SEARCH

SEARCH BY CITATION

Keywords:

  • deficit syndrome;
  • diffusion tensor imaging;
  • myelin;
  • schizophrenia;
  • uncinate fasciculus

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES

Aims:  Schizophrenia is a psychiatric disorder manifesting with heterogeneous symptom clusters and clinical presentations. The deficit syndrome is the condition defined by the existence of primarily negative symptoms, and patients with the deficit syndrome differ from non-deficit patients on measures of brain structure and function. In the current study, by using diffusion tensor imaging (DTI), we investigated the frontotemporal connectivity that is hypothesized to differ between deficit and non-deficit schizophrenia.

Methods:  Twenty-nine patients and 17 healthy controls were included in the study. The patients had deficit (n = 11) or non-deficit (n = 18) schizophrenia and they were evaluated clinically with the Schedule for Deficit Syndrome (SDS) and Positive and Negative Syndrome Scale (PANSS). Diffusion-based images were obtained with a 1.5T Siemens Magnetic Resonance Imaging machine and analyses were carried out with Functional Magnetic Resonance Imaging of the Brain Library Software – Diffusion tool box software.

Results:  The fractional anisotropy values in the left uncinate fasciculus of schizophrenia patients with the deficit syndrome were lower than those of non-deficit patients and the controls. There were no differences between non-deficit schizophrenia patients and controls.

Conclusion:  These findings provide evidence of left uncinate fasciculus damage resulting in disrupted communication between orbitofrontal prefrontal areas and temporal areas in deficit schizophrenia patients.

A LARGE NUMBER of neuroimaging and neurophysiology studies have shown anatomical and functional abnormalities in multiple brain regions of schizophrenia patients.1–5 Among them, frontal and temporal regions contain the most frequently reported abnormalities. These two regions function in concert to perform many functions, such as speech, social cognition, decision-making, and emotional learning; functions which are defective in many schizophrenia patients. Recent data suggest that localized defects in either region, or abnormal connectivity leading to a disintegration of neuronal dynamics between the regions, leads to characteristic deficits in schizophrenia.6–8 Abnormal connectivity might be at the synaptic level or at the axonal level interconnecting the regions. Such axons are myelinated, and recent genetic and pathological studies indicate myelin abnormalities in interregional bundles in schizophrenia.9,10 In recent years, multiple studies have tested frontotemporal disconnectivity via diffusion tensor imaging (DTI), which allows quantification of water molecule movement in the white matter tracts. The ratio between the diffusion along the axonal fiber and the amount of diffusion perpendicular to it, is called diffusion anisotropy. If the anisotropy is high, then most of the diffusion occurs in the axonal direction, indicating a high level orientation in the underlying structure. A DTI index called fractional anisotropy (FA) is thought to be a marker of the structural integrity of fibers,11 the degree of myelination,12 coherence of fiber tracts,13 and fiber diameter and packing density;11 changes in this index could indicate changes in any of these characteristics of the white matter microstructure or a combination of them. In prefrontal areas of schizophrenia patients, reduced FA has been related to positron emission tomography (PET) evidence of reduced metabolism,14 to negative symptoms,15 and to poor clinical outcome.16 The uncinate fasciculus (UF) is the largest of three fasciculi which interconnect the frontal lobes to temporal regions (Fig. 1). However, the results of studies measuring fractional anisotropy (FA) in the UF are contradictory. An initial report17 showed no difference in UF FA but pointed out the disturbance of the normal asymmetry in schizophrenia patients. That report was followed by others showing reduced or even increased FA values in the UF compared to healthy controls.18–21

image

Figure 1. Uncinate fasciculus connecting frontal lobe to limbic regions of temporal lobe. (a) Axial view; (b) sagittal view.

Download figure to PowerPoint

One reason for the contradictory results might be clinical heterogeneity among schizophrenia subjects. It is likely that schizophrenia includes a heterogeneous group of disorders differing in cause and pathophysiology, and manifesting clinically with diverse symptoms, clinical courses, and treatment responses. Some specific symptom clusters might be more related to UF abnormality than others. Indeed, Szeszko et al.22 reported that reduced FA values were correlated with negative symptoms in recent-onset schizophrenia. This finding supported a structural study23 showing reduced white matter volume in the left UF in schizophrenia patients with primary negative symptoms. Previously, Friston and Frith8 had shown in a functional magnetic resonance imaging (fMRI) study that schizophrenia patients with psychomotor poverty, which is a prominent cluster of negative symptoms, have an abnormal pattern of functional connectivity between their left prefrontal and temporal areas. Based on these findings, we hypothesized that patients with prominent negative symptoms (deficit schizophrenia) would have reduced UF FA, which would not be seen in other schizophrenia patients with less prominent negative symptoms (non-deficit schizophrenia).

