HIPPOCAMPUS 24:1524–1532 (2014)
Hippocampal-Parietal Dysconnectivity and Glutamate Abnormalities in Unmedicated Patients With Schizophrenia Nina Vanessa Kraguljac,1 David Matthew White,1 Jennifer Hadley,1,2 Meredith Amanda Reid,1,2 and Adrienne Carol Lahti1*
ABSTRACT: Abnormalities in resting state connectivity in schizophrenia (SZ) are now well established, but the biological substrates of these functional alterations remain to be elucidated. We performed a combined functional magnetic resonance imaging and magnetic resonance spectroscopy study in 22 unmedicated patients with SZ and 22 matched healthy controls (HCs) to evaluate resting state functional connectivity of the hippocampus and Glx/Cr (a combined glutamate 1 glutamine peak normalized to creatine) in the hippocampus and investigate functional and neurometabolic abnormalities and examine the relationship between these. Functional connectivity between the left hippocampus and bilateral precuneus was significantly decreased in unmedicated patients with SZ when compared to HCs [t(4.22), cluster extent (kE) 5 751, PFDRcorr 5 0.001, Montreal Neurological Institute coordinates: x 5 24, y 5 256, z 5 44]. Glx/Cr in the hippocampus was significantly elevated in SZ (HC: mean 5 0.601/20.10 SZ: 0.671/20.10; F 5 5.742; P 5 0.02), but was not correlated with functional connectivity deficits (P > 0.05). In this study, we found hippocampal resting state functional connectivity deficits to the precuneus in unmedicated patients with SZ and an increase of Glx/Cr in the hippocampus, but did not observe a direct relationship between these abnormalities. However, our findings do not exclude the possibility of a shared underlying pathology, C 2014 Wiley Periodicals, Inc. which warrants further investigation. V KEY WORDS: functional magnetic resonance imaging; magnetic resonance spectroscopy; resting state; precuneus; default mode network; memory
INTRODUCTION Resting state functional connectivity allows determining the level of synchronicity between spatially distinct brain areas in the absence of a task, assessing large scale networks of the brain (Biswal et al., 1995). Abnormalities in resting state connectivity in schizophrenia (SZ) are now well established (reviewed in Kuhn and Gallinat, 2013), but biological substrates of these functional alterations remain to be elucidated.
1
Department of Psychiatry and Behavioral Neurobiology, University of Alabama at Birmingham; 2 Department of Biomedical Engineering, University of Alabama at Birmingham, Birmingham, Alabama. Grant sponsor: National Institute of Mental Health; Grant number: RO1 MH081014. *Correspondence to: Adrienne C. Lahti, MD, Department of Psychiatry and Behavioral Neurobiology, The University of Alabama at Birmingham, SC 501, 1720 7th Ave S Birmingham, AL, USA. E-mail: [email protected] Accepted for publication 13 July 2014. DOI 10.1002/hipo.22332 Published online 17 July 2014 in Wiley Online Library (wileyonlinelibrary.com). C 2014 WILEY PERIODICALS, INC. V
Given preclinical findings of correlations between glutamatergic neurotransmitter flux, neuronal firing rate, and blood oxygen level dependent signals (BOLD; Hyder et al., 2002; Smith et al., 2002), human studies reporting associations between glutamate levels and resting state connectivity patterns in healthy subjects (Enzi et al., 2012; Scheidegger et al., 2012; Duncan et al., 2013), and our recent findings of elevations in hippocampal glutamate in vivo measured with Proton Magnetic Resonance Spectroscopy (1H-MRS) in unmedicated patients with SZ (Kraguljac et al., 2013), we sought to investigate hippocampus functional connectivity patterns in relationship to glutamate abnormalities in SZ. In this multimodality neuroimaging study, we obtained functional magnetic resonance imaging (fMRI) during a resting state using a seed based approach to describe a functional hippocampus network and used 1H-MRS to measure Glx/Cr levels in the left hippocampus in 22 unmedicated patients with SZ and 22 healthy controls (HC) matched on gender, age, and parental socioeconomic status. Based on the literature and our previous findings, we hypothesized that we would find (1) abnormal hippocampus resting state connectivity patterns in SZ, (2) elevations of Glx/Cr in the hippocampus in SZ, and (3) a correlation between hippocampus functional connectivity abnormalities and Glx/Cr elevations.
