Potential impact of a 32-channel receiving head coil technology on the results of a functional MRI paradigm

Original article

AbstractPurpose the authors investigated the potential of a 32-channel (32ch) receiving head coil for functional magnetic resonance imaging (fMri) compared to a standard eight-channel (8ch) coil using a motor task.Material and Methods Brain activation was analyzed in 14 healthy right-handed subjects performing finger tap-ping with the right index finger (block design) during two experimental sessions, one with the 8ch and one with the 32ch coil (applied in a pseudorandomized order). addition-ally, a phantom study was performed to compare signal-to-noise ratios (Snrs) of both coils.Results During both fMri sessions, analysis of motor con-ditions resulted in an activation of the left “hand knob” (precentral gyrus). application of the 32ch coil obtained additional activation clusters in the right cerebellum, left superior frontal gyrus (SMa), left supramarginal gyrus, and left postcentral gyrus. the phantom study revealed a significantly higher SNR for the 32ch coil compared to the 8ch coil in superficial cortical areas located near the sur-face of the brain.

Conclusion the 32ch technology has a potential impact on fMri studies, especially in paradigms that result in activa-tion of cortical areas located near the surface of the brain.

Keywords  Finger tapping · fMri · Motor paradigm · Phantom study · Signal-to-noise ratio

Potentieller Einfluss einer 32-Kanal-Empfangsspulen-Technologie auf die Ergebnisse eines funktionellen MRT-Paradigmas

ZusammenfassungZiel In diesem Artikel wurde der Einfluss einer 32-Kanal-Empfangsspule im Vergleich zu einer Acht-Kanal-Stan-dardspule auf die ergebnisse einer funktionellen Magnetre-sonanztomographie-(fMrt-)Studie, in der ein motorisches Paradigma genutzt wurde, untersucht.Material und Methodik Die Hirnaktivierung von 14 gesun-den rechtshändern, die eine Fingertapping-aufgabe mit dem rechten Zeigefinger (Blockdesign) ausführten, wurde während zweier experimenteller Durchläufe in pseudoran-domisierter abfolge untersucht. Während eines Durchlaufs wurde eine 32-Kanal-Spule, während eines zweiten Durch-laufs eine Acht-Kanal-Spule benutzt. Außerdem wurde eine Phantomstudie durchgeführt, um das Signal-Rausch-Verhältnis beider empfangsspulen zu vergleichen.Ergebnisse Die analyse der Motorbedingung während beider Durchläufe resultierte in einer aktivierung des lin-ken Handareals (gyrus praecentralis). Die Verwendung der 32-Kanal-Spule resultierte in einer Aktivierung von zusätz-lichen Hirnarealen wie dem rechten cerebellum, linken gyrus frontalis superior (SMa), linken gyrus supramar-ginalis und linken gyrus postcentralis. Die Phantomstudie zeigte ein signifikant höheres Signal-Rausch-Verhältnis in

clin neuroradiol (2010) 20:223–229DOi 10.1007/s00062-010-0029-2

Potential Impact of a 32-Channel Receiving Head Coil Technology on the Results of a Functional MRI Paradigm

J. Albrecht · M. Burke · K. Haegler · V. Schöpf · A. M. Kleemann · M. Paolini · M. Wiesmann · J. Linn

J. albrecht () · K. Haegler · V. Schöpf · A. M. Kleemann · M. Paolini · M. Wiesmann · J. linnDepartment of neuroradiology, ludwig Maximilians University, Munich, germanye-mail: [email protected]

M. Burkege Healthcare, Solingen, germany

M. WiesmannDepartment of Diagnostic and interventional neuroradiology, rWtH, aachen, germany

received: 5 april 2010 / accepted: 10 august 2010 / Published online: 21 September 2010© Urban & Vogel 2010

224 J. albrecht et al.

Arealen nahe der Hirnoberfläche für die 32-Kanal-Spule im Vergleich zur Acht-Kanal-Spule.Schlussfolgerung Die 32-Kanal-Empfangsspulen-Techno-logie hat einen potentiellen Einfluss auf die Ergebnisse von fMrt-Studien, im Speziellen auf Paradigmen, die in Hirnaktivierung von kortikalen arealen nahe der Hirnober-fläche resultieren.

