Multimodal imaging of human early visual cortex by combining functional and molecular measurements with fMRI and PET

www.elsevier.com/locate/ynimg NeuroImage 41 (2008) 204 – 211

Multimodal imaging of human early visual cortex by combining functional and molecular measurements with fMRI and PET Florian Gerstl,a Christian Windischberger,a Markus Mitterhauser,b Wolfgang Wadsak,b Alexander Holik,c Kurt Kletter,b Ewald Moser,a Siegfried Kasper,c and Rupert Lanzenbergerc,⁎ a

MR Centre of Excellence, Centre for Biomedical Engineering and Physics, Medical University of Vienna, Austria Department of Nuclear Medicine, PET Centre, Medical University of Vienna, Austria c Department of Psychiatry and Psychotherapy, Medical University of Vienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria b

Received 21 December 2007; revised 26 February 2008; accepted 27 February 2008 Available online 6 March 2008

Receptor distribution patterns of neurotransmitters and distinct functional fields of the human brain appear to be tightly connected with respect to their topological allocation along the cerebral cortex. There is, however, considerable lack of human data directly demonstrating this association in vivo. Here, we assessed the relationship between the distribution of the major inhibitory serotonergic neurotransmitter receptor, the 5-HT1A subtype, and the functional organization within early visual cortex defined by retinotopic mapping. The 5HT1A receptor-binding potential was quantified by positron emission tomography (PET) using the highly selective and specific radioligand [carbonyl-11C]WAY-100635 in seven healthy subjects. The retinotopic maps and borders determined by functional magnetic resonance imaging (fMRI) were compared to the receptor distribution employing surface-based region of interest analysis in each of these subjects. We found a significant difference in receptor-binding potential in the functionally defined primary (V1) compared to secondary (V2) visual area, as V1 exhibits only 68% of receptor binding found in V2 in both hemispheres, which is consistent with postmortem data. Our in vivo findings clearly support prior assumptions of a link between receptor distribution and functional fields of the human cortex. © 2008 Elsevier Inc. All rights reserved. Keywords: Retinotopic mapping; 5-HT1A; Serotonin; fMRI; PET

Introduction Recent studies in brain research have suggested a close connection between distribution patterns of neurotransmitter receptors

⁎ Corresponding author. E-mail address: [email protected] (R. Lanzenberger). Available online on ScienceDirect (www.sciencedirect.com). 1053-8119/$ - see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2008.02.044

and functional organization in the cerebral cortex (Zilles et al., 2002; Scheperjans et al., 2005; Wohlschlager et al., 2005; Eickhoff et al., 2007). Although such a correlation would represent a key factor for understanding the physiological and pathological processes underlying psychiatric and neurological disorders (Buhot, 1997; Buhot et al., 2000), there is considerable lack of in vivo data demonstrating such correspondence. The primary objective of this study was, therefore, to investigate the analogy of distribution patterns of a selected neurotransmitter receptor subtype and functional specificity in the human brain. The visual cortex is particularly suited to examine such a relationship as (1) detailed functional organization of visual areas has been extensively investigated using functional imaging techniques (Van Essen et al., 1992; Engel et al., 1994; Sereno et al., 1995; Tootell et al., 1998); (2) besides the well-established delineation of functional areas, the primary visual cortex is also anatomically defined by the presence of a myelinated layer, the stria of Gennari, that has successfully been used for establishing a correspondence between anatomy and function employing magnetic resonance imaging (Bridge et al., 2005); (3) the retinotopic organization of the visual cortex, initially described by T. Inouye and G. Holmes based on war injuries (for a review, see Fishman, 1997), is assessable by functional magnetic resonance imaging (fMRI) in vivo (Engel et al., 1994; Sereno et al., 1995; DeYoe et al., 1996) and, therefore, the extent and functional borders of cortical subdivisions can be mapped in detail. This allows a robust non-invasive differentiation between primary and secondary as well as higher functional areas in the visual cortex on a per-subject basis, which is of importance as different subdivisions of the visual cortex may vary considerably in size (Amunts et al., 2000). In contrast to the large number of neurotransmitter subtypes accessible via postmortem studies, only very few radioligands are suitable for PET-based in vivo receptor mapping of the human cortex. An ideal radioligand for the objectives of this study should be (1) highly selective for a specific receptor subtype, (2) not