MATERIALS AND METHODS Subjects Patients with SZ were recruited from the psychiatry emergency services and outpatient clinics at the University of Alabama at Birmingham. HC, matched on age, sex, and parental occupation, were recruited by advertisements in flyers and the university’s newspaper. Approval by the local Institutional Review Board was obtained. Written informed consent to participate was obtained after the subjects’ ability to give informed consent was evaluated (Carpenter et al., 2000). Diagnoses were established by review of medical records and consensus of two clinicians and then confirmed with the Diagnostic Interview for Genetic Studies (DIGS) (Nurnberger et al., 1994). Exclusion
HIPPOCAMPUS DYSCONNECTIVITY AND GLUTAMATE IN SCHIZOPHRENIA criteria were major medical conditions, substance abuse within 6 months prior to enrolment (except for nicotine), pregnancy, neurological disorders, and head injury with loss of consciousness for more than 2 min. Personal or family history in a firstdegree relative of an Axis I disorder was an exclusion criterion for HC. We included 22 SZ subjects who were off medications for at least two weeks prior to enrolment and 22 HC in this analysis. All subjects had been recruited for a prospective multimodality neuroimaging study with the goal to explore biomarkers of response to antipsychotic treatment. Neurometabolite measurements of some subjects and have been included in earlier reports (Hutcheson et al., 2012; Kraguljac et al., 2012b).
Clinical Variables Clinical assessments were obtained the day the scan was performed. The Brief Psychiatric Rating Scale (BPRS) and its positive and negative subscales were used to assess symptom severity (Flemenbaum and Zimmermann, 1973). Cognitive function was characterized using the repeatable battery for the assessment of neuropsychological status (RBANS; Randolph et al., 1998).
Image Acquisition All imaging was performed on a 3T head-only scanner (Magnetom Allegra, Siemens Medical Solutions, Erlangen, Germany), equipped with a circularly polarized transmit/receive head coil. A high-resolution structural scan was acquired for anatomic reference using the three-dimensional T1-weighted magnetization prepared rapid acquisition gradient-echo (MPRAGE) sequence (TR/TE/inversion time [TI] 5 2300/ 3.93/1100 ms, flip angle5 12 , 256 3 256 matrix, 1-mm isotropic voxels).
Proton Magnetic Resonance Spectroscopy A series of sagittal, coronal, and axial T1-weighted anatomical scans (gradient-recalled echo sequence, TR/TE 5 250/3.48 ms, flip angle 5 70 , 5 mm slice thickness, 1.5 mm gap, 512 3 512 matrix) were acquired for spectroscopic voxel placement. Data were collected from a voxel in the hippocampus (2.7 3 1.5 3 1 cm; Fig. 1). To control for head tilt, slices were realigned to midline anatomical landmarks in transverse and coronal orthogonal planes. To facilitate prescription of the hippocampal volume, axial MR images were obtained in an orientation tilted along the long axis of the hippocampi, as viewed in the sagittal images (Fig. 1). The voxel was placed in a region of the left hippocampus such that the amount of gray matter (GM) was maximized while avoiding major vessels. Manual shimming was performed to optimize field homogeneity across the voxel, and chemical shift selective (CHESS) pulses were used to suppress the water signal. Water-suppressed spectra were acquired using the point-resolved spectroscopy sequence [PRESS; TR/TE 5 2000/80 ms to optimize the Glx signal and minimize macromolecule contribution to the spec-
1525
FIGURE 1. On the left is an example of 1H-MRS voxel placement in the left hippocampus (2.7 3 1.5 3 1 cm). Image is displayed in radiological convention (right side of image is subject’s left side). On the right side is an example of a proton 1H-MRS spectrum. The black line is a spectrum (640 averages) obtained from the left hippocampus voxel, the line is an overlay of spectral fit. Cho: choline; Cr: creatine; Glx: glutamate 1 glutamine; NAA: N-acetyl-aspartate; ppm: parts per million. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
tra; 1200 Hz spectral bandwidth; 1024 points; number of averages 5 640 (21 min 20 s)] (Schubert et al., 2004). To limit possible effects of nicotine intoxication or withdrawal, patients were allowed, but not encouraged, to smoke up to one hour prior to acquisition of images. 1H-MRS data were analyzed in jMRUI (version 3.0). The residual water peak was removed using the Hankel-Lanczos singular values decomposition (HLSVD) filter. Spectra were quantified in the time domain using the AMARES (Advanced Method for Accurate, Robust, and Efficient Spectral fitting) algorithm (Vanhamme et al., 1997). As previously described, prior knowledge derived from in vitro and in vivo metabolite spectra was included in the model (Kraguljac et al., 2013). All peak shapes were fixed at Lorentzian. Cramer-Rao lower bounds (CRLB), used as an estimate of uncertainty, were calculated for each peak. Exclusion criteria were (1) line width of the magnitude signal during manual shimming water greater than 25 Hz at full width at half maximum (FWHM) (2) CRLB >20%, and (3) failure of the fitting algorithm. Based on these criteria, we excluded one SZ. Glx was quantified with respect to Cr. As previously reported, coefficient of variability of MRS measurements on five consecutive days in one HC was 11.93% for Glx/Cr (Hutcheson et al., 2012). We determined GM, white matter (WM), and cerebrospinal fluid (CSF) voxel content by segmenting subjects’ structural scans in SPM8. 1H-MRS voxel images, created from the 1H-MRS raw data headers, were used to mask each of the tissue classes, and the volumes were calculated in MATLAB.