Schlüsselwörter Fingertapping · fMrt · Motorisches Paradigma · Phantomstudie · Signal-rausch-Verhältnis

Introduction

as the data acquisition and processing capabilities of modern magnetic resonance imaging scanners advance, there is a trend toward the use of multielement coils with a higher number of coil elements for magnetic resonance imaging (Mri) in general [1–11] and functional magnetic resonance imaging (fMri) in particular [12]. Two possible benefits for the use of higher-element coils have been discussed in the scientific community. On the one hand, they should enable higher acceleration factors for parallel imaging techniques [1, 3, 4, 8, 9] which, in turn, lead to shorter examination times, reduction of acoustic scanner noise as well as redu-ced susceptibility artifacts for fMri due to shorter echo train durations during echo-planar imaging (ePi) [13, 14]. On the other hand, one expects a higher signal—expressed as signal-to-noise ratio (Snr) [2, 3, 6, 10, 12]—which would be reflected in the ability of improved object discrimination holding different signal intensities.

a major limitation of the fMri methodology is the extre-mely low physiological Snr of the blood oxygen level-dependent (BOlD) signal which is caused by physiological signal variations. We do not expect that the use of a better head coil technique will reduce physiological noise of the BOlD signal, however, since it is known that the thermal noise is larger than the physiological noise in functional images with a high spatial resolution, an improvement of the SNR would be particularly beneficial for those fMRI sequences [15, 16].

to evaluate new Mr techniques [17] and to perform fMri experiments comparing brain activation of healthy subjects and patients [18–21] in a standardized way, finger tapping tasks are widely used as standard motor fMri para-digms [22–24].

Here we aimed to (1) evaluate the potential of a 32-chan-nel receiving (32ch) head coil in comparison to a standard eight-channel (8ch) coil for fMri using a standard motor paradigm, and to (2) compare Snrs for both coils during phantom measurements.

 Material and Methods

Design of Head coils

the 8ch coil utilized during this experiment was the stan-dard head coil of a 3-t clinical Mri scanner (Signa HDx, ge Healthcare, Milwaukee, Wi, USa).

the 32ch coil was manufactured by Mr instruments (Minnetonka, Mn, USa) and consisted of two rings of 16 parallel-arranged array elements, which were axially spaced for imaging in the xy-plane (Fig. 1). each of the parallel array elements was isolated from right to left neighboring elements, as well as respective superior or inferior neighbo-ring elements. For the design, low-input impedance pream-plifiers were used to deliver preamplifier isolation to each of the coil elements to further reduce coupling between ele-ments that cannot make use of traditional isolation methods (e.g., geometric overlap). Such an arrangement of coil ele-ments provides the basis for parallel acceleration factors to be applied in both the xy- and z-planes. element size was dictated by an empirical study between the trade-off of the coil’s patient cavity size and achievable Snr at the center of the imaging volume. this resulted in an element loop size of approximately 65 cm2. the 32ch coil had an inner diameter of 23.6 cm and an outer diameter of 29.4 cm.

Participants

Fourteen volunteers (six females, eight males, mean age 28.0 years, standard deviation [SD] 5.7 years, range 21–43 years) participated in the experiment. all participants

Fig. 1 32ch receiving head coil

225Potential impact of a 32-channel receiving Head coil technology on the results of a Functional Mri Paradigm

were right-handed (laterality quotient: mean = 98.6, SD = 5.4 [25]), healthy, and not taking any medication. all of them had given their written informed consent to participate in the experiment prior to the beginning of the study. the local ethics committee has approved the study. therefore, the study meets the criteria of the Declaration of Helsinki.