F. Gerstl et al. / NeuroImage 41 (2008) 204–211

displaceable by endogenous neurotransmitters, and (3) appropriate for low test–retest variability of PET measurements (Parsey et al., 2000; Rabiner et al., 2002; Heiss and Herholz, 2006). Based on these criteria and the availability of detailed postmortem data of neurotransmitter receptor densities in the human primary and secondary visual cortices (Eickhoff et al., 2007), we chose the serotonin1A (5-HT1A) receptor for this study. Its receptor distribution is assessable in vivo using the highly selective and specific high-affinity radioligand [carbonyl-11C]WAY-100635 (Gunn et al., 1998; Rabiner et al., 2002; Turner et al., 2005; Lanzenberger et al., 2007). The 5-HT1A receptor expression in cortical regions is areaspecific, showing high levels in the limbic system and low levels in primary cortical areas (see Fig. 3). It is the major inhibitory serotonergic receptor subtype in the human brain. A central role of the 5-HT1A receptor for modulation of neural plasticity as well as the development of the columnar structure was demonstrated by studies in kittens and rats (Kirkwood et al., 1995; Bear and Rittenhouse, 1999). The importance of serotonin and 5HT1A receptor-mediated effects in human cerebral ontogenesis was also strongly indicated, as 5-HT1A receptor concentration in humans has been reported to be three to four times higher during fetal development than in the adult cortex (Bar-Peled et al., 1991). Whether serotonergic innervation plays a similar role for regulating ongoing neuronal plasticity in adult visual areas (Furmanski et al., 2004; Maertens and Pollmann, 2005) or has different physiological functions (Seeburg et al., 2004) remains a matter of ongoing debate (Karmarkar and Dan, 2006). The importance of this receptor for brain function is further emphasized by several studies showing significant alterations in 5-HT1A receptor binding in psychiatric disorders, including depression (Owens and Nemeroff, 1998; Drevets et al., 1999; Stockmeier, 2003; Meltzer et al., 2004), anxiety disorders (Kasper, 2001; Neumeister et al., 2004; Lanzenberger et al., 2007) and schizophrenia (Tauscher et al., 2002). Multimodal studies using simultaneously fMRI and electrical recordings showed that the blood oxygenation level-dependent (BOLD) signal used in fMRI is an indirect measure of glutamatergic dendrite activity (Logothetis and Pfeuffer, 2004). It is this activity which is modulated by serotonin1A receptors as 5-HT1A receptors are primarily expressed on glutamatergic neurons (Palchaudhuri and Flugge, 2005). Based on this close link and further supported by electrophysiological studies (Kojic et al., 2001), we hypothesized a significant relation between the level of serotonergic modulation imposed by 5-HT1A receptors and the hierarchical segregation of functional areas. We performed multimodal neuroimaging combining established fMRI-based retinotopic mapping procedures and PET in healthy subjects to obtain both functional and neurochemical data of the early visual cortex, in order to directly test for a putative relationship between function and neurochemistry in the visual cortex.

205

Methods Subjects Multimodal imaging including PET, structural and functional MRI was performed in 7 healthy subjects (6 females, 1 male) aged 24 ± 2.2years (mean ± SD). None of the subjects had a history of brain injury, neurological or psychiatric disorders or substance abuse. Study participants gave written informed consent prior to the experiment after the procedures had been fully explained at the screening visit. The study was approved by the ethics committee of the Medical University of Vienna and the General Hospital of Vienna. Structural and functional magnetic resonance imaging MR measurements were performed using a 3-T Medspec whole-body MR scanner (Bruker BioSpin, Ettlingen, Germany) and the standard birdcage headcoil. In order to perform cortical flattening high-resolution T1-weighted structural MR images (MPRAGE sequence, 256 × 256 matrix, 0.78 × 0.86-mm voxel size, slice thickness 1.56mm, 128 slices) were acquired. Using SPM5, structural data were normalized to MNI space and corrected against intensity-bias caused by RF inhomogeneities (Ashburner and Friston, 2005). For cortical surface reconstruction, image size was reduced by cropping anterior parts of the brain, leaving only the occipital region. Segmentation of structural images, building and flattening of mid-cortical surface models from segmented image data was done with the CARET 5.5 software package (Van Essen et al., 2001). Functional imaging was performed by employing high-resolution gradient-recalled echo planar imaging (EPI) where 128 slabs of ten oblique axial slices parallel to the calcarine sulcus were acquired using asymmetric k-space sampling (matrix size 128 × 128, voxel size 1.65 × 1.95 × 3mm, slice gap = 0.5mm TR = 1000ms, TE = 31ms, flip angle = Ernst angle) resulting in a total scan time of approximately 2 min for each run. Functional volumes were reconstructed from raw k-space data and were converted to Analyze-format for further processing in SPM5. Data sets underwent correction for slice timing, as well as motion correction through rigid body registration of each timeframe to a mean image. Functional data sets were acquired at three to four different sessions per subject. Each functional data set was coregistered to its corresponding anatomical scan, and normalization parameters as derived from the anatomical scan were applied on the functional volumes. Normalized functional data sets were then averaged across sessions and were smoothed with a Gaussian kernel (6-mm FWHM). Volumetric analysis of phase-encoded data including signal-to-noise (SNR) estimation was performed with self-written routines in Interactive Data Language (RSI, Boulder, USA). SNR was

Fig. 1. The rotating double-wedge paradigm employed fulfilled a duty cycle of two complete rotations (2 × 360°), thus stimulating each visual hemifield 4 times during the 128 s of stimulus presentation by the presentation of a single wedge. To achieve maximal visual stimulation, the wedge contained a flickering checkerboard pattern that reversed its polarity at 8 Hz.

