Introduction
Generalized anxiety disorder (GAD) is a common, debilitating condition characterized by long-term pathological worry and anxiety concerning different everyday activities and events (Kessler, Reference Kessler2000; Nutt et al. Reference Nutt, Argyropoulos, Hood and Potokar2006). It is associated with somatic and psychological symptoms, which may progressively lead to impairments in interpersonal and occupational functioning (Nutt et al. Reference Nutt, Ballenger, Sheehan and Wittchen2002; Whalen et al. Reference Whalen, Johnstone, Somerville, Nitschke, Polis, Alexander, Davidson and Kalin2008). Although GAD is one of the most prevalent anxiety disorder affecting the general population with a lifetime prevalence of about 2–3% across different countries (Lieb et al. Reference Lieb, Becker and Altamura2005; Lim et al. Reference Lim, Ng, Chua, Chiam, Won, Lee, Fones and Kua2005; Comer et al. Reference Comer, Blanco, Hasin, Liu, Grant, Turner and Olfson2011), its neurobiology is still not completely elucidated (Gorman et al. Reference Gorman, Hirschfeld and Ninan2002; Mathew et al. Reference Mathew, Price, Mao, Smith, Coplan, Charney and Shungu2008, Reference Mathew, Price, Shungu, Mao, Smith, Amiel and Coplan2010). Nonetheless, recent magnetic resonance imaging (MRI) studies have tried to explore the neural correlates of GAD (Cannistraro & Rauch, Reference Cannistraro and Rauch2003). In particular, abnormally increased superior temporal gyrus and amygdala volumes in the right hemisphere along with increased amygdalar activation in response to fear stimuli have been shown in paediatric GAD populations (De Bellis et al. Reference De Bellis, Casey, Dahl, Birmaher, Williamson, Thomas, Axelson, Frustaci, Boring, Hall and Ryan2000, Reference De Bellis, Keshavan, Shifflett, Iyengar, Dahl, Axelson, Birmaher, Hall, Moritz and Ryan2002; Thomas et al. Reference Thomas, Drevets, Dahl, Ryan, Birmaher, Eccard, Axelson, Whalen and Casey2001; McClure et al. Reference McClure, Monk, Nelson, Parrish, Adler, Blair, Fromm, Charney, Leibenluft, Ernst and Pine2007). Disturbances in activation of the amygdala and medial frontal regions have also been reported by recent functional MRI studies in adult GAD patients (Etkin et al. Reference Etkin, Prater, Schatzberg, Menon and Greicius2009; Paulesu et al. Reference Paulesu, Sambugaro, Torti, Danelli, Ferri, Scialfa, Sberna, Ruggiero, Bottini and Sassaroli2010). Moreover, increased N-acetylaspartate:creatine ratio in the right dorsolateral prefrontal cortex and higher relative metabolism in the right posterior temporal lobe and in the right precentral frontal gyrus have been found in GAD (Wu et al. Reference Wu, Buchsbaum, Hershey, Hazlett, Sicotte and Johnson1991; Mathew et al. Reference Mathew, Mao, Coplan, Smith, Sackeim, Gorman and Shungu2004). Together, these findings suggest structural, functional and biochemical abnormalities in subjects suffering from GAD, particularly affecting the right hemisphere.
Therefore, it is important to understand whether white-matter communication is impaired in GAD. In this regard, diffusion weighted imaging (DWI) is a non-invasive technique that allows the exploration of white-matter integrity in vivo (Assaf & Pasternak, Reference Assaf and Pasternak2008; Pasternak et al. Reference Pasternak, Sochen, Gur, Intrator and Assaf2009), providing information on the microstructure organization based on the apparent diffusion coefficient (ADC) of water molecules (Wu et al. Reference Wu, Field, Chung, Badie and Alexander2004; Alexander et al. Reference Alexander, Lee, Lazar and Field2007). In the presence of unaltered axonal membranes, diffusion of water molecules follows the same direction of dominant fibre tracts (anisotropic diffusion) and ADC is small (Lazar et al. Reference Lazar, Lee and Alexander2005). Conversely, abnormalities in size or number of myelinated axons, as well as loss of directional coherence of fibre tracts, induce free water distribution (isotropic diffusion), thus increasing ADC (Le Bihan et al. Reference Le Bihan, Mangin, Poupon, Clark, Pappata, Molko and Chabriat2001). Usually, ADC is increased in the case of white-matter disruption, for example in multiple sclerosis (Nusbaum et al. Reference Nusbaum, Tang, Wei, Buchsbaum and Atlas2000).