Data acquisition

fMRI Experiment

Subjects were lying in the Mr scanner with their head carefully fixed and their eyes closed. They were instruc-ted to perform intermittent finger tapping with a frequency of approximately 1 Hz with their right index finger on a response pad (lumina lP-400 response Pads for fMri, cedrus corporation, San Pedro, ca, USa) during the fMri session. Subjects were told to begin and end the tapping fol-lowing commands (german monosyllabic words “Start” and “Stopp”, for “start” and “stop”, respectively) given through Mri compatible earphones. eight alternating blocks of “tapping” and “no tapping” followed the first condition (“no tapping”). the duration of each condition was 24 s.

Functional images were acquired in two series on a 3-t standard clinical Mr scanner (Signa HDx, ge Healthcare, Milwaukee, Wi, USa) with 32 receiving channels. Functio-nal images were obtained using echoplanar imaging (ePi) with a t2*-weighted gradient-echo multislice sequence (echo time [te] = 35 ms, repetition time [tr] = 2,100 ms, flip angle = 90°, field of view (FOV) = 240 mm, voxel size = 3.75 × 3.75 × 4.00 mm3, matrix size = 64 × 64, array Spatial Sensitivity encoding technique (aSSet)–parallel accele-ration factor 2). During one series, the 8ch coil, during the other series, the 32ch coil was used. the order of both series was pseudorandomized. thirty-seven slices covering the whole brain of the subject were positioned parallel to the connection line between anterior commissure and posterior commissure utilizing a sagittal localizer image. Within each fMri series, 200 image volumes were acquired, whereof the first five images of each image series were discarded to eli-minate magnetization equilibration effects. the acquisition time of each fMri series was approximately 7 minutes.

Since coil architecture of the 32ch coil did not allow for sliding ear pads in order to fixate the subject’s head, we instead used a special cushion of the size and shape of a head (see Fig. 1). additionally, we used small cushions that were brought into the coil on both sides of the head to faci-litate head fixation.

after each functional series, a structural image volume was acquired using a t1-weighted three-dimensional fast spoiled gradient-echo recalled (FSPgr) sequence (te = 3.2 ms, tr = 7.2 ms, flip angle = 15°, slice thick-ness = 1.2 mm, overlap = 0.6 mm, FOV = 220 mm, matrix

size = 288 × 288, neX = 1, aSSet–parallel acceleration fac-tor 2). acquisition time of the structural series was approxi-mately 8 minutes.

Phantom Studies

For the phantom studies, a standard head phantom (Model 2359877, Dielectric corp., raymond, Me, USa) was used. Prior to the beginning of the phantom study, four fiducial markers were positioned onto the phantom in order to be able to correctly place the phantom in the center of the respective coil for both runs. For Snr measurements, a series of 200 volumes was acquired (ePi, te = 40 ms, tr = 2,000 ms, matrix size 64 × 64, FOV = 240 mm × 240 mm, flip angle 90°, slice thickness = 4.0 mm, 27 slices, aSSet–parallel acceleration factor 2).

Data analysis

fMRI Experiment

imaging data were processed using statistical parametric mapping (SPM5) [26] implemented in Matlab 7.5 (the MathWorks inc., Sherborn, Ma, USa). Motion correction was performed by realigning each volume to the first of each scanning series [27]. afterwards, the structural image volume was co-registered to the mean image of the corre-sponding functional series. the structural image volume was segmented and the parameters for spatial normalization into the standard space defined by the Montreal Neurologi-cal institute (Mni) were estimated [26]. these parameters were subsequently used to reslice the functional and the structural images. the resulting voxel size of the functional image volumes was 3 × 3 × 3 mm3. Datasets were smoothed using an 8-mm full-width at half-maximum (FWHM) iso-tropic gaussian kernel to compensate for individual gyral variability and to attenuate high-frequency noise in order to improve Snr. For single subject analyses, statistical parametric maps were calculated using the general linear model (glM) [26] with regressors corresponding to the onset times of the tapping blocks convolved with the cano-nical hemodynamic response function (HrF). additionally, individual realignment parameters accounting for potential head movement were included as regressors. as primary contrast, the activation in response to finger tapping was investigated. the resulting contrast images of the subjects underwent a random-effects group analysis. P-values of the random-effects group analysis were corrected using family-wise error rate (p < 0.05 FWe correction). FWe correction procedures are powerful Bonferroni-related procedures cor-recting for multiple comparisons across whole brain volume that are described in detail elsewhere [28]. contrast estima-tes within the precentral gyrus activation (–36, –12, 48) of