206

F. Gerstl et al. / NeuroImage 41 (2008) 204–211

Fig. 2. Regions of interest are shown on a flat map and an inflated cortical surface reconstruction. ROIs for V1 are encircled light blue, those for V2 are encircled red and whole V1 is marked as spotted green area. The sizes of V1 and V2 ROI are of comparable size.

estimated in each voxel in order to assess the reliability of the calculated phase φ following the approach described by (Warnking et al., 2002). fMRI paradigm Retinotopic stimuli (Fig. 1) used in this study were based on previously published research (Engel et al., 1997; Tootell et al., 1998; Warnking et al., 2002; Dumoulin et al., 2003). Experimental procedures involved the presentation of two opposing checkerboard wedges, each about an eight of the display in size. These wedges were presented in a “propeller” configuration that reversed contrast polarity at 8Hz and rotated clockwise around a fixation

cross at a rate of 1 cycle per 64s. Visual stimuli were projected from a video beamer outside the scanner room via a mirror to a screen positioned in the head-side gantry of the MR scanner. Study participants watched the stimulus via a small mirror mounted on the head coil. Stimulus timing was controlled with the Presentation software package (Neurobehavioural Systems Inc., Albany, CA). Subjects were instructed to fixate the middle of the screen and asked to maintain fixation during the whole stimulus presentation. This layout produced a stimulation of approximately the central 30° of the visual field in both directions. The stimulus employed creates a traveling wave of activation in areas associated with different angels of polarity in the visual field. Since upper and lower vertical meridian are in the same axis as the calcarine sulcus but are represented at opposing borders of the primary visual cortex (V1), this activation wave travels from border to border and was used to identify the borders of V1 to the secondary visual (V2) cortex. Retinotopic data were assigned to the cortical surface model node enclosed by the functional voxel at its stereotactic position. V1/V2 borders are located at phase map maxima and minima, respectively, and were manually defined on the flattened phase maps. Regions of interest were defined on both sides of the retinotopically defined V1/V2 border. Regions of interest (ROIs) within V1 and V2 were chosen to be equal in size in individual hemispheres. An additional ROI enclosing the total primary visual cortex area (V1ext) of the medial occipital cortex was defined between the dorsal and ventral V1 borders. All of these regions were situated at the medial part of the occipital lobe. Fig. 2 shows a set of delineated regions of interest. An inflated representation of the occipital lobe accompanies the flat map to increase visibility. The black lines indicate the retinotopically defined ventral and dorsal V1/V2 border while the light blue lines circumscribe ventral and dorsal V1 ROIs. Likewise, red lines demarcate V2 ROIs. The green spotted area indicates the whole V1 region of interest.

Fig. 3. 5-HT1A receptor distribution map superimposed on dorsal (A), lateral (B), posterior (C) view of the 3D brain and on a mid-sagittal section (D). A coregistered structural (T1-weighted) MRI scan from the same subject underlies the parametric PET data. Yellow arrows indicate anatomical regions of primary cortical areas (primary visual cortex, primary motor and somatosensory cortices). The color table indicates the 5-HT1A receptor-binding potential values.

F. Gerstl et al. / NeuroImage 41 (2008) 204–211

PET acquisition Quantitative molecular data were acquired using a GE Advance PET scanner (General Electric Medical Systems, Milwaukee, WI) as described in detail recently (Lanzenberger et al., 2007). [Carbonyl-11C] WAY-100635 was prepared using a 11C methylation synthesizer from Nuclear Interface (now GE Medical Systems) as reported in detail previously (Wadsak et al., 2007). Briefly, a 5-min transmission scan was performed in two-dimensional mode using a retractable 68Ge ring source to correct for tissue attenuation. Dynamic PET scans were acquired in 3D mode commencing simultaneously with intravenous bolus injection of the radioligand [carbonyl-11C]WAY-100635 in phosphate-buffered saline. An average dose of 5.86 ± 0.55MBq/kg body weight was administered. Thirty consecutive timeframes (15 × 1min, 15 × 5min) were acquired, leading to a total measurement time of 90min. An iterative filtered back-projection algorithm (FORE-ITER) was used for reconstruction of 35 contiguous slices (matrix 128 × 128) with a thickness of 4.25mm each. Spatial resolution of the reconstructed volume was 4.36-mm full width at half maximum (FWHM) at the center of the FOV. Calculation of the binding potential (5-HT1A BPND) and distribution estimation of

207

the 5-HT1A receptor was done with the PMOD 2.6 software (http:// www.pmod.com). Kinetic modelling was based on the simplified reference tissue model with the cerebellum as reference region (Lammertsma and Hume, 1996). Coregistration with structural MRI data was done with SPM5 (Fig. 3). Surface-mapping procedures Receptor binding in V1 and V2 was examined by mapping quantitative volumetric PET data onto the cortical surface representation. In order to consider possible effects of parameter mapping methodology, four different surface-mapping algorithms available in the CARET software were evaluated: (1) Enclosing voxel mapping (EV) used one-to-one correspondence between each node of the mid-cortical surface representation and the PET voxel enclosing it. (2) Averaging between the node's and its connected neighbours' enclosing voxel values (AN). (3) Assigned the value obtained from a Gaussian weighting kernel (4-mm bounding box, surface normal σ: 4mm, surface

Fig. 4. The native surface-mapping of PET data (A) employing an “enclosing voxel” algorithm exhibits strong dependency to local folding patterns resulting from partial voluming. The insensitivity to the partial volume effect (iPVE) can be quantified by mapping a grey matter segmentation smoothed with the resolution of the PET camera (D) to the surface (B). By dividing the native PET mapping by the iPVE map a representation of the topographical distribution of the neurotransmitter receptor in scope unbiased by partial voluming can be achieved. Colorscales in panels A and C indicate the calculated receptor binding potential, the colorscale in panel B represents the iPVE, ranging from zero to one.

