset depending on the complexity of the movement) to app. ½ s prior to movement onset and the late component BP2 lasts from ½ s before to the onset of movement itself (s 0 in Fig. 1). BP1 is symmetrical even for unilateral movement, while BP2 is larger over the contralateral hemisphere.
DC-EEG
Electroencephalography originally was the only method fast enough to
record movement-related cortical events. Attempts to record movement-related
potentials date back to as early as 1951, when Bates [6] probed into the
pre-movement period using a cathode-ray oscilloscope on a photographic
plate superimposing several movement-related epochs. He did not find any
pre-movement potential, for the BP escaped him due to the short time constant
([tau]) necessary for the superposition method. Kornhuber and Deecke [34,35]
were lucky enough to select a [tau] of 1.2 s in order to enable discovery
of the BP. Certainly, for modern standards, this still is too short - attenuating
BP amplitude by more than 1/3 - and subsequently we used longer and longer
time constants ([tau] = 3, 5, 10, 20, etc.) until, from 1988, we have employed
only recordings with [tau] = [infinity], ie true DC-EEG (64-channel computer-assisted
DC-EEG amplifier [42].
Fig. 1: Movement-related potentials in different recording locations
of a subject performing flexions of the right index finger (N=1100 self-initiated
movements). Bereitschaftspotential (BP) precentro-parietally negative;
frontally positive. BP1 (early component of the BP) app. 1½ to ½
s prior to movement onset. BP2 (late component of the BP) ½ s through
movement onset (s 0, vertical line). Pre-motion positivity (PMP, from culmination
point 90 to 80 ms prior to movement onset in all centro-parietal leads,
type C-subject). Motor potential (MP) in bipolar recording L/R precentral
60 to 50 ms prior to movement onset. Reafferent potentials (RAP) after
movement onset. From Deecke 1974 [13]
Figure 1 gives an example of the BP preceding voluntary right index finger flexions. We see the slow negativity of the BP starting already 1½ s prior to the onset of movement (movement onset is at s 0, vertical line). We also see the two components of the BP. As indicated at the top of Fig. 1, we see an early, more shallow negative slope, BP1, and a late, steeper negative slope, BP2. Barrett et al. [4] reported on three components of the BP, introducing an 'intermediate component' between BP1 and BP2. However, we never found an 'intermediate conponent' and believe that the pre-voluntary movement period is best described postulating two principal BP components corresponding to two principal generators. It is our hypothesis, supported by many pieces of experimental evidence, that the BP1 principal generator is the mesial prefrontal cortex including the supplementary motor area (SMA) and probably also the cingulate motor area (CMA, see below), while the BP2 principal generator is the primary motor cortex (MI, area 4, precentral gyrus, 'motor strip', etc.).
The BP as such has an early start of 1 to 2 s prior to movement onset, for complex movements even as much as 3 s. For simple movements, as in Fig. 1, the BP is positive in frontal leads, and negative in central and parietal leads as well as at the vertex. According to theory of EEG, negativity can be related to activity of the cortical areas under study, while positivity is related to inactivity [11]. The BP voluntary movement paradigm investigates internally-produced movements, ie those performed voluntarily [12,14,20,21]. For these volitional movements we postulate an 'act of will' necessary to self-initiate the movement [36]. This certainly is a 'frontal act', and the decision of 'when to move' is thought to be elaborated by the mesial fronto-central cortex. This contains the SMA. However, there is evidence that not only SMA but also adjacent parts of the cingulate gyrus are activated prior to the onset of movement ('cingulate motor area', CMA) [47].
The situation is thus with unilateral voluntary movements. However,
what is the nature of bilateral movements? Experiments investigating unilateral
versus bilateral movements can tell us something about the temporal order
of the moments when activation starts in the two principal generators that
produce the BP, ie SMA as the principal generator for BP1 and MI as the
principal generator for BP2. When does activation start? There are three
possibilities: do both generators start activation at the same time, as
the Cleveland group believes [32], or is MI earlier than SMA, which nobody
believes or is SMA activation earlier than MI activation, which we believe
[20,22]. Our hypothesis that SMA is upstream in the final motor cascade
when it comes to channeling motivation, intention or the act of will into
motor execution is demonstrated in Fig. 2.
Fig. 2: Movement-related potentials preceding bilateral
voluntary movements.