One caveat for testing our hypothesis concerned the heterogeneity of the negative symptoms. These symptoms may be either primary, or secondary to identifiable sources, such as positive symptoms, treatment with antipsychotics or social isolation. The distinction between secondary and primary negative symptoms bears important therapeutic implications, since the former are susceptible to improvement following treatment, while the latter likely persist in spite of treatment.24–26 Furthermore, patients with primary negative symptoms are globally more neuropsychologically impaired in neurocognitive tests and have reduced regional blood flow or glucose metabolism in the frontal and temporal cortex.27,28 Therefore, we used the concept of deficit schizophrenia defined by Carpenter et al.26 to identify a relatively homogenous subgroup of schizophrenia patients with primary negative symptoms and compared their UF FA values with non-deficit schizophrenia patients and healthy controls.

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES

Subjects

A total of 29 clinically stable schizophrenia outpatients treated with antipsychotic medication (11 deficit and 18 non-deficit) and 17 healthy volunteers participated in this study. A trained psychiatrist confirmed the diagnoses of schizophrenia using the Turkish version of the Structural Clinical Interview for DSM-IV (SCID). The diagnoses of deficit and non-deficit schizophrenia were reached with the help of the Turkish version of the Schedule for Deficit Syndrome (SDS).29,30 The SDS is a semi-structured instrument for the diagnoses of deficit and non-deficit schizophrenia patients, and was developed for the purpose of studying primary, enduring negative symptom patients. The six symptoms evaluated by the SDS (restricted affect, diminished emotional range, poverty of speech, curbing of interest, diminished sense of purpose and diminished social drive) are rated from normal 0 to severely affected 4. To fulfill criterion 1, a score of 2 or more must be obtained from two or more of the six symptoms. Criterion 2 defines the duration of the items from criterion 1, which must have been present continually for the last 12 months. Criterion 3 qualitatively evaluates the primary or secondary nature (depression, antipsychotic treatment, etc.) of the negative symptoms present.

All healthy volunteers were recruited via local advertisements. We included those who were of similar age and sex to the schizophrenia patients. Healthy controls were screened with the non-patient version of the SCID, and those with any axis I disorder or a first-degree relative with schizophrenia or bipolar disorder were excluded from the study. The other exclusion criteria for all healthy and patient volunteers were: (i) current or past neurological condition; (ii) history of head trauma with more than 3 min of unconsciousness; (iii) DSM-IV substance abuse or dependence other than nicotine; and (iv) being left-handed. Volunteers with schizophrenia were evaluated for their ability to provide informed consent before signing consent documents. All subjects gave written informed consent before participation in the study. This study was approved by the Ege University Ethics Committee.

Patients' symptoms were evaluated with the Positive and Negative Symptom Scale (PANSS) within 24 h of MRI-DTI scanning. Chlorpromazine equivalent doses31 of the patients' current antipsychotic medications are presented in Table 1.

Table 1.  Demographic and clinical variables of patients and controls
 Deficit schizophrenia (n = 11)Non-deficit schizophrenia (n = 18)Controls (n = 17)Comparison
  • Chlorpromazine equivalents.

  • a-a': P < 0.016, b-b': P = 0.016; b-b: P < 0.001 (post hoc Mann–Whitney U-test).

  • PANSS, positive and negative syndrome scale.

Age32.36 ± 8.2340.77 ± 12.2733.82 ± 10.11H = 5 d.f. = 2 P = 0.08
Sex (male/female)7/49/99/8X2 = 1.57 d.f. = 2 P = 0.5
Education (years)7.88 ± 4.88a9.23 ± 3.81a15.56 ± 3.6a'H = 17.9 d.f. = 2 P < 0.001
Duration of illness (months)81 ± 33.8146.88 ± 95.41U = 34.5 P = 0.05
Age of onset24.25 ± 8.3129.23 ± 13.37U = 48 P = 0.26
PANSS    
 Positive13.1 ± 5.3613.4 ± 6.52 U = 89.5 P = 0.98
 Negative31.4 ± 9.116.2 ± 7.35 U = 25.5 P = 0.001
 General41.2 ± 6.434 ± 9.58 U = 42 P = 0.021
 Psychopathology total89.5 ± 15260.94 ± 15.3U = 11.5 P < 0.001
 Antipsychotic doses543.57 ± 613.7390. 81 ± 277.1U = 54 P = 0.92
Fractional anisotropy in uncinate fasciculus    
 Left0.3 ± 0.01b0.35 ± 0.06b'0.36 ± 0.06b″H = 12.1 d.f. = 2 P = 0.002
 Right0.37 ± 0.020.38 ± 0.050.39 ± 0.04H = 1.17 d.f. = 2 P = 0.5