208

F. Gerstl et al. / NeuroImage 41 (2008) 204–211

tangent σ: 4mm, above surface cutoff: 1mm) centered at the node to that node. (4) Assigned the value obtained from a cubic weighting kernel (4-mm bounding box, surface normal σ: 4mm, surface tangent σ: 4mm, above surface cutoff: 1mm) centered at the node to that node The delineation of cortical areas based on quantitative neurochemical data can suffer from systematic errors due to partial volume effects caused by the low spatial resolution of PET imaging in conjunction with the complex structure of the cerebral surface. Thus, a recently described approach of surface-based partial volume correction was applied (Park et al., 2006) where a simulated PET image is mapped to the surface model using exactly the same mapping algorithm as used for measured PET data. Simulation of the PET image was based on the assumption that 5-HT1A receptor binding is limited to cortical brain areas. Thus, a grey matter segmentation of the cortex was convolved with a Gaussian kernel (4.36mm) to reproduce the changes occurring in a PET camera operating at 4.36-mm spatial resolution. The resulting surface mappings reflect partial volume effects at each surface node and can be used for correction by dividing the native PET mapping by the simulated mapping (Fig. 4). Mean 5-HT1A receptor-binding potentials were obtained for V1, V2 and V1ext ROIs, and binding potential ratios V1/V2 and V1ext/V2 were calculated. Left and right hemisphere ratios were averaged in each subject. In order to visualize the difference in 5-HT1A receptor binding across retinotopically defined V1/V2 borders, lines were drawn connecting opposing V2/V1 and V1/V2 border nodes. Forty equispaced points along each line were sampled, and lines were extrapolated into

ventral and dorsal V2 areas where twenty points were sampled each, resulting in a total of 80 samples per line. 5-HT1A-binding potentials were averaged across lines to obtain a single curve for each subject. Subject-specific binding potentials were then normalized to the individual mean V2-binding potential to eliminate effects from radiochemical variability between subjects and were averaged. Results Definition of dorsal and ventral functional borders delineating primary and secondary visual cortex was possible in all of our 7 subjects. Region of interest size was constrained by the individual extent of high SNR retinotopic phase data. Across subjects V1 ROI area was 405 ± 170mm2, V2 was 376 ± 152mm2 and V1ext ROI was 1099 ± 356mm2 (mean ± standard deviation). Both real and simulated PET data were projected to the surface model using all four mapping algorithms, and each PET mapping was corrected for partial volume effects with its corresponding simulated PET data set (Fig. 5). Table 1 shows the 5-HT1A receptor-binding potentials comparing the values in four different surface-mapping procedures. Receptor-binding potentials without correction for partial volume effects are also given. The ratios for V1/V2 and V1ext/V2 of partial volume corrected 5-HT1A receptor-binding potentials were subjected to one-way ANOVAs revealing no significant difference between the four mapping strategies (V1/V2: F(3,24) = 0.07, p = 0.976; V1ext/V2: F(3,24) = 0.05, p = 0.985). Further analysis was thus confined to enclosing voxel (EV) mapping results. Binding potentials in V1 and V1ext compared to V2 were tested using paired t-tests and showed significant differences (V1/V2: T = 3.9, p b 0.01; V1ext/V2: T = 2.9, p b 0.03). The V1/V2 ratio in 5-HT1A receptor-binding potential

Fig. 5. Two examples of phase-coded maps acquired by retinotopic fMRI (left) used to define ventral (bottom black line) and dorsal (top black line) V1/V2 borders. The color semicircle at the bottom indicates the position of the wedge stimulus in the left visual hemifield corresponding to the phase mapping on the flat maps. Along lines of equal proportional distance (right, blue lines) to these borders neurochemical data were sampled to derive average receptor binding perpendicular to the V1/V2 border.

F. Gerstl et al. / NeuroImage 41 (2008) 204–211

209

Table 1 5-HT1A receptor-binding potential values (mean ± SD) after PVE correction (left part of the table) and before PVE correction (right four columns of the table) in the regions of interest (ROI) are shown ROI

V1 V2 V1ext

PVE-corrected data

Uncorrected data

AN

EV

Gauss

AV

AN

EV

Gauss

AV

2.6 ± 1.5 3.8 ± 2.2 2.5 ± 1.1

2.3 ± 1.5 3.4 ± 2.2 2.2 ± 1.0

2.3 ± 1.5 3.4 ± 2.1 2.1 ± 1.1

2.3 ± 1.5 3.4 ± 2.1 2.1 ± 1.1

1.6 ± 0.6 1.9 ± 0.7 1.5 ± 0.5

1.3 ± 0.5 1.6 ± 0.6 1.3 ± 0.4

1.3 ± 0.5 1.6 ± 0.6 1.3 ± 0.4

1.3 ± 0.5 1.6 ± 0.6 1.3 ± 0.4

Results of four different surface-mapping procedures (AN, EV, Gauss, AV) are given for the primary (V1), secondary (V2) and extended primary visual (V1ext) cortex.