A: Cerebral potentials preceding bilateral simultaneous simple monophasic index finger flexions. Monophasic means as in all our experiments in movement that the subject performed self-initiated brisk voluntary flexions of the two index fingers that remained in flexed position returning to indifferent position only during ITI (inter trial interval). Bereitschaftspotential (BP) earliest and largest at vertex. Note that at the fronto-central midline (vertex, Cz), negativity starts already at -1.1 s, while at the primary motor cortex (MI, C'3, C'4), the BP starts significantly later, at -0.7 s (movement onset as usual at 0 s, vertical line).128 artifact-free trials. Monopolar recordings against linked mastoids. C'3 or C'4 1cm anterior of C3 or C4. Detail of Fig. 1 from Kristeva and Deecke, 1980 [37]
B: Cerebral potentials preceding bilateral simultaneous voluntary
self-initiated index finger extensions with different loads. Note similar
finding as in A: Early BP onset over the supplementary motor area (SMA,
fronto-central midline, FCz and Cz) at -1.8 s; significantly later BP onset
over primary motor cortex (MI, C3 and C4) at -0.8 s. The generally earlier
BP onset times are due to the fact that the loaded movements are complex
movements as opposed to the simple movements of A. From experiments of
Cui, Huter, Lindinger, Lang and Deecke (1997)
In Fig. 2A, the BP preceding simple monophasic bilateral index finger flexions is shown. At the fronto-central midline (vertex, Cz), negativity already starts at -1.1 s, while at the primary motor cortex (MI, C'3, C'4), the BP starts significantly later, at -0.7 s (movement onset at 0 s, vertical line). In Fig. 2B a nice confirmation of such thinking can be seen: experiments in movement involving bilateral simultaneous voluntary self-initiated index finger extensions with different loads were carried out by Cui et al. (in preparation). We see an early BP onset over the supplementary motor area (SMA, fronto-central midline, FCz and Cz) at -1.8 s, and a significantly later BP onset over primary motor cortex (MI, C3 and C4) at -0.8 s. The generally earlier BP onset times are due to the fact that the loaded movements of Fig. 2B are complex movements as opposed to the simple movements of Fig. 2A. We have shown that BP onset times ('latencies') increase with the level of complexity of the movement, the longest onset times having been observed with speaking [19] or writing and drawing [46].
MEG
Experiments in movement using the magneto-encephalogram (MEG) started
in 1982. Generally, all graphoelements of the EEG can also be found in
the MEG. Thus, the equivalent of the Bereitschaftspotential BP has been
recorded in the MEG and has been termed Bereitschaftsfeld (BF) or readiness
field (RF) [17]. The advantage of MEG over EEG is better localization.
The BF late component (BF2) has been localized in MI (area 4) [17,18].
It reveals the typical somatotopic organization described by Penfield and
Rasmussen [44], ie for movements of different parts of the body, the equivalent
current dipoles corresponding to BF2 were distributed in a homuncular pattern
[10], from which Fig. 3 is taken).
Fig. 3: Homuncular organization of the Bereitschaftsfeld´s
(BF) or readiness magnetic field´s (RF) late component (BF2/RF2).
A: Three-dimensional dipole source locations for 7 different voluntary movement conditions projected onto lateral and posterior views of the subject´s head. Head outlines are drawn from digitized points taken over the sagittal and coronal midlines using the 3-D digitizing system also used to produce the head coordinate system. Dipole locations are indicated by filled circles and are numbered with respect to the voluntary movement performed, and the muscle from which the EMG trigger was recorded as indicated in the legend below (tibialis anterior, etc.).
B: Source locations projected onto a coronal magnetic resonance
image (MRI) taken from the same subject. A thin tube filled with contrast
agent (gadolinium) was placed over the anatomical landmarks used to establish
the head coordinate system (pre-auricular points and nasion) in order to
align the dipole coordinate system to that of the MR images. The MR plane
shown here lies parallel to the vertical-lateral plane of the head coordinate
system at the level indicated by the thin vertical line labelled `MRI'
anterior of the vertical line of the head coordinate system. Note homuncular
distribution of foot-, wrist-, digit-, face- and tongue movements indicating
that the generators of BF2/RF2 lie in primary motor cortex (area 4). From
Cheyne, Kristeva & Deecke, 1991 [10].
