SEARCH

SEARCH BY CITATION

Keywords:

  • antisaccade;
  • fMRI;
  • hyperfrontality;
  • saccade;
  • schizophrenia

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. ACKNOWLEDGMENT
  8. REFERENCES

Aims:  Antisaccadic eye movements, requiring inhibition of a saccade toward a briefly appearing peripheral target, are known to be impaired in schizophrenia. Previous neuroimaging studies have indicated that patients with schizophrenia show diminished activations in the frontal cortex and basal ganglia. These studies used target fixation as a baseline condition. However, if the levels of brain activities at baseline are not compatible between patients and healthy subjects, between-group comparison on antisaccade-related activations is consequently invalidated. One possibility is that patients with schizophrenia may present with greater activation during fixation than healthy subjects. In order to examine this possibility, here we investigated brain activities associated with antisaccade in the two groups without using target fixation at baseline.

Methods:  Functional brain images were acquired during prosaccades and antisaccades in 18 healthy subjects and 18 schizophrenia patients using a box-car functional magnetic resonance imaging design. Eye movements were measured during scanning.

Results:  In the patient group, the elevated activities in the dorsolateral prefrontal cortex (DLPFC) and thalamus, normally seen in antisaccade tasks relative to saccade tasks, were no longer observed. Moreover, in normal subjects, activities in the DLPFC and thalamus were greater during the antisaccade task than during the saccade task. In patients, no such difference was observed between the two tasks, suggesting that these brain regions are likely to be highly activated even by a simple task such as fixation. In particular, the DLPFC and thalamus in patients were not activated at a level commensurate with the difficulty of the tasks presented.

Conclusions:  From these results, it is suggested that schizophrenia entails dysfunctions in the fronto-striato-thalamo-cortical network associated with motor function control.

SACCADIC EYE MOVEMENTS are the primary mechanism used by primates to visually explore their environments. A visually guided reflexive saccade can be defined as an automatic orienting response to a novel visual target in the peripheral field. Patients with schizophrenia perform prosaccades normally, making rapid and accurate eye movements to targets.1–3 In contrast, the inhibition of automatic saccades is impaired in schizophrenic patients.4 One task used to investigate saccade inhibition is the antisaccade task, which requires subjects to inhibit a saccade toward a briefly appearing peripheral target, and to instead immediately generate a saccade to a point in the opposite direction.5 Antisaccade deficits have high sensitivity and high specificity for the diagnosis of schizophrenia and are thought to be a genetic marker for the illness. Reported rates of antisaccade deficits range from 24% to 71% in patients with schizophrenia and from 2% to 27% in normal controls.6–9

Several comparison studies to date have examined the brain regions associated with antisaccade tasks in schizophrenic patients and normal control subjects.10–12 Most of these have reported reduced activity in the basal ganglia and the cortex, including the prefrontal area, in the schizophrenic group. As we will discuss further below, we question whether the activities of these brain regions were in fact reduced or not. Functional magnetic resonance imaging (fMRI) is a specialized MRI scan that measures hemodynamic responses related to neural activity in the brain. When two actions that generate neural activity are compared in fMRI, an analysis is based on the difference between a baseline signal and a signal measured at the time of task execution. Therefore, when comparing two groups, it is important to be able to assume that the baseline levels in the two groups are equivalent. Most previous fMRI research on antisaccade and saccade tasks used a target that required subjects to focus on a central fixation point during baseline imaging. One possibility is that patients with schizophrenia may exhibit greater cerebral activities during the fixation condition than healthy subjects. In order to examine this baseline effect, here we compared schizophrenic patients and normal subjects using a blank screen on which subjects were not required to focus at baseline.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. ACKNOWLEDGMENT
  8. REFERENCES

Subjects

Eighteen patients with schizophrenia (11 men and 7 women; mean age 34.8 ± 7.9) and 18 healthy subjects (9 men and 9 women; mean age 37.6 ± 4.8) participated in this study. All patients met the criteria for schizophrenia according to the DSM-IV. The mean duration of education was significantly longer (P < 0.05) in the healthy subject group than in the schizophrenia group. In the latter group, the mean age at onset of psychosis was 25.8 years old, the mean Brief Psychiatric Rating Scale total score was 41.9, and the mean total dose of antipsychotic medication per patients converted to haloperidol equivalency was 16.0 mg. All healthy subjects were free from neurological or psychiatric illness, and no abnormalities were observed on brain structural MRI. Written informed consent was obtained from all participants. All participants were right-handed according to the Edinburgh Handedness Inventory.13 This project was conducted in accordance with the Declaration of Helsinki and approved by the Ethical Committee of Nihon University School of Medicine.