226 J. albrecht et al.

both random-effects group analyses were calculated using SPM5. Paired t-tests were used to compare the effects of fin-ger tapping between the groups “8ch coil” and “32ch coil” (p < 0.05 FWe correction).

Areas of significant activation at the group level were anatomically categorized using the Mni coordinates and the parcellation method along with the anatomy toolbox for SPM5 described by eickhoff et al. [29]. The defined struc-tural sites of activation clusters were then verified using an anatomical brain atlas [30].

in order to evaluate physiological Snr, β-values for both, the tapping condition and the constant (rest or base-line condition) in left precentral gyrus (–36, –12, 48) were investigated on single-subject level. SPSS (version 16.0 for Windows, SPSS inc., chicago, il, USa) was used for sta-tistical evaluation of these values, which were submitted to a paired t-test. the α-level for this test was set at 0.05.

Phantom Studies

For calculation of Snr maps, the signal intensity time course for each voxel was detrended by fitting a second-order polynom to the time course and subtracting the first- and second-order terms from the measured signal intensities. in order to obtain Snr maps of the images from each coil, the standard deviation of the detrended signal intensity time course of each voxel served as noise level. this was done

because signal intensity contribution of regions of interest (rOis) outside the objects might be differently weighted by the different coil elements rendering coil comparisons impossible. therefore, for Snr data the mean signal inten-sity of the detrended time course of a voxel was divided by its standard deviation and thereby normalized. Mean Snr was calculated for concentric spheres for each of the two coils and plotted against the spheres’ distance to the surface of the phantom. the center of each concentric sphere was equidistantly placed at radii 1.125 cm, 3.375 cm, 5.625 cm, and 7.875 cm, respectively.

Results

fMri experiment

the random-effects group analysis of the fMri experiment performed with the 8ch coil yielded activation only in the left precentral gyrus (“hand knob”, four activated voxels; p < 0.05, FWe; Fig. 2, table 1). By utilizing the 32ch coil, an activation cluster of 36 voxels in the left precentral gyrus was obtained. in addition, this coil also yielded activation clusters in the right cerebellum (lobule 6), the left superior frontal gyrus (supplementary motor area [SMa]), the left supramarginal gyrus, and the left postcentral gyrus (p < 0.05, FWe; Fig. 2, table 1).

Fig. 2 fMri activation associa-ted with finger tapping during the session utilizing the 8ch coil (a) and the 32ch coil (b). acti-vation maps showing significant increases in BOlD signal obtai- ned by statistical group analysis for the contrast finger tapping using SPM5. activations are projected onto a standard template brain (group analysis, n = 14; p < 0.05 FWe-corrected for whole brain volume; L left). Sagittal, coronal and axial slices are shown

Local activationmaximum:left precentral gyrus(–36, –12, 48)

Local activationmaximum:left precentral gyrus(–36, –12, 48)

8

6

4

2

0

8

10

6

4

2

0

a 8-channel head coil b 32-channel head coil

L

L

L

L

227Potential impact of a 32-channel receiving Head coil technology on the results of a Functional Mri Paradigm

the paired t-test utilizing the contrast “8ch coil > 32ch coil” and the contrast “32ch coil > 8ch coil” revealed no results on the p < 0.05 FWe-corrected level.

contrast estimates within the precentral gyrus activa-tion of both random-effects group analyses (–36, –12, 48) were 1.01 (standard error of the mean [SeM] = 0.12) of the session utilizing the 8ch coil, and 1.08 (SeM = 0.10) of the session utilizing the 32ch coil. the t-value in this activation maximum was 8.63 for the 8ch coil session and 10.70 for the 32ch coil session.

