Hemispheric asymmetry of spatial working memory deficit in schizophrenia

https://doi.org/10.1016/S0167-8760(99)00088-4Get rights and content

Abstract

Spatial working memory function was assessed in schizophrenia patients, hypothetically ‘psychosis-prone’ individuals who report unusual perceptual experiences and normal control subjects with an oculomotor delayed response task. Past studies point to the important role of dorsolateral prefrontal system in spatial working memory deficits of schizophrenia patients. In order to better understand the processes precipitating in working memory deficit, two types of working memory errors were examined: never-corrected vs. immediately-corrected errors. In schizophrenia patients, the loss of spatial representation in working memory, as captured by the presence of never-corrected errors, was much more severe when the target was presented in the right visual hemifield than when the target was presented in the left visual field. The same pattern was observed in healthy, psychometrically ascertained ‘psychosis-prone’ subjects. Therefore, the observed asymmetry of spatial working memory deficit seems unlikely to be a mere side-effect of medication or hospitalization. Normal control subjects did not show hemispheric asymmetry in error patterns. These results suggest that the loss of spatial representation during a delay period may be more severe in the left hemisphere in patients with schizophrenia and in ‘psychosis-prone’ individuals.

Introduction

The role of dorsolateral prefrontal cortex in spatial working memory function has been studied extensively in animals using the delayed response paradigm (e.g. Jacobsen, 1935, Blum, 1952, Funahashi et al., 1989, Funahashi et al., 1990, Quintana and Fuster, 1992, Funahashi et al., 1993). Lesions in the prefrontal cortex also result in behaviors that resemble major clinical features of schizophrenia, such as increased distractibility, perseverations and anergia (e.g. see Goldman-Rakic, 1991). Based on neuroanatomical and neurophysiological data, Goldman-Rakic (1991) has proposed that one fundamental deficit of schizophrenia may be conceptualized as a dysfunction of working memory that leads to a disintegration of all behaviors guided by internal representations. In neuroanatomical studies of spatial working memory in the monkey, it is possible to map out the regions of ‘mnemonic scotoma’ resulting from focal lesions to the area 46 (Funahashi et al., 1993). Funahashi et al. observed that working memory loss is confined to the specific regions corresponding to the site of the lesion in area 46. Therefore, depending on the region of abnormality, the spatial working memory deficit may be precisely localized using the oculomotor delayed response paradigm.

Although occasional cases of initial misdiagnosis of patients with frontal lobe tumors as schizophrenic (e.g. Hall and Young, 1992) are suggestive of structural frontal lobe abnormalities in schizophrenia, it is unclear whether there are apparent, gross structural abnormalities of prefrontal cortex in patients with schizophrenia (see Wible et al., 1995). However, recent studies suggest that there may be a reduced frontal asymmetry (Falkai et al., 1995a) or small frontal volume reductions (Turetsky et al., 1995) in schizophrenia patients. These findings suggest that prefrontal deficits in schizophrenia may involve structurally subtle but functionally significant disturbances in prefrontal cortical cells and circuits (Benes et al., 1991, Breier et al., 1992, Selemon et al., 1995, Daviss and Lewis, 1995). For example, cell density is increased in the area 46 in schizophrenia patients (Daviss and Lewis, 1995, Lewis and Anderson, 1995, Selemon et al., 1995) but it is not known whether these relatively subtle structural abnormalities are lateralized to left or right prefrontal regions.

Behaviorally, patients with schizophrenia show consistent deficits in spatial working memory, regardless of the modality of the response (Park and Holzman, 1992, Spitzer, 1993, Keefe et al., 1995, Carter et al., 1996, McDowell and Clementz, 1996, Spindler et al., 1997). In contrast, bipolar patients show no impairments on the same delayed-response tasks (Park and Holzman, 1992, Park and Holzman, 1993). Spatial working memory deficit has also been observed in the first-degree relatives of schizophrenia patients (Park et al., 1995a) and in psychometrically ascertained schizotypic subjects (Park et al., 1995b, Park and McTigue, 1997), suggesting that working memory deficit may be a potential marker for schizophrenia.

Given that the delayed response task is deceptively simple, it is not clear why anybody, schizophrenic or not, would make an error on this task. Parsing the spatial delayed response task suggests that successful performance on the delayed response task depends on several hypothetical cognitive components (Park and O’Driscoll, 1996). These components include the ability to maintain the spatial representation of the target during the delay period, inhibition of irrelevant distractors at the response stage, and the initiation and execution of appropriate motor responses (e.g. a saccade or a hand movement). Failure to mediate or facilitate any of these hypothetical components may lead to an error in the delayed response task; not all errors are created equal.