MRI acquisition

MRI data were acquired using a 1.5T Siemens Symphony system (Siemens, Erlangen, Germany). Gradient-recalled echo planar imaging (EPI) was used for the fMRI sequence to obtain blood oxygen level-dependent contrast. Interleaved multi-slice gradient EPI was used to produce 40 continuous, 3-mm thick axial slices encompassing the entire brain (echo time = 62 ms, repetition time = 4000 ms, flip angle = 90 degrees, field of view = 192 mm, 64*64 matrix). Each series comprised 104 scans with a complete duration of 416 s. The run began with four dummy volumes to allow for T1 equilibration effects. The head of the subject was fixed using cushions to minimize motion artifacts.

Behavioral methods

Saccade and antisaccade performance was recorded outside of the magnet. Horizontal and vertical eye movements and target position were measured using electro-oculography (EOG) (NEC) and a goggles-type display (SONY).

Stimulus projection

The stimulus was generated using a personal computer (OS: Windows 98) and made to order software. The stimulus was projected on a small screen attached to a head coil, using a liquid crystal display projector system customized to our MRI machine (Kiyohara Optics, Tokyo).

Prosaccade task

Each trial began with the target in central fixation (0 degrees) for a random duration of 500–1500 ms. The target then shifted randomly to left or right horizontal peripheral locations (10 degrees from the center position), where it remained for 1000 ms. The target size was 1 degree of the visual angle. The number of left and right saccadic eye movements was the same. Participants were instructed to follow the target as quickly and accurately as possible, alternating between 40 s of control condition task and 40 s of prosaccade condition, completing 10 sets of trials in all. During the baseline condition, subjects were in total darkness and were asked to maintain fixation and not blink.

Antisaccade task

The parameters for the antisaccade task were identical to those for the prosaccade task. The antisaccade task required participants to fixate the target in the central position and to redirect their gaze in the opposite direction of the target as soon as it shifted to the periphery. Participants performed 10 sets of trials in total, alternating antisaccade and control conditions.

fMRI data analysis

Image analysis was performed using an Ultra5 workstation (Sun Microsystems, Palo Alto, CA, USA) using MATLAB (Mathworks Inc., Natick, MA, USA) and statistical mapping (SPM99, Wellcome Department of Cognitive Neurology, London, UK; http://www.fil.ion.ucl.ac.uk/spm). Before statistical parametric maps were calculated, EPI images for each time series were realigned to the first functional image to remove residual head movement. Images were then coregistered and normalized to the Montreal National Institute template. Confounding effects of global volume activity and magnetic noise were removed using linear regression and cosine functions (up to a maximum of 1 cycle per 40 scans). Removing the latter confounds corresponds to high-pass filtering of the time series to remove low-frequency artifacts that can arise due to aliased cardiac and other cyclical components. After normalization, three-dimensional spatial smoothing was applied to each volume using a Gaussian kernel of 8 × 8 × 8 mm. Alternating periods of baseline and activation were modeled using a simple delayed box-car reference vector to account for delayed cerebral blood flow after stimulus presentation. Significantly activated pixels were searched for using the General Linear Model approach for time-series data.

To create the subtraction activation image between saccade and antisaccade, data was analyzed using random-effect analysis. Statistical significance was set at the level of P < 0.001, uncorrected for multiple comparisons.

Intra-individual comparisons between saccade and antisaccade were analyzed using paired t-tests, and statistical significance was set at the level of P < 0.005, uncorrected for multiple comparisons.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. ACKNOWLEDGMENT
  8. REFERENCES

Behavioral data

Demographic and performance data are summarized in Table 1. The analysis of EOG revealed no differences in prosaccades between the patients and normal controls. In contrast, error rates in antisaccades were higher and latencies of prosaccades and antisaccades were longer in the patient group than in the control group.

Table 1.  Subjects and eye movement performance
 Patients with SchizophreniaControl
  1. Statistical analysis (T-test) *P < 0.05.

  2. BPRS, Brief Psychiatric Rating Scale; HPD, haloperidol.

Number of cases (male/ female)18 (12/6)18 (9/9)
Age (year)34.8 ± 7.937.6 ± 4.8
Education (year)*11.2 ± 2.915.3 ± 2.2
Age at the onset (year)25.8 ± 6.4
HPD equivalence (mg)16.0 ± 16.1
BPRS total score41.9 ± 7.9
Saccade error (%)*0.5 ± 0.670.00 ± 0.00
Saccade latency (ms)*212.2 ± 30.1174.2 ± 11.8
Anti-saccade error (%)*1.1 ± 1.60.14 ± 0.35
Anti-saccade latency (ms)*244.9 ± 48.4205.6 ± 18.5

fMRI data

Activated areas in the normal control group are shown in Fig. 1a for the saccade tasks and in Fig. 1b for the antisaccade tasks (P < 0.001, uncorrected for multiple comparisons). During the saccade tasks, regional activations were observed bilaterally in the frontal eye fields (FEF), supplementary eye fields (SEF), and parietal eye fields (PEF), left lenticular nucleus, and bilateral occipital cortices (V1). During the antisaccade tasks, activations were observed in the same regions as during saccade tasks, as well as bilaterally in the inferior parietal lobules (IPL), thalami, right lenticular nucleus, inferior frontal gyrus (IFG), and left dorsolateral prefrontal cortex (DLPFC) (Table 2).