Mean signal intensity change in the precentral gyrus acti-vation (–36, –12, 48) was 0.63 (SeM = 0.08) for the session utilizing the 8ch coil and 0.75 (SeM = 0.07) for the session utilizing the 32ch coil. These values did not differ signifi-cantly [t(13) = 1.71; p = nS].

Phantom Studies

Mean Snr determined by averaging the obtained Snr values within a rOi including the whole object of the 32ch coil did not differ from that of the 8ch coil (727 [SD = 195] vs. 758 [SD = 212], t(423) = 1.56, p = nS). However, the Snr maps showed spatial differences with respect to the distance of the voxels from the surface of the coils. With regard to the 32ch coil, SNR was significantly higher for regions close to the coils’ elements (radial position 1; p < 0.001), whereas it was lower for regions close to the center of the coil com-pared to the 8ch coil (radial positions 3 and 4; p < 0.001; Fig. 3). Furthermore, for the 32ch coil there was a stronger Snr drop in voxels at the rim of the phantom compared to voxels located in the object center.

Discussion

During the fMRI experiment, tapping of the right index fin-ger induced brain activation in the contralateral precentral gyrus independent of the utilized receiver coil. this result is

in accordance with the results of previous functional brain imaging studies investigating neural correlates related to motor tasks of the fingers or the hand [18, 22, 23, 31–36]. in addition to this, activations in the ipsilateral cerebellum, the contralateral superior frontal gyrus (SMa), the supra-marginal gyrus, and the postcentral gyrus have been obser-ved using the 32ch coil. all of these areas have also been reported in at least one of the above-mentioned studies on finger tapping.

Based on visual inspection of the results of the random-effects group analyses, application of the 32ch coil resul-ted in more activated brain areas as well as larger clusters of brain activation. However, comparing signal intensity change in an rOi (precentral gyrus) on the individual level,

Table 1 fMRI activation associated with finger tapping during the session utilizing the 8ch coil (random-effects group analysis, n = 14; p < 0.05 FWe-corrected for whole brain volume)Brain region Mni coordinates (mm) activated

voxels (n)Maximum Z

X y z8ch coilleft precentral gyrus −36 −12 48 4 4.9032ch coilright cerebellum (lobule 6) 33 −51 −27 46 5.43

18 −54 −24 4.95left precentral gyrus −36 −12 48 36 5.36left superior frontal gyrus (supplementary motor area) −3 6 60 4 5.20left supramarginal gyrus −45 −30 21 14 5.18left postcentral gyrus −51 −12 45 9 5.178ch eight-channel; 32ch 32-channel, fMRI functional magnetic resonance imaging, FWE family-wise error rate; MNI Montreal neurological institute

Fig. 3 Dependence of Snr on the concentric spheres of the 8ch and 32ch head coil. the colors indicate the different concentric spheres for which the Snr was calculated. the center of each concentric sphere was equidistantly placed at radii 1.125 cm, 3.375 cm, 5.625 cm, and 7.875 cm, respectively. Snr is indicated in arbitrary units (AU)

850

800

750

700

650

600

SN

R

1.0 1.5 2.0 2.5 3.0 3.5 4.0

Distance to surface (AU)

32ch8ch

228 J. albrecht et al.

no difference between the data of both sessions could be observed. therefore, as expected, the physiological Snr observed in this regions did not differ significantly between the two coils, representing the reliability and stability of our results across the two coils.

