Verbal memory and 5-HT1A receptors in healthy volunteers – A PET study with [carbonyl-11C]WAY-100635

Author’s Accepted Manuscript Verbal memory and 5-HT1A receptors in healthy volunteers – a PET study with [carbonyl11C]WAY-100635 Jani Penttilä, Jussi Hirvonen, Lauri Tuominen, Ville Lumme, Tuula Ilonen, Kjell Någren, Jarmo Hietala www.elsevier.com/locate/euroneuro

PII: DOI: Reference:

S0924-977X(15)00411-3 http://dx.doi.org/10.1016/j.euroneuro.2015.12.028 NEUPSY11188

To appear in: European Neuropsychopharmacology Received date: 17 December 2014 Revised date: 5 December 2015 Accepted date: 12 December 2015 Cite this article as: Jani Penttilä, Jussi Hirvonen, Lauri Tuominen, Ville Lumme, Tuula Ilonen, Kjell Någren and Jarmo Hietala, Verbal memory and 5-HT1A receptors in healthy volunteers – a PET study with [carbonyl-11C]WAY1 0 0 6 3 5 , European Neuropsychopharmacology, http://dx.doi.org/10.1016/j.euroneuro.2015.12.028 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Verbal memory and 5-HT1A receptors in healthy volunteers – a PET study with [carbonyl-11C]WAY-100635 Jani Penttilä, MD, PhD1, Jussi Hirvonen, MD, PhD2,3, Lauri Tuominen, MD, PhD 3,4, Ville Lumme, MD4, Tuula Ilonen, PhD4, Kjell Någren, PhD3, and Jarmo Hietala, MD, PhD3,4 Author affiliations: 1 Department of Adolescent Psychiatry, Päijät-Häme Central Hospital, Lahti, Finland; 2 Department of Radiology, University of Turku, Turku, Finland 3 Turku PET Centre, University of Turku and Turku University Central Hospital, Turku, Finland 4 Department of Psychiatry, University of Turku, Turku, Finland Short title: Serotonin 5-HT1A receptors and verbal memory Key words: Serotonin, 5-HT1A receptor, memory, positron emission tomography, imaging Corresponding author: Jarmo Hietala, M.D., Professor of Psychiatry Department of Psychiatry, University of Turku Kunnallissairaalantie 20, Building 9 FIN-20700 Turku, FINLAND tel: +358-2-269 2520; fax: +358-2-269 2528 E-mail: [email protected] Abstract The serotonin 5-HT1A receptor is a putative drug development target in disorders with cognitive and in particular memory deficits. However, previous human positron emission tomography (PET) studies on 5-HT1A receptor binding and memory functions have yielded discrepant results. We explored the association between verbal memory and 5-HT1A receptor binding in 24 healthy subjects (14 male, 10 female, aged 18-41 years). The cognitive tests included the Wechsler Memory Scale-Revised (WMS-R), Wechsler Adult Intelligence Scale-Revised (WAIS-R) and Wisconsin Card Sorting Test (WCST). 5-HT1A receptor binding was measured with PET and the radioligand [carbonyl-11C]WAY-100635, which was quantified with the gold standard method based on kinetic modeling using arterial blood samples. We found that global 5-HT1A receptor binding was positively correlated with measures of verbal memory, such that subjects who had higher receptor binding tended to have better verbal memory than subjects who had lower receptor binding. Regional analyses suggested significant correlations in multiple neocortical brain regions and the raphe nuclei. We did not find significant correlations between 5-HT1A receptor binding and executive functions as measured with WCST. We conclude that neocortical as well as raphe 5-HT-1A receptors are involved in verbal memory function in man. Introduction

