- Top of page
Aim: The aim of this study was to investigate whether magnetoencephalographic oscillations underlying working memory dysfunction in the dorsolateral prefrontal cortex (DLPFC) are related to psychopathological disturbance in patients with schizophrenia-like psychosis of epilepsy (SLPE).
Methods: Twelve patients with SLPE and 14 non-psychotic epilepsy controls participated in this study. Magnetoencephalography was recorded while patients performed a visual working memory (WM) task. Psychopathology was assessed using a four-factor structure of the Brief Psychiatric Rating Scale, and regression analyses were carried out to examine the relative impact of severity of psychopathology on WM-induced activation of the DLPFC.
Results: We found that activation of the WM-compromising DLPFC, as indicated by increased alpha desynchronization in patients with SLPE compared with their non-psychotic counterparts, showed a positive linear correlation with disorganization symptom scores. This association remained significant after controlling for confounding factors, including age, task performance, IQ, and duration of psychosis.
Conclusion: Our results indicate that abnormal activation in prefrontal areas engaged during working memory may be critical to domains of psychopathology, in particular disorganized thought-processing in patients with SLPE.
CHRONIC INTERICTAL PSYCHOSIS (CIP) clinically closely resembles schizophrenia with a typical presentation as a paranoid hallucinatory syndrome with other psychopathological disturbances, which is why it is also referred to as ‘schizophrenia-like psychosis of epilepsy’ (SLPE). In addition to psychopathology, cognitive deficits, in particular working memory (WM) dysfunction, are cardinal abnormalities in the two disorders.1,2 It has been proposed that analysis of WM is critical to understanding not only memory systems but also thought itself and psychopathological disturbance.3 Evidence for this view comes from functional MRI studies that related brain activation during WM processing to the cardinal symptoms of schizophrenia.4–6 However, despite the idea that similar pathophysiological mechanisms may underlie both schizophrenia and SLPE,2 the neuroimaging approach combining WM-related activation and clinical symptomatology has not yet been applied to SLPE for a better understanding of the correlation between the two disorders.
Using magnetoencephalography (MEG), we previously identified cortical regions, namely the right dorsolateral prefrontal cortex (DLPFC) and the left inferior temporal gyrus that showed activation deficits during visual-object WM-retention in patients with SLPE.7 Unlike dysfunction of the inferior temporal cortex that is explained by the involvement of this region in visual memory processing of the ventral stream specifically for object representation,7,8 the DLPFC is known to be implicated in both WM deficits, regardless of memory storage material, and the pathophysiology of schizophrenia.5,8
The neurophysiological basis for these WM deficits localized to the DLPFC was power changes in the alpha frequency band, in particular an alpha event-related desynchronization (ERD) phenomenon. Alpha ERD has been linked to active sensory and cognitive stimulus processing, and is considered as an electrophysiological correlate of cortical activation. This notion comes from the fact that alpha rhythm typically attenuates or suppresses in the occipital (visual) cortex by opening the eyes or visual stimulus.9 In this study we tested the hypothesis that WM-induced activation deficits in the DLPFC are also associated with domains of psychopathology in patients with SLPE.
- Top of page
All patients achieved an accuracy rating of 75% or higher on the WM task, although patients with SLPE performed slightly worse than their non-psychotic counterparts (SLPE 90.5 ± 7.1%, nPE 96.7 ± 1.7%; P < 0.05). The performance of the psychotic epilepsy patients was also lower than that of the reference healthy subjects (97.4 ± 2.8%; P < 0.01). The analysis of cortical activation, as indicated by power changes during memory retention revealed significant differences (threshold: t = 3.74, d.f. = 24, P < 0.001) exclusively in the alpha frequency band: patients with SLPE showed greater decrease in power or ERD in the right DLPFC compared to the nPE group (SLPE −4.31 ± 2.50%, nPE −0.28 ± 2.11%; t-maxima −4.39) (Fig. 1). Similar results were obtained from the exploratory comparison of MEG power changes between patients with SLPE and healthy subjects (alpha ERD −0.74 ± 2.37%; t-maxima −3.75, P < 0.001).