regarding the phantom study and the comparison of Snr distribution across the diameter of the phantom, we found that, on average, Snr between images obtained by both coils did not differ significantly for images obtained by the 32ch coil compared to that obtained by the 8ch coil for Mri parameters used in this study. However, Snr varied in radial direction for images from both coils. this is usually observed within multielement coils manufactured from mul-tiple surface coils [3, 6, 10, 12]. Voxels close to the surface of the 32ch coil had a stronger contribution to the detected signal than voxels at the center of the coil. Our results show that the 32ch coil was more sensitive in cortical areas near the surface of the coil. this might be ascribed to the fact that the smaller elements of the 32ch coil (in comparison to the 8ch coil) are more sensitive for the detection of signals arising from smaller object volumes. thus, it might be less sensitive to contributions of noise arising from voxels far away from the volume the respective coil element is most sensitive to.

the results of the phantom study corresponded well to the fMRI results, i.e., the higher confidence level for corti-cal activation observed with the 32ch coil during the motor paradigm. as expected, the motor task resulted in activati-ons of cortical areas, which are predominantly located near the surface of the brain (e.g., the precentral gyrus and the SMa). these are the areas in which the 32ch coil yielded a significantly better SNR compared to the 8ch coil. In con-clusion, the changes detected between both coils after com-parison of the two random-effects analyses and observing the results of the paired t-test reflect the differences in sen-sitivity of the two coils for the same physiological BOlD signal intensity changes.

These results provide a valuable benefit for fMRI studies performed for presurgical planning in patients with tumors, e.g., in the region adjacent to the motor cortex [37–39], which is a rather common clinical application of fMri. Furthermore, the potential of the coil to provide better Snr in areas near its surface could be especially useful in studies with event-related designs, which typically cope with extre-mely low Snrs [40]. these issues should be addressed in further studies.

One important limitation of this study is that we tested only one 32ch coil, with a specific coil design. Thus, the results cannot necessarily be transferred to 32ch coils provi-ded by other manufacturers.

Conclusion

the 32ch coil has the potential to yield a higher statistical significance of brain activation in fMRI studies, especially in cortical areas near the coil’s surface. Based on the results of the phantom study, we must conclude, that the 32ch coil presented here, might be of limited value for fMri studies investigating deep brain regions. thus, the decision about the usage of the 32ch coil should be made in dependence of the cortical region of interest.

 Acknowledgment We thank Dr. thomas Stephan for invaluable comments on the manuscript and proofreading the manuscript.

Conflict of Interest Statement  the authors declare that there is no actual or potential conflict of interest in relation to this article.

the 8ch and the 32ch coil were provided by ge Healthcare, Solingen, germany. the authors who were not employees of ge Healthcare had full control of inclusion of any data and information that might have presented a conflict of interest for those authors who were employees of ge Healthcare at every time point of the study.

References

1. Otazo r, tsai SY, lin FH, Posse S. accelerated short-te 3D pro-ton echo-planar spectroscopic imaging using 2D-SenSe with a 32-channel array coil. Magn reson Med. 2007;58:1107–16.

2. tsai SY, Otazo r, Posse S, lin Yr, chung HW, Wald ll, Wiggins gc, lin FH. accelerated proton echo planar spectroscopic ima-ging (PePSi) using graPPa with a 32-channel phased-array coil. Magn reson Med. 2008;59:989–98.

3. Wiggins gc, triantafyllou c, Potthast a, reykowski a, nittka M, Wald ll. 32-channel 3 tesla receive-only phased-array head coil with soccer-ball element geometry. Magn reson Med. 2006;56:216–23.

4. Zhu Y, Hardy CJ, Sodickson DK, Giaquinto RO, Dumoulin CL, Kenwood G, Niendorf T, Lejay H, McKenzie CA, Ohliger MA, rofsky nM. Highly parallel volumetric imaging with a 32-ele-ment rF coil array. Magn reson Med. 2004;52:869–77.

5. McDougall MP, Wright SM. 64-channel array coil for single echo acquisition magnetic resonance imaging. Magn reson Med. 2005;54:386–92.

6. de Zwart Ja, ledden PJ, van gelderen P, Bodurka J, chu r, Duyn JH. Signal-to-noise ratio and parallel imaging performance of a 16-channel receive-only brain coil array at 3.0 tesla. Magn reson Med. 2004;51:22–6.