image

Figure 1. Brain regions displaying greater activities during (a) saccade and (b) antisaccade conditions than during control condition in healthy subjects. In the rightmost image, the activation map is overlaid onto a T1 SPM normalized brain image. The height threshold is set at P < 0.001, uncorrected.

Download figure to PowerPoint

Table 2.  Brain regions more active during visually guided saccades and antisaccades than during control tasks in healthy subject
Brain region Saccade vs rest Coordinatet-valueAntisaccade vs rest Coordinatet-value
XYZXYZ
  1. DLPFC, dorsolateral prefrontal cortex; FEF, frontal eye fields; IPL, inferior parietal lobule; L, left; NS, not significant; PEF, parietal eye fields; R, right; SEF, supplementary eye fields.

DLPFCRNS504084.01
LNS445044.24
FEFR466505.34402505.87
L406506.08384526.52
SEFR66624.2088524.20
L44605.87210465.56
PEFR3254483.802658546.70
L3056564.282660527.91
IPLRNS6436286.14
LNS6440345.75
ThalamusR−12−18103.96101488.30
L−10−18−26.53101686.29

Activation areas in the patient group are shown in Fig. 2a for the saccade tasks and in Fig. 2b for the antisaccade tasks (P < 0.001, uncorrected for multiple comparisons). During the saccade task, regional activation was observed bilaterally in the FEF, SEF, and PEF, left lenticular nucleus, and V1. These regions are the same as those seen in the normal subject group. However, the patient group also showed activations in the IFG, DLPFC, IPL, lenticular nucleus and thalamus during saccade tasks. During the antisaccade tasks, activation was observed in the same regions as in the saccade tasks (Table 3).

image

Figure 2. Brain regions displaying greater activities during (a) saccade and (b) antisaccade conditions than during control conditions in patients with schizophrenia. In the rightmost image, the activation map is overlaid onto a T1 SPM normalized brain image. The height threshold is set at P < 0.001, uncorrected.

Download figure to PowerPoint

Table 3.  Brain regions more active during visually guided saccades and antisaccades than during control tasks in patients with schizophrenia
Brain region Saccade vs rest Coordinatet-valueAntisaccade vs rest Coordinatet-value
XYZXYZ
  1. DLPFC, dorsolateral prefrontal cortex; FEF, frontal eye fields; IPL, inferior parietal lobule; L, left; PEF, parietal eye fields; R, right; SEF, supplementary eye fields.

DLPFCR3656268.80425686.05
L−3854144.263644125.00
FEFR342648.522604812.06
L444589.91366468.99
SEFR1216385.38104485.30
L822385.23120465.53
PEFR3054487.2522605412.32
L28525010.7128525612.89
IPLR56−32228.636238185.09
L−58−40223.776238184.27
ThalamusR12−1426.70101847.09
L−12−14−26.92121626.23

Furthermore, in the normal control group, comparing brain activity during the antisaccade task with that during the saccade task revealed that antisaccade eye movements induced elevated activities in the bilateral FEF, PEF, IPL, ACC, IFG, and DLPFC (P < 0.005, uncorrected for multiple comparisons). In the patient group, however, only bilateral activation in the PEF was observed.

Correlation between fMRI activation and eye movement performance

In order to assess the effect of performance on brain activity, we analyzed the correlation between error rate and brain activity. Figure 3 shows the correlation between fMRI activation and eye movement performance in patients with schizophrenia. fMRI activation is calculated from each peak voxel. No significant correlation was observed between two parameters.

image

Figure 3. The correlation between brain activation and the number of antisaccade errors. The horizontal axis represents the number of antisaccade errors, and the vertical axis represents the estimated magnetic resonance imaging signal. rFEF, right frontal eye fields; rPEF, right parietal eye fields; rPFC, right prefrontal cortex; rthalamus, right thalamus.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. ACKNOWLEDGMENT
  8. REFERENCES

Our understanding of human cortical control of saccades is derived from observations of cerebral lesions14–16 and from transcranial magnetic stimulation,17,18 positron emission tomography,19,20 and fMRI.21–24 Previous studies in these areas have indicated that saccadic eye movements are controlled by a cortical network that includes the PEF, located in the intraparietal sulcus and superior parietal lobule, the FEF, located in the precentral gyrus, and the SEF, located in the upper medial wall of the frontal lobe. Activation has also been observed in the bilateral dorsolateral prefrontal cortices, supramarginal gyri, anterior cingulate cortices, and thalami during antisaccade tasks.25 In short, in normal subjects no activation of DLPFC, IFG, striatum, and thalamus were observed during the saccade tasks.