Serotonin (5-HT) system has been consistently implicated in cognitive functions such as memory as well as cognitive flexibility that refers primarily to different dimensions of executive functioning and attention (for review, see Schmitt et al., 2006). 5-HT system may be particularly important for verbal memory as 5-HT depletion robustly impairs verbal memory consolidation (Helmbold et al., 2013; Mendelsohn et al., 2009) and carriers of the S-allele in the 5-HT transporter promoter region perform worse in verbal memory task (Zilles et al., 2012). The serotonin 5-HT1A receptor may have crucial role at conveying these effects, as both 5-HT1A receptor agonists and partial agonists impair verbal memory functions ((Riedel et al., 2002; Wingen et al., 2007; Yasuno et al., 2003); for review, see (Meneses and Perez-Garcia, 2007)). Correspondingly, drugs acting on 5-HT1A could be useful in the treatment of memory disorders (Roth et al., 2004; Schechter et al., 2002; Yasuno et al., 2003). In addition to effects on verbal memory, there is preliminary evidence that 5-HT1A receptor agonism could also influence cognitive flexibility (Depoortere et al., 2007). Positron emission tomography (PET) with the radioligand [carbonyl-11C]WAY100635 offers a tool to investigate the 5-HT1A receptors in humans in vivo (Pike et al., 1996). Between-subject variability in 5-HT1A receptor binding is large in normal subjects, the biological basis of which is mostly unclear (Borg et al., 2006; Parsey et al., 2002; Rabiner et al., 2002). Previous [carbonyl-11C]WAY-100635 PET studies on human cognitive functions are sparse and have yielded discrepant results. Yasuno et al. (2003) reported a negative correlation between explicit memory function and 5HT1A receptor binding in hippocampus, but not in other cortical regions or in the raphe nuclei. In contrast, Borg et al. (2006) found no significant correlations between regional 5-HT1A receptor binding and a wide range of cognitive tests, including tasks of various aspects of memory and cognitive flexibility. This negative finding was later replicated in a larger data set by the same authors (Borg et al., 2009). In a more recent study, a positive correlation was found between hippocampal 5-HT1A receptor binding and memory performance in patients with temporal lobe epilepsy with another radioligand (Theodore et al., 2012). We sought to further characterize the association between serotonin 5-HT1A receptor binding and verbal memory in healthy subjects using [carbonyl-11C]WAY-100635 PET. As a gold standard method of quantification, we used arterial plasma samples and kinetic modeling instead of reference tissue modeling which may induce bias (Hirvonen et al., 2007). In an exploratory manner, we also evaluated the relationship between 5-HT1A receptor binding and the cognitive flexibility.

Experimental procedures The study protocol was approved by the Joint Ethical Committee (EC) of the University of Turku and the Turku University Central Hospital. After having received all the relevant information in written form from the investigators, all study subjects gave EC-approved written consents. This study followed the ethical guidelines of the Declaration of Helsinki. This study was part of a larger study examining the role of serotonin 5-HT1A receptors in cognitive function and personality traits in healthy subjects (Hirvonen et al., 2015).

Subjects 24 healthy subjects (14 male, 10 female) were recruited for the study. All subjects were right-handed and non-smoking. Subjects were free of any somatic or psychiatric illness, illicit drug abuse and alcoholism, as confirmed by blood and urine tests, somatic assessment, and a Structured Clinical Interview for DSM-IV Axis I disorders (SCID-I) carried out by a psychiatrist. The age, height, weight, body mass index, and years of education of the subjects were 25.2 years (6.1; 18-41), 174 cm (8; 158-186), 69 kg (11; 35-87), 22.7 kg/m2 (2.2; 19-27), and 14 years (2; 9-17), respectively (mean, SD; range). To rule out structural brain abnormalities and to obtain an anatomical reference for quantification of PET images, all subjects underwent T1-weighted magnetic resonance imaging at 1.5 T. Cognitive testing All subjects underwent a comprehensive neuropsychological evaluation, which included tests of intelligence, memory, learning and executive function. The Wechsler Adult Intelligence Scale - Revised (WAIS-R) was used to assess general cognitive competence (Wechsler, 1992). The Wechsler Memory Scale - Revised (WMS-R) was used to measure immediate and delayed verbal memory and learning functions (Wechsler, 1987). Executive function was assessed through the Wisconsin Card Sorting Test (WCST Heaton et al., 1993). The full-scale IQ according the WAIS-R was 119 ± 9 (verbal: 118 ± 9, non-verbal: 117 ± 10; means ± SD).

Positron emission tomography The radioligand [carbonyl-11C]WAY-100635 was prepared as previously described (Hirvonen et al., 2007). We used the GE Advance PET scanner (GE, Milwaukee, Wis., USA) in 3D mode as previously described using 35 slices of 4.25 mm thickness (Hirvonen et al., 2007). Briefly, an intravenous rapid bolus of [carbonyl-11C]WAY 100635 was given to each subject (radioactivity 264 ± 80 MBq, specific radioactivity 68 ± 32 MBq/nmol, mass of radiotracer 1.9 ± 0.8 g, means ± SD), and radioactivity was measured in brain for 57 minutes using 14 time frames of increasing duration. Serial samples of arterial blood were drawn from a radial artery to measure plasma radioactivity and cleared from radioactive metabolites according to previously published protocol (Hirvonen et al., 2007). Image analysis and modeling of radioligand kinetics An automated region of interest (ROI) analysis was carried out to calculate regional time-activity curves (Hirvonen et al., 2008). Individual integral (summed over time) PET images were first spatially normalized onto a ligand-specific template. These transformation parameters were then used to normalize individual dynamic PET frames into standard stereotactic space. ROIs were then applied in the standard space onto for the following brain regions: anterior cingulate cortex, dorsal raphe nuclei, lateral temporal cortex, mesial temporal cortex, parietal cortex, and prefrontal cortex. Cerebellar white matter was used as a reference region (Hirvonen et al., 2007; Parsey