Resting State Functional MRI Resting state functional MRI scans were acquired during a 5-min gradient recalled echo-planar imaging sequence (repetition time/echo time [TR/TE] 5 2100/30 ms, flip angle5 70 , field of view 5 24 3 24 cm2, 64 3 64 matrix, 4-mm slice thickness, 1-mm gap, 26 axial slices, 150 acquisitions). Subjects Hippocampus
1526
KRAGULJAC ET AL.
were instructed to keep their eyes open and stare passively ahead. Using SPM8, resting state fMRI data were slice timing corrected, resliced to 2 mm3 voxels, realigned using rigid-body motion transforms, coregistered to the high-resolution structural scan, normalized to MNI (Montreal Neurological Institute) space and spatially smoothed using diffeomorphic anatomical registration through exponentiated lie algebra (DARTEL) with a 6 mm3 at full-width half-maximum Gaussian kernel (Ashburner, 2007). To assess motion effects on connectivity data, we calculated mean absolute displacement of the brain from one timeframe to the next for each subject; Mean absolute displacement did not differ between groups (HC 5 0.191/20.13 mm, SZ 5 0.19 mm1/20.09 mm; t 5 0.110; P 5 0.91).To remove potential sources of noise, a nuisance regression was conducted using the six motion parameters identified during the realignment step and their first derivatives as regressors. Scan time points that were likely contaminated by motion were identified using the methods by Power et al. (2012) and interpolated prior to applying a second order band-pass filter (0.009 < f < 0.08 Hz) to the data (Carp, 2011). A principle component analysis was used to extract the components of WM and CSF necessary to explain 90% of signal variance from those regions. These were used as regressors in a second nuisance regression (Behzadi et al., 2007). Following data preprocessing and motion scrubbing, the mean BOLD time series from each voxel was extracted for all subjects. Each region’s mean BOLD time series was correlated to all other voxels in the brain to produce a functional connectivity map from each voxel for all subjects. The seed region for functional connectivity analysis was defined as the left hippocampus from the Automated Anatomical Labeling (AAL) atlas, implemented in the WFU PickAtlas software V.2 (tzourio-mazoyer automated anatomical labeling; Tzourio-Mazoyer et al., 2002; Maldjian et al., 2003). Following data preprocessing and motion scrubbing, the first eigenvariate of the BOLD time series from each region was extracted and correlated to all other voxels in the brain to produce functional connectivity maps. Pearson’s correlations were converted to normally distributed values using Fisher’s r-to-Z transform.