Figure 1. Correlation between working-memory-induced activation of the right dorsolateral prefrontal cortex (DLPFC) and disorganization scores in patients with schizophrenia-like psychosis of epilepsy. Top panel shows the area with significant alpha power change (threshold t = 3.74, P < 0.01) and the voxel with maximum activation (crosshair) within the DLPFC. The color bar represents t-values for the percentage of decrease (blue/green) and increase (red/yellow) in alpha power. Bottom panel shows scatter plot of DLPFC activation against disorganization symptom scores. A, anterior; P, posterior; R, right; L, left.
Download figure to PowerPoint
The analysis of psychopathology scores in patients with SLPE revealed a BPRS total score of 49.3 ± 7.8 (Table 1). The mean BPRS syndrome scores were: (i) positive symptoms, 11.17 ± 2.89 (range: 7–18); (ii) negative symptoms, 10.42 ± 3.26 (range: 6–17); (iii) disorganization, 7.42 ± 1.56 (range: 5–10); and (iv) affect, 14.42 ± 3.68 (range: 11–23).
The regression analyses indicated a positive linear correlation between disorganization syndrome and WM-related activation of the DLPFC (β coefficient = 0.63), with this model (F = 6.30, d.f. = 1,10, P = 0.029) accounting for 40% of the variance in alpha power-change magnitude, as an index of cortical activation. This association remained significant after controlling for confounding variables, namely age, task performance, IQ, and psychosis duration (Table 2). Figure 1 shows a scatter plot of DLPFC activation against disorganization scores showing a linear correlation. No other psychopathology domains were independently associated with DLPFC activation, although there was a significant difference for negative symptoms only when task performance was taken into account.
Table 2. Linear regression analyses for region of maximal activation in the prefrontal cortex and Brief Psychiatric Rating Scale syndrome scores in patients with schizophrenia-like psychosis of epilepsy
| ||Disorganization symptoms||Positive symptoms||Negative symptoms||Affect domain|
|R2/F||β (p)‡||R2/F||β (p)‡||R2/F||β (p)‡||R2/F||β (p)‡|
|maxαERD†||0.40/6.30||0.63 (0.029)*||0.08/0.89||0.28 (0.36)||0.22/2.90||0.47 (0.12)||0.14/1.68||−0.38 (0.22)|
|maxαERD +Age||0.49/4.39||0.84 (0.016)*||0.09/0.42||0.28 (0.39)||0.24/1.46||0.49 (0.13)||0.19/1.10||−0.47 (0.18)|
|maxαERD +Performance||0.59/6.58||0.67 (0.012)*||0.31/2.01||0.40 (0.19)||0.50/4.65||0.61 (0.031)*||0.22/1.30||−0.27 (0.40)|
|maxαERD +IQ||0.45/3.66||0.60 (0.038)*||0.14/0.75||0.24 (0.45)||0.41/3.07||0.59 (0.056)||0.26/1.54||−0.41 (0.18)|
|maxαERD +Psychosis duration||0.43/3.49||0.58 (0.047)*||0.29/1.80||0.43 (0.17)||0.39/2.91||0.53 (0.073)||0.25/1.50||−0.37 (0.24)|
To examine the potential influence of antipsychotic medication on source-power changes in patients with SLPE, the maximum alpha ERD values were correlated with chlorpromazine equivalent dose. No statistically significant correlation was found (P = 0.31).
- Top of page
In the present study, we examined the correlation between WM-related activation of the DLPFC and psychopathological disturbance in patients with SLPE. We found that cortical activation, as indicated by increased MEG alpha desynchronization, underlying WM deficits in the right DLPFC in psychotic epilepsy patients relative to controls showed a positive correlation with the severity of disorganization symptoms (Fig. 1, Table 2). This finding indicates that dysfunction in prefrontal areas engaged during active maintenance of information in WM may be critical to some domains of psychopathology, in particular disorganized thought-processing or disorganized behavior in patients with SLPE.