Contrary to previous reports, the present study showed the activations in the DLPFC, IFG, striatum, and thalamus during both the saccade and antisaccade tasks in patients with schizophrenia. In addition, differential activation maps between the antissacade and saccade tasks exhibited the bilateral activation of the FEF, PEF, IPL, ACC, IFG, and DLPFC in normal subjects, whereas only the PEF were activated bilaterally in the patient group. These results show that normal subjects process the saccade and antisaccade tasks in different brain regions, whereas patients with schizophrenia likely use virtually the same regions when processing both tasks. In the patient group, therefore, when brain activations during eye movement tasks were compared directly, the elevated activations of the DLPFC and thalamus normally seen in antisaccade tasks relative to saccade tasks were no longer observed. In comparing the patients to the normal controls, the present study demonstrated higher activity of the thalamus and broad cortical regions (including the prefrontal area), especially during saccade tasks. This suggests hyperactivation, not reduced activation, in the prefrontal cortex and thalamus in patients with schizophrenia. Taken together, though the antisaccade task is cognitively more demanding than the saccade task, these regions in the patients with schizophrenia did not seem to be activated at a level that corresponded to the degree of difficulty of the tasks presented.

The tasks used in most of the previous studies required subjects to focus on a gazing point, and the reduced activities of the DLPFC and thalamus were observed during the antisaccade tasks in patients with schizophrenia. Three recent studies using fMRI revealed reduced activation in the right DLPFC and reduced activation in the striatum in schizophrenia.10–12

In contrast, our results showed higher activities in broad cortical and subcortical regions during the saccade and antisaccade tasks in the patient group as compared with the normal control group. This suggests that these regions could already be activated by the time the schizophrenic patient focuses on the gazing point; therefore, the difference in activation levels between baseline and eye movements becomes smaller in the patient group.

In the present study, we demonstrated the activations in the DLPFC and thalamus during the saccade task in the patient group. The fronto-striato-thalamo-cortical network,26–28 including the prefrontal cortex and thalamus, is important for control of antisaccades. Schizophrenia presents with dysfunction in dopaminergic neural networks29 and the fronto-striato-thalamic circuit.30,31 Dysfunction in the striato-thalamo-cortical dopaminergic circuitry may reduce inhibition of reflexive saccade and thus facilitate saccades for the target direction during the antisaccade task in schizophrenics. Our results indicate that this dysfunction has an important influence on subtle motor control and therefore affects antisaccade generation through both the direct and indirect basal ganglia pathways. These findings suggest that patients with schizophrenia who display antisaccade inhibition errors may present with dysfunction in the fronto-striato-thalamo-cortical network.

Given that previous studies have targeted patients with schizophrenia with poor performance in cognitive tasks, a bias toward reduced brain activations may have been present.32–36 In order to assess the effect of performance on brain activity, we analyzed the correlation between error rate and brain activity. No significant correlation was observed between the two variables. Therefore, we conclude that the performance did not directly affect the results.

CONCLUSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. ACKNOWLEDGMENT
  8. REFERENCES

In order to examine baseline effect, we employed an eye movement task that did not require subjects to focus on a fixation point during the baseline condition, and compared brain activity between patients with schizophrenia and normal control subjects. In normal subjects, activities in the DLPFC and thalamus were greater during antisaccade tasks than during saccade tasks, whereas no significant difference was observed in patients with schizophrenia. These results suggest that the brains of patients with schizophrenia did not seem to be activated at a level that corresponded to the degree of difficulty of the tasks presented. Previous studies that used target fixation at baseline assessment showed reduced activities of the DLPFC and thalamus in patients. In contrast, our study demonstrated hyperactivation of the DLPFC and thalamus in patients, suggesting that in patients with schizophrenia these brain regions were already activated by the time patients viewed a fixed target at baseline. We think that these results reflect the symptom that patients of schizophrenia can not adapt to the environment. Finally, we suspect that patients with schizophrenia may be affected by a defect in the fronto-striato-thalamo-cortical network associated with motor function control.

ACKNOWLEDGMENT

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. ACKNOWLEDGMENT
  8. REFERENCES

We thank Professor Makoto Uchiyama of the Department of Neuropsychiatry, Nihon University School of Medicine for providing helpful suggestions. We gratefully acknowledge the contributions of the members of the Tamagawa University Brain Science Institute. This work was supported by Grants-in-Aid for Scientific Research #19790838 & #18300275 for T.M. from Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONCLUSION
  7. ACKNOWLEDGMENT
  8. REFERENCES