In memory research, how an error is elicited and how the subject might attempt to correct a mistake, are rarely examined. We have argued elsewhere that the most informative data on abnormal processes leading to spatial working memory deficits may lie in the patterns of guessing response after an error has been made and that it is possible to make inferences about the pathological processes underlying the error response (Park and O’Driscoll, 1996). If the spatial representation was maintained during the delay period but the subject was temporarily disinhibited at the response stage, the subject should be able to make an immediate correction after the incorrect response has been elicited. In other words, as long as the representation was maintained in working memory, even if an error is produced, it will be corrected if the subject is given a chance to do so. On the other hand, if the spatial representation was disrupted or lost during the delay period, it will be impossible to correct the error response (see Fig. 1).

Park and O’Driscoll (1996) examined spatial working memory errors to see if subsequent attempts to correct a mistake were successful. Normal control subjects and bipolar patients made very few errors, but if they did, they were able to correct their mistakes immediately. In addition, the majority of the initial error responses lay within the same quadrant as the correct target position. This pattern of immediate error-correction suggests that normal control subjects and bipolar patients were able to maintain spatial representation of the target during the delay period but they may have been temporarily disrupted at the response stage. Given a chance to correct themselves, they immediately chose the correct target position from memory. If subjects were not able to maintain spatial representation of the target during the delay and that is why they made an error, any subsequent attempts to correct errors would result in random guessing. Normal control subjects and bipolar patients made very few never-corrected errors. In contrast, schizophrenia patients made large numbers of both immediately-corrected and never-corrected errors. Thus, schizophrenia patients seemed to have deficits in the maintenance of representation during the delay, as well as in initiating correct responses at the response stage. The same pattern of deficit was observed in schizotypic subjects. In sum, it is important to bear in mind that what may differentiate schizophrenia patients from the bipolar and normal control subjects, apart from the sheer number of errors, is how these errors arise. All subjects are susceptible to making some errors due to distracting or disinhibitory processes at the response stage but only schizophrenia patients and a sub-group of schizotypic subjects seem to lose the spatial representation during the delay period.

In neuroanatomical studies of spatial working memory, ‘mnemonic scotomas’ and ‘mnemonic hemianopoeias’, corresponding to the site of the dorsolateral prefrontal lesion, have been described (see Funahashi et al., 1993). For example, a focal lesion in the right principal sulcus area led to the ‘mnemonic scotoma’ in a small corresponding region of the left visual field but not in any other areas of the visual field. The errors elicited after the focal lesion was made, were not confined to the immediate neighborhood of the correct target position. Instead, these errors tended to be scattered all over the visual field, suggesting the loss of spatial representation of the target. In other words, if the cells which support the maintenance of the internal representation during the delay period are destroyed by lesions, working memory for a specific spatial region seems to be lost. In addition to the lesion data, Funahashi et al., 1989, Funahashi et al., 1990, Funahashi et al., 1993 also showed that incorrect responses on the oculomotor delayed response task were accompanied by an absence of cellular activity in the principal sulcus cells during the delay period. Therefore, any events that disrupt the maintenance of spatial representation during the delay period will result in errors that cannot be corrected. By examining the ‘never-corrected’ errors, it may be possible to specifically probe the left and right dorsolateral prefrontal functions. This strategy departs from the traditional neuropsychological method of administering tasks that are presumed to tap the left or right hemisphere functions to examine asymmetry (e.g. verbal vs. spatial tasks).

In previous studies of spatial working memory, we did not observe any hemispheric asymmetry in the overall accuracy scores or in reaction times (e.g. Park and Holzman, 1992) but we did not examine asymmetry in relation to the types of errors. In this study, immediately-corrected and never-corrected errors in left and right visual fields were examined in schizophrenia patients and normal control subjects. In addition, data from a previous study of psychometric schizotypes (Park et al., 1995b) were re-analyzed to see if there is a hemispheric differences in the incidence of never-corrected errors in individuals who may carry latent liability for schizophrenia.

Section snippets

Schizophrenia patients

Thirty-three patients with schizophrenia were recruited from a private psychiatric hospital. These subjects met the criteria for a DSM-III-R diagnosis of schizophrenia, as determined from the SCID (Spitzer and Williams, 1985). The mean age of the patients was 31.5 years (S.E.=1.4) and they had been ill for an average of 13.2 years (S.E.=1.5). The mean years of education was 14.7 (S.E.=0.4) and the mean WAIS score was 110.7 (S.E.=3.0). All subjects were receiving anti-psychotic medications. No

Schizophrenia patients and age-matched control subjects

Multifactorial, repeated measures ANOVA was conducted to compare the differences between the diagnostic groups in the types of errors made in the two visual hemifields. There was a main effect of diagnostic group (F1,60=29.4, P<0.0001). Schizophrenia patients made more working memory errors than did the normal control subjects but the two groups did not differ on the sensory control task. There was also a main effect of the type of errors (F1,60=6.2, P<0.02). Immediately-corrected errors were

Discussion

Spatial working memory function in schizophrenia patients may be characterized, not only by the sheer number of errors, but also by the nature of the errors. Loss of spatial representation, mediated by the dorsolateral prefrontal system, features prominently in the way schizophrenia patients and high PerAb subjects forget. When normal control subjects make errors, they are able to correct their errors immediately. This pattern of errors suggests that the control subjects are able to maintain

Acknowledgements

This work was supported by grants from NIMH NARSAD and Scottish Rite Schizophrenia Research Program. I am very grateful to Mark Lenzenweger for allowing me to study the high PerAb individuals and two anonymous reviewers for their thoughtful suggestions.