Figure 3A shows three-dimensional dipole source locations for seven
different movement conditions as indicated. Sources of neural activity
- identified using the non-invasive measurements of cerebral magnetic fields
(MEG) - were found to confirm the somatotopic organization of primary motor
cortex for movements of different parts of the body in normal human subjects.
The somatotopic map produced with this technique revealed slight differences
to the classic homunculus obtained from studies using invasive cortical
stimulation in epileptic patients [44]. In Fig. 3B these source locations
were projected onto a coronal magnetic resonance image (MRI) taken from
the same subject. The typical homuncular distribution of foot-, wrist-,
digit-, face- and tongue movements indicated that the generators of BF2/RF2
lie in area 4. An exception is the dipole for voluntary tongue movements.
It showed more variability due to mechanical triggering and other causes.
So far so good for the MI generator (BF2). MEG localization of the BF1 generator was less readily achieved. In the healthy subject, both SMA generators are also active in case of unilateral movement. Due to the anatomical localization of the SMA on the mesial surface of the two hemispheres, the two SMAs face and partially cancel eachother. Therefore, we made experiments in movement in a patient with a right SMA lesion [40]. Unilateral SMA lesion was selected in order to overcome the cancellation problem in MEG of the two opposing SMA dipoles in the normal subject. The patient having only one (the left) SMA, performed voluntary flexions with his right index finger. The results were quite convincing: in the early phase of the readiness magnetic field (1.2 to 0.8 s prior to the onset of movement, corresponding to BP1) field lines left the head at the vertex (Cz) and entered the head at a frontal position between F3 and Fz, thus enveloping an electric dipole on the mesial surface of the left hemisphere in the left (intact) SMA. A similar dipole was found during the period 0.8 to 0.6 s prior to the onset of movement, still corresponding to the BP1-SMA system. However, in the late phase of the readiness field (during BP2), in this case measured between 0.2 and 0 s prior to movement onset, activity had shifted: magnetic field lines now left the head at FC3 and entered it at FCz. This indicated an electrical dipole in the left area-4 hand representation (MI). This again supports our hypothesis that the SMA leads the MI motor cortex activation prior to human voluntary movement. We believe that any self-initiated movement is preceded by SMA activity [14] and largely based on our BP results[25].
Thus, we can state: Activation of SMA/CMA leads activation of MI in time. For this and other reasons we attribute the starting function of voluntary action to the SMA. Patients with chronic unilateral lesions of one SMA were examined using a voluntary movement paradigm involving different kinds of bimanual motor sequences [3]. Clinically, these patients presented with bradykinesia contralateral to the lesion, deceleration of initial movement, switching from sequential performance into simultaneity and frequent failure to initiate or inhibit a movement on one or other side. The clinical syndrome of SMA lesion is termed motor dysrhythmia or dyschronokinesia. BP (prior to movement) and negativity of performance (N-P, post movement onset) were larger over the lesioned hemisphere. This seemingly paradoxical result is due to the orientation of the dipole of the remaining SMA on the mesial surface of the hemisphere.
The bradykinesia of these frontal patients resembled that seen in basal ganglia disorder (Parkin-son's disease, PD). In the latter, the BPs are normal over the SMA but reduced and delayed over MI. On closer examination, in PD BP1 is reduced, while a normal BP2 helps the patient to pull up, at movement onset reaching almost the same amplitude as in normal controls - even overshooting after movement onset. The BP changes in PD as compared to healthy controls reveal that postsynaptic changes occur as well as presynaptic changes (dopamine depletion) in PD, ie dysfunction of cortical structures as a `hodological' consequence of the dopaminergic deficiency! That is why PD patients have trouble in self-initiation (action) while responding to external stimuli (re-action) is less impaired [15,16,24,31,48,49]. These postsynaptic dysfunctions seem to be partially reversible with L-Dopa substitution [24,27].
Analyses of BP topography revealed that the contribution of the SMA to the BP was reduced in PD [24]. PET studies confirmed a diminution of movement-related SMA activation [33,45]. In primate models of PD, it now became clear that the feedback circuits between cortex, basal ganglia, and thalamus are disturbed in this disease [1,2,23]; ie there is a lack of inhibition of the inhibitory action of the globus pallidum internum (Gpi) on the excitatory (glutamatergic) thalamo-cortical pathways. This dopaminergic lack of inhibition of inhibition is the intrinsic pathophysiology of the disease: Parkinson patients are akinetic because they are caught in the state of this enormous Gpi-`hyperinhibition', which in the untreated severe cases downregulates their motor capacity to a complete `rien ne va plus.'