et al., 2005). Before modeling, contributions of total blood radioactivity to regional tissue time-activity curves was eliminated by assuming 5 % blood volume in ROI and subtracting it directly from regional radioactivity. Distribution volumes (VT) for the standard two-tissue compartmental model (Innis et al., 2007) were estimated using the metabolite-corrected arterial plasma time-activity curve as input function (Hirvonen et al., 2007). Binding potential (BP) values were then calculated from the VT values of ROIs and the reference region. Two commonly used estimates of specific binding were considered: BPP = VT-VND and BPND = VT/VND-1, where VND is the distribution volume of the non-displaceable tissue compartment (approximated by VT in the reference region). BPP represents specific binding relative to arterial plasma, and BPND represents specific binding relative to the reference region (Innis et al., 2007).

Statistical parametric mapping An independent voxel-based statistical parametric mapping (SPM) analysis was done to confirm the ROI-based findings. Spatial normalization, smoothing, and statistical analyses of parametric images were performed using SPM8 (http://www.fil.ion.ucl.ac.uk/spm) running on Matlab R2011a (The Mathworks Inc., Sherborn, MA, USA). First, parametric images were calculated for each subject by estimating BPP separately for each voxel in the brain using cerebellar white matter as reference region. Second, these images were normalized to a ligand-specific [carbonyl-11C]WAY 100635 template that was created in-house. Finally, normalized parametric images were smoothed with an 8-mm Gaussian kernel (full-width halfmaximum). Within each voxel, statistical significance of the association between the composite verbal memory score and BPP was tested using one-sample t-test. Voxellevel uncorrected p-value <0.005 and cluster-level family-wise error (FWE) corrected p-value <0.05 was considered statistically significant. In addition to mapping the magnitude of the effects in terms of voxel-level t values, we also calculated the coefficient of determination (R2) for each voxel, and projected those values on a brain surface template using Caret software version 5.62 (http://www.nitrc.org/projects/caret/). Statistical analyses of ROI data Statistical analyses were performed with SPSS 13.0 for Windows (Release 13.0.1, copyright SPSS Inc., 1989-2004). All variables were normally distributed according to the Shapiro-Wilk test, except for Information (p=0.022), Digit span (p=0.014), Picture completion (p=0.033), Delayed verbal paired associates, (p<0.001), where the distributions were negatively skewed. Thus, we applied both parametric and nonparametric statistical tests. Since the 5-HT1A binding cross-correlates highly among the brain regions, a repeated measures analysis of variance (rmANOVA) was carried out, with sex, age, brain region, and each cognitive parameter as predictors of 5HT1A binding. This main analysis considers all brain regions simultaneously by including brain region as a repeated within-subject variable. The analysis is especially powerful to detect persistent changes across brain regions, since regional values are

highly intercorrelated (R>0.8). Importantly, analyzing all brain regions simultaneously obviates the need of post hoc corrections for multiple brain regions, since we test for the main effect of cognitive performance on 5-HT1A across all regions. This analysis can be used to determine whether or not there are regiondependent contributions (Hirvonen et al., 2008) since all brain regions are considered behaving in a similar manner if the interaction term (cognitive test by brain region) is not statistically significant. . Partial correlation with sex and age as covariates was carried out regionally as a post-hoc test following the rmANOVA, followed by nonparametric Spearman correlation coefficients. As a methodological validation, cerebellar white matter VT values were correlated against cognitive measures to justify whether to use BPP or BPND. Partial correlation analysis revealed positive associations of VND with Vocabulary (r=0.43, p=0.047), Arithmetics (r=0.54, =0.010) and Similarities (r=0.44, p=0.039). Thus, BPP was considered the main outcome measure of interest as in our previous studies (Hirvonen et al., 2008; Martikainen et al., 2007) because it is less confounded by changes in VND than is BPND. P values below 0.05 were considered to denote statistical significance. Results