was 0.68 ± 0.11 (mean ± SD), the V1ext/V2 ratio was 0.68 ± 0.13 (mean ± SD). Fig. 6 shows the curve reflecting the change in 5-HT1A receptorbinding potential from ventral V2 over ventral and dorsal V1 to dorsal V2. Error bars indicate the standard error across subjects. A phase uncertainty occurring in retinotopic mapping due to local variability in the hemodynamic response to neural activation (Kastrup et al., 1999) when presenting the rotating visual stimulus in only one orientation (counterclockwise in our experimental setup) can introduce uncertainty about the exact location of the V1/V2 border. This is indicated in Fig. 6 as a color gradient, where white and grey colors indicate V1 and V2, respectively. Note also that actual changes in receptor-binding potential appear smoothed due to the limited spatial resolution of PET. Nevertheless, it can clearly be seen that 5-HT1A receptor-binding potential rapidly changes in the transition from V2 to V1 while exhibiting constant binding in central V1. Discussion The aim of this combined molecular and functional study was to bridge the gap between postmortem and in vivo investigations on

Fig. 6. Normalized serotonin1A receptor-binding potential in medial V1 and V2 in the study group (13 hemispheres) with error bars indicating the standard error. X-axis is in relative units based on vV1/V2 to dV1/V2— distance. Y-axis indicates relative receptor-binding potential. Average V2binding potential in the sampled curve is normalized to one to illustrate relative BP difference in V1.

neurochemical distribution patterns in areas with distinct functionalities, in particular the human early visual cortex. Due to the pronounced interindividual variability in the extent of human primary and secondary visual cortices (Amunts et al., 2000), fMRI retinotopic mapping techniques were used to specify the functional borders in each subject separately. Individual distribution patterns of the major serotonergic inhibitory neurotransmitter 5-HT1A were obtained using PET. The main result of this study is a highly significant difference in 5-HT1A receptor-binding potential between the retinotopically defined primary and secondary visual cortex. We found a ratio of 0.68 between V1 and V2 in 5-HT1A-binding potential. This is in excellent agreement with recently published postmortem data based on four human hemispheres (Rottschy et al., 2007). Rottschy et al. identified V1 and V2 using established cyto- and myeloarchitectonic criteria and reported a mean receptor concentrations of 186.3 fmol/ mg and 260.5 fmol/mg in V1 and V2 postmortem, respectively. Their resulting ratio of 0.71 between V1 and V2 in 5-HT1A receptor concentration fits perfectly to the ratio of 0.68 found in our in vivo study. As such, this is the first study showing a correspondence of functional specificity and neurotransmitter receptor distribution in the human brain in vivo. One particular advantage of our in vivo approach in humans is that the living brain does not experience the changes in microstructural and neurochemical tissue properties possibly occurring in the postmortal period. Additionally, in most cases, the study sample of in vivo imaging is different to postmortem studies mainly based on subjects with an average age over 50 years (which was also true for the study of Eickhoff et al. quoted above) and biased by factors such as antemortal agony that are reported to have an influence on neurochemical parameters (Hynd et al., 2003). Therefore, several aspects need to be considered when postmortem data are compared with in vivo results. The mean age of our study population was 24 years; therefore, we had a significant difference in age to the postmortem population of Eickhoff et al. Age-related depletion in receptor density has been indicated for several receptor types including 5-HT1A with receptor densities declining by approximately 10% per decade (Dillon et al., 1991; Tauscher et al., 2001). Other studies, however, question this age-related decline. For example, Rabiner et al. found no age-related decrease in 5-HT1A receptor-binding potential within an age range of 24–53 years (Rabiner et al., 2002). The effect of age on neurotransmitter receptor density may, therefore, be regarded as an issue of ongoing debate. It appears, however, highly doubtable that a putative age effect may affect the correspondence of functionality and neurochemical properties, given the excellent agreement of existing postmortem examinations of elderly populations (Scheperjans et al., 2005; Eickhoff et al., 2007) with the data presented here. By demonstrating that receptor-binding ratios obtained by multimodal in vivo imaging methods are compatible to postmortem