Internal pallidotomy can markedly improve akinesia [23,38]. Why is it effective? Now we can understand, why: It removes this `all paralyzing' hyperinhibition from the thalamocortical output circuits. The SMA is one of the major projection areas of these thalamo-cortical circuits receiving even stronger input than MI. It has been shown that post pallidotomy, movement-related SMA activity as assessed by PET is restored in PD patients [9,28]. It is likely that the Parkinson-typical changes of the Bereitschaftspotential will also normalize post pallidotomy, though this has not been shown so far. However, it is clear that Parkinson akinesia is the result of dysfunction of a complicated feedback circuit system, in which not only the basal ganglia but also the SMA plays an important role.
Emission CT (SPECT)
Experiments in movement have also been carried out using emission CT. Of the two methods of this family (PET and SPECT), SPECT has been employed. Using the compound HMPAO that shows a trapping effect in brain tissue, rCBF as the one physiological parameter and movement-related DC-potentials as the other were measured in the same 17 normal Ss. A visuomotor tracking (T) paradigm was employed [39]. Trajectorial learning was required in a conflicting situation: a visual target moved on a screen and had to be track-ed by moving a light stylus in the right hand on a photo-detector plate in an inverted fashion (IT), eg movements of the target to the right side required hand movement to the left and vice versa. Compared to a normal non-inverted control task (T). IT required the development of a novel motor program and the prevention of returning to routine direct persuit.
These additional demands in IT caused a relative hyperperfusion in regions including the middle frontal gyri, frontomedial cortex (containing the SMA), right basal ganglia, and left cerebellum. Correlations of rCBF values between the middle frontal gyrus and basal ganglia indicated a functional relation between these two brain structures. Trajectorial performance was accompanied by slow negative DC-potential shifts before tracking (BP) and during tracking (negativity of performance, N-P). In frontolateral and frontomedial (containing the SMA) recordings, amplitudes of DC-negativity were higher in IT than they were in T. This additional frontal negativity covaried with the success in trajectorial learning. These results substantiated, using a dual approach, our previous suggestion that the frontal lobe plays an important role in trajectorial learning in providing a learning memory for novel motor skills. Cognitive skills relevant for performing music were investigated as well, also emphasizing a role of the SMA in timing.
Other studies dealt with the investigation of motor imagery, that is we are able to see something `in our mind´s eye.' For example athletes make use of motor imagery (`mental rehearsal') before the trial in order to improve performance. Their `mental representation of motor acts' is so good that they know beforehand whether a stroke for instance in golf will be a hit or a miss. We conducted a topographical study on the mental imagery of motor and other tasks [30,50].
The imagery paradigm. An experiment was carried out in which human mental activity was recorded and mapped through scalp electrodes using 12-channel DC-amplifiers (Fig. 4) [50].
The problem: Mental activity usually occurs spontaneously and unpredictably. The recording of cortical DC-potentials requires averaging. How can the averaging process be triggered by an unpredictable event?
Fig. 4: Experiments in movement imagery.
A: Imagery paradigm. Negative DC-shifts (upper trace) were recorded from 4 s before to 9 s after the volitional initiation of the trial (subjects presented themselves with a slide by self-initiated button presses with both index fingers simultanously). Centre trace, time marker indicating presentation of slide revealing the condition (COLOUR, FACE or MAP). The lower trace depicts the moment of the auditory item presentation (particular politician, or particular colour or route from 1 to 2 on the spatial map). The hatched area marks the time range from s 7 to 9 of epoch for calculating the mean amplitude of DC-negativity during mental imagery.
B: DC potential waveforms averaged across all 28 subjects for imagining COLOURS (left), FACES (right) and the spatial MAP (bottom). Negativity up. The vertical line represents the moment when subjects initiated the stimulus programme by pressing buttons with both index fingers simultaneously. Note that COLOURS created higher amplitudes (generators on the convexity) than did FACES (generators at the brain base, infratemporal and infraoccipital locations). Highest amplitudes with MAP due to occipito-parietal generators (separate visual pathways for object vision and spatial vision according to Mishkin et al. [43]). Modified from Uhl et al. 1990.