Statistical Analyses Statistical analyses were performed using SPSS 12.0 for Microsoft Windows (SPSS, Chicago, Il). For analyses of group differences in demographic and clinical variables independent t-tests and v2 analyses were performed. Differences in metabolite ratios were investigated with multivariate analyses of covariance (MANCOVA), using metabolites as within-group factors, disease state as between-group factor, and age, smoking (packs per day), voxel GM, and WM content as covariates. Planned contrasts were created to assess group differences in individual metabolites. Group-level functional connectivity maps were obtained by performing one-sample t-tests on each group’s participant level functional connectivity maps. Group differences in functional connectivity were assessed using a two-sample ttest on the groups’ participant-level functional connectivity Hippocampus
maps. All analyses were corrected for multiple comparisons at the cluster level using false discovery rate corrections (FDR), P< 0.01. In an exploratory fashion, we investigated correlations of hippocampus-precuneus connectivity strength with clinical variables. These analyses were conducted by extraction of first eigenvariate from our regions of interest – left hippocampus and precuneus. The time course extractions were then correlated with each other as a measure of functional connectivity, producing Pearson’s correlations coefficients for each individual and each region of interest pair. Before performing group comparisons, a Fisher’s r-to-Z transformation was done. The precuneus region of interest was defined in two ways: the area of deficit from the between groups analysis of left hippocampal connectivity, and a mask created of the precuneus from AAL pickatlas. Given structural heterogeneity and heterogeneity of resting state functional connectivity patterns of the precuneus (Margulies et al., 2009), we decided to do a post hoc analysis correlating clinical variables with connectivity strength between the hippocampus and precuneus subregions [Regions of interest were based on a report by Margulies et al. (2009)]; we separately extracted connectivity strength to the hippocampus in the anterior sensorimotor region (seeds 5 and 6), central cognitive/associative region (seeds 14 and 15), posterior visual region (seeds 17–19), and posterior cingulate/retrosplenial region (seeds 1–4, 8, 12, 20, and 21) of the precuneus. A P value of 0.05). Also, no correlation between connectivity strength between hippocampus and the area of functional connectivity deficit and Glx/Cr was observed (P > 0.05). Visual inspection of scatterplots also did not suggest a potential relationship between functional connectivity and Glx/Cr (plots not shown but available from authors per request).
7 24.95 (47.13)
DISCUSSION 47.77 (9.32) 9.09 (3.28) 6.73 (2.80) 69.27 75.64 67.00 85.59 77.27 73.45
(12.05) 99.64 (14.21) (17.95) 100.32 (10.83) (13.35) 93.50 (18.36) (10.10) 99.59 (14.47) (19.56) 106.91 (18.34) (16.15) 98.27 (8.29)
Hippocampal-Parietal Dysconnectivity and Glutamate Abnormalities in Unmedicated Patients With Schizophrenia Nina Vanessa Kraguljac,1 David Matthew White,1 Jennifer Hadley,1,2 Meredith Amanda Reid,1,2 and Adrienne Carol Lahti1*
ABSTRACT: Abnormalities in resting state connectivity in schizophrenia (SZ) are now well established, but the biological substrates of these functional alterations remain to be elucidated. We performed a combined functional magnetic resonance imaging and magnetic resonance spectroscopy study in 22 unmedicated patients with SZ and 22 matched healthy controls (HCs) to evaluate resting state functional connectivity of the hippocampus and Glx/Cr (a combined glutamate 1 glutamine peak normalized to creatine) in the hippocampus and investigate functional and neurometabolic abnormalities and examine the relationship between these. Functional connectivity between the left hippocampus and bilateral precuneus was significantly decreased in unmedicated patients with SZ when compared to HCs [t(4.22), cluster extent (kE) 5 751, PFDRcorr 5 0.001, Montreal Neurological Institute coordinates: x 5 24, y 5 256, z 5 44]. Glx/Cr in the hippocampus was significantly elevated in SZ (HC: mean 5 0.601/20.10 SZ: 0.671/20.10; F 5 5.742; P 5 0.02), but was not correlated with functional connectivity deficits (P > 0.05). In this study, we found hippocampal resting state functional connectivity deficits to the precuneus in unmedicated patients with SZ and an increase of Glx/Cr in the hippocampus, but did not observe a direct relationship between these abnormalities. However, our findings do not exclude the possibility of a shared underlying pathology, C 2014 Wiley Periodicals, Inc. which warrants further investigation. V KEY WORDS: functional magnetic resonance imaging; magnetic resonance spectroscopy; resting state; precuneus; default mode network; memory
INTRODUCTION Resting state functional connectivity allows determining the level of synchronicity between spatially distinct brain areas in the absence of a task, assessing large scale networks of the brain (Biswal et al., 1995). Abnormalities in resting state connectivity in schizophrenia (SZ) are now well established (reviewed in Kuhn and Gallinat, 2013), but biological substrates of these functional alterations remain to be elucidated.