In the verbal memory parameters of WMS-R, the median scores (range) of Immediate and Delayed logical memory, representing logical memory, were 30.5 (18–38) and 29 (15–38), respectively. The respective scores of Immediate and Delayed verbal paired associates, representing verbal learning, were 21 (11–24) and 8 (4–8). A composite verbal memory score was constructed as previously described (Lumme et al., 2010), and had a median of 107 and range of 82–123 in this sample. We found that 5-HT1A receptor binding was positively correlated with measures of verbal memory, such that subject with higher binding tended to perform better than subject with lower binding (Fig. 1). Significant main effects of global 5-HT1A receptor binding were seen for measures of verbal memory and immediate learning in WMS-R, and for measures of verbal functions and language skills as well as concept formation and reasoning in WAIS-R (Table 1). There were no statistically significant interactions with the region term, consistent with a similar effect across brain regions. We found no significant associations between 5-HT1A receptor binding and performance in WCST. As expected based on the pattern of the rmANOVA results suggesting a global main effect, partial correlations were seen throughout the neocortex and in the raphe nucleus, but less in the mesial temporal cortex including the hippocampus (Table 1). Strongest neocortical correlations were seen in the composite verbal memory score. Age or sex did not significantly correlate with 5HT1A binding in any of the rmANOVA models (all p>0.2). Association between 5-HT1A receptor binding and composite verbal memory was confirmed with an independent voxel-based SPM analysis showing a significicant cluster including frontal, parietal, temporal and occipital lobes (Fig. 2). Voxel-wise coefficient of determination (R2) between composite verbal memory score and BPND had typical peak voxel values around 35–40%, highest values being found in bilateral temporal and parietal cortices and in left dorsal anterior cingulate cortex (Fig. 3). Positive correlation between the composite verbal memory score and 5-HT1A binding were confirmed by the non-parametric Spearman’s correlation analysis, which showed significant correlations in most brain regions (ranging from ρ=0.51, p=0.011 in the anterior cingulate cortex to ρ=0.59, p=0.001 in the raphe nuclei), except for the

mesial temporal cortex (ρ=0.34, p=0.11). This pattern of results suggest that the slight non-normality of distributions did not confound the results. We also explored associations between verbal memory and BPND, a measure of specific binding relative to reference region. The rmANOVA did not show statistically significant associations between composite verbal memory and BPND, although a significant correlation was seen in the raphe nuclei (R=0.45, p=0.028). Because previous studies had used cerebellar gray matter as reference region, we sought to replicate those previous results with BPND estimated with the simplified reference tissue model (SRTM) and gray matter as reference region. This analysis did not show any significant correlations, even in the raphe nuclei (R=0.18, p=0.39). This likely occurred because verbal memory correlated with VT in the cerebellar gray matter (R=0.47, p=0.022), but not with VT in cerebellar white matter (R=0.34, p=0.109).

Discussion We found a positive correlation between verbal memory and 5-HT1A receptor binding in healthy subjects, consistent with the central role of this receptor subtype in normal memory functions (Meneses and Perez-Garcia, 2007). The involvement of 5HT1A receptor -mediated neurotransmission in verbal memory was not limited to specific brain regions, but was rather found to cover most cortical regions as well as the raphe nuclei. In previous studies, serotonin has the most consistently been related to long-term memory functions (Schmitt et al., 2006). However, we detected significant correlations also between immediate logical and explicit episodic memory performance and 5-HT1A receptor binding. Our current results lend further support to the rationale that drugs acting on 5-HT1A receptors might be beneficial in the treatment of cognitive dysfunction (Roth et al., 2004; Schechter et al., 2002). For example, Sumiyoshi et al. (Sumiyoshi et al., 2001) have shown that the addition of a 5-HT1A partial agonist (tandospirone) to the ongoing treatment of patients with schizophrenia improves their verbal memory scores in WMS-R. On the other hand, 5-HT1A receptor blockade has been demonstrated to cause a selective verbal memory impairment in healthy volunteers treated with an SSRI (Wingen et al., 2007). In addition, we found a relation between 5-HT1A receptor binding and performance in the Arithmetics and Similarities tasks in WAIS-R, which are orally administered concept formation and reasoning tests. As Arithmetics and Similarities are both known to demand immediate memory – together with concentration, conceptual manipulation and tracking (Lezak, 1995) – the association with 5-HT1A receptor expression probably arises from the memory functions that are required in these tests. Serotonin is though to participate in performance monitoring and flexible, goaldirected behavior (Evers et al., 2006; Schmitt et al., 2006). Nevertheless, we found no associations between 5-HT1A receptor binding and cognitive flexibility as measured using WCST. This suggests that the role of serotonin is less prominent in cognitive flexibility than in memory functions, or that the cognitive flexibility is modulated by serotonin via receptor subtypes other than 5-HT1A. However, the exploration of relationship between executive functioning and 5-HT-1A receptor binding in healthy subjects using the WCST error numbers and number of completed categories as

neuropsychological outcomes may not be optimal due to skewed distribution of WCST performance i.e. majority of healthy subjects perform well in this test.