This is the first neuroimaging study to address the correlation of brain activation deficits during WM with clinical symptoms in SLPE. Our findings are consistent with those of a previous fMRI study by Perlstein et al. in patients with the classic form of schizophrenia.5 They used a verbal WM paradigm and noted that although several cortical areas engaged during WM, including Broca's area and the parietal cortex, were active during the task, significant WM load-related differences were observed only in the right DLPFC in the patients compared to healthy controls. Interestingly, this WM-related dysfunction was found to be selectively associated with disorganization symptoms, as measured by the Positive and Negative Syndrome Scale (PANSS). Furthermore, other fMRI investigations of schizophrenia have reported that in addition to disrupted activation of the prefrontal cortex, impaired DLPFC functional connectivity also correlated with disorganization symptom dimension.24 The fact that the DLPFC activation in our study was detected specifically in the right hemisphere may be explained partly by task-related factors, as visual-object memory storage (e.g. numbers) has been proposed to produce more activation in the right prefrontal cortex compared to storage of other materials (e.g. verbal, visuospatial).8 Altogether, these findings implicate a strong link between DLPFC dysfunction, WM deficits and disorganization dimension in patients with both schizophrenia and SLPE.
Neuropsychological evidence of WM deficiency correlating with the schizophrenic disorganization domain lends further support to our findings. For instance, Takahashi et al. using the Advanced Trail Making Test (ATMT) and the BPRS demonstrated that visuospatial WM deficits were associated with disorganization symptoms as well as with social functioning in patients with schizophrenia.25 Similarly, Daban et al. found deficits on WM tests sensitive to prefrontal dysfunction in patients with recent-onset schizophrenia, and noted that these cognitive abnormalities were highly correlated with the disorganization syndrome.26
It is noteworthy that in addition to source localization of WM-related dysfunction and its correlation with clinical parameters, MEG-beamformer revealed that the neurophysiological basis for these WM deficits was increased alpha ERD in patients with SLPE compared to their non-psychotic counterparts, most likely reflecting abnormal overactivation to perform the cognitive task. This is further supported by the fact that the psychotic epilepsy patients also showed a significantly larger alpha ERD in the DLPFC relative to the reference healthy subjects. This argument is consistent with evidence from previous reports demonstrating that desynchronization of alpha activity plays an important role in cognitive functions, especially memory processing,27,28 and that abnormal patterns of overactivation can underlie working memory deficits in some patients with schizophrenia.29
Interestingly, although overactivation represented the abnormal pattern of DLPFC activity in psychotic epileptic patients, greater activation was associated with less disorganization symptoms, and therefore with a better clinical state. A possible explanation for this seemingly conflicting finding might be linked to performance-related factors. It has been demonstrated that DLPFC dysfunction in schizophrenia can manifest as both overactivation and underactivation depending on behavioral performance and task demands; low-performing patients (accuracy < 75%) exhibit underactivation (i.e. hypofrontality) suggestive of a failure to maintain the prefrontal network that processes the cognitive information, whereas those with high performance show primarily overactivation as a possible compensatory neural mechanism.30,31 This reflects a compromised neural strategy for handling WM information mediated by the DLPFC, and led to the notion of WM-related DLPFC ‘inefficiency’ rather than simple underactivation or overactivation in schizophrenia.30,32 Although patients with SLPE in our study had a slightly lower performance level compared to nPE patients and healthy subjects, they performed highly accurately (>75%), with the majority of the patients having a correct response rate over 90% (n = 8). In this context, the linear regression results may indicate that among patients with SLPE, those with less disorganization symptoms produce more overactivation of the DLPFC in an attempt to compensate the underlying abnormality in WM information processing and achieve a better performance on the task.
Our results should be considered in light of the limitation of a small sample size, which is not uncommon in SLPE neuroimaging research,1 and a potential confounding effect of antipsychotic medication. However, based on the outcome of the correlation analysis of chlorpromazine equivalents it is unlikely that these pharmacological agents influenced our results. Overall findings together with previous reports of schizophrenic prefrontal dysfunction underlying WM deficiency and severity of psychopathology5 indicate that WM deficits and disorganization symptoms may share common neural substrates in both schizophrenia and SLPE. Furthermore, the finding of convergent cognitive abnormalities and psychopathological disturbance in the DLPFC suggests a specific role for this cortical region in the pathophysiology of SLPE.