210

F. Gerstl et al. / NeuroImage 41 (2008) 204–211

results, we established a first step towards reversing the classical approach of interpreting molecular imaging as PET and SPECT data based on prior parcellation. A detailed account of recent approaches for subdividing the human cerebral cortex including parcellation based on neurotransmitter receptor distribution patterns has been given by Zilles et al. (2002). They showed area-specific ratios between several receptor subtypes in postmortem data suitable as parcellation criteria. In contrast to postmortem data, in vivo studies might be restricted to few receptor subtypes, because (1) the number of PET scans in single subjects is limited by ethical regulations to three or four scans, and (2) currently, there are few radioligands available that selective for receptor subtypes in cortical areas with low receptor levels. Additionally, in vivo imaging methods available are not capable of delivering information on the laminar structure of receptor distribution patterns, a feature which is known to be highly variable between different regions despite comparable mean receptor binding as assessed with PET (Scheperjans et al., 2005). Despite these limitations, our results suggest that certain radiotracers such as [carbonyl-11C]WAY-100635 for quantification of the inhibitory 5-HT1A receptor, [18F]altanserin for the excitatory 5-HT2A receptor, and [11C]DASB for the serotonin transporter might be applicable for cortical parcellation of areas with distinct functional specificity. For example, Eickhoff et al. found a 39.6% higher 5-HT1A concentration in V2 than in V1 consistent with our findings, while for 5-HT2A receptor concentration a 8.7% lower concentration than in V1 was reported. Given these differences in 5-HT1A receptor binding between V1 and V2 and the relatively homogeneous distribution of the excitatory 5-HT2A receptor in these regions, the influence of serotonin release on neural activity via inhibitory 5-HT1A and excitatory 5-HT2A is area-specific (Adams et al., 2004; Hurlemann et al., 2005; Eickhoff et al., 2007). However, even within functional areas of the same hierarchical level, patterns of receptor distributions need not necessarily be identical. As reported by Scheperjans et al. (2005), neurotransmitter receptor patterns in extrastriate areas V2 and V3 change with the distance to the fovea indicating subtle but measurable relations between function and receptor distribution even for one receptor subtype. A major limitation of our study is the focus on one receptor subtype and the low PET resolution. However, our results might serve as a proof of principle for future studies with multireceptor imaging and high-resolution PET. There is a broad range of possible applications: it has been shown that several neuropathological conditions change receptor distribution patterns along with functional alterations (Savic et al., 2004). As already emphasized prior in this article, the distribution pattern of the 5-HT1a receptor that was used as target structure for molecular imaging in this study is of special relevance for clinical practice, since changes in its topological distribution are related to functional alterations occurring within several important psychiatric disorders such as depression, anxiety and schizophrenia. The methods applied in this study may serve to investigate systematically in vivo the impact of physiological and pathological processes and medication on neural plasticity comparing changes in function and protein distributions. Conclusion The distribution pattern of the major inhibitory serotonergic neurotransmitter receptor, the 5-HT1A subtype, measured by PET is associated with the functional organization of the primary and secondary visual cortex defined by retinotopic mapping with fMRI. This study showed that multimodal neuroimaging combining PET

and fMRI can be used to investigate the relationship between neurotransmitter receptor distribution and functional specialization in human cortical areas in vivo. We regard this as an essential first step towards applying neurochemical criteria for classification and parcellation of the living human brain. Acknowledgments We thank C. Spindelegger, U. Moser, P. Stein and L.K. Mien for their support. This research was funded by a grant from the Austrian National Bank (OeNB P11468) to R. Lanzenberger. References Adams, K.H., Pinborg, L.H., Svarer, C., Hasselbalch, S.G., Holm, S., Haugbol, S., Madsen, K., Frokjaer, V., Martiny, L., Paulson, O.B., Knudsen, G.M., 2004. A database of [(18)F]-altanserin binding to 5-HT (2A) receptors in normal volunteers: normative data and relationship to physiological and demographic variables. Neuroimage 21, 1105–1113. Amunts, K., Malikovic, A., Mohlberg, H., Schormann, T., Zilles, K., 2000. Brodmann's areas 17 and 18 brought into stereotaxic space-where and how variable? Neuroimage 11, 66–84. Ashburner, J., Friston, K.J., 2005. Unified segmentation. Neuroimage 26, 839–851. Bar-Peled, O., Gross-Isseroff, R., Ben-Hur, H., Hoskins, I., Groner, Y., Biegon, A., 1991. Fetal human brain exhibits a prenatal peak in the density of serotonin 5-HT1A receptors. Neurosci. Lett. 127, 173–176. Bear, M.F., Rittenhouse, C.D., 1999. Molecular basis for induction of ocular dominance plasticity. J. Neurobiol. 41, 83–91. Bridge, H., Clare, S., Jenkinson, M., Jezzard, P., Parker, A.J., Matthews, P.M., 2005. Independent anatomical and functional measures of the V1/V2 boundary in human visual cortex. J. Vis. 5, 93–102. Buhot, M.C., 1997. Serotonin receptors in cognitive behaviors. Curr. Opin. Neurobiol. 7, 243–254. Buhot, M.C., Martin, S., Segu, L., 2000. Role of serotonin in memory impairment. Ann. Med. 32, 210–221. DeYoe, E.A., Carman, G.J., Bandettini, P., Glickman, S., Wieser, J., Cox, R., Miller, D., Neitz, J., 1996. Mapping striate and extrastriate visual areas in human cerebral cortex. Proc. Natl. Acad. Sci. U. S. A. 93, 2382–2386. Dillon, K.A., Gross-Isseroff, R., Israeli, M., Biegon, A., 1991. Autoradiographic analysis of serotonin 5-HT1A receptor binding in the human brain postmortem: effects of age and alcohol. Brain Res. 554, 56–64. Drevets, W.C., Frank, E., Price, J.C., Kupfer, D.J., Holt, D., Greer, P.J., Huang, Y., Gautier, C., Mathis, C., 1999. PET imaging of serotonin 1A receptor binding in depression. Biol. Psychiatry 46, 1375–1387. Dumoulin, S.O., Hoge, R.D., Baker Jr., C.L., Hess, R.F., Achtman, R.L., Evans, A.C., 2003. Automatic volumetric segmentation of human visual retinotopic cortex. Neuroimage 18, 576–587. Eickhoff, S.B., Rottschy, C., Zilles, K., 2007. Laminar distribution and codistribution of neurotransmitter receptors in early human visual cortex. Brain Struct. Funct. 212, 255–267. Engel, S.A., Rumelhart, D.E., Wandell, B.A., Lee, A.T., Glover, G.H., Chichilnisky, E.J., Shadlen, M.N., 1994. fMRI of human visual cortex. Nature 369, 525. Engel, S.A., Glover, G.H., Wandell, B.A., 1997. Retinotopic organization in human visual cortex and the spatial precision of functional MRI. Cereb. Cortex 7, 181–192. Fishman, R.S., 1997. Gordon Holmes, the cortical retina, and the wounds of war. The seventh Charles B. Snyder Lecture. Doc. Ophthalmol. 93, 9–28. Furmanski, C.S., Schluppeck, D., Engel, S.A., 2004. Learning strengthens the response of primary visual cortex to simple patterns. Curr. Biol. 14, 573–578. Gunn, R.N., Sargent, P.A., Bench, C.J., Rabiner, E.A., Osman, S., Pike, V.W., Hume, S.P., Grasby, P.M., Lammertsma, A.A., 1998. Tracer kinetic

F. Gerstl et al. / NeuroImage 41 (2008) 204–211 modeling of the 5-HT1A receptor ligand [carbonyl-11C]WAY-100635 for PET. Neuroimage 8, 426–440. Heiss, W.D., Herholz, K., 2006. Brain receptor imaging. J. Nucl. Med. 47, 302–312. Hurlemann, R., Matusch, A., Eickhoff, S.B., Palomero-Gallagher, N., Meyer, P.T., Boy, C., Maier, W., Zilles, K., Amunts, K., Bauer, A., 2005. Analysis of neuroreceptor PET-data based on cytoarchitectonic maximum probability maps: a feasibility study. Anat. Embryol. (Berl) 210, 447–453. Hynd, M.R., Lewohl, J.M., Scott, H.L., Dodd, P.R., 2003. Biochemical and molecular studies using human autopsy brain tissue. J. Neurochem. 85, 543–562. Karmarkar, U.R., Dan, Y., 2006. Experience-dependent plasticity in adult visual cortex. Neuron 52, 577–585. Kasper, S., 2001. Depression and anxiety–separate or continuum? World J. Biol. Psychiatry 2, 162–163. Kastrup, A., Kruger, G., Glover, G.H., Neumann-Haefelin, T., Moseley, M.E., 1999. Regional variability of cerebral blood oxygenation response to hypercapnia. Neuroimage 10, 675–681. Kirkwood, A., Lee, H.K., Bear, M.F., 1995. Co-regulation of long-term potentiation and experience-dependent synaptic plasticity in visual cortex by age and experience. Nature 375, 328–331. Kojic, L., Gu, Q., Douglas, R.M., Cynader, M.S., 2001. Laminar distribution of cholinergic- and serotonergic-dependent plasticity within kitten visual cortex. Brain Res. Dev. Brain Res. 126, 157–162. Lammertsma, A.A., Hume, S.P., 1996. Simplified reference tissue model for PET receptor studies. Neuroimage 4, 153–158. Lanzenberger, R.R., Mitterhauser, M., Spindelegger, C., Wadsak, W., Klein, N., Mien, L.K., Holik, A., Attarbaschi, T., Mossaheb, N., Sacher, J., Geiss-Granadia, T., Kletter, K., Kasper, S., Tauscher, J., 2007. Reduced serotonin-1A receptor binding in social anxiety disorder. Biol. Psychiatry 61, 1081–1089. Logothetis, N.K., Pfeuffer, J., 2004. On the nature of the BOLD fMRI contrast mechanism. Magn. Reson. Imaging 22, 1517–1531. Maertens, M., Pollmann, S., 2005. fMRI reveals a common neural substrate of illusory and real contours in V1 after perceptual learning. J. Cogn. Neurosci. 17, 1553–1564. Meltzer, C.C., Price, J.C., Mathis, C.A., Butters, M.A., Ziolko, S.K., MosesKolko, E., Mazumdar, S., Mulsant, B.H., Houck, P.R., Lopresti, B.J., Weissfeld, L.A., Reynolds, C.F., 2004. Serotonin 1A receptor binding and treatment response in late-life depression. Neuropsychopharmacology 29, 2258–2265. Neumeister, A., Bain, E., Nugent, A.C., Carson, R.E., Bonne, O., Luckenbaugh, D.A., Eckelman, W., Herscovitch, P., Charney, D.S., Drevets, W.C., 2004. Reduced serotonin type 1A receptor binding in panic disorder. J. Neurosci. 24, 589–591. Owens, M.J., Nemeroff, C.B., 1998. The serotonin transporter and depression. Depress. Anxiety 8 (suppl 1), 5–12. Palchaudhuri, M., Flugge, G., 2005. 5-HT1A receptor expression in pyramidal neurons of cortical and limbic brain regions. Cell Tissue Res. 321, 159–172. Park, H.J., Lee, J.D., Chun, J.W., Seok, J.H., Yun, M., Oh, M.K., Kim, J.J., 2006. Cortical surface-based analysis of 18F-FDG PET: measured metabolic abnormalities in schizophrenia are affected by cortical structural abnormalities. Neuroimage 31, 1434–1444. Parsey, R.V., Slifstein, M., Hwang, D.R., Abi-Dargham, A., Simpson, N., Mawlawi, O., Guo, N.N., Van Heertum, R., Mann, J.J., Laruelle, M., 2000. Validation and reproducibility of measurement of 5-HT1A receptor parameters with [carbonyl-11C]WAY-100635 in humans: comparison of arterial and reference tissue input functions. J. Cereb. Blood Flow Metab. 20, 1111–1133. Rabiner, E.A., Messa, C., Sargent, P.A., Husted-Kjaer, K., Montgomery, A., Lawrence, A.D., Bench, C.J., Gunn, R.N., Cowen, P., Grasby, P.M.,