Comparison to previous studies and methodological issues The positive correlation between verbal memory and 5-HT1A receptor binding is consistent with a report on a positive correlation between hippocampal 5-HT1A receptor binding and memory in patients with temporal lobe epilepsy with another radioligand (Theodore et al., 2012), but in contrast to previous [carbonyl-11C]WAY100635 PET studies in healthy subjects showing either no correlation (Borg et al., 2006; Borg et al., 2009) or a negative correlation between verbal memory and hippocampal 5-HT1A binding (Yasuno et al., 2003). Methodological issues may in part explain this discrepancy. We employed the gold standard method of quantification based on arterial blood samples and kinetic modeling of [carbonyl-11C]WAY-100635 to derive estimates of specific binding to 5-HT1A receptors. In addition, we used cerebellar white matter as reference region, instead of cerebellar gray matter, which is confounded by specific binding in humans (Parsey et al., 2010), although non-specific binding may be different in gray and white matter. Previous studies looking at [carbonyl-11C]WAY100635 binding and cognitive measures have used BPND with cerebellar gray matter as the outcome measure (Borg et al., 2006; Yasuno et al., 2003). Use of cerebellar gray matter to estimate BPND is problematic because reference region VT appears as the denominator in BPND and small absolute error in that denominator may propagate large proportional error in BPND. We found significant correlations between verbal memory and cerebellar gray matter VT, but not cerebellar white matter, which likely explains the lack of finding when using BPND and confirms previous notions about the problems when using this reference region (Hirvonen et al., 2008; Parsey et al., 2010). Thus, the choice of outcome measure (BPP vs. BPND) and reference region (cerebellar white vs. gray matter) may contribute to the discrepancy between the results from the current study and those from previous studies (Borg et al., 2006; Borg et al., 2009; Yasuno et al., 2003). Correlating multiple cognitive variables and brain regions carries the risk of type I errors due to multiple comparisons. However, it appears unlikely that the current results would represent false positive findings, because the rmANOVA demonstrated global effects of 5-HT1A binding where the model collapses over multiple brain regions (which, in fact, are not independent observations). Moreover, the current findings make up a logical pattern of results, where significant associations were seen with 5-HT1A binding in the brain and cognitive variables in the domain of verbal, but not of non-verbal, memory.

Conclusion Our data from 24 healthy subjects, using PET with [carbonyl-11C]WAY-100635 and the gold standard of quantification based on arterial blood samples, suggest that verbal memory functions involve serotonin 5-HT1A receptor-mediated neurotransmission widely in the neocortex and in the raphe nuclei. These data are

consistent with previous animal studies and lend support to targeting the 5-HT1A receptor in treating memory deficits in neurological and psychiatric disorders.

References Borg, J., Andree, B., Lundberg, J., Halldin, C., Farde, L., 2006. Search for correlations between serotonin 5-HT1A receptor expression and cognitive functions-a strategy in translational psychopharmacology. Psychopharmacology (Berl) 185, 389-394. Borg, J., Henningsson, S., Saijo, T., Inoue, M., Bah, J., Westberg, L., Lundberg, J., Jovanovic, H., Andrée, B., Nordstrom, A.L., Halldin, C., Eriksson, E., Farde, L., 2009. Serotonin transporter genotype is associated with cognitive performance but not regional 5-HT1A receptor binding in humans. Int J Neuropsychopharmacol 12, 783792. Depoortere, R., Auclair, A.L., Bardin, L., Bruins Slot, L., Kleven, M.S., Colpaert, F., Vacher, B., Newman-Tancredi, A., 2007. F15063, a compound with D2/D3 antagonist, 5-HT 1A agonist and D4 partial agonist properties. III. Activity in models of cognition and negative symptoms. Br J Pharmacol 151, 266-277. Evers, E.A., van der Veen, F.M., van Deursen, J.A., Schmitt, J.A., Deutz, N.E., Jolles, J., 2006. The effect of acute tryptophan depletion on the BOLD response during performance monitoring and response inhibition in healthy male volunteers. Psychopharmacology (Berl). 187, 200-208. Epub 2006 May 2019. Heaton, R.K., Chelune, G.J., Talley, J.L., Kay, G.G., Curtiss, G., 1993. Wisconsin Card Sorting Test Manual, revised and expanded. Psychological Assessment Resources, Inc, Lutz (FL). Helmbold, K., Bubenzer, S., Dahmen, B., Eisert, A., Gaber, T.J., Habel, U., Konrad, K., Herpertz-Dahlmann, B., Zepf, F.D., 2013. Influence of acute tryptophan depletion on verbal declarative episodic memory in young adult females. Amino Acids 45, 1207-1219. Hirvonen, J., Kajander, J., Allonen, T., Oikonen, V., Någren, K., Hietala, J., 2007. Measurement of serotonin 5-HT1A receptor binding using positron emission tomography and [carbonyl-11C]WAY-100635—considerations on the validity of cerebellum as a reference region. Journal of Cerebral Blood Flow & Metabolism 27, 185-195. Hirvonen, J., Karlsson, H., Kajander, J., Lepola, A., Markkula, J., Rasi-Hakala, H., Nagren, K., Salminen, J.K., Hietala, J., 2008. Decreased brain serotonin 5-HT1A receptor availability in medication-naive patients with major depressive disorder: an in-vivo imaging study using PET and [carbonyl-11C]WAY-100635. Int J Neuropsychopharmacol 11, 465-476. Hirvonen, J., Tuominen, L., Någren, K., Hietala, J., 2015. Neuroticism and serotonin 5-HT1A receptors in healthy subjects. Psychiatry Res 234(1), 1-6.