References (38)

  • R.A. Blum

    Effects of subtotal lesions of frontal granular cortex on delayed reaction in monkeys

    AMA Arch. Neurol. Psychiatry

    (1952)
  • A. Breier et al.

    Brain morphology and schizophrenia: an MRI study of limbic, prefrontal cortex and caudate structures

    Arch. Gen. Psychiatry

    (1992)
  • L.J. Chapman et al.

    Body-image aberration in schizophrenia

    J. Abnorm. Psychology

    (1978)
  • T.J. Crow

    Temporal lobe asymmetries as the key to the etiology of schizophrenia

    Schizophr. Bull.

    (1990)
  • P. Falkai et al.

    Reduced frontal and occipital lobe asymmetry on the CT-scans of schizophrenic patients. Its specificity and clinical significance

    J. Neural Transm. Gen. Sect.

    (1995)
  • S. Funahashi et al.

    Mnemonic coding of visual cortex in the monkey’s dorsolateral prefrontal cortex

    J. Neurophysiol.

    (1989)
  • S. Funahashi et al.

    Visuospatial coding in primate prefrontal neurons revealed by oculomotor paradigms

    J. Neurophysiol.

    (1990)
  • S. Funahashi et al.

    Dorsolateral prefrontal lesion and oculomotor delayed response performance: evidence for mnemonic ‘scotomas’

    J. Neurosci.

    (1993)
  • P.S. Goldman-Rakic

    Prefrontal cortical dysfunction in schizophrenia: the relevance of working memory

  • Cited by (27)

    • Insights to the schizophrenia continuum: A systematic review of saccadic eye movements in schizotypy and biological relatives of schizophrenia patients

      2017, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      Studies were excluded if they did not include a prosaccade, antisaccade or memory-guided saccade paradigm, did not make group performance comparisons, or were not peer-reviewed empirical articles. Articles were also excluded if they contained populations comprised entirely of people diagnosed with SPD (Brenner et al., 2001; Cadenhead et al., 2002; Tsunoda et al., 2005; Wonodi et al., 2006), any family members other than first degree relatives (McDowell and Clementz, 1997; McDowell et al., 1999), used a non-horizontal saccadic paradigm (Park et al., 1995a,b; Park, 1999), used a measurement of saccadic performance with low sensitivity, such as frequency of hypometric and hypermetric saccades (Schreiber et al., 1997) or reported a dependent variable that was a composite of saccadic and non-saccadic paradigms (Myles-Worsley et al., 1999). A flowchart documenting the inclusion process is shown in Fig. 1.

    • Correlation of neurological soft signs and neurocognitive performance in first episode psychosis

      2014, Psychiatry Research
      Citation Excerpt :

      Spatial working memory requires a normal functioning of the dorsolateral prefrontal cortex, the parietal system, and the orbitofrontal locations (Park, 1999; Park et al., 1999). Spatial working memory dysfunction has been proposed to be a key feature of schizophrenia (Park, 1999; Park et al., 1999). Previous studies (Pantelis et al., 2009; Pantelis et al., 2001) reported a significant association between poorer spatial working memory capacity and negative symptoms.

    • Impaired processing of binaural temporal cues to auditory scene analysis in schizophrenia

      2013, Schizophrenia Research
      Citation Excerpt :

      As our paradigms will examine sounds that are perceived as being located in either right or left hemispace, there is an opportunity to examine whether any deficits in patients are lateralized. Previous research has provided evidence of lateralized deficits for information presented in the right-hemispace in individuals with schizophrenia on a range of cognitive abilities such as; attention (Posner et al., 1998), working memory (Park, 1999) and mental number line bisection (Cavezian et al., 2007). The present study will explore whether this hemi-neglect is evident at the level of operation of the auditory scene analysis processes and thus whether it influences the quality of information individuals with schizophrenia can extract from stimuli presented to the right side of space.

    • Hemisphere asymmetry in schizophrenia as revealed through line bisection, line trisection, and letter cancellation

      2007, Brain Research
      Citation Excerpt :

      Some imaging studies have noted abnormalities of the left cerebral hemisphere (Friston et al., 1992) while other studies have argued for a right hemispheric dysfunction (Mayer et al., 1985) and even deficits in interhemispheric transfer. One study (Park, 1999) found what appears to be an asymmetry in spatial working memory in which schizophrenia patients appear to exhibit a greater loss of spatial representation by the left prefrontal system than the right. Left hemisphere dysfunction in particular has been considered by many as a central deficit present in the illness (Roberts, 1991).

    View all citing articles on Scopus
    View full text