MR Imaging

MR scanning was performed on a 1.5 Tesla Siemens Symphony (Vision to Symphony Upgrade, Siemens Numaris/4 Syngo MR 2004a Erlangen, Germany) using a two-channel circularly polarized head coil. DTI data were acquired with a spin echo single-shot, echo-planar imaging sequence with sensitivity (SENSE = 2) encoding (2 × 2 × 2.2 mm voxels, 256 × 256 mm FOV, 128 × 128 acquired matrix and no gap), TR/TE = 10070/103 ms, with diffusion gradients applied along 60 non-collinear directions at a b factor of 700 s/mm2. A minimally weighted image with b = 0 s/mm2 was also acquired.

Image processing

Images were transferred to a Linux workstation to perform data construction and analysis. DICOM images were converted to 4D Nifti file format, and applied eddy current correction was performed using the relevant programs contained in the Functional Magnetic Resonance Imaging of the Brain (FMRIB) Software Library (FSL) (http://www.fmrib.ox.ac.uk/fsl/). After creating a binary brain mask, diffusion tensor models fit each voxel using DTI-FIT software of FDT v2.0, FSL's diffusion toolbox.32 To perform voxelwise analysis of multi-subject diffusion data, we preferred a Tract Based Spatial Statics (TBSS) approach to improve sensitivity because of non-linear registration and tract projection of the mean FA skeleton and FA values of these voxels. All FA data were aligned to a 1x1x1mm MNI152 standard space with non-linear registration, and aligned using the most representative subject data as a guideline for all subjects to generate a study-specific model. FA values of all subjects were projected onto the mean FA skeleton derived from all subjects, with a threshold of 0.2 to exclude FA values which were not significant. After orientation and registration of all subjects' data, a binary mask was created to asses FA values of the UF. Both left and right UF Region of Interest (ROI) masks were created from the Johns Hopkins Hospital white matter tractography atlas (http://www.cmrm.med.jhmi.edu)33,34 and a mean FA skeleton mask was weighted as 1 for 0.2 thresholded FA tracts and 0 for other regions. Mean FA values were calculated using fslmeants software of FSL with this mask on all subjects. The analyses were restricted to UF based on our a priori hypotheses mentioned in the introduction.

Data analyses

Due to the small number of subjects in the groups and non-homogeneous distribution of FA values in the schizophrenia group, we preferred non-parametric statistical tests for the data analyses. For multiple group comparisons, we used the Kruskal–Wallis test, while the Mann–Whitney U-test was preferred for the two-group comparisons. Spearman's rank correlation coefficient was used for testing the correlation between the clinical and imaging variables. The alpha value was accepted as 0.05 for group comparisons and 0.016 for post-hoc analyses.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES

Comparison of demographic and clinical variables

There were no age or sex differences between schizophrenia patients and healthy controls but as expected, there was a significant difference in the years of education (Table 1). Both deficit and non-deficit patients had shorter durations of education compared to controls.

Within the schizophrenia group, the duration of illness was longer in the non-deficit schizophrenia group but the age of onset was similar in the two groups. According to PANSS, both deficit and non-deficit patients had similar scores for positive symptoms but deficit patients had higher scores for negative symptoms, general psychopathology, and PANSS total (Table 1).

Correlation between the demographic and clinical variables with uncinate FA values

Right and left uncinate FA values did not correlate with age or education in either schizophrenia or control groups (Table 2). Neither were there any correlations between UF FA values and duration of illness, age of onset, PANSS scores or antipsychotic doses (Table 2). Repeating the correlation analyses for each schizophrenia subgroup again revealed no significant correlations (data not shown).

Table 2.  The correlation matrix between the demographic and clinical values with uncinate FA values
 Right uncinate FALeft uncinate FA
  • Chlorpromazine equivalents.