Innis, R.B., Cunningham, V.J., Delforge, J., Fujita, M., Gjedde, A., Gunn, R.N., Holden, J., Houle, S., Huang, S.C., Ichise, M., Iida, H., Ito, H., Kimura, Y., Koeppe, R.A., Knudsen, G.M., Knuuti, J., Lammertsma, A.A., Laruelle, M., Logan, J., Maguire, R.P., Mintun, M.A., Morris, E.D., Parsey, R., Price, J.C., Slifstein, M., Sossi, V., Suhara, T., Votaw, J.R., Wong, D.F., Carson, R.E., 2007. Consensus nomenclature for in vivo imaging of reversibly binding radioligands. J Cereb Blood Flow Metab 27, 1533-1539. Lezak, M.D., 1995. Neuropsychological Assessment. Oxford University Press, New York. Lumme, V., Hirvonen, M.M., Ilonen, T., Hirvonen, J., Nagren, K., Hietala, J., 2010. Cortical dopamine D2/D3 receptors and verbal memory in man. Neuroimage 51, 918922. Martikainen, I., Hirvonen, J., Kajander, J., Hagelberg, N., Mansikka, H., Nagren, K., Hietala, J., Pertovaara, A., 2007. Correlation of human cold pressor pain responses with 5-HT1A receptor binding in the brain. Brain Research 1172, 21-31. Mendelsohn, D., Riedel, W.J., Sambeth, A., 2009. Effects of acute tryptophan depletion on memory, attention and executive functions: a systematic review. Neurosci Biobehav Rev 33, 926-952. Meneses, A., Perez-Garcia, G., 2007. 5-HT(1A) receptors and memory. Neurosci Biobehav Rev 31, 705-727. Parsey, R.V., Arango, V., Olvet, D.M., Oquendo, M.A., Van Heertum, R.L., John Mann, J., 2005. Regional heterogeneity of 5-HT1A receptors in human cerebellum as assessed by positron emission tomography. Journal of Cerebral Blood Flow & Metabolism 25, 785-793. Parsey, R.V., Ogden, R.T., Miller, J.M., Tin, A., Hesselgrave, N., Goldstein, E., Mikhno, A., Milak, M., Zanderigo, F., Sullivan, G.M., 2010. Higher Serotonin 1A Binding in a Second Major Depression Cohort: Modeling and Reference Region Considerations. Biological Psychiatry 68, 170-178. Parsey, R.V., Oquendo, M.A., Simpson, N.R., Ogden, R.T., Van Heertum, R., Arango, V., Mann, J.J., 2002. Effects of sex, age, and aggressive traits in man on brain serotonin 5-HT1A receptor binding potential measured by PET using [C-11]WAY100635. Brain Res. 954, 173-182. Pike, V.W., McCarron, J.A., Lammertsma, A.A., Osman, S., Hume, S.P., Sargent, P.A., Bench, C.J., Cliffe, I.A., Fletcher, A., Grasby, P.M., 1996. Exquisite delineation of 5-HT1A receptors in human brain with PET and [carbonyl-11 C]WAY-100635. Eur J Pharmacol. 301, R5-7. 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., 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.