In this DWI study, for the first time to the best of our knowledge, white-matter microstructure coherence in patients with GAD was explored, expecting disrupted white-matter organization particularly in the right hemisphere.
Method
Subjects
A total of twelve patients with GAD who met the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) criteria were recruited from the out-patient Psychiatric Clinic of the University Hospital of Udine, Italy, as diagnosed with Structured Clinical Interview for Axis I DSM-IV Disorders (SCID-I). Diagnoses were confirmed by the clinical consensus of two staff psychiatrists. Of the 12 patients, seven were receiving regular antidepressants at the time of MRI scanning: three were taking venlafaxine, one sertraline, one amitriptiline, one paroxetine and one citalopram. In addition, six patients were also taking as-needed benzodiapines, whereas five patients were medication free. Patients with co-morbid Axis I disorders, including current major depressive episodes, alcohol or substance abuse, history of traumatic head injury with loss of consciousness, neurological or medical illnesses were excluded. A total of 15 healthy control subjects matched for age, gender, handedness and ethnicity to patients were also studied. They were recruited from the local community and had no DSM-IV Axis I disorders, alcohol or substance abuse, head trauma or neurological or major medical illness. The Brief Psychiatric Rating Scale (BPRS 24-item version) (Ventura et al. Reference Ventura, Nuechterlein, Subotnik, Gutkind and Gilbert2000), the Hamilton Rating Scale for Anxiety (HAMA; Hamilton, Reference Hamilton1959) and the Global Assessment Functioning (GAF) scale (Endicott et al. Reference Endicott, Spitzer, Fleiss and Cohen1976) were administered to all participants. Handedness was determined by the Edinburgh Inventory (Oldfield, Reference Oldfield1971). All the relevant sociodemographic and clinical characteristics of the subjects are shown in Table 1.
Sample characteristics
GAD, Generalized anxiety disorder; GAF, Global Assessment Functioning; BPRS, Brief Psychiatric Rating Scale; HAMA, Hamilton Anxiety Rating Scale.
Data are given as mean (standard deviation) or as number of patients.
This study was approved by the biomedical ethics committee of the Azienda Ospedaliero-Universitaria (AOU) of Udine. Written informed consent was obtained from all participants before participation.
MRI protocol
Subjects were scanned using a 1.5 T Avanto magnetic resonance scanner (Siemens, Germany). The standard head coil was applied for radio frequency transmission and reception of nuclear magnetic resonance signals. Head motion was reduced with restraining foam pads. First, a set of three orthogonal T1-weighted images was used as localizers [echo time (TE)=5 ms, repetition time (TR)=20 ms, matrix=512×512]. The presence of any brain focal lesion was investigated using a two-dimension proton density T2-weighted sequence in the axial orientation (TR=2500 ms, TE1=24 ms, TE2=125 ms, flip angle=150°, slice thickness=5 mm, matrix=512×512), parallel to the anterior commissure–posterior commissure line. For diffusion weighting, a two-dimension scan using a T2-weighted spin echo-echo planar scanning sequence was performed (TR=3200 ms, TE=94, flip angle=90°, slice thickness=5 mm, matrix=192×192; these parameters were the same for b=0, b=1000, and the ADC maps).