The solution: The S starts the trial by a self-initiated voluntary movement (preceded by a BP), ie the S presses buttons of a slide projector presenting himself with a slide. This initiating movement provides the trigger for time-locked potential averaging (at s 0 in Fig. 4). Instantaneously, a slide appears on the screen in front of the S for 200 ms (ending at s 0.2 ). By the contents of the slide, the S was instructed which of the 3 categories - colours, faces or a spatial map - would follow. After 3 s, a recorded voice announced the item that was to be imagined by mental imagery for a duration of 0.2 s (from s 3.2 to 4.4 in Fig. 4). Ss were instructed to create a visual image of the item, ie to see it in their mind's eye. They were asked to generate this image as soon as possible after they had heard the word and to hold it until the experimenter announced the end of the trial. This was done at irregular intervals in order to prevent the occurence of an expectancy wave (CNV) [51].
When taking the measurement of the cortical DC-level between s 7 and 9, one has a high probability that the cooperative S actually performs the mental task of imagery, which can then be recorded and mapped over the scalp. The results showed that the bilateral movement of pressing buttons, occuring at s 0, was preceded by a BP. This showed the usual central maximum and was bilaterally symmetrical. It was frontally absent, since no learning was involved in this paradigm. However, it was also present in occipital leads, where the BP is usually absent preceding simple finger movement tasks.
Thus, already in the distribution of the BP preceding it, the present visual task displayed itself in a modality-specific manner. After the onset of movement, a visual evoked potential in response to the self-presentation of the slide occured. Thereafter, we see a typical expectancy wave (CNV) in anticipation of the loudspeaker stimulus announcing the item to be imagined, followed by the auditory (verbal) evoked potential in response to the item presentation. While the early potential complex seemed to be rather normal, the late components appeared to be quite remarkable: the peak that appeared at s 4.4 (1.2 s after the onset of the loudspeaker signal) showed reversal; it was negative over frontal areas (F3 to C4) and positive over posterior areas (P3) at s 4.4. Thus, it pointed towards a generator process in deeper structures that is perhaps related to word comprehension and memory storage.
DC-shifts during imagery: Since Ss were instructed to start immediately imagining the item announced and maintain the image, the subsequent DC-potential shifts are undoubtedly related to creating and maintaining mental imagery. While over frontal and central areas, negative shifts declined, there was sustained DC-negativity over the retrorolandic brain, that is only at areas showing sustained negativity we assumed imagery to actually occur. Since frontal and central areas showed large initial DC-negativity it seems reasonable to assume that these areas, in particular frontal ones, participate in the generation of imagery. The maintainance of the imagery finally seems to occur in retrorolandic brain regions only, measurements being taken between s 7 and 9 (hatched area in Fig. 4). And here we found inter-esting topographical differences between the 3 categories of imagined items. The level of negativity was lowest for faces, medium for colours and highest for the map (Fig. 4). It is surprising that imagination of images that complex as faces generated less negativity than did imagining of colours (plain unicoloured sheets).
The explanation might be that face imagery is more inferotemporal than is that of colours and - in particular - the map, which is more parietal. Negative DC-shifts in basal cortical areas were not seen by our electrodes arranged over the convexity of the skull, and, what is more, these areas can even inject a positive bias on the DC-shift recorded via the linked ears reference electrodes. The fact that the map caused maximum overall DC-negativity also fits this assumption because now imagery particularly involved areas of the convexity, namely the parietal lobes. The cortical DC-potential (prevailing between s 7 and s 9 of the epoch - imagery period) was mapped using current source density mapping algorithms (Laplacian transformation and spline interpolation) [42]. With this display, the 2 experimental conditions involving visual imagery (faces and colours) differed from the condition spatial imagery (map). The faces´ and colours´ regional DC-negativities were distributed over the occipital cortex and over temporal areas, showing a preponderance over the left hemisphere. For the spatial map the topography was different, regional DC-negativity being distributed occipitally (but somewhat more anteriorly) and parietally (not temporally). A slight left preponderance was observed here also.
We conclude from these observations that visual imagery and spatial imagery differ in their regional DC-negativity pattern, the first having an occipital and temporal distribution the latter having an occipital and parietal topography. Since the same difference was found for visual and spatial perception, we considered our data as subserving further evidence for the postulate that imagery takes place at the same areas that elaborate the percepts.