  • FA, fractional anisotropy; PANSS, positive and negative syndrome scale.

Controls  
 Age−0.1−0.3
 Education (years)−0.1−0.14
Schizophrenia patients  
 Age−0.3−0.03
 Education (years)0.20.16
 Age of onset−0.06−0.18
 Duration of illness (months)−0.270.19
PANSS  
 Positive0.040.1
 Negative0.2−0.02
 General  
 Psychopathology0.0460.056
 Total0.02−0.1
 Antipsychotic doses−0.0070.02

Comparison of uncinate fasciculus values

Left UF FA values of schizophrenia patients were significantly lower than those of controls (schizophrenia FA = 0.33 ± 0.05, controls FA = 0.36 ± 0.06; U = 142, P = 0.02) while the right UF FA values were similar in the groups (schizophrenia FA = 0.37 ± 0.03, controls FA = 0.38 ± 0.04; U = 199, P = 0.28).

Comparing schizophrenia patients subgrouped into deficit and non-defcit groups with healthy controls, left UF FA levels showed a difference while right UF FA levels did not (Table 1, Fig. 2). Post-hoc analyses showed that deficit patients had lower left FA values compared to non-deficit patients (U = 46, P = 0.016) and controls (U = 18, P < 0.001). There were no differences between non-deficit patients and controls (U = 124, P = 0.31).

image

Figure 2. Uncinate fasciculus fractional anisotropy (FA) values of patients and controls. **Deficit schizophrenia patients showed reduced fractional values only in the left uncinate fasciculus compared to the controls (U = 18, P < 0.001) and non-deficit patients (U = 46, P = 0.016) while FA values of non-deficit schizophrenia patients were comparable with controls (U = 124, P = 0.31). DSZ, deficit schizophrenia; NDSZ, non-deficit schizophrenia; HC, healthy controls.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES

Impaired frontotemporal functioning has long been hypothesized to be the basis of many schizophrenia symptoms. Turetsky et al.35 reported abnormal temporal lobe asymmetry only in patients with deficit schizophrenia, in whom a selective increased left temporal lobe cerebrospinal fluid volume was observed in comparison with both non-deficit patients and healthy controls. In a single-photon emission computed tomography study, Gonul et al.27 found that deficit patients had lower blood perfusion in frontal and temporal regions as well as in parietal areas.

The results of the present study suggest that frontotemporal communication via the UF might be specifically impaired in schizophrenia patients suffering from the deficit syndrome. The term ‘deficit syndrome’ was introduced by Carpenter et al.26 to identify a relatively homogenous subgroup of patients with the diagnoses of schizophrenia, characterized by the presence of primary and enduring negative symptoms. Extensive research has shown that the deficit syndrome has a degree of stability, and deficit patients are different from non-deficit patients in aspects of neurocognition, brain imaging, and electrophysiological findings.36 Although the results of several studies37,38 suggest an involvement of frontoparietal brain circuits in deficit schizophrenia, others imply that frontotemporal circuits might also be involved differentially in the two schizophrenia subgroups. Sigmundsson et al.23 found decreased white matter in the left UF of schizophrenia patients with primary negative symptoms, and the present FA findings reinforce this idea.

It is generally accepted that fibers running through the UF connect the orbitofrontal cortex (OFC) with the temporopolar region, the rostral parahippocampal gyrus (entorhinal/perirhinal region), and the amygdala.39,40 Fibers from Brodmann Areas 10 & 11 (prefrontal association cortex) and 32 (prefrontal limbic association cortex) course through the UF to terminate in the rostral temporal cortex and the amygdala.41 Thus, UF connects temporal regions involved in sound and object recognition (superior and inferior temporal gyri) and recognition/episodic memory (entorhinal, perirhinal, and parahippocampal cortices) with frontal areas involved in cognition and emotion. Moreover, the interconnection of OFC and amygdala is important for the evaluation of environmental stimuli, especially for threats, assigning emotional significance to sensory experience, and the prediction of the outcome of actions.42,43 The evaluation based on associative information, and particularly information about the value of likely outcomes, is manipulated in representational memory and integrated with information concerning subsequent behavior, current context and internal state. The resultant expectancies then influence processing in downstream limbic areas and the ventral striatum as well as other prefrontal regions. Such interactive processes thereby promote voluntary, cognitive, and goal-directed behavior. These abilities are impaired in schizophrenia and are defined as deficit (negative) symptoms. In recent years, fMRI studies have shown that patients with schizophrenia have impaired functional coupling of OFC and amygdala during emotion recognition, goal-directed activity, and social decision-making.44–46 Moreover, functional or structural alteration in these areas is related to negative symptoms and flat affect.46–48 Thus, our finding of reduced FA values in the left UF of patients manifesting the deficit syndrome, but not of non-deficit patients, leads us to believe that UF malfunction might contribute to the deficit state in schizophrenia patients via decreasing the frontotemporal connection.