Image analysis
The ADCs of water molecules for the corpus callosum and the cortical white matter were detected by using in-house developed software written in MatLab (version 7; The Mathworks Inc., USA). Circular regions of interest (ROIs) were placed in the left and right side on the non-diffusion weighted (b=0) echoplanar images and were then automatically transferred to the corresponding maps to obtain the ADC of water molecules in accordance with well-established methods from our group (Brambilla et al. Reference Brambilla, Cerini, Gasparini, Versace, Andreone, Vittorini, Barbui, Pelizza, Nose, Barlocco, Marrella, Gregis, Tournikioti, David, Keshavan and Tansella2005; Andreone et al. Reference Andreone, Tansella, Cerini, Versace, Rambaldelli, Perlini, Dusi, Pelizza, Balestrieri, Barbui, Nose, Gasparini and Brambilla2007; Mengotti et al. Reference Mengotti, D'Agostini, Terlevic, De Colle, Biasizzo, Londero, Ferro, Rambaldelli, Balestrieri, Zanini, Fabbro, Molteni and Brambilla2011). For all measures, a trained rater blind to study hypotheses, group assignment and sociodemographic/clinical data measured all scans, after being trained by an experienced rater. They achieved high reliability, as defined by intra-class correlation coefficients over 0.90, established by tracing 10 training scans. The ADC maps were obtained from the diffusion images with b=1000, according to the following equation:
![-b{\rm ADC } \equals {\rm ln }\ \left[ {A\lpar b\rpar \sol \lpar 0\rpar } \right]\comma](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20160927005007949-0675:S0033291711001255:S0033291711001255_eqnU1.gif?pub-status=live){kind=small}
where A(b) is the measured echo magnitude, b is the measure of diffusion weighting and A(0) is the echo magnitude without diffusion gradient applied (Basser, Reference Basser and Atlas2002). The resulting ADC was expressed in units of 10−5 mm2/s.
For the corpus callosum, circular ROIs standardized at five pixels (corresponding to an area of 0.16 cm2) were placed in the genu and in the splenium, after checking adjacent slices to ensure that partial volume effects from cerebrospinal fluid were minimized. Landmarks for callosal subregions were adapted from Witelson (Reference Witelson1989). Two slices in the axial plane were chosen for detecting the most anterior part of the corpus callosum, the first one at the level of the lateral ventricle anterior horns and of the third ventricle and the second one at the level of the anterior and posterior horns of the lateral ventricles. For measuring the posterior section of the corpus callosum, one slice in the axial plane at the level of the of the lateral ventricle anterior horns and of the third ventricle was taken. The two resulting subsections corresponded approximately to two anatomical subdivisions of the corpus callosum (i.e. rostrum+genu and splenium).
For cortical white matter, ADCs were obtained by placing, bilaterally, the circular ROIs in the frontal, temporal, parietal and occipital cortex in reference to standard brain atlases (Jackson & Duncan, Reference Jackson and Duncan1996; Patel & Friedman, Reference Patel and Friedman1997) and according to previous studies (Sun et al. Reference Sun, Wang, Cui, Breeze, Du, Wang, Cong, Zhang, Li, Hong and Zhang2003; Wolkin et al. Reference Wolkin, Choi, Szilagyi, Sanfilipo, Rotrosen and Lim2003; Kumra et al. Reference Kumra, Ashtari, McMeniman, Vogel, Augustin, Becker, Nakayama, Gyato, Kane, Lim and Szeszko2004; Kitamura et al. Reference Kitamura, Matsuzawa, Shioiri, Someya, Kwee and Nakada2005). Specifically, for the frontal cortex the ROIs were positioned in the axial slice at the level of the genu of corpus callosum (standardized at 43.5 mm2), then in the inferior one (standardized at 43.5 mm2) and in the two superior slices (standardized at 84.4 mm2), posteriorly and medially to the frontal horns of the lateral ventricles. For the temporal cortex, the ROIs (standardized at 43.5 mm2) were positioned in the axial slice at the level of the lateral fissure and in the inferior slice, posteriorly and laterally to the lateral fissure. For the parietal cortex, the ROIs (standardized at 84.4 mm2) were placed in the axial slice when the lateral ventricles first disappeared and were positioned posteriorly to the postcentral sulcus. For the occipital cortex, the ROIs (standardized at 43.5 mm2) were placed in the two inferior axial slices where the occipital horns of the lateral ventricles become visible, posteriorly to the occipital horns.
Statistical analyses
Data were analysed using the Statistical Package for the Social Sciences (SPSS v. 15.0; SPSS Inc., USA). The level of significance was placed at 0.05. The Kolmogorov–Smirnov test was used to check for normality of the data. Group differences in sociodemographic and clinical variables were examined using the independent-group t test, Mann–Whitney test, and Pearson χ2, or Fisher's exact test, as appropriate.