HMPAO-SPECT and mental imagery
30 Ss - 28 of whom were identical with those of the DC-potential study - participated in the experiments [30]. The distribution of rCBF was assessed during a resting state and during imagining either colours or faces - visual imagery - or a route on a map -spatial imagery- similar as in the DC-potential experiment [50]. Ss were in a supine position, and were blindfolded. An intravenous line was laid to a cubital vein before the experiment. Ss wore earphones connected to a tape recorder, which announced the items to be imagined at a rate of 15 s. Imagery experiments were compared to resting state sessions during which the tape was silent and Ss were advised to lie relaxed and 'to think of nothing'.
The results showed that - as compared to the resting state - the imagination of faces caused distinctly more regional activation than with the other two imagery conditions. Statistically, imagining faces was also the only condition that led to an increase of activity in - particularly the left - inferior occipital region which has been suggested by previous studies as being a crucial area for visual imagery [5,26,29,30,50]. Also for colours, basal temporal and occipital regions are activated but less so than for faces.
The imagery of routes on a map, on the contrary, showed a tendency towards more activation in visual association areas on the convexity of the brain, particularly in parietal areas. This is another piece of evidence for a 'visual/spatial dichotomy' as suggested by Mishkin et al. [43]. We now know that the visual modality has many submodalities such as form, disparity, colour, motion perception and others that are envisaged of being represented in different visual subsystems, some of which occupying distinct areas of the visual association cortex V1 to V7. Mishkin's distinction between object vision and spatial vision being represented in two separate visual pathways is a particularly intriguing concept which is partially corroborated by the two imagery studies reported here.
Furthermore, the finding in our SPECT study [30] that imagining faces involves basal cortical areas to a greater extent than does colour and map imagery - the latter rather activating the parietal convexity - can explain the above paradoxical result of our DC-potential study [50] that the imagination of faces yielded less overall negativity than did colours. Since the information content of faces is by orders of magnitude higher than the one of plain unicoloured sheets, this paradox can be explained by the greater degree of 'basality' (inferotemporal, inferooccipital) of faces as compared to colours. It is important to consider that there are large cortical regions which the electroencephalographer does not see. This again calls for the joint studies. From the SPECT study on imagery [30] we can conclude as well that imagery occurs in those brain areas where perception takes place.
FMRI
The technique of functional magnetic resonance imaging allows the measurement of activation-related cerebral blood flow changes occuring with specific tasks. However, the spatial relationship between neuronal activity and functional cerebral blood flow changes is not known yet. A study was carried out in order to compare the centre of neuronal activation (measured by MEG) with that of the blood flow responses (measured by FMRI). Eight Ss participated in a typical BP paradigm (ie voluntary movement paradigm) of tapping with their right index finger [7,8]. The results are shown in Fig. 5.
Fig. 5: Experiments in movement using FMRI and MEG.
Coronal section of MRI of a subject. Right is left. The common centre
of neuronal activation (which is the mean of the motor and sensory cortex
MEG centres) is shown by the triangle. The centre of the blood flow response
is shown by the square. Note that `neuronal signal' and `cerebrovascular
signal' are not identical. From: Beisteiner et al. [7]
Experiments in movement are shown using FMRI and MEG. Fig. 5 shows a coronal section of an MRI of a subject. The common centre of neuronal activation (which is the mean of the motor and sensory cortex MEG centres) is shown by the triangle. The centre of the blood flow response is shown by the square. As can be seen, the `neuronal signal' (MEG) and the `cerebrovascular signal' (FMRI) are not at the same location. Rather a difference in topography between the two signals is seen. On the average, the two signals were about 1½ cm separate from eachother, which calls for methodological improvements, ie both signals lay in the left MI area but 1½ cm apart from each other. One source of localization error may be pixels with large signal amplitudes, since these pixels may be expected to stem from larger vessels that may be even remote from the centre of neuronal activation (large vessel effect). Using MEG - thought to be very close to the `neuronal signal' - in addition to the `cerebrovascular signal' (FMRI) in a simple finger tapping task, such multimethodological approach should help to improve FMRI brain mapping results. It was in favour of this hypothesis that results showed, indeed, a deterioration of FMRI localization quality with increasing signal amplitudes indicating increased contribution of the large vessel effect. Thus, for FMRI evaluation it is recommended pixels with large signal amplitudes be excluded in order to reduce the large vessel effect, and to get closer to the neuronal signal, which is the important parameter in functional topography of the brain.
REFERENCES