An initial report17 found no FA difference between schizophrenia patients and controls. However, the same group later reported reduced FA values in recent-onset schizophrenia patients.49 Although there are studies (including those by Price et al.,20 Szeszko et al.22 and particularly Peters et al.19) supporting reduced UF FA in early stage schizophrenia, others suggest that chronicity of the illness might increase the white matter pathologies.50,51 It should be noted that in our study group, deficit patients had a shorter duration of illness, and illness duration did not show any correlation with UF FA values in either group. Thus, our findings supported the view that reduced FA in UF might be a characteristic of specific patient group and present even at early stages of the illness.

It is unfortunate that many UF DTI studies do not present the patients' clinical symptoms in detail,20,52–54 and others include schizophrenia patients with minor negative symptoms.51 None of these specifically evaluated the patients for a deficit syndrome or psychomotor poverty. Our study indicates that a clinical deficit state is associated with reduced UF FA values and that the reduction is specific to the left hemisphere. In another study, Rowland et al.38 found reduced FA values in the right superior longitudinal fasciculus which interconnects frontal and parietal lobes in deficit patients. They proposed that their lateralized finding was due to specialization of the right hemisphere in emotion perception and interpretation. One should note that the study by Rowland et al. and our study tested two different a priori hypotheses and regions, which obviate direct comparison. Thus, further studies should include larger patient groups and explore both fasciculi.

DTI methodology is still evolving and therefore, methodological differences in postprocessing and quantification of DTI images may play a role in the conflicting results as discussed here and elsewhere.57,58 In our study, we used a hybrid method combining tract-based spatial statistics and an extracted mean FA value in a specific region, the UF. In this aspect, our results should be evaluated as a tractography study.

This study had several limitations. We restricted our analysis of FA specifically to the UF, consistent with our narrow hypothesis. Our sample sizes were small, especially for deficit patients (n = 11). Our study had a power of 0.7 and the F value for effect size was 0.6. Although this power suggests that our study should be viewed as preliminary, it demonstrates the importance of clinically subtyping schizophrenia patients. In our study, the education levels of patients were lower than those of controls. Education levels have been associated with gray and white matter changes,59–61 and thus should be viewed as a possible confounding factor. However, we could not find any correlation between the education levels and FA values of either patient or control groups (data not shown). Moreover, the reduced FA levels of deficit patients cannot be attributed to education levels because the latter did not differ significantly between the deficit and non-deficit patient subgroups.

All the patients in our study were receiving antipsychotic medications. Although some studies22,55 have indicated that antipsychotics may have effects on FA values, not all studies19 have agreed. Even those studies showing an association found it regionally rather than globally.22 A second reason for incongruent results might be differences in calculating antipsychotic dosages among the studies. It is known that antipsychotics might increase negative symptoms.56 However, in our study, six of 11 deficit patients had clozapine and only four patients (of 11) had even mild extrapyramidal signs. Thus, high negative symptom scores could not be the result of extrapyramidal signs in our deficit group.

In conclusion, the negative symptoms forming the deficit syndrome in schizophrenia have long been thought to be manifestations of impaired functioning of frontal and temporal areas. Previously we reported reduced blood flow to frontal and temporal regions in deficit syndrome patients, but not in their non-deficit peers.27 We now report reduced FA values in the UF of deficit but not of non-deficit patients, and these reduced values were restricted to the left hemisphere. Our findings add weight to the view that the deficit syndrome reflects frontal and temporal impairment, and further suggest that communication between these regions is degraded by UF pathology in the left hemisphere.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES

The authors wish to thank Koksal Alptekin and Baybars Veznedaroglu for their support during the recruitment of patients. This study is a part of the Standardization of Computational Anatomy Techniques for Cognitive and Behavioral Sciences (SoCAT) project supported by Ihsan Dogramacı Foundation, Ankara, Turkey and Ege University Science Project # 2005TIP028.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. REFERENCES