7. McDougall MP, Wright SM. investigation of coil phase compen-sation in 3D imaging at very high acceleration factors. J Magn reson imaging. 2007;25:1305–11.

8. Sodickson DK, Hardy CJ, Zhu Y, Giaquinto RO, Gross P, Ken-wood G, Niendorf T, Lejay H, McKenzie CA, Ohliger MA, Grant AK, Rofsky NM. Rapid volumetric MRI using parallel imaging with order-of-magnitude accelerations and a 32-ele-ment rF coil array: feasibility and implications. acad radiol. 2005;12:626–35.

9. Hardy cJ, Darrow rD, Saranathan M, giaquinto rO, Zhu Y, Dumoulin CL, Bottomley PA. Large field-of-view real-time MRI with a 32-channel system. Magn reson Med. 2004;52:878–84.

229Potential impact of a 32-channel receiving Head coil technology on the results of a Functional Mri Paradigm

10. Porter Jr, Wright SM, reykowski a. a 16-element phased-array head coil. Magn reson Med. 1998;40:272–9.

11. Schoth F, Kraemer N, Niendorf T, Hohl C, Gunther RW, Krom-bach ga. comparison of image quality in magnetic resonance imaging of the knee at 1.5 and 3.0 tesla using 32-channel receiver coils. eur radiol. 2008;18:2258–64.

12. Bodurka J, ledden PJ, van gelderen P, chu r, de Zwart Ja, Morris D, Duyn JH. Scalable multichannel Mri data acquisition system. Magn reson Med. 2004;51:165–71.

13. de Zwart JA, van Gelderen P, Kellman P, Duyn JH. Application of sensitivity-encoded echo-planar imaging for blood oxygen level-dependent functional brain imaging. Magn reson Med. 2002;48:1011–20.

14. de Zwart JA, van Gelderen P, Kellman P, Duyn JH. Reduction of gradient acoustic noise in Mri using SenSe-ePi. neuroimage. 2002;16:1151–5.

15. Frederick Bd, Wald ll, Maas lc 3rd, renshaw PF. a phased array echoplanar imaging system for fMri. Magn reson ima-ging. 1999;17:121–9.

16. Hesselmann V, girnus r, Wedekind c, Hunsche S, Bunke J, Schulte O, Sorger B, Lasek K, Krug B, Sturm V, Lackner K. Functional Mri using multiple receiver coils: BOlD signal chan-ges and signal-to-noise ratio for three-dimensional-PreStO vs. single shot ePi in comparison to a standard quadrature head coil. J Magn reson imaging. 2004;20:321–6.

17. Kim SG, Tsekos NV, Ashe J. Multi-slice perfusion-based func-tional Mri using the Fair technique: comparison of cBF and BOlD effects. nMr Biomed. 1997;10:191–6.

18. Brooks BR, Bushara K, Khan A, Hershberger J, Wheat JO, Bel-den D, Henningsen H. Functional magnetic resonance imaging (fMri) clinical studies in alS—paradigms, problems and pro-mises. amyotroph lateral Scler Other Motor neuron Disord. 2000;1 Suppl 2:S23–32.

19. Lee HK, Kim JS, Hwang YM, Lee MJ, Choi CG, Suh DC, Lim tH. location of the primary motor cortex in schizencephaly. aJnr am J neuroradiol. 1999;20:163–6.

20. Marshall RS, Perera GM, Lazar RM, Krakauer JW, Constantine rc, DelaPaz rl. evolution of cortical activation during reco-very from corticospinal tract infarction. Stroke. 2000;31:656–61.

21. Chen XK, Xiao YY, Zheng WB, Chen FY, Wu RH. Functional magnetic resonance mapping of motor cortex in patients with mass lesions near primary motor and sensory cortices. conf Proc ieee eng Med Biol Soc. 2006;1:1877–80.

22. Gordon AM, Lee JH, Flament D, Ugurbil K, Ebner TJ. Functio-nal magnetic resonance imaging of motor, sensory, and posterior parietal cortical areas during performance of sequential typing movements. exp Brain res. 1998;121:153–66.