Riedel, W.J., Klaassen, T., Griez, E., Honig, A., Menheere, P.P., van Praag, H.M., 2002. Dissociable hormonal, cognitive and mood responses to neuroendocrine challenge: evidence for receptor-specific serotonergic dysregulation in depressed mood. Neuropsychopharmacology 26, 358-367. Roth, B.L., Hanizavareh, S.M., Blum, A.E., 2004. Serotonin receptors represent highly favorable molecular targets for cognitive enhancement in schizophrenia and other disorders. Psychopharmacology (Berl) 174, 17-24. Schechter, L.E., Dawson, L.A., Harder, J.A., 2002. The potential utility of 5-HT1A receptor antagonists in the treatment of cognitive dysfunction associated with Alzheimer s disease. Curr Pharm Des 8, 139-145. Schmitt, J.A., Wingen, M., Ramaekers, J.G., Evers, E.A., Riedel, W.J., 2006. Serotonin and human cognitive performance. Curr Pharm Des 12, 2473-2486. Sumiyoshi, T., Matsui, M., Yamashita, I., Nohara, S., Kurachi, M., Uehara, T., Sumiyoshi, S., Sumiyoshi, C., Meltzer, H.Y., 2001. The effect of tandospirone, a serotonin(1A) agonist, on memory function in schizophrenia. Biol Psychiatry. 49, 861-868. Theodore, W.H., Wiggs, E.A., Martinez, A.R., Dustin, I.H., Khan, O.I., Appel, S., Reeves-Tyer, P., Sato, S., 2012. Serotonin 1A receptors, depression, and memory in temporal lobe epilepsy. Epilepsia 53, 129-133. Wechsler, D., 1987. Wechsler Memory Scale-Revised Manual. The Psychological Corporation, Harcourt Brace Jovanovich, Inc, New York. Wechsler, D., 1992. Wechsler Adult Intelligence Scale–Revised Manual. The Psychological Corporation, Cleveland, Ohio. Wingen, M., Kuypers, K.P., Ramaekers, J.G., 2007. Selective verbal and spatial memory impairment after 5-HT1A and 5-HT2A receptor blockade in healthy volunteers pre-treated with an SSRI. J Psychopharmacol 21, 477-485. Yasuno, F., Suhara, T., Nakayama, T., Ichimiya, T., Okubo, Y., Takano, A., Ando, T., Inoue, M., Maeda, J., Suzuki, K., 2003. Inhibitory effect of hippocampal 5-HT1A receptors on human explicit memory. Am J Psychiatry 160, 334-340. Zilles, D., Meyer, J., Schneider-Axmann, T., Ekawardhani, S., Gruber, E., Falkai, P., Gruber, O., 2012. Genetic polymorphisms of 5-HTT and DAT but not COMT differentially affect verbal and visuospatial working memory functioning. Eur Arch Psychiatry Clin Neurosci 262, 667-676.

Figure legends Figure 1. Serotonin 5-HT1A receptor binding (BPP) correlates positively with composite verbal memory score. Solid line is the linear regression line, and curved dashed lines represent the bounds for the 95% confidence interval of the regression. Figure 2. Statistical parametric mapping (SPM) analysis shows positive correlation between serotonin 5-HT1A receptor binding (BPP) and composite verbal memory score at voxel level. A large cluster encompassing neocortical regions comprised of 47881 voxels and had a cluster-level FWE-corrected p value of <0.001. Voxel-level uncorrected p <0.005 (T value > 2.82) and cluster-level FWE-corrected p <0.05 were thresholds for statistical significance. Color bar represents T values at voxel level. The significant cluster is projected on sagittal, axial and coronal slices of a template T1-weighed MR image in the standard stereotactic space, and X, Y, and Z values indicate the corresponding MNI co-ordinates. Figure 3. Surface projections of coefficient of determination (R2) of the association between serotonin 5-HT1A receptor binding (BPP) and composite verbal memory score at voxel level. Coefficient of determination reveals the percentage of variation in verbal memory that is explained by 5-HT1A receptor binding. R2 data is thresholded at >10%.

Acknowledgements We thank the staff of the Turku PET Centre for assistance.

Conflicts of interest All authors declare that they have no conflicts of interests.

Contributors Jarmo Hietala designed the study, Jani Penttilä, Jussi Hirvonen and Ville Lumme collected the PET data, Tuula Ilonen did the cognitive ratings, Kjell Någren was responsible for the production of the tracer, Jussi Hirvonen and Lauri Tuominen analyzed the data, and Jani Penttilä, Jussi Hirvonen and Lauri Tuominen wrote the draft of the manuscript. All authors contributed to and have approved the final manuscript.

Role of Funding Source The study was supported by Turku University Hospital grant #P3848. The funding source had no involvement in the collection, analysis, writing of the report, or in the decision to submit the paper for publication.

Table 1 Correlations between 5-HT1A receptor BPP and performance in cognitive tests for 24 healthy subjects Whole brain analysis