1
Department of Psychiatry and Behavioral Neurobiology, University of Alabama at Birmingham; 2 Department of Biomedical Engineering, University of Alabama at Birmingham, Birmingham, Alabama. Grant sponsor: National Institute of Mental Health; Grant number: RO1 MH081014. *Correspondence to: Adrienne C. Lahti, MD, Department of Psychiatry and Behavioral Neurobiology, The University of Alabama at Birmingham, SC 501, 1720 7th Ave S Birmingham, AL, USA. E-mail: [email protected] Accepted for publication 13 July 2014. DOI 10.1002/hipo.22332 Published online 17 July 2014 in Wiley Online Library (wileyonlinelibrary.com). C 2014 WILEY PERIODICALS, INC. V
Given preclinical findings of correlations between glutamatergic neurotransmitter flux, neuronal firing rate, and blood oxygen level dependent signals (BOLD; Hyder et al., 2002; Smith et al., 2002), human studies reporting associations between glutamate levels and resting state connectivity patterns in healthy subjects (Enzi et al., 2012; Scheidegger et al., 2012; Duncan et al., 2013), and our recent findings of elevations in hippocampal glutamate in vivo measured with Proton Magnetic Resonance Spectroscopy (1H-MRS) in unmedicated patients with SZ (Kraguljac et al., 2013), we sought to investigate hippocampus functional connectivity patterns in relationship to glutamate abnormalities in SZ. In this multimodality neuroimaging study, we obtained functional magnetic resonance imaging (fMRI) during a resting state using a seed based approach to describe a functional hippocampus network and used 1H-MRS to measure Glx/Cr levels in the left hippocampus in 22 unmedicated patients with SZ and 22 healthy controls (HC) matched on gender, age, and parental socioeconomic status. Based on the literature and our previous findings, we hypothesized that we would find (1) abnormal hippocampus resting state connectivity patterns in SZ, (2) elevations of Glx/Cr in the hippocampus in SZ, and (3) a correlation between hippocampus functional connectivity abnormalities and Glx/Cr elevations.
MATERIALS AND METHODS Subjects Patients with SZ were recruited from the psychiatry emergency services and outpatient clinics at the University of Alabama at Birmingham. HC, matched on age, sex, and parental occupation, were recruited by advertisements in flyers and the university’s newspaper. Approval by the local Institutional Review Board was obtained. Written informed consent to participate was obtained after the subjects’ ability to give informed consent was evaluated (Carpenter et al., 2000). Diagnoses were established by review of medical records and consensus of two clinicians and then confirmed with the Diagnostic Interview for Genetic Studies (DIGS) (Nurnberger et al., 1994). Exclusion
HIPPOCAMPUS DYSCONNECTIVITY AND GLUTAMATE IN SCHIZOPHRENIA criteria were major medical conditions, substance abuse within 6 months prior to enrolment (except for nicotine), pregnancy, neurological disorders, and head injury with loss of consciousness for more than 2 min. Personal or family history in a firstdegree relative of an Axis I disorder was an exclusion criterion for HC. We included 22 SZ subjects who were off medications for at least two weeks prior to enrolment and 22 HC in this analysis. All subjects had been recruited for a prospective multimodality neuroimaging study with the goal to explore biomarkers of response to antipsychotic treatment. Neurometabolite measurements of some subjects and have been included in earlier reports (Hutcheson et al., 2012; Kraguljac et al., 2012b).
Clinical Variables Clinical assessments were obtained the day the scan was performed. The Brief Psychiatric Rating Scale (BPRS) and its positive and negative subscales were used to assess symptom severity (Flemenbaum and Zimmermann, 1973). Cognitive function was characterized using the repeatable battery for the assessment of neuropsychological status (RBANS; Randolph et al., 1998).
Image Acquisition All imaging was performed on a 3T head-only scanner (Magnetom Allegra, Siemens Medical Solutions, Erlangen, Germany), equipped with a circularly polarized transmit/receive head coil. A high-resolution structural scan was acquired for anatomic reference using the three-dimensional T1-weighted magnetization prepared rapid acquisition gradient-echo (MPRAGE) sequence (TR/TE/inversion time [TI] 5 2300/ 3.93/1100 ms, flip angle5 12 , 256 3 256 matrix, 1-mm isotropic voxels).
Proton Magnetic Resonance Spectroscopy A series of sagittal, coronal, and axial T1-weighted anatomical scans (gradient-recalled echo sequence, TR/TE 5 250/3.48 ms, flip angle 5 70 , 5 mm slice thickness, 1.5 mm gap, 512 3 512 matrix) were acquired for spectroscopic voxel placement. Data were collected from a voxel in the hippocampus (2.7 3 1.5 3 1 cm; Fig. 1). To control for head tilt, slices were realigned to midline anatomical landmarks in transverse and coronal orthogonal planes. To facilitate prescription of the hippocampal volume, axial MR images were obtained in an orientation tilted along the long axis of the hippocampi, as viewed in the sagittal images (Fig. 1). The voxel was placed in a region of the left hippocampus such that the amount of gray matter (GM) was maximized while avoiding major vessels. Manual shimming was performed to optimize field homogeneity across the voxel, and chemical shift selective (CHESS) pulses were used to suppress the water signal. Water-suppressed spectra were acquired using the point-resolved spectroscopy sequence [PRESS; TR/TE 5 2000/80 ms to optimize the Glx signal and minimize macromolecule contribution to the spec-
1525
FIGURE 1. On the left is an example of 1H-MRS voxel placement in the left hippocampus (2.7 3 1.5 3 1 cm). Image is displayed in radiological convention (right side of image is subject’s left side). On the right side is an example of a proton 1H-MRS spectrum. The black line is a spectrum (640 averages) obtained from the left hippocampus voxel, the line is an overlay of spectral fit. Cho: choline; Cr: creatine; Glx: glutamate 1 glutamine; NAA: N-acetyl-aspartate; ppm: parts per million. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
tra; 1200 Hz spectral bandwidth; 1024 points; number of averages 5 640 (21 min 20 s)] (Schubert et al., 2004). To limit possible effects of nicotine intoxication or withdrawal, patients were allowed, but not encouraged, to smoke up to one hour prior to acquisition of images. 1H-MRS data were analyzed in jMRUI (version 3.0). The residual water peak was removed using the Hankel-Lanczos singular values decomposition (HLSVD) filter. Spectra were quantified in the time domain using the AMARES (Advanced Method for Accurate, Robust, and Efficient Spectral fitting) algorithm (Vanhamme et al., 1997). As previously described, prior knowledge derived from in vitro and in vivo metabolite spectra was included in the model (Kraguljac et al., 2013). All peak shapes were fixed at Lorentzian. Cramer-Rao lower bounds (CRLB), used as an estimate of uncertainty, were calculated for each peak. Exclusion criteria were (1) line width of the magnitude signal during manual shimming water greater than 25 Hz at full width at half maximum (FWHM) (2) CRLB >20%, and (3) failure of the fitting algorithm. Based on these criteria, we excluded one SZ. Glx was quantified with respect to Cr. As previously reported, coefficient of variability of MRS measurements on five consecutive days in one HC was 11.93% for Glx/Cr (Hutcheson et al., 2012). We determined GM, white matter (WM), and cerebrospinal fluid (CSF) voxel content by segmenting subjects’ structural scans in SPM8. 1H-MRS voxel images, created from the 1H-MRS raw data headers, were used to mask each of the tissue classes, and the volumes were calculated in MATLAB.
Resting State Functional MRI Resting state functional MRI scans were acquired during a 5-min gradient recalled echo-planar imaging sequence (repetition time/echo time [TR/TE] 5 2100/30 ms, flip angle5 70 , field of view 5 24 3 24 cm2, 64 3 64 matrix, 4-mm slice thickness, 1-mm gap, 26 axial slices, 150 acquisitions). Subjects Hippocampus
1526
KRAGULJAC ET AL.
were instructed to keep their eyes open and stare passively ahead. Using SPM8, resting state fMRI data were slice timing corrected, resliced to 2 mm3 voxels, realigned using rigid-body motion transforms, coregistered to the high-resolution structural scan, normalized to MNI (Montreal Neurological Institute) space and spatially smoothed using diffeomorphic anatomical registration through exponentiated lie algebra (DARTEL) with a 6 mm3 at full-width half-maximum Gaussian kernel (Ashburner, 2007). To assess motion effects on connectivity data, we calculated mean absolute displacement of the brain from one timeframe to the next for each subject; Mean absolute displacement did not differ between groups (HC 5 0.191/20.13 mm, SZ 5 0.19 mm1/20.09 mm; t 5 0.110; P 5 0.91).To remove potential sources of noise, a nuisance regression was conducted using the six motion parameters identified during the realignment step and their first derivatives as regressors. Scan time points that were likely contaminated by motion were identified using the methods by Power et al. (2012) and interpolated prior to applying a second order band-pass filter (0.009 < f < 0.08 Hz) to the data (Carp, 2011). A principle component analysis was used to extract the components of WM and CSF necessary to explain 90% of signal variance from those regions. These were used as regressors in a second nuisance regression (Behzadi et al., 2007). Following data preprocessing and motion scrubbing, the mean BOLD time series from each voxel was extracted for all subjects. Each region’s mean BOLD time series was correlated to all other voxels in the brain to produce a functional connectivity map from each voxel for all subjects. The seed region for functional connectivity analysis was defined as the left hippocampus from the Automated Anatomical Labeling (AAL) atlas, implemented in the WFU PickAtlas software V.2 (tzourio-mazoyer automated anatomical labeling; Tzourio-Mazoyer et al., 2002; Maldjian et al., 2003). Following data preprocessing and motion scrubbing, the first eigenvariate of the BOLD time series from each region was extracted and correlated to all other voxels in the brain to produce functional connectivity maps. Pearson’s correlations were converted to normally distributed values using Fisher’s r-to-Z transform.
Statistical Analyses Statistical analyses were performed using SPSS 12.0 for Microsoft Windows (SPSS, Chicago, Il). For analyses of group differences in demographic and clinical variables independent t-tests and v2 analyses were performed. Differences in metabolite ratios were investigated with multivariate analyses of covariance (MANCOVA), using metabolites as within-group factors, disease state as between-group factor, and age, smoking (packs per day), voxel GM, and WM content as covariates. Planned contrasts were created to assess group differences in individual metabolites. Group-level functional connectivity maps were obtained by performing one-sample t-tests on each group’s participant level functional connectivity maps. Group differences in functional connectivity were assessed using a two-sample ttest on the groups’ participant-level functional connectivity Hippocampus
maps. All analyses were corrected for multiple comparisons at the cluster level using false discovery rate corrections (FDR), P< 0.01. In an exploratory fashion, we investigated correlations of hippocampus-precuneus connectivity strength with clinical variables. These analyses were conducted by extraction of first eigenvariate from our regions of interest – left hippocampus and precuneus. The time course extractions were then correlated with each other as a measure of functional connectivity, producing Pearson’s correlations coefficients for each individual and each region of interest pair. Before performing group comparisons, a Fisher’s r-to-Z transformation was done. The precuneus region of interest was defined in two ways: the area of deficit from the between groups analysis of left hippocampal connectivity, and a mask created of the precuneus from AAL pickatlas. Given structural heterogeneity and heterogeneity of resting state functional connectivity patterns of the precuneus (Margulies et al., 2009), we decided to do a post hoc analysis correlating clinical variables with connectivity strength between the hippocampus and precuneus subregions [Regions of interest were based on a report by Margulies et al. (2009)]; we separately extracted connectivity strength to the hippocampus in the anterior sensorimotor region (seeds 5 and 6), central cognitive/associative region (seeds 14 and 15), posterior visual region (seeds 17–19), and posterior cingulate/retrosplenial region (seeds 1–4, 8, 12, 20, and 21) of the precuneus. A P value of 0.05). Also, no correlation between connectivity strength between hippocampus and the area of functional connectivity deficit and Glx/Cr was observed (P > 0.05). Visual inspection of scatterplots also did not suggest a potential relationship between functional connectivity and Glx/Cr (plots not shown but available from authors per request).
7 24.95 (47.13)
DISCUSSION 47.77 (9.32) 9.09 (3.28) 6.73 (2.80) 69.27 75.64 67.00 85.59 77.27 73.45
(12.05) 99.64 (14.21) (17.95) 100.32 (10.83) (13.35) 93.50 (18.36) (10.10) 99.59 (14.47) (19.56) 106.91 (18.34) (16.15) 98.27 (8.29)