23. Kawashima R, Itoh H, Ono S, Satoh K, Furumoto S, Gotoh R, Koyama M, Yoshioka S, Takahashi T, Takahashi K, Yanagisawa T, Fukuda H. Changes in regional cerebral blood flow during self-paced arm and finger movements. A PET study. Brain Res. 1996;716:141–8.

24. Habas c, axelrad H, cabanis ea. the cerebellar second homun-culus remains silent during passive bimanual movements. neuro-report. 2004;15:1571–4.

25. Oldfield rc. the assessment and analysis of handedness: the edinburgh inventory. neuropsychologia. 1971;9:97–113.

26. Friston KJ, Holmes AP, Worsley KJ, Poline JP, Frith CD, Fracko-wiak rSJ. Statistical parametric maps in functional imaging: a general linear approach. Hum Brain Mapp. 1994;2:189–210.

27. Friston KJ, Williams S, Howard R, Frackowiak RS, Turner R. Movement-related effects in fMri time-series. Magn reson Med. 1996;35:346–55.

28. Brett M, Penny WD, Kiebel S. An introduction to random field theory. in: Frackowiak rSJ, ashburner Jt, Penny WD, Zeki S, Friston KJ, Frith CD, Dolan RJ, Price CJ, editors. Human brain function. 2nd ed. San Diego: elsevier; 2004. p. 867–879.

29. Eickhoff SB, Stephan KE, Mohlberg H, Grefkes C, Fink GR, Amunts K, Zilles K. A new SPM toolbox for combining proba-bilistic cytoarchitectonic maps and functional imaging data. neu-roimage. 2005;25:1325–35.

30. Mai JK, Assheuer J, Paxinos G. Atlas of the human brain. Amster-dam: elsevier academic Press; 2004.

31. Yousry ta, Schmid UD, alkadhi H, Schmidt D, Peraud a, Buett-ner a, Winkler P. localization of the motor hand area to a knob on the precentral gyrus. a new landmark. Brain. 1997;120:141–57.

32. van gelderen P, ramsey nF, liu g, Duyn JH, Frank Ja, Wein-berger Dr, Moonen ct. three-dimensional functional magnetic resonance imaging of human brain on a clinical 1.5-t scanner. Proc natl acad Sci U S a. 1995;92:6906–10.

33. rao SM, Binder Jr, Bandettini Pa, Hammeke ta, Yetkin FZ, Jesmanowicz, a, lisk lM, Morris gl, Mueller WM, estkow-ski lD, Wong ec, Haughton VM, Hyde JS. Functional magnetic resonance imaging of complex human movements. neurology. 1993;43:2311–8.

34. roland Pe, larsen B, lassen na, Skinhoj e. Supplementary motor area and other cortical areas in organization of voluntary movements in man. J neurophysiol. 1980;43:118–36.

35. Nakai T, Kato C, Glover GH, Toma K, Moriya T, Matsuo K. A functional magnetic resonance imaging study of internal modu-lation of an external visual cue for motor execution. Brain res. 2003;968:238–47.

36. Rubia K, Overmeyer S, Taylor E, Brammer M, Williams S, Simmons a, andrew c, Bullmore e. Prefrontal involvement in “temporal bridging” and timing movement. neuropsychologia. 1998;36:1283–93.

37. Tieleman A, Deblaere K, Van Roost D, Van Damme O, Ach-ten e. Preoperative fMri in tumour surgery. eur radiol. 2009;19:2523–34.

38. chakraborty a, Mcevoy aW. Presurgical functional mapping with functional Mri. curr Opin neurol. 2008;21:446–51.

39. Moller M, Freund M, greiner c, Schwindt W, gaus c, Heindel W. real time fMri: a tool for the routine presurgical localisation of the motor cortex. eur radiol. 2005;15:292–5.

40. Haller S, Bartsch aJ. Pitfalls in fMri. eur radiol. 2009;19: 2689–706.