Regional analyses

Lateral temporal cortex F p r p r p r p r p Wechsler Memory Scale-Revised (WMS-R); tests of verbal memory 0. 7.2 0.01 0. 0.0 0.5 0.0 0. 00 0.01 Verbal memorya 4 4 56 07 4 10 57 6 0.52 2 0. Logical memory, 5.5 0.02 0. 0.0 0.4 0.0 0. 01 0.03 immediate 5 9 50 19 8 24 52 4 0.46 2 0. Logical memory, 7.5 0.01 0. 0.0 0.5 0.0 0. 00 0.02 delayed 4 2 53 11 1 15 56 7 0.49 1 Verbal paired 0. associates, 4.8 0.03 0. 0.0 0.4 0.0 0. 02 0.02 immediate 8 9 50 17 8 23 47 7 0.49 2 Verbal paired associates, 3.3 0.4 0.0 0.4 0.0 0. 0.0 delayed 7 0.08 1 6 2 5 40 7 0.40 0.06 Wechsler Adult Intelligence Scale-Revised (WAIS-R) 0. 3.2 0.4 0.0 0.3 0.1 0. 04 Full scale IQ 6 0.09 1 6 6 0 44 1 0.41 0.06 0. 3.1 0.4 0.0 0.3 0.1 0. 04 Verbal IQ 7 0.09 1 6 6 0 44 3 0.40 0.06 Prefront Anterior Parieta al cortex cingulate l cortex

1.9 0 0.18

0.3 0.1 1 6

0.2 0.2 9 0

0. 0.1 35 1 0.33 0.13

Information

0.9 6 0.34

0.1 0.4 9 0

0.2 0.2 5 7

0. 0.2 24 8 0.19 0.41

Digit span

0.6 0 0.45

0.2 0.2 6 4

0.1 0.5 5 1

Vocabulary

4.2 0.05 2 3

0. 0.0 46 32

0.4 0.0 3 47

Arithmetics

4.5 0.04 8 5

0. 0.0 46 30

0.3 0.0 7 9

Comprehension

3.4 9 0.08

0. 0.0 44 42

0.3 0.0 8 8

7.3 0.01 4 4

0. 0.0 53 11

0.5 0.0 1 15

0. 0.2 24 8 0.22 0. 0. 02 49 2 0.46 0. 0. 02 47 9 0.44 0. 0. 02 47 8 0.41 0. 0. 00 58 4 0.56

0.5 3 0.48

0.1 0.3 9 9

0.1 0.4 7 5

0. 0.4 19 1 0.23 0.31

Performance IQ Verbal WAIS scales:

Similarities Performance WAIS scales: Picture completion

Mesial temporal cortexb r p

0.35 0.11 0.32 0.15 0.41 0.06 0.32 0.15 0.26 0.24

0.27 0.22 0.23 0.30 0.26 0.25

Raphe r

p

0. 5 6 0. 5 4 0. 5 9 0. 3 8 0. 3 8

0. 00 7 0. 01 0 0. 00 4

0. 3 5 0. 4 2 0. 1 7

0. 08 0. 08

0. 11 0. 05 0. 43

0.34

0.09 0.69

0.03 2

0.28 0.22

0.04 1

0.37 0.09

0.06

0.22 0.32

0.00 7

0.36 0.10

0. 3 2 0. 1 3 0. 4 2 0. 4 1 0. 4 7 0. 5 7

0.08 0.72

0. 1 0. 6 49

0.12 0.59

0. 15 0. 55 0. 05 0 0. 06 0. 02 9 0. 00 6

Picture arrangement

1.1 9 0.29

0.2 0.3 1 6

0.2 0.3 1 5

0. 0.2 26 5 0.23 0.30

Block design

1.6 2 0.22

0.2 0.2 8 1

0.2 0.1 9 9

Object assembly

3.2 4 0.09

0.4 0.0 0 7

0.3 0.0 9 8

0. 0.1 34 2 0.30 0.17 0. 0. 03 45 8 0.41 0.06

0.4 0.1 0.5 Digit symbol 1 0.53 5 0 Wisconsin Card Sorting Test (WCST)

0.1 0.5 3 6

0. 0.4 17 5 0.15 0.50

0.23 0.30 0.20 0.38 0.32 0.15

0.17 0.44

0. 2 1 0. 2 2 0. 2 1 0. 0 2

0. 34 0. 32 0. 35 0. 92

0. 0 0. Errors 0.27 0.22 8 73 0. Perseverative 0.3 0.1 0.5 0.1 0.4 0. 0.7 0 0. responses 6 0.56 3 8 7 6 08 1 0.13 0.57 0.19 0.39 4 86 0. Perseverative 0.9 0.2 0.3 0.2 0.2 0. 0.4 0 0. errors 1 0.35 0 6 4 8 17 6 0.21 0.34 0.25 0.26 3 90 0. Number of 0.6 0.1 0.3 0.1 0.4 0. 0.5 0 0. categories 0 0.45 9 9 9 0 13 5 0.17 0.45 -0.19 0.39 5 81 aVerbal memory = Logical memory, immediate + Verbal paired associates, immediate. bMesial temporal cortex = hippocampus + amygdala. Bold font denotes statistically significant (p<0.05) correlations. 0.8 7 0.36

0.1 0.4 7 6

0.2 0.3 0 7

0. 0.5 12 9 0.20 0.38