A general linear model (GLM) for repeated measures (hemisphere as the repeated-measures factor) was performed to compare the ADCs between patients with GAD and healthy control subjects. The assumptions that the vector of the measures followed a multivariate normal distribution (Shapiro–Wilk test) and the variance–covariance matrices were circular in form (Mauchly's test) were verified. Univariate GLM was performed to compare the ADC measures between patients with GAD and healthy subjects.
Pearson's correlation and Spearman's correlation analyses were used to explore possible associations between age and clinical variables, respectively, and ADC values, applying the Bonferroni correction for multiple comparisons.
Results
Patients with GAD had significantly greater ADC values compared with healthy individuals [GLM for repeated measures, hemisphere as the repeated-measures factor: F=4.29, degrees of freedom (df)=1, 25, p=0.04], with greater values for the right splenium (univariate GLM: F=13.89, df=1, 25, p=0.001) and the right parietal cortex (univariate GLM: F=11.93, df=1, 25, p=0.002) (Table 2).
ADC measures in healthy controls and patients with GAD
ADC, Apparent diffusion coefficient; GAD, generalized anxiety disorder; GLM, general linear model.
Data are given as mean (standard deviation).
a GLM for repeated measures was performed to compare ADC values between patients with GAD and healthy controls (F 4=0.29, degrees of freedom=1,25, p=0.04). In the Table the statistics of the univariate GLM are reported.
Patients receiving antidepressant drugs (n=7), of whom five were also on benzodiazepines, did not significantly differ in any ADC values compared with drug-free patients (n=5) (GLM for repeated measures, hemisphere as the repeated-measures factor: F=4.29, df=1, 25, p>0.05). Also, ADC values did not significantly correlate with age (Pearson's correlations), BPRS total score, BPRS depressed mood, HAMA total score, HAMA depressed mood, GAF scores (Spearman's correlations) in both groups and with antidepressant dosages in patients with GAD (Spearman's correlations) after Bonferroni correction (p>0.05).
Discussion
This study showed, for the first time, that white-matter connectivity is impaired in patients with general anxiety disorder (GAD), particularly in the posterior right hemisphere including the parietal lobe and the callosal splenium. It has been shown that the parietal lobe discriminates the self from others and regulates selective attention and visuospatial processing ensuring environmental monitoring (Witte et al. Reference Witte, Villareal and Marrocco1996; Fogassi et al. Reference Fogassi, Ferrari, Gesierich, Rozzi, Chersi and Rizzolatti2005; Cunnington et al. Reference Cunnington, Windischberger, Robinson and Moser2006; Bellani et al. Reference Bellani, Ferro and Brambilla2010a). In particular, the posterior parietal cortex is crucial to recognize and process the location of salient stimuli (Saxe et al. Reference Saxe, Xiao, Kovacs, Perrett and Kanwisher2004; Constantinidis & Steinmetz, Reference Constantinidis and Steinmetz2005), which plays a key role for detecting and encoding social and emotional stimuli (Kret et al. Reference Kret, Denollet, Grezes and de Gelder2011). Also, the splenium of the corpus callosum, which is primarily formed by smaller-diameter axons (Aboitiz et al. Reference Aboitiz, Scheibel, Fisher and Zaidel1992), participates in high-order cognitive functions and allows communication between the associative parietal and temporal areas (Pfefferbaum et al. Reference Pfefferbaum, Sullivan and Carmelli2001; Baloch et al. Reference Baloch, Brambilla and Soares2009; Bellani et al. Reference Bellani, Marzi and Brambilla2009, Reference Bellani, Marzi, Savazzi, Perlini, Cerruti, Ferro, Marinelli, Sponda, Rambaldelli, Tansella and Brambilla2010b; Fabri et al. Reference Fabri, Polonara, Mascioli, Salvolini and Manzoni2011). Interestingly, an MRI study has recently found an altered functional network in GAD, including the amygdala, the dorsolateral prefrontal cortex (DLPFC) and the posterior parietal cortex (Etkin et al. Reference Etkin, Prater, Schatzberg, Menon and Greicius2009), which may be part of neural circuitry controlling emotional and non-emotional stimuli (Duncan et al. Reference Duncan, Knapp and Breese1996; Ochsner & Gross, Reference Ochsner and Gross2005). Structural and functional abnormalities have also been found in other regions of the right side of the brain in paediatric patients with GAD, such as the amygdala and the ventrolateral prefrontal cortex (De Bellis et al. Reference De Bellis, Casey, Dahl, Birmaher, Williamson, Thomas, Axelson, Frustaci, Boring, Hall and Ryan2000, Reference De Bellis, Keshavan, Shifflett, Iyengar, Dahl, Axelson, Birmaher, Hall, Moritz and Ryan2002; Thomas et al. Reference Thomas, Drevets, Dahl, Ryan, Birmaher, Eccard, Axelson, Whalen and Casey2001; Monk et al. Reference Monk, Nelson, McClure, Mogg, Bradley, Leibenluft, Blair, Chen, Charney, Ernst and Pine2006, Reference Monk, Telzer, Mogg, Bradley, Mai, Louro, Chen, McClure-Tone, Ernst and Pine2008; McClure et al. Reference McClure, Monk, Nelson, Parrish, Adler, Blair, Fromm, Charney, Leibenluft, Ernst and Pine2007), supporting the hypothesis that right hemisphere alterations may be crucial for the pathophysiology of GAD, even during neurodevelopment. Furthermore, there are some electrophysiological and imaging studies in anxious individuals sustaining altered right hemisphere functions, particularly to the parietal cortex (Heller et al. Reference Heller, Nitschke, Etienne and Miller1997), such as greater electrical activity (Heller et al. Reference Heller, Etienne and Miller1995; Davidson et al. Reference Davidson, Abercrombie, Nitschke and Putnam1999), higher rates of glucose metabolism (Stapleton et al. Reference Stapleton, Morgan, Liu, Yung, Phillips, Wong, Shaya, Dannals and London1997), and dominant activation during conditioned fear (LaBar et al. Reference LaBar, Gatenby, Gore, LeDoux and Phelps1998). In general, parietal white matter has been underinvestigated in psychiatric disorders. Nonetheless, deficits have been observed in schizophrenia by some (Kyriakopoulos et al. Reference Kyriakopoulos, Perez-Iglesias, Woolley, Kanaan, Vyas, Barker, Frangou and McGuire2009; White et al. Reference White, Magnotta, Bockholt, Williams, Wallace, Ehrlich, Mueller, Ho, Jung, Clark, Lauriello, Bustillo, Schulz, Gollub, Andreasen, Calhoun and Lim2011), although not all diffusion imaging studies (Andreone et al. Reference Andreone, Tansella, Cerini, Versace, Rambaldelli, Perlini, Dusi, Pelizza, Balestrieri, Barbui, Nose, Gasparini and Brambilla2007; Brambilla & Tansella, Reference Brambilla and Tansella2007), possibly relating to misattribution of symptoms and of intentions to others (Antonius et al. Reference Antonius, Prudent, Rebani, D'Angelo, Ardekani, Malaspina and Hoptman2011; Vistoli et al. Reference Vistoli, Brunet-Gouet, Lemoalle, Hardy-Bayle and Passerieux2011). In contrast, callosal abnormalities have consistently been reported in autism, bipolar disorder and schizophrenia (Brambilla et al. Reference Brambilla, Nicoletti, Sassi, Mallinger, Frank, Kupfer, Keshavan and Soares2003, Reference Brambilla, Nicoletti, Sassi, Mallinger, Frank, Keshavan and Soares2004, Reference Brambilla, Cerini, Gasparini, Versace, Andreone, Vittorini, Barbui, Pelizza, Nose, Barlocco, Marrella, Gregis, Tournikioti, David, Keshavan and Tansella2005; Mengotti et al. Reference Mengotti, D'Agostini, Terlevic, De Colle, Biasizzo, Londero, Ferro, Rambaldelli, Balestrieri, Zanini, Fabbro, Molteni and Brambilla2011), suggesting that inter-hemispheric miscommunication may be independent of diagnosis, partly sustaining impaired cognition and psychopathological symptoms in major psychiatric disorders (Kubicki et al. Reference Kubicki, Styner, Bouix, Gerig, Markant, Smith, Kikinis, McCarley and Shenton2008; Brambilla et al. Reference Brambilla, Bellani, Yeh, Soares and Tansella2009; Whitford et al. Reference Whitford, Kubicki, Schneiderman, O'Donnell, King, Alvarado, Khan, Markant, Nestor, Niznikiewicz, McCarley, Westin and Shenton2010).
Considering the role of the parietal cortex in modulating social material, it can be argued that right posterior regions may be engaged in the cognitive regulation of excessive anxiety arising from environmental and affective stimuli. Therefore, loss of coherence in right posterior hemisphere white matter in GAD may interfere with cognitive strategies in processing and modulating external and internal stimuli (Wells, Reference Wells1999). Hypothetically, altered right hemisphere dysconnectivity may represent a trait marker for increased sensitivity to social/threat cues with anxious arousal (Engels et al. Reference Engels, Heller, Mohanty, Herrington, Banich, Webb and Miller2007). In accordance to cognitive theories of anxiety, there might be in patients suffering from GAD an impaired ability to manage the impact of (negative) emotions and daily stress. This would reinforce negative thought patterns, such as rumination, worries, self-blame and catastrophizing, ultimately resulting in general anxiety (Brantley et al. Reference Brantley, Mehan, Ames and Jones1999; Aikins & Craske, Reference Aikins and Craske2001; Legerstee et al. Reference Legerstee, Garnefski, Verhulst and Utens2011). In this regard, it is interesting to outline that dysfunctional coping cognitions have been reported in both adult and child patients with GAD who, for instance, have low esteem of their capacity to deal with stressors and negatively interpret ambiguous situations (Bradley et al. Reference Bradley, Mogg, Millar and White1995; Bogels & Zigterman, Reference Bogels and Zigterman2000). However, these hypotheses clearly remain speculative.
Limitations of the study
It should be considered that the population of this study was relatively small, partially restricting the generalization of the results. However, it was in line with the sample size of other MRI studies in this field (Mohlman et al. Reference Mohlman, Price, Eldreth, Chazin, Glover and Kates2009; Paulesu et al. Reference Paulesu, Sambugaro, Torti, Danelli, Ferri, Scialfa, Sberna, Ruggiero, Bottini and Sassaroli2010) and it was composed of very homogeneous patients with pure GAD and matched healthy controls. Second, seven out of 12 patients were treated with antidepressants plus benzodiazepines (n=5) or with antidepressant monotherapy (n=2). Therefore, although no differences between treated and drug-free patients were shown, it cannot completely be excluded that diffusion coefficients may have potentially been modulated by such psychotropic drugs, although it has been found that ADC values were minimally affected by antidepressants (Taylor et al. Reference Taylor, Kuchibhatla, Payne, Macfall, Sheline, Krishnan and Doraiswamy2008). Third, ADC was calculated on ROIs including water diffusivity of both intra-axonal and extracellular compartments; therefore, increased ADC values may be explained as consequence of either intracellular (such as disruption of microtubular integrity within axons, altered permeability of axonal membranes) or extracellular (such as deposition of extracellular matrix, reduced density of axons, gliosis, demyelination) anomalies (Beaulieu & Allen, Reference Beaulieu and Allen1994; Basser, Reference Basser and Atlas2002). Fourth, based on our method, the diffusion tensor sequence was not collected and we could not detect any specific antero-posterior white-matter tract.
Conclusions
In conclusion, this is the first study showing altered posterior right hemisphere white-matter tissue in GAD, located in the parietal cortex and callosal splenium. The findings were independent of age, clinical severity and, potentially, treatment, suggesting that this may represent a structural marker of the disease and may potentially sustain altered cognitive control over excessive anxiety. Future diffusion imaging investigations, combined with functional MRI, are expected to better elucidate by the means of fibre tracking and functional connectivity the communication between the parietal cortex and other key regions in the right side, such as the dorsolateral prefrontal cortex and amygdala, in sustaining cognitive processing of social and emotional stimuli in patients with GAD.
Acknowledgements
P.B. was partly supported by the American Psychiatric Institute for Research and Education (APIRE), by the Italian Ministry for University and Research and by the Italian Ministry of Health (IRCCS ‘E. Medea’).
Declaration of Interest
None.