211

2002. A database of [(11)C]WAY-100635 binding to 5-HT(1A) receptors in normal male volunteers: normative data and relationship to methodological, demographic, physiological, and behavioral variables. Neuroimage 15, 620–632. Rottschy, C., Eickhoff, S.B., Schleicher, A., Mohlberg, H., Kujovic, M., Zilles, K., Amunts, K., 2007. Ventral visual cortex in humans: Cytoarchitectonic mapping of two extrastriate areas. Hum. Brain Mapp. 28, 1045–1059. Savic, I., Lindstrom, P., Gulyas, B., Halldin, C., Andree, B., Farde, L., 2004. Limbic reductions of 5-HT1A receptor binding in human temporal lobe epilepsy. Neurology 62, 1343–1351. Scheperjans, F., Palomero-Gallagher, N., Grefkes, C., Schleicher, A., Zilles, K., 2005. Transmitter receptors reveal segregation of cortical areas in the human superior parietal cortex: relations to visual and somatosensory regions. Neuroimage 28, 362–379. Seeburg, D.P., Liu, X., Chen, C., 2004. Frequency-dependent modulation of retinogeniculate transmission by serotonin. J. Neurosci. 24, 10950–10962. Sereno, M.I., Dale, A.M., Reppas, J.B., Kwong, K.K., Belliveau, J.W., Brady, T.J., Rosen, B.R., Tootell, R.B., 1995. Borders of multiple visual areas in humans revealed by functional magnetic resonance imaging. Science 268, 889–893. Stockmeier, C.A., 2003. Involvement of serotonin in depression: evidence from postmortem and imaging studies of serotonin receptors and the serotonin transporter. J. Psychiatr. Res. 37, 357–373. Tauscher, J., Verhoeff, N.P., Christensen, B.K., Hussey, D., Meyer, J.H., Kecojevic, A., Javanmard, M., Kasper, S., Kapur, S., 2001. Serotonin 5-HT1A receptor binding potential declines with age as measured by [11C]WAY-100635 and PET. Neuropsychopharmacology 24, 522–530. Tauscher, J., Kapur, S., Verhoeff, N.P., Hussey, D.F., Daskalakis, Z.J., Tauscher-Wisniewski, S., Wilson, A.A., Houle, S., Kasper, S., Zipursky, R.B., 2002. Brain serotonin 5-HT(1A) receptor binding in schizophrenia measured by positron emission tomography and [11C]WAY-100635. Arch. Gen. Psychiatry 59, 514–520. Tootell, R.B., Hadjikhani, N.K., Vanduffel, W., Liu, A.K., Mendola, J.D., Sereno, M.I., Dale, A.M., 1998. Functional analysis of primary visual cortex (V1) in humans. Proc. Natl. Acad. Sci. U. S. A. 95, 811–817. Turner, M.R., Rabiner, E.A., Hammers, A., Al-Chalabi, A., Grasby, P.M., Shaw, C.E., Brooks, D.J., Leigh, P.N., 2005. [11C]-WAY100635 PET demonstrates marked 5-HT1A receptor changes in sporadic ALS. Brain 128, 896–905. Van Essen, D.C., Anderson, C.H., Felleman, D.J., 1992. Information processing in the primate visual system: an integrated systems perspective. Science 255, 419–423. Van Essen, D.C., Drury, H.A., Dickson, J., Harwell, J., Hanlon, D., Anderson, C.H., 2001. An integrated software suite for surface-based analyses of cerebral cortex. J. Am. Med. Inform. Assoc. 8, 443–459. Wadsak, W., Mien, L.K., Ettlinger, D., Lanzenberger, R., Haeusler, D., Dudczak, R., Kletter, K., Mitterhauser, M., 2007. Simple and fully automated preparation of [carbonyl-11C]WAY-100635. Radiochim. Acta 95, 417–422. Warnking, J., Dojat, M., Guerin-Dugue, A., Delon-Martin, C., Olympieff, S., Richard, N., Chehikian, A., Segebarth, C., 2002. fMRI retinotopic mapping–step by step. Neuroimage 17, 1665–1683. Wohlschlager, A.M., Specht, K., Lie, C., Mohlberg, H., Wohlschlager, A., Bente, K., Pietrzyk, U., Stocker, T., Zilles, K., Amunts, K., Fink, G.R., 2005. Linking retinotopic fMRI mapping and anatomical probability maps of human occipital areas V1 and V2. Neuroimage 26, 73–82. Zilles, K., Palomero-Gallagher, N., Grefkes, C., Scheperjans, F., Boy, C., Amunts, K., Schleicher, A., 2002. Architectonics of the human cerebral cortex and transmitter receptor fingerprints: reconciling functional neuroanatomy and neurochemistry. Eur. Neuropsychopharmacol. 